Xerox University Microfilms 300 North Zaab Road Ann Arbor, Michigan 48106 77-2364

INFORMATION TO USERS

This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.

The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.

1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.

2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You w ill find a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again - beginning below the first row and continuing on until complete.

4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.

5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.

Xerox University Microfilms 300 North Zaab Road Ann Arbor, Michigan 48106 77-2364

CARMODY, Michael John, 1950- THERMAL ISOMERIZATIONS OF cis-9,10- DIHYDRONAPHTHALENE DERIVATIVESt AND THERMAL AND SOLVOLYTIC CHEMISTRY OF TRICYCLO[4.2.0.02,5]0CTA-3,7-DIENE DERIVATIVES. The Ohio State University, Ph.D., 1976 Chemistry, organic

Xerox University MicrofilmsAnn , Arbor, Michigan 48106 THERMAL ISOMERIZATIONS OF cls-9,10-DIHYDRONAPHTHALENE DERIVATIVES, AND THERMAL AND SOLVOLYTIC CHEMISTRY OF TRICYCLOL^.2*0.02,5]OCTA-3,7-DIENE DERIVATIVES

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Michael John Carmody, B.S, # # # # *

The Ohio State University

1976

Reading Committee: Approved by

Leo A. Paquette Harold Shechter Robert J. Oullette c x . Advisor Department of Chemistry DEDICATION

To Ann and Emily, my girls.

ii ACKNOWLEDGMENTS

The author would like to thank Dr. Charles Cottrell for the recording of all 13C and 1H (FT) pmr spectra,

Including the variable temperature data in Part II. Mr. J,

Michael Geckle is acknowledged for the LAOCOON treatments and patient instruction in the use of DNMR programs.

In addition, the author expresses gratitude to all of his coworkers whose advice and criticism shaped the content of this manuscript as the experimental work and writing progressed. The growth of these coworkers into friends made the pursuit of knowledge emotionally as well as intellectually rewarding.

Special thanks are accorded to Dr. Leo A. Paquette whose imagination and enthusiasm provided the impetus for this work and the authors development as an organic chemist. A great scientist who can relate to people on such a human level is rare indeed.

Finally the author wishes to thank his wife, Ann, not only for typing the manuscript but also for the en couragement and understanding without which this work could not have been completed.

iii VITA

May 14, 1950 Born - Detroit, Michigan

1972 ...... B.S., University of Detroit, Detroit, Michigan

1972 - 1974 Teaching Assistant, Department of Chemistry, The Ohio State University, Columbus, Ohio

1974 - 1976 Research Assistant, Depart ment of Chemistry, The Ohio 'State University, Columbus, Ohio

PUBLICATIONS

L.A. Paquette and M.J, Carmody, "Thermal Isomerization Reactions of cis-9,10-Dihydronaphthalene Derivatives," J. Amer. Chem. Soc., ^1* 5941 (1975).

L.A. Paquette and M.J. Carmody, "Thermally Promoted Ring Cleavage Reactions of Tetracyclo[4.3*0.02,5,07»9]non- 3-enes, Pentacyclo[5•3•0.02,0 ,03,5.0°,10]decanes, and Their Epoxide Counterparts," J. Amer. Chem. Soc.. submitted.

L.A. Paquette, M.J. Carmody, J.M. Geckle, "DNMR Analysis of Ring Inversion in anti_-l ,5-Bis homo eye looct ate traenes, " Tetrahedron Lett,, in press.

iv TABLE OF CONTENTS

Page

DEDICATION ...... 11

ACKNOWLEDGMENTS ...... Ill

VITA ...... lv

LIST OF FIGURES ...... vll

LIST OF ILLUSTRATIONS ...... Vlii

LIST OF TABLES ...... ix

THERMAL REARRANGEMENTS OF cls-9,10-DIHYDR0- NAPHTHALENE DERIVATIVES ...... 1

Introduction ...... 2

Results ...... 15

r( Discussion ...... 31

Experimental ...... 43

THERMAL REACTIONS OF TRICYCLO[4.2.0.02 »5]OCTA- 3.7-DIENE DERIVATIVES ...... 59

Introduction ...... 60

Results ...... *...... 72

Discussion ...... 97

Experimental ...... 102

SOLVOLYTIC STUDIES OF TRICYCLOC4.2.0.02*53-OCTA- 3.7-DIENE DERIVATIVES ...... 115

v Introduction ...... 116

Results ...... 123

Discussion ...... 133

Experimental ...... 1^3

APPENDIX ...... 160

REFERENCES ...... 164

vi LIST OF FIGURES

Figure Page

1 Variable temperature jjMR of 107 .... 88

2 Variable temperature NMR of 109 ...... 90 ^ ^ 3 Variable temperature NMR of 111 ..... 93

4 Energy profiles for ring inversion in 107 109 i H I ...... 777 9*

5 Diagram of orbital splitting due to through bond interaction ...... 119

vii LIST OF ILLUSTRATIONS

Scheme Page

I Orbital symmetry allowed interconver sions on the ( C H ^ q energy surface ...... 5

II Retro Dlels-Alder mechanism for rearrange ment of 13 and 35 ...... 32

III Cope rearrangement mechanism for 13 and 35 ” ...... 34

IV [l.SlSigmatropic shift mechanism for rear rangement mechanism of 13 and 35 .... 35

V [10]Annulene mechanism for rearrangement of 13 and 35 ...... ,39

VI Conversion of 132 to 142 138 “ w- “m

viii LIST OF TABLES

Table Page

I Isomerization - Dehydrogenation Studies of 22 and 23 at 80° ...... 19

II Isomerization - Dehydrogenation Studies of 13, 22, and 23 at 100°...... 22

III Vapor Phase Thermolysis of 42 ...... 26

IV Rate Constants and Activation Parameters for Thermal Rearrangement of 13 and 35 . . • 26 ^ - V •S* V ^-H NMR Data for Tetracyclononanes ...... 75

VI 1H NMR Data for Pentacyclodecanes ...... 76

VII !3c NMR Data for 100 - 103 ...... 77

VIII Rate Constants and Activation Parameters for the Thermal^Rearrangement of Penta- cyclo[5* 3.0.02 , 6 . * 5 . o u,iu]aecanes .. o3

IX Rate Constants and Activation Parameters for the Thermal Rearrangement of the Tetracyclo[.3.0.o2»5.o3,5,o? >9 jnonenes and Related Compounds ...... 85

X Thermodynamic Data and Activation Para meters for the Ring Inversions In 107, 109, 111 ...... 91

XI Rates of Acetolysis of 131 - 134 in Acetic Acid ...... 125

XII Example of a Solvolytic Rate Calculation . 152

Ix THERMAL REARRANGEMENTS OF

9 ,10-DIHYDRONAPHTHALENE DERIVATIVES INTRODUCTION

Monocyclic polyenes have held chemists' attention since

* f~i s~ X Kekule's proposal for the structure of benzene in 1865 .

The uniqueness of benzene's "aromatic sextet" was placed into 2 a more general context when Huckel published a quantum mechanically based theory which predicted enhanced stability for those monocyclic fully conjugated polyenes which contained a closed shell of *Jn + 2 electrons. The 4n homolog of benzene, cyclooctatetraene, had been synthesized by Wllstatter^ in 1911 and has been shown not to be blessed with such stability.

The next higher ^n + 2 analog of benzene is cyclodecapentaene ([10]annulene) which has proven, until recently, to be remarkably resistant to attempts by synthetic chemists to achieve it**^, Evidence for the transient existence of

[10]annulene derivatives formed upon thermolysis of substi tuted cls-9,10-dlhydronaphthalenes is to be presented in the present section of this thesis.

The name [10]annulene actually encompasses a variety of geometric isomers such as the all-cis 1, mono-trans 2 and di-trans 3. A discussion of aromatic properties, then is dependent on the structure and geometric and electronic

2 3

factors inherent to it.

The questions of steric strain (in 3) and bond angle strain (in 1^) have been addressed by Mislow^ and \\ Burkoth , respectively. In MIslow's model each of the two internal 1,6 hydrogen atoms resides well within the van der Waals radius of the other in the planar form, thus demanding significant distortions from planarity. That considerable aromatic character can be retained in spite

1 5 of such distortions is evidenced by Allinger’s? recent

synthesis of a substituted [7]paracyclophane (4) which

exhibits characteristic aromatic spectral properties even though the para carbons of the benzene ring were ap proximately 17° out of the plane of the other four benzene carbons.

Removal of the 1,6 interaction by placement of a monoatomic bridge between those two carbon atoms leads to stable derivatives. Vogel^ and Sondheimer^ have prepared derivatives such as 5 in which X = CH2 » NH,

NAc and 0. Debate over the extent of aromaticity in these species continues^1 H.

If 1 were to exist as a planar structure, then the carbon-carbon bond angles would be increased fromthe strainless 120° to lM°, The destabilization due to this strain may be overcome If the resonance energy is great enough, a situation which obtains for the IOtt cyclooctatetraene d i a n i o n ^ (< CCC, 135°) and the cyclo- nonatetraenyl anionl2 »13 (< CCC, 140°).

The most fruitful synthetic approaches to [10]annulene have been based on the "valence bond isomer" approach. The working premise here has been to gain access to a different CioH10 isomer capable of thermal or photo chemical transmutation into [10]annulene by mechanisms which are in accord with the Woodward-Hof fmann selection rules1^. Scheme I shows some "allowed" interconversions among these systems.

SCHEME I

hv hv 1

8 6

[10]Annulenes are seen to be available not only from cij>-9*10- dihydronaphthalene (6) and trans-9,10-dihydronaphthalene

(/) but also from bicyclo[6,2.0]decatetraene (8). Available • m * 0 pathways are also predicted for the interconversions 6^7, * m 6£8, and 7 ^ 8 which [10]annulenes may serve as reactive intermediates. Of these, the valence bond systems involving conversion of dihydronaphthalenes to [10]annulenes is the most studied. Disrotatory thermal opening of the central bond in 6 leads to the all-cis eyclodecapentaene 1 or the bis- - * 0 trans 3 while the allowed photochemical conrotatory process yields 2 which contains a single trans tt bond. Formation of 1 or 3 depends on the relative importance of the afore- * ■ * 0 mentioned steric and angle strains. The converse situation then holds for 7 with 1 and 3 forming photochemically, ^ * 0 * m while 2 is the thermally allowed product.

Both 6 and 7 have been synthesized and their properties examined in light of the above discussion. Pappas and van Tamelenl? were the first to prepare 6 from the quinone- butadiene adduct which could be converted into a mixture

0 Br Br Ll/Hg

0 Br Br 7 of tetrabromides which yielded 6 upon treatment with lithium analgatn. The pmr of 6 indicated a static structure not in equilibrium with a [10]annulene to a detectable extent at room temperature and the ultraviolet spectrum was typical of a homoannular diene. Upon heating to 150-200° in a sealed tube, aromatization with loss of hydrogen was kinetically dominant giving naphthalene as the only observed product. Irradiation of 6 through Vycor yielded nothing which exhibited absorption in the ultraviolet.

Rosenthal and Doeringl^ later showed that 6 exhibited no temperature dependent NMR spectrum and that the for mation of naphthalene upon thermolysis was the result of a disproportionation process which also forms various hexalins and tetralins. Their photochemical studies which were further elaborated by Jones-*-? show that 6 yields blcyclo[*l.2 .2]deca-2,4 ,7 ,9-tetraene (9) upon irradiation in • w

6 9 8 pentane at 0°. This in turn isomerizes to bullvalene

(10) which yields tricyclic triene 11 in a further conversion on the (CH)^q energy surface,

Burkoth^® synthesized 7 by a route analogous to that for 6 and found that its NMR revealed no tautomerisra to

[10]annulene. Upon thermolysis in a flow system at 250°

1-phenylbutadiene was isolated as the sole product.

The photochemical results were less lucid but at -190° in a solid solution 7 was inferred by ultraviolet evidence to have been converted to a [10]annulene. Upon warming and recooling, 6 was found to result.

Some of the most revealing work in the area has been the result of Masamune's investigations. Together with coworkers^ he found that 8 could be obtained as one of the products from the photolysis of a salt of the tosylhydra- zone of bicyclo[6.1. 0 ] n o n a - 2 ,6-triene-trans-9-carboxalde- hyde if the mixture was worked up at 0°. The thermal pro duct derived quantitatively from 8 at *10° was trans-9»10- dihydronaphthalene. An obvious proposal for the resulting stereochemistry is a controtatory opening of C^-Cg to give 2 as an unstable intermediate which can further under go allowed disrotatory closure to produce 7. 9

hv V - A

CHNNTs ^ ' Na 8

The [10]annulenes 1 and 2 have finally been synthesized and isolated by Masamune20. Irradiation of 6 at -60° followed by removal of tetracyclo[4.k.0.02 .0^]. deca-3,8-diene (12) which crystallized at -80° yielded crystalline 1 and 2 which did not interconvert thermally and which could be separated on alumina at -80°. Low temperature ultraviolet spectroscopy showed 1 to be nonplanar with iMeOH at 257 nra (e 2000). Both cmr and pmr max however, show single signal spectra invariant from -160° to -40°. These results are rationalized by an auto- merization of a nonplanar structure reminiscent of the interconversion between twist-boat cyclohexane conformers.

Isomerization of 1 to 6 Is the sole reaction observed thermally at -14°.

Likewise 2 exhibits a single signal pmr at -40° but at -100° five separate signals appear. A process in which the "trans” double bond in 2 migrates around the ring is invoked to explain these observations. The ultra violet spectrum of 2 shows 257 and 265 nm with IJlaA intensities of 2.9 and 2.0x10^ , respectively. At -40° 2 isomerizes to 7. 'V Neither 1 nor 2, then, appear to be aromatic compounds.

The reaction pathways predicted by orbital symmetry (see

Scheme I) seem to be adhered to for the most part, however.

The lability of [10]annulenes would seem to preclude their isolation from thermolysis reactions. However, their transient formation under thermal conditions is not dis counted and it has been of interest to examine this question more closely.

Vogel, Meckel, and Grimme*^ considered the possibility that replacement of the sp3-bound hydrogens in 6 by car- 11

bomethoxy groups would preclude dehydrogenation and

[1,5]slgmatropic shifts from competing with isomerizatlons to or through the [10]annulene„ These investigators noted that when 13 was heated at 90° for 2 hr and the result-

ing mixture of dihydronaphthalenes directly dehydrogenated at 150°, diesters 1^-16 were isolated together with other unidentified products.

c o 2c h 3

CO2c h 3 c o 2c h 3

13 lH

co2c h3

15 16

The formation of 1*1 and 15 which have the substituents symmetrically disposed was taken as inferrentlal evidence for the transient intervention of an unstable [10]annulene.

On the other hand, the isolation of 16 has been widely A* interpreted22 as the probable result of intramolecular 4 + 2 cycloaddition to give 17 followed by a retro-Diels-Alder reaction in the opposite direction yielding 16.

H 0 CH NC 3 3 19

The facile thermal isomerizations of parent diene to 6^7 and dinitrile 19 to 1 ,2-dicyanonaphthalene22 have been cited as mechanistic analogy.

A further study into the nature of the carbon scramb ling mechanisms available to 6 was undertaken by Paquette2^.

If 6 opens to a [lOjannulene such as 1 and recloses in a ■ w * w like manner with bond formation between two different carbon atoms, a degenerate rearrangement results. The

2.3-dideuterio derivative of 6 (6') was synthesized and pas sed through a flow system at 510°. In addition to dispro- portlonation products, 6' was isolated with a near statis- tical scrambling of deuterium indicating that some degenerate process had indeed resulted, the mechanism of which shall be further elucidated (see Discussion).

D D

It was the intent of the present work to ascertain whether or not the kinetic parameters for thermal rearrange ment of various 9>lQ-disubstituted-9,10-dlhydronaphthalenes were indeed consistent with the intermediacy of ClOjan- nulenes. The effects of substitution on the course of reaction was also investigated. Because the temperature used for dehydrogenation in Vogelts^-L investigation was conducive to further rearrangement, a reinvestigation of the thermal behavior of 13 was warranted. The here- “ t. * * tofore unknown, 9,10-dimethyl-9,10-dihydronaphthalene

(35) was also investigated in the same manner. RESULTS

cls-9 ,10-Dicarbomethoxy-9,10-dihydronaphthalene 13.

The synthesis of 13 began with readily available anhydride2^

20 and was effected in poor yield by sequential bromination with two equivalents of N-bromosuccinimide followed by dehydrobomination with quinoline yielding tetraene anhydride 21. Various other bromination-dehydrobromination procedures proved totally ineffective. Two step esterification according to the method of Vogel21 gave 13. Thermal

co2ch3

C 0 2 C H 3 co2ch3 H 3C02C

H3C02C 22

I co2ch3 co2ch3 rearrangement of 13 in dimethoxyethane solution at 80° for 19 hr gave rise to two new dihydronaphthalene diesters

22 and 23 which could be separated by careful chromato- I V V graphy on silica gel. The compound of higher Rf was identified as 1 ,5-dicarbomethoxy isomer 22 on the basis of its pmr, ir and uv spectra (see Experimental) and its conversion exclusively to 14 when dehydrogenated with DDQ in dimethoxyethane at 50°• Diester 14 was also prepared independently by sodium dichromate oxidation of 1,5- dimethylnaphthalene^ antj esterificatlon of the resulting diacid. Product identity in the case of 23 likewise followed from its spectral features and clear aromatization to methyl a-naphthoate on treatment with DDQ.

Additionally, the properties of 22 and 23 are in reasonable agreement with data previously reported by Meckel2®.

The problem of defining unequivocally the cis stereo chemistry of 22 and 23 persists. Although the uv spectra of 13 [Xmax 27® nm (e 4072)] and trans-9,10-dihydronaph thalene 7 Umax 276 nm U 3850) ]27 do differ widely, direct extrapolation to such dicarbomethoxy derivatives is fraught with serious limitations^®. Nmr methods are likewise unsatisfactory. Diester 22 as well as its trans isomer are symmetrical molecules; ^3c measurements are conse quently not capable of resolving the issue. A further complication lies in the fact that the signal due to the

sp3-bound protons in 22 is overlapped partially by the

intense methoxyl singlet. Detailed analysis of this portion

of the spectrum is thereby precluded in the absence of

suitable deuterium substitution. However, in view of the

experimental findings relating 9,10-dimethyl-9,10-dihydro-

naphthalene (35) to 42 and 43 (vide infra) whose cis stereo-

chemistry is no longer equivocal, the ring junctures in

22 and 23 are assigned cis by analogy. This matter is under-

standably of mechanistic import and the assumption that 13

and 35 rearrange by identical pathways may seem unjustified.

However, the great similarity in the kinetic behavior of

these two systems (vide infra) should substantially lessen

the arbitrariness of this comparison.

When both 22 and 23 were individually warmed to 80°

for 7 hr in sealed nmr tubes (tetrachloroethylene solution),

only minor proton spectral changes were noted. Extension

of the heating period to 24 hr, followed by treatment of

the solutions with DDQ at 50°, resulted in formation of

those product mixtures detailed in Table I, These reactions are of course indicative of the tendency of such

9,10-dihydronaphthalene diesters to further skeletal 18

rearrangement. From the product composition data, It follows that 22 (30/E of 1*+ isolated) is approximately twice as •V -V reactive as 13 (62£ of 24 obtained), barring the incursion - m — - » of undetectable degenerate isomerizations. In either case, the energy barrier which requires surmounting is necessarily larger than that facing 13. ■ w ■*»

c o 2c h 3

c o 2c h 3 25 26

As concerns the behavior of 22, a preference is seen for a return to vicinal placement of the two carbomethoxyl groups although now at C^, Cg (20% of 16) rather than at

Ci, Cq (trace of 24). During passage to 15 (5%)» the substituents have remained on diametrically opposite corners of the naphthalene core. Conversion to 26 (13%) > on the 4 Hf other hand, results perhaps from a more deep-seated re alignment of these groups. In the case of 23, the pattern is dismayingly simple, the two carbomethoxy functions remaining attached to adjacent carbon centers during 19

Table I. Isomerization-Dehydrogenation Studies of 22 and 23 at 80° (C12C=CC12 Solution)**

Naphthalene Products Diester 1-C02CH3 1,5-(C02CH3)2 1, 2-(C02CH3)2 2j3-(C02CH3)2 (24) J (14) (16) (15) - V 4 t

22 trace 29.8 20.1 trace

23 61.5 trace 22.8 14.9 4

Naphthalene Products Diester 2,6-(C02CH3)2 1,7- ■(c o ?ch3)2 Unknown(s) (25) <26/

22 4.8 13.2 32 •tr **

23 4 4 * - - a Data obtained by manual integration of vpc traces from duplicate runs. The percent composition figures are average values. 20

conversion both to 16 (23%) and 25 (15%)* ^ > w % ^ It should be noted here that authentic samples of 25 ■*> and 26 were synthesized by dichromate oxidation of 2,3- <*w and 1,7-dimethylnaphthalenes^( and esterificatlon of the

resulting diacids. A new route to 1 ,7-dimethylnaphthalene

was undertaken beginning with 7-methyl-l-tetralone (27)29.

Treatment of 27 with methyltriphenylphosphonium iodide and

n-butylllthium yielded methylene derivative 28, The

desired 29 was then obtained in'good yield from 28 upon

heating with elemental sulfur.

27 28 29

In the l,5-dicarbomethoxy-9,10-dihydronaphthalene experiments, an apparently homogeneous substance was also produced in appreciable amounts (32SE). On the basis of the pmr spectrum [6CDC13 7.6 (br s), 7.2 (m), 3.8 (s), TMS and 2.8 - 4,0 (br m)], its structure was recognized not to be that of a fully aromatized naphthalene derivative. 21

The possibility was considered that the isomerization and

subsequent dehydrogenation of 22 had given rise to 30^5 t

31^0, or 323°# but these diesters proved to be quite

different in their pmr features. Nor does the compound

appear to be 3 3 ^ or 3^ At this point, structural

assignment to the unknown has not been pursued further

because of its difficult accessibility.

C02CH3

j o a c o 2ch3 31■*v c o 2c h 3 30 C°2ch3 ^ J ^ co2ch3

c h 3o 2c XXX

Although thermal activation of 13, 22, and 23 at still ^ ^ ^ “v “w more elevated temperatures was expected to promote yet

more extensive rearrangement, their isomerization at 100°

(dioxane solution) and 150° (in diglyme) was examined

to determine if the available reaction manifolds were Table II. Isomerization-Dehydrogenation Studies of 13* and il at 100° (dioxane solution) and 150° (diglyme solutionja

Temp, ■ Na pntnaiene x roaucts Diester °c I-COOCH3 1 ,5-(COOCH3)2 1,2-(COOCH3)2 2,3-(COOCH3)2 (24) (14) (16) (25) Mr m . ■W -s. M> Mr Mr —m

3 100 25.0 10.8 20.6 26.6 M. 150 2.7 6.9 28.0 29.6

12 100 3.1 36.8 10.6 13.5 M r Mr 150 15.3 26.7 20.8 11.1

13 100 17.4 6.7 26.7 42.9 m m ^ 150 35.9 5.1 22.5 25.3

Temp . Napivtnaiene iroQucts Diester' °C 2,6-(COOCH3)2 i,7-(cooch3 )2 Unknown(s), % (15) (26) “v Mr

3 100 3.2 6.9 6.5 ■w 150 9.2 3.1 20.2 r 22 100 12.6 24.2 r 150 9.2 9.8 7.1

23 100 6.3 150 — * 11.2 a Data obtained by manual Integration of vpc traces from duplicate runs. The 23

entirely overlapping. As before, dehydrogenation with DDQ

was effected at 50° prior to product analysis by vpc.

The results are summarized in Table II. The same general pattern as seen at 80° emerges here with 22 giving chiefly ^ “V products with substituents diametrically opposed, while those derived from 23 show a preference for vicinal placement. It should be noted that significant amounts of 1,2- and 2,3-disubstituted products are obtained from 22 at

150°.

cis-9,10-Dimethy1-9,1O-dihydronaphthalene (35)* Pre paration of hydrocarbon 35 began with conversion of readily available dimesylate 36 to diiodide 37 by reaction with sodium iodide in hexamethylphosphoramide. Lithium aluminum hydride reduction of 37 in dimethoxyethane solution gave in 74? yield a semisolid distillate which was shown by vpc to consist of 38 33 an(j 393^ in a 9:1 ratio. Separation * w <*w was effected only after conversion to the respective dienes 35 and 40. Interestingly, a sequence involving bromination of 38 with pyridinium hydrobromide perbromide and dehydrohalogenation of 4l with anhydrous lithium chloride and lithium carbonate in hexamethylphosphoramide provided not 35, but chiefly cis-l,5-dimethyl-9,10-dihydronaph- - w • ■ thalene (42, 10JE) admixed with lesser amounts of 40, 45 **pi -w 2k

CH2OMs c h 2i

Nal _HMPA) CD CH2OMs CH2I

36 37

LiAlHi,

38 CH3 39

CH- CH3 LiCl H H Br Li2CO } Br Br HMPA 25 ch3 ch3 kl 1 A CH: H CHo DDQ

43 +

k2 25

and an unidentified monobromide. Skeletal rearrangement

could be effectively deterred, however, by sequential use

of N-bromosuccinimide and potassium tert-butoxide (in

tetrahydrofuran). The pmr spectrum of 35 is characterized

by an AA'BB' multiplet of area 8 centered at <5 6.52 and

a six-proton singlet at 1.03. Its catalytic reduction

provided cis-9,10-dimethyIdeealin exclusively. The features

of the pmr spectrum of 42 are equally revealing, with

absorptions appearing in the olefinic [6 5-91 (dd, 2),

5.66 (d, 2), and 5.25 (d, 2)3 doubly allylic [3.04 (m, 2)], and methyl regions [1.83 (s, 6)]. The electronic

spectra of both 35 [Aisooctane 244 nm (e 5575)] and 42 max [Xisooctane 250 nm ( e 6470)] are so similar to that of max cis-9,10-dihydronaphthalene (vide supra) that the likelihood

of trans ring fusion in either compound can be dismissed with certainty. The ready DDQ dehydrogenation of 42 to 1,5- dlmethylnaphthalene (45) serves as added proof of structure.

At 80° in tetrachloroethvlene solution, 35 was quanti tatively isomerized to a mixture of 42 (53*555) and 43

(46. 555) which could be separated by vpc methods at 80° using a 2-ft 7 - 555 SE-30 column. The more rapidly eluted component was assigned the cis-1,9-dimethyl structure on the basis of its spectra and dehydrogenation chiefly to 1- methylnaphthalene (44). This Isomer shows pmr signals 26 due to 7 olefinic protons (6 5*22 - 5-96), one doubly allylic hydrogen (2.92), and two distinctly different methyl groups (1.66 and 1.17). According to the uv [\ls°octane 240 max (e 3910) and 246 nm (3920)], a cis juncture is likewise present in 43. Suitable control experiments showed 42 and 43 to be stable to the reaction conditions.

The gas phase pyrolysis of 42 at several temperatures was investigated briefly. In these experiments, the pyrolysate was dehydrogenated directly to facilitate pro duct analysis since all of the methyl and dimethylnaphthalenes were available for comparison.

The data which are summarized in Table III reveal an

Table III. Vapor Phase Thermolysis of 42a

Percent Dimethylnaphthalenesb, rel % conversion T, °C 1,5- 1,7- 2,6- 2,3-

15.3 250 84.7 10.8 2.8 1.7

40.1 280 59.9 29.4 5.8 4.8

65-9 310 34.1 51.2 11.2 3.5 a b Contact time of 1-2 sec. Produced upon DDQ oxidation of the pyrolysates in dimethoxyethane solution at 50°; analysis performed by planimeter tracing of unweighted vpc peaks. 27

initial dominance of 1 ,5-dimethylnaphthalene (45) in the mixture. As the temperature was increased, the relative concentrations of 1,7- and 2 ,6-dimethylnaphthalenes were seen to increase at the expense of 45; significantly, 1,2- dimethylnaphthalene formation was not observed.

Kinetic Results

The isomerizations of 13 and 35 were studied kinetically.

Good first-order plots were obtained for the disappearance of both reactants and were reproducible when duplicate runs were made. The rate constants and activation para meters are listed in Table IV. Although dimethyl deri vative 35 is seen to rearrange approximately 14 times faster than diester 13 at 71.1°, the values of k^ determined at m m three temperatures show the two compounds to differ in significantly in their Ea » log A, AH*, and AS* terms.

cis-9»10-Dicarbomethoxy-9,10-dihydronaphthalene (13) rearranged smoothly to give a distribution of 22 (25.7?) and 23 (74.3?) which remained invariant during the period of rearrangement and was seemingly independent of tempera ture. Factoring for the disappearance of 13 by these percentage values provided quantitative data for the rates of appearance of these 9,10-dihydronaphthalenes (22: k^^

0.36x10“5 sec-1; £3 1 .2° - 1.04x10~5 sec-1; Table IV. Rate Constants and Activation Parameters for Thermal Rearrangement of 13 and 35 (C^C^CCls-Cf^COC^, 3:2 >a

Temp, kixl05, Ea, log A, AH*, AS*, Compd °C sec”1 kcal/mol sec-1, kcal/mol cal/deg mol

13 71.1 1 .39*0.06 1.38±0.05

81,2 4.07*0.11 26.7*0.7 12.1*0.4 25.7*0.4 *6 .3*1.2 4.04*0.14

91.0 11.50*0.50 11.50*0.49

35 51.0 1 .77*0.10 1 .76*0.08

61.3 5 .62*0.21 26.0*0.8 12.8*0.6 25.4*0.5 -2 .1*1.6 5 .65*0.12

71.1 18.90*1,10 18.70*0.91

a For 35 vpc analysis was employed. In the case of 13 ( pmr methods were utilized and the acetone solvent was dg. 29

*£91.0° = 2.96x10“5 sec-1; EA = 26.1 kcal/mol; log A = 11.2. 23: k-^ q0 ® 1.03x10-5 sec-1; *S8l.2° = 3*02

x10‘5 sec"1 ; i£91to0 = 8 *5il iO'^sec-1; Ea = 26.4 kcal/mol; log A = 11.8). These computations assume that we are not dealing with two consecutive first-order reactions of the type 13-*2 2-*23 or 13 "*■ 23 -* 22, However, the constancy of the percentage composition and the demonstrated stability of both 22 and 23 to the rearrangement conditions do not allow for this possibility. Rather, all indications are that 22 and 23 result directly from 13 at very similar rates.

The kinetic behavior of 13 is strikingly similar to '■w that of diester 46 which had been determined in perchloro- butadiene solution at 50-80°28. On the basis of the activation parameters for ring opening in 38 (Ea = 24.5 kcal/mol; log A = 12.0; AH+ = 23.8 kcal/mol; and AS+ =

-5.8 kcal/mol), Martin and Hekman favored that unimolecular decomposition pathway involving homolytic cleavage of the central bond and generation of 47^8 ^ From analogy, initial disrotation of one of the cyclobutene rings is expected 35 to be more energy demanding (EAi 2? kcal/mol and log A=14)J .

cis-9,10-Dimethyl-9,10-dihydronaphthalene (35) was shown to give rise only to 42 and 43 whose relative percentages

(53.5 vs 46.5) also remained constant with time at tne 30 three temperatures studied. As previously discussed, both and ^3 are unreactive to the conditions of rear- >V rangement. Factoring as before is seen to provide k^ values for the appearance of these hydrocarbons which are approximately half those given for 35 in Table IV.

Expectedly, all three sets of activation parameters are comparable.

Recently, ^8 was reported to have a half-life of 20 min at 1^5°^^. No additional kinetic data was provided.

Suitable extrapolation of our data for 35 to this temperature leads to an estimate of ^ / 2 s “ 5 sec! Judging from this comparison, we see that 35 is some 250-fold more reactive than ^8 despite the appreciably higher level of ring strain in the latter.

c o 2c h 3

CO2CH3 DISCUSSION

The question now arises as to how the thermal rear rangements of cis-9,10-dihydronaphthalenes take place.

The experiments conducted at higher temperatures reveal that several mechanisms may be operating simultaneously in a complex fashion so the present discussion will deal mainly with the lower temperature reactions. As a working hypothesis we have assumed that one or more of four symmetry allowed processes could account for the observed scrambling of substituents. They are: (a) intramolecular Diels-

Alder cycloaddition, followed by retrograde Diels-Alder reaction; (b) Cope rearrangement; (c) [1,5]-sigmatroplc shifting of one or more sp^-hybridized ring carbons; and

(d) disrotatory ring opening with formation of a cis^- cyclodecapentaene and subsequent disrotatory reclosure of this transient intermediate in an alternative sense.

Distinctions between these schemes are now possible since the present experimental findings do not satisfy certain of the conditions individually required by these four

31 32 options.

According to the first alternative, thermal activation of a 9,10-disubstituted cis-9,10-dihydronaphthalene could h p trigger intramolecular (tths +tt s) bonding in which an olefin in one of the cyclohexadiene rings acts as dienophile

(Scheme II). As previously noted1"^’22 (see Introduction)4 the new diene arising in this manner (42) can ring open under retro Diels-Alder control, a process which leads to 50, the precursor to 1,2-disubstituted naphthalene

SCHEME II

50 m m

Tl4s + tt2 s O X * 52 51 m m « ■ derivatives. Repetition of this particular pathway with

50 leads via 51 ultimately to 52. However, neither 1,2- nor 2 ,3-disubstituted naphthalenes are encountered upon thermal rearrangement-dehydrogenation of 13 or 35. Since these isomers appear at more elevated temperatures, we are of the opinion that 50 and 51, if formed, would be stable to the reaction conditions. However, this point has not yet been assessed directly by experiment. Import antly, access to 1,5 and 2,6 substitution plans cannot be gained readily In this reaction manifold.

Alternatively, the cls-9,10-dlhydronaphthalenes could be experiencing symmetry-allowed [3.3Jsigmatropic shifting

(Scheme III). Such Cope rearrangements have the effect of distributing the substituents at Co and C^q to positions

2,6 and 1,5 in a manner which excludes formation of other isomers, i.e., 9,10^2,6^1,5. 2,6-Isoiner 53 is a requisite intermediate, but no evidence for this species has been uncovered. Furthermore, 22 has been found to exhibit little tendency for interconversion with its 9,10- and

2,6-isomers (Table I) as expected of these relationships.

The reason for the absence of this pathway as a low energy process very likely arises from the- fact that L3 - 33 — sigmatropic shifting in this fashion necessarily occurs in 34

SCHEME III

R FT R

an "antara-antara" manner^, for which there are few bona

fide examples despite a seemingly adequate search^.

When the data for thermal activation of ljj and 35 are closely examined, a decided preference for initial isomeriza tion to 1,9- and 1,5-isomers is observed. Wnile formation of the 1 ,9-disubstituted products is the result of continued vicinal positioning of the pair of R groups, conversion to the 1,5 series necessitates placement of these labels at diametrically opposite positions on the cis-9,10-dlhydronaphthalene frame. These changes cannot be plausibly accounted for in terms of a single reaction channel; rather, a two pronged pathway involving mechanistic categories (c) and (d) which must be occuring with similar activation parameters best fits the observations without the need for special assumptions.

Suprafacial [1,5] migration of one of the four equi valent sp^-hybridized a-carbons in 13 or 35 leads directly to a 1,9-disubstituted 9,10-dihydronaphthalene of correct stereochemistry (cis ring fusion; Scheme IV). Note, too

SCHEME IV that traversing this pathway one step further leads directly

to the 1,5-isomer, a process which must not be operative

in view of the fact that 22 and *12 are inert to the conditions

under which 13 and 35 rearrange. Thermal activation of

57 could give rise to 1,7-isomer 58 and 2,6-isomer 53

a situation which may well obtain for the vapor phase

pyrolysis of *12. The 1,7- and 2 ,6-dimethylnaphthalenes “V are obtained as major products in yields which increase

in proportion to the temperature. That a [1,5]-sigmatropic

shift proceeds to the exclusion of possible competitive

migration by the bridgehead carbomethoxy or methyl groups

reo'd^s’1 " t-h^t w w’ 'dnvi v nanhon »nr>ssess Inherently — better migratory aptitude in this particular structural situation.

Although prior examples of comparable vinyl carbon migration

are indeed few^9j several investigators have observed

related isomerizations.

The most telling analogy is the kinetic analysis by

Schiess and Fiinfschilling of the rearrangements of 1- methylcyclohexadienes 59 to the isomeric dienes 60 in the

temperature range 150-360°They have shown that formyl

(Ea = 31.0 kcal/mol; log A * 11.5) migrates more than one

hundred times faster than hydrogen (EA « 35-9 kcal/mol; 37

59 A V log A = 11.2) which is very comparable to acetyl (EA *

35.9 kcal/mol; log A = 12.1). The most sluggish reaction was migration of carbomethoxy (Ea *= ^0.8 kcal/mol; log

A = 11.8) by a factor of 70 compared to hydrogen. All four groups migrated faster than methyl. The latter two facts would seem to indicate that Ll,5Jsigmatropic shifts of angular carbomethoxy (in 13) or methyl (in 35) could not compete as viable pathways for initial scrambling of labels while the data taken as a whole show that sp2- hybridized carbon can undergo sigmatropic shifts.

In their study of the thermal reorganization of substituted indenes, Miller and co-workers noted the pre ference order to be operative in those [l,5]sigmatropic shifts which occurred**-1-. Though additional work is required to expand on this limited view, the evidence currently available underscores the greater migratory capability of trigonal carbon centers relative to their

sp3-hybridized counterparts. That 9,10-disubstituted

cis-9,10-dlhydronaphthalenes would selectively undergo

preferential a-carbon migration as shown in Scheme IV

need therefore not be considered puzzling or demanding of

special interpretation.

The Arrhenius parameters for the appearance of

1,9-isomers 26 kcal/mol; log A = 11-12) are consistent

with a symmetry allowed reaction. In the 9,10-disubstituted

series, [l,5]sigmatropic shifting is facilitated due to relief of strain arising from removal of the 9,10 inter action. Continued rearrangement of intermediate 56 should not be as favorable since the steric driving force is

lacking.

A mechanism which could account for the positioning of the substituents in a diametrically opposed relationship

(as in 57) would be disrotatory opening to an all cis-

[10]annulene Intermediate which closes in one of two alternative paths to give products (Scheme V). Two as sumptions must be made If this mechanism is operative. First,

the [10]annulene must close to products much faster than it is formed. Secondly, placement of the labels In the 1,5-positions must take place to the exclusion of any 39

SCHEME V

O —

path a b * XX? R

other possibilities.

Several apparently analogous processes are known in which fused cyclohexadienes open to all-cis trienes cis-Blcyclo[4,2.0]octadiene (62)^2 opens to 0153-1,3,5 cyclo- octatrlene and tricyclic 63**3 opens to 64 at moderate - ** ^ ■ « temperatures and with comparable activation enthalpies 40

(26.6 and 25.1 kcal/mol, respectively) to those determined in this work for 22 and 42 (approximately 26 kcal/mol).

Masamune has shown^®, however, that parent hydrocarbon

1 closes to 6 with AH+ = 20 kcal/mol (AS^ = -3 eu )• Thus, analogy indicates that [10jannulenes should close thermally more readily than they are formed.

Arguments for selective positioning of the R groups at Ci and originate chiefly from thermodynamic considera tions; however, it remains entirely possible that kinetic control factors operate in the same direction. Information on this last point Is indeed scanty. In their study of the kinetics of electrocyclic ring closure In alkyl 1,3»5- hexatrienes, Spangler and co-workers observed that substi tution at the 3 position enhances the cyclization rate, whereas placement of the same group at the 1 position has a marginal effectUnfortunately, the consequences of

C2 substitution were not examined. Along other lines,

Photis has observed that 1,2,3>^>6-pentamethylcyclooctatetraene

(65) undergoes valence tautomerism preferentially to 66 rather than 67 The two bicyclic trienes differ only

H CH 3 CH V CH

CH CH

CH CH CH 66 in relative positioning of the fifth methyl group located on the conjugated diene moiety and consequently have comparable structural strain. However, the substituent prefers bonding to the terminus of the conjugated if system where direct interaction can operate. Cross conjugation as in 2,6- dimethyl-9,10-dihydronaphthalenes and 67 is certain to be less effective for stabilization purposes. It is especially noteworhy that thermal extrusion of nitrogen from tetracy clic azo compound 68 leads after dehydrogenation to a single product identified as 1,5-dimethylnaphthalene ^6,

In this instance, central bond cleavage is demanded by the

cheletropic reaction.

N , CH, CH

In summary, then, the kinetic parameters associated with thermolysis nf 9,10“disubstituted-9,10-dihydronaDhthalenes seems entirely consistent with formation of 1,6-disubstituted-cls^-cyclodecapentaenes as reactive Intermediates.

Formation of 1 ,5-disubstituted-9,10-dihydronaphthalenes further bears this out as does the strikingly similar behavior of 13 and 35- The formation of 1 ,9-disubstltuted products, however, makes it necessary to invoke a novel mechanism involving a [1,5]sigmatropic shift of a trigonal carbon atom to account for the formation of these products.

Furthermore, if one assumes that both mechanisms proceed with similar activation parameters, then an entirely reasonable and self-consistent reaction scheme evolves. EXPERIMENTAL

Infrared spectra were recorded on a Perkin-Elmer

Model ^67 Instrument. The pmr spectra were determined with Varian A-60A and Jeolco MH-100 instruments and apparent splittings are given in all cases. Mass spectra were measured with an AEI-MS9 spectrometer at an Ionizing potential of 70 eV. Microanalyses were performed by the

Scandinavian Microanalytical Laboratory, Herlev, Denmark.

9,10-Dihydronaphthalene-9,10-dicarboxylic Acid Anhydride (21).

A mixture of >6_hexalin-9 ,10- 0 dicarboxylic acid anhydride

(20)211 (25 g, 0.12 mol), N-

bromosuccinimide (50 g, 0.28 mol),

0 and (0.2 - 0.3 g) benzoyl peroxide

In 100 ml of carbon tetrachloride was heated at reflux for

1 hr with stirring. The mixture was cooled, the insoluble succinimide separated by filtration, and the solvent evaporated to leave a viscous yellow oil. This material was dissolved in 100 ml of quinoline and the solution stirred

43 at 1^*5° for 15 min. The mixture was poured onto Ice (250 g) and hydrochloric acid (50 ml) and extracted with ether

(3 x 175 ml). The combined organic layers were washed with

10? sodium thiosulfate, 10? hydrochloric acid, and saturated sodium bicarbonate solutions before drying and concentration.

The residual brown oil was chromatographed on silica gel

[elution with hexane-ether (3 :1)3 to give 4.7 g (19?) of colorless crystals, mp 74 - 75° (lit^l mp 74 -75°), whose pmr spectrum was identical to that previously reported. cis-9,10-Dicarbomethoxy-9,10-dihydronaphthalene (13)•

A solution of 8.0 g (40 mmol)

of 21+* «•* in 200 ml of methanol was heated at 45° with stirring for

C02CH.j ^ ^r * cooled, and evaporated to leave 9*3 g (100?) of monoacid ester; ^CDCl^ 5 ^ _ TMS (m, 8 , olefinic) and 3.60 (s, 3, methoxyl). This sub stance was slowly added to an ethereal solution of diazomethane (0.4 mol) which after 2 hr at room temperature was treated with formic acid (to neutrality) and evaporated.

The pasty white solid was recrystallized from ether-hexane to furnish 6.3 g (64?) of white needles, mp 102.5 - 104°

(lit21 mp 103-104°) whose pmr was the same as that 45 reported; ^isooctane 256 nm (e 6850). max

Thermal Rearrangement of 13 in Dlmethoxyethane Solution (80°).

A magnetically stirred solution CO CH 2 3 of 13 (2.0 g 8.13 mmol) in dry

| 'j y dlmethoxyethane (20 ml) was heated

at reflux for 19 hr, cooled, and CO CH evaporated^in3vacuo. This mixture was chromatographed on

silica gel (300 g). Elution with dichloromethane at a rate of 75 ml/hr gave the following1results (50 ml fractions); fr 18-33 , 270 mg, recovered 13 and minor aromatics; fr 34-

45, 520 mg, pure 22 ; fr 46-54 380 mg, mixture of 22 and 23;

fr 55-81, 600 mg, pure 2 3 . The total weight of 1700 mg represents an 8 9 % recovery.

For 22: vccl4 1730, 1710, and 1280 cm"1; *lsooctane 275 max max (£7390), 240 (4110), and 232 nm (3990); <5CDCl3 7.1 (d, J = 5-5 TMS “ Hz, 2), 5 .6-6.3 (m, 4), 2.7-4.2 (m, 2), and 3*80 (s, 6); calcd for CxijH^Oij m/e 246.0892, found 246.0895.

For 23: vccl4 1750, 1720, 1235, and 910 cm-1; xisooctane —~ max max 283 sh (e 4190), 257 (7720), and 252 nm (7890); 0 ^ 1 3 (<*» jL Mo J “ 5.5 Hz, 1), 5.5-6.2 (m, 6), 3.72, 3.69 (s, 3 each), and

3.54-4.0 (m, 1); calcd for m/e 246.0892, found

246.0895.

When both 22 and 23 were individually sealed in nmr tubes (tetrachloroethylene solutions) and heated at 80° for 7 hr, only minor spectral changes were noted. 46

Dehydrogenation of 22. A solution of 22 (63 mg, 0.25 mmol)

in 10 ml of dlmethoxyethane

^ was treated with 300 mg (1.3

^ mmol) of 2 ,3-dichloro-2 ,3-dicyano-

benzoquinone (DDQ) and heated

with stirring at 50° under

nitrogen for 24 hr. Chromatography on alumina (dichloro-

methane elution) led to the isolation of 1 ,5-biscarbo methoxy-

naphthalene (14) contaminated with residual 22. Repetition

of the process for an additional 24 hr gave only 14, mp 117 -

118° (lit ^ mp 119°), whose ir and pmr spectra were super-

imposable upon those of an authentic s a m p l e 2 ^.

Dehydrogenation of 2 3 . A 187 mg sample (0.75 mmol) of 23

was heated with 500 mg (2.2 mmol) CH0O2C 1 2 3 * I nnci of DDQ and 10 ml of dlmethoxyethane

under nitrogen at 50° for 24 CD hr. Purification by chromato

graphy on alumina (dichloromethane

elution) and vpc purification afforded methyl a-naphthoate

(24)29 accompanied by small quantities of 1,2- and 2,3-

dlcarbomethoxynaphthalenes. **7

Thermal Rearrangement of 22 and 23 In Tetrachloroethylene

Solution at 80°. Samples of 22 and 23 dissolved in tetra-

chloroethylene were heated

in sealed nmr tubes for A 24 hr at 80°. The solution

was cooled to 50°, DDQ added, c o 2c h 3 and the mixture stirred for CO^CH^ H„C0„C 24 hr. Purification was achieved

by filtration through a

short column of alumina.

Product composition data-

(Table I) were obtained by comparison of vpe retention times (12 ft x 0.25 in., 1556 QF-1 on Chromosorb

W) with those of authentic samples and by manual planimeter integration of the traces.

Thermal Rearrangement of 13* 22, and 23 at 100° and

150°. A magnetically stirred solution of 50 mg (0.2 ^ ^ % mmol) of a given dihydro - COjCHj A diester in 10 ml of dry dio-

> xane (for the 100° experi-

2 3 ments) or 10 ml of puri- °2ch3 A fied diglyme (for the runs at 150°) was heated for 24

hr under an argon atmos c o 2c h 3 phere. The solution was H,C0?C C02CH3 3 & cooled to 50°» DDQ (150 mg,

0.6 mmol) was added, and

the mixture was stirred

* at this temperature for 24 hr. Purification was

achieved as above giving the data in Table II.

l-Methylene-7-methyltetralin^(28). To a mixture of

11.6 g (28.7 mmol) methyltri-

phenylphosphonium iodide

In dry tetrahydrofuran under

nitrogen at 0° was added 22

ml (28.7 mmol) of 12.9% n-

butyllithium in hexane

and the mixture stirred for 1 hr. To this was added dropwise 4.6 g (28.7 mmol) 7-methyl-l-tetralone In 75 ml of dry tetrahydrofuran. The solution was allowed to warm to room 49 temperature, refluxed for 20 hr, cooled to 0°, filtered, and concentrated to leave 3.8 g (83?) of the hydrocarbon as a clear oil: 7.43 (s, 1H, aromatic), 6.97 (s, 2H, CDC1 ^ aromatic), 5.45 (s, 1H, vinyl), 4.92 (s, 1H, vinyl),

2.73 (t, J = 6.0 Hz, 2H), 2.50 (t, J = 8.0 Hz, 2H),

2.28 (s, 3H, methyl), 1.80 (t, J = 6.0 Hz, 2H).

1.7-Dimethylnaphthalene (29). A mixture of 1.1 g (6.9 IV IV IV *W "V « «v IV

elemental sulfur was heated at

210° for 5 hr. Chromatography

of the resulting oil on neutral alumina (elution with 25? ether-petroleum ether) yielded

820 mg (80?) 1,7-dimethylnaphthalene whose NMR was the same as the known (Sadtler spectrum #128llM).

1 .7-Dicarbomethoxynaphthalene (30). 1 ,7-Dimethylnaphthalene

(2.4 g, 15.4 mmol), sodium di- COpCHo , % , | * J chromate (13*1 g)» 44 mmol), CO2CH3 and 25 ml water were heated In a

rocking autoclave for 18 hrs at

250°. The vessel was cooled, the contents suction filtered, and the residue washed copiously with water. The combined

filtrates were acidified with 1:1 hydrochloric acid:

water and the resulting diacid precipitate filtered and

vacuum dried. The diacid was cautiously added to an excess

of ethereal diazomethane, the excess dlazomethane quenched

with formic acid, and the solvent rotary evaporated to

leave 980 mg (52?) of 1 ,7-dicarbomethoxynaphthalene mp

88 - 890 (lit26 90°).

cis-9,10-Bis(iodomethyl)-A2 *^-hexalln (37). A mixture

of 25 g (0,072 mol) of dimesylate

22^7 and 107 g (0.71 mol) of

CH2I hexamethylphosphoramide was heated at 130 - 135° with

stirring for 3 days. The black solution was cooled, treated with water (500 ml), and extracted with ethyl acetate

(5 x 300 ml). The combined organic layers were washed with water (1.5 1), followed by 10? sodium thiosulfate,

10? hydroclorlc acid, 10? sodium bicarbonate, and saturated salt solutions before drying. Removal of solvent under reduced pressure and recrystallization of the residue from

95? ethanol afforded 22.0 g (74?) of white needles, mp

95 - 99° (lit mp 97 - 98°32, 99°48). 51

cls-9i10-Dimethylhexalln (38). A solution of 20.2 g (0.05

mmol) of unpurified 37 in 100 ml

of dlmethoxyethane was introduced

CO dropwise into a magnetically

Cf*3 stirred slurry of lithium aluminum hydride (5.7 g, 0.15 mol) in the same solvent (200 ml) and the mixture was kept at reflux for 3 days. The contents were cooled and water was carefully added to destroy excess hydride before pouring into 500 ml of cold 10? hydrochloric acid. Ex-

* traction with petroleum ether (3 x 500 ml), washing of the combined organic layers with water (600 ml) and brine

(500 ml), drying, and concentration by distillation through an 18 in. Vigreux column at atmospheric pressure followed by vacuum distillation yielded 5.9 g (74?) of a semisolid distillate, bp 98 - 100° (13 mm). Vpc analysis (6 ft 10?

XF-1150 on Chromosorb G at 130°) showed this material to be composed of 90? 38 and 10? of 39. For 38:vneat 3100-2900, max 1450, 1*135, 1240, and 1030 cm-1; 6CDCI3 5.55 (m, 4 , olefinic), TMS I .83 (m, 8, methylene), and 0.9 (s, 6, methyl). This spectrum compares well with that previously described by Scott82, cls-9,10-Dlmethy1-9.10-dihydronaphthalene (35). A mixture

of N-bromosuccinimide (1.18 g,

6.7 mmol), hydrocarbons 38 and 39

9 0 :1 0 ) (500 mg, 3.1 mmol), 52 carbon tetrachloride (10 ml), and a few grains of AIBN was heated on a steam bath at reflux for 1 hr. The cooled mixture was rid of succinimide by filtration and the filtrate was concentrated on a rotary evaporator to leave a pale yellow oil (quantitative); 6 CDC13 5,5 (mj 4 ) # 4,8 (pr m , 2 ), TMS 2.1 (m, 4), 0.9 - 1.4 (series of methyl singlets, total area 6 ); a singlet at 2.65 due to the ethano bridged im purity was also clearly evident.

This oil was dissolved in 15 ml of tetrahydrofuran and the solution was transferred to a 25 ml 3-necked round bot tom flask equipped with a mechanical stirrer, condenser, and nitrogen inlet. After cooling to -78°, potassium t_- butoxide (3.36 g, 31 mmol) was added in one portion and the mixture was allowed to stir at room temperature for 16 hr.

The contents were poured into 250 ml of ice water and ex tracted with pentane (3 x 200 ml).The combined organic layers were washed with water (2 x 400 ml), saturated sodium bicarbonate (400 ml) and brine solutions (400 ml), dried, filtered, and evaporated to leave a yellow oil. Preparative vpc purification (6 ft x 0.25 in. 5% SE-30 on Chromosorb

G at 85°) furnished 46 mg (9 -2?) of 35 and 10 mg (2%) of

40. For 35;lisooctane 244 nm (e 5575) j SCDC13 6.52 (AA'BB', max TMS 8 , olefinic) and 1.03 (s, 6 , methyls); calcd for Ci2Hi4 H/®. ,

158.1095, found 158.1097. 53

The pmr spectrum of was identical to that reported^** .

Catalytic Hydrogenation of 35. A 10 mg sample of 35 in 5 ml

CHq of ether was hydrogenated at 1

H2/Pd atmosphere over 10? palladium on

carbon (50 mg) for 1 hr. The

ch3 solution was filtered and evaporated to give aclear camphoraceous semisolid whose pmr spectrum was identical to that of the known cis-9 ,10-di- V methyldecalin 33t Analogous catalytic reduction of diene 2*1 produced the same saturated hydrocarbon. cls-1,5-Dlmethyl-9»10-dihydronaphthalene(*12). To a mechanically

CH3 stirred solution of dienes 38 and

■v39 -« (9 0 :10 ) (3.0 g, 0.018 mol) in 60 ml of acetic acid was added

23.7 g (0.07 mol) of freshly prepared pyridinium hydrobromide perbromide and the mixture was stirred at room temperature for 1 hr. Water (500 ml) was added and the mixture was extracted with ether (2 x 200 ml). The combined organic layers were washed with 10? sodium hydroxide solution (500 ml) and water (500 ml), dried, and evaporated. After vacuum drying, there remained

8.22 g (91?) of powdery white tetrabromlde mixture which was used without further purification. 54

A 1.0 g (2,1 mmol) sample of tetrabromide was dissolved

in 25 ml of hexamethylphosphoramide. To this solution was

added 0.9 g (21 mmol) of anhydrous lithium chloride and 1.55 g

(21 mmol) of dry lithium carbonate, and the mixture was heated

at 90° for 5 hr. After cooling, water (150 ml) and pentane

(150 ml) were introduced and the organic phase was washed with

water (3 x 300 ml) and brine (200 ml), dried, and concentrated.

The residual yellow oil consisted of four components; these were separated in a pure state toy preparative vpc on 5JE SF-

96 at 115° (Chromosorb G). The initial hydrocarbon to elute was 42 (100 mg, 10%): 6CDCI3 5 .9 1 (dd, J = 6.0 and 4.0 Hz, 2), TMS 5.66 (d, J = 8.0 Hz, 2), 5.25 (d, J = 10.0 Hz, 2), 3.04 (m,

2), and I .83 (s, 6)49; xisooctane 250 (e 6470) and 277 sh nm max (2130); Calcd for Ci2Hil] m/e 158.1097, found 158.1095.

The second fraction (80 mg, 856) was identified as 40 on Oil the basis of its pmr spectrum-* .

The third component (19.2 mg, 1.$%) was shown to be

1,5-dimethylnaphthalene (45) by direct comparison with an authentic sample.

The final product was a monobromide (m/e 238) which remains unidentified (58 mg, 5*8/!).

Thermal Rearrangement of 35 in Tetrachloroethylene Solution

(80°). A 40 mg (0.25 mmol) sample of 35 was dissolved in 0.5 I * 4 ^ I V ml of tetrachloroethylene and the resulting solution sealed 55

under vacuum in a Pyrex tube. The tube was immersed in a pre-heated (80°) oil bath for 5 hr. The components of the product mixture were isolated by preparative vpc (2 ft x

0.25 In. 7-5% SE-30 on Chromosorb P) at 80°. There was obtained 14 mg (3*150 of *12 which proved identical to the above sample and 9 mg (2350 of *13: 6CDCI3 5.22 - 5.96 (m, TMS 7, olefinic), 2.92 (m, 1), 1.66

Kinetic Determinations, A standard solution was prepared by dissolving 3.00 g (12.4 mmol) of 13 in 25 ml of 3:2 tetra- chluroethyiene-acelone-U6 (0.49 M). Aliquots (200 m1) were transferred via syringe to 13 x 0.5 cm Pyrex tubes which had previously been treated with dilute aqueous nitric acid, dilute aqueous ammonium hydroxide, water, and acetone before drying at 75°. The sample tubes were sealed under a vacuum of 30 mm. For each run 10 tubes were immersed in an oil bath thermostated at the appropriate temperature. At various suitable time intervals, a tube was removed, immediately immersed in ice water, and stored at -10° for short periods until pmr analysis was possible.

Progress of the reaction was monitored by manual

Integration of the ester peaks of 22 ( 63.80) and 23 (3.72 56

only; the 3.67 singlet falls under that due to 13) relative “V ^ to that of 13 (3.65) on expanded spectra recorded at 60 MHz.

The percent of 13 present at a given time was calculated ac

cording to the following equation:

Q-3] area of 3.65 singlet-area of 3*72 singlet 13 = ------— ------=------[l3l+[22]+[23] total area

Interestingly, that ratio of 22:23 remained invariant at

25.7:7*1.3 in all runs. The percentage of 13 was plotted

vs time and the resulting data was analyzed by the method of

least squares. A tabulation of these findings is given in

Table IV.

For the dimethyl example, a standard solution was pre pared by dissolving 46 mg of 35 and 40 mg of n-decane in 10 “S * *-> “ “ ml of 3:2 tetrachloroethylene-acetone. Sealed tubes were prepared as before (100 ul aliquots). In this instance, progress of the rearrangement was monitored by flame ionization vpc (2 ft 5% SE-30 on Chromosorb G) at 65°.

Again, the ratio of 42 and 43 remained invariant at

53.5:46.5 from run to run. The kinetic data are summarized

In Table IV.

Dehydrogenation of 42 and 43. A solution of 4 mg of the hydrocarbon in 2 ml of dry dlmethoxyethane was treated with 57

15 mg of DDQ and heated at 50° for 24 hr. After chroma

tography on a short alumina column (dichloromethane elution)

and careful solvent evaporation, product analysis was

achieved by vpc methods (12 ft x 0.25 in. 5£ Carbowax

20 M-1SS potassium hydroxide at 140° )* While 42 gave 1,5-

dimethylnaphthalene (45) as the only detectable volatile,

4jg yielded an 80:20 mixture of a-methylnaphthalene (44)

and 45.

V Vapor Phase Pyrolysis of 42. In a typical experiment, 10 mg

of 42 was passed through a quartz tube packed with quartz ^ '■>* chips in the vapor phase with a nitrogen stream at 30 mm pressure and the pyrolysate (5-8 mg) was collected in a

trap cooled with Dry Ice-isopropyl alcohol. This material was

dissolved in 1 ml of dlmethoxyethane,and dehydrogenated with

DDQ (25 mg) as before (20 hr at 50°). After passage of this

solution through a short alumina column (dichloromethane elution), product analysis was conducted on a 12 ft x 0.25 in. vpc column packed with Carbowax 20M-1JK KOH on Chromosorb G at l4o°. The breakdown on the levels of dimethylnaphthalenes formed is given in Table III and partially in the accompany ing graph. 90 -

I I 1 ,5-Dimethyl- 80 - naphthalene p TJ O *1 70 TJ ET 3J P CD M M 60 - A 1,7-Olmethyl- P naphthalene C t H- tr < CD CD ►l 50- 3 *n o cd t—1 q 2 ,6-Dimethyl- 30 - I -Cr < f\J 20 -

"l— 250 THERMAL REACTIONS OP TRICYCLO[4.2.0.02 *5 ]OCTA-

3 ,7-diene derivatives INTRODUCTION

Orbital symmetry considerations^ predict that

the thermal concerted cleavage of cyclobutanes ought

to occur by a a2a + c2s pathway. This requires that

two carbon atoms destined to become 7r-bonded must rotate

l80° while the other two carbon atoms remain fixed.

A severely twisted, high energy geometry is required for

adherence to this pathway and as a result little stereo

selectivity has been observed in the pyrolytic cleavage of

stereochemically labeled cyclobutanes.

Historically, experimental work has focused on

attempts to redirect this reaction pathway towards Its

non-concerted alternative. Incorporation of the four- membered ring into increasingly more rigid polycyclic

fused ring systems, for example, was expected to enhance

60 61

this bias. Fusion of first one, and then a second cyclobutane

nucleus has led, however, to products whose structures increas

ingly seem to implicate adherence to the "allowed" pathway.

However, conformational effects in putative diradical inter

mediates can also account for the observed products. Seemingly,

however, these compensatory effects could be inhibited by a

further increase in rigidity. A greater understanding of the

mechanism of cyclobutane might thereby be made available.

This has been accomplished herein by incorporation of the four-

membered ring into sterioisomeric pentacyclo[5.3-0.02*^.

03>5,o^>i°]decane structures.

In an example supplied by Srinivasan^1, the pyrolysis of

6^ led at low conversions to a near equilibrium mixture of

420-30° >

(44jC) 69

products in support of a non-concerted mechanism. Pretwisting of the cyclobutane into the required

geometry has also failed to promote 2a + 2S cyclorever-

sion. Paquette and K u k l a ^ 2 synthesized 70, 71 and £2 and

found that thermolysis, yielded mixtures of the three

possible 2 ,6-octadienes and that 70, 71 and 72 could

be interconverted under the reaction conditions. A1 62

CH3 — CHot c h 3

c h 3 A70 ch3 72

c h 3 73 self-consistent rationale was formulated in terms of a diradical mechanism.

In addition to examples of the type mentioned above, extensive kinetic data35,53 have proven consistent with the intermediacy of singlet 1 ,^-butanediyl biradicals.

Bicyclo[2.2.0]hexane, which consists of two fused four-membered rings, has proven even less conducive to concerted cleavage to 1,5-hexadiene products since both rings must be severely distorted in the transition state.

The experimentally determined5** activation energy for thermolysis of the parent hydrocarbon (36 kcal/mol) corresponds very closely to theoretical estimates based on diradical intermediates.

Yet, the experimental findings show that high levels of antarafaciality are in fact attained by the double 55 bonds. Goldstein and Benzon"^ studied the thermal chemistry of exo-2 ,3 »5 ,6-d^-bicyclo[2 .2 .0 ]hexane and found that, after permanganate oxidation of the resulting tetradeuterio-1 ,5-hexadlene product(s), only d^- meso-succinlc acid was formed. This finding corresponds to essential adherence to thea2 s + a2a pathway.

The 2,3-dicarbomethoxy derlvatives56 7^ 75 and 76 also showed single inversion of stereochemistry on heating, with 77 as the major product from 7k and 75 (80JE in both cases) and 78 resulting from 76. Although such results

£ ^ C 0 2 C H 3 djf *1. CO-iCH-i ^ CO 64 would normally argue convincingly for the concerted alternative, a two-step process involving homolytic rupture of the center bond seems more consistent. The boat cyclohexanediyl initially formed can undergo

79 80

conformational Inversion to the more stable chair form and thus deliver 1,5-hexadiene products of appropriate stereochemistry.

This rationale has also been used by Roth^7 to explain the fact that identical product mixtures were obtained from azo compounds 8l and 82. Again, boat to chair interchange allows for facile generation of the three possible octa-2,6-dienes.

The additional strain associated with fusion of a third four-membered ring makes concerted cleavage of the central cyclobutane bond in syn- and anti-tricyclo- 65

[4. 2 .0. 0^ *5 ]octanes highly unlikely. Initial thermolysis of the parent hydrocarbons 83 and 84 by Nenitzescu^S yielded only cis^-cycloocta-1,5-diene(8£) . The "allowed"

K v ~

83 84 66

CH3O2C C02CH3

c h 3°2C^ v/ / N^-CO 2c h 3 4

CH302C^\ X^°2 CH- CH302C c°2CIi3 87 86

reaction, therefore, seemed to be energetically unfavorable.

Vogel59 found that the tetraester 86 also exhibited the

same reaction pattern yielding the cis^-cyclooctadlene 8 7 .

The first example of the isolation of a cis,trans

diene from a tricyclooctane derivative was provided by

Bellus^0 who observed that tetracyano isomer 88 was "m * v transformed upon sublimation to the stable cis,trans

diene 8 9 . Reaction of 89 with furan or cyclopentadiene Aw A. -W *W produced adducts having trans-fused rings. Further,

NC CN 89 88 i 67 thermolysis of 89 at 270° resulted in isomerization of the trans double bond to its cis form.

A subsequent reinvestigation of the behavior of

83 84

91 93

i I

85 90

1 the parent hydrocarbon by Martin and Eisenmann^1 at 150° in solution showed that both els,els- and els,trans- l,5- cyclooctadiene (90) could be obtained from either stereo isomer. No trans,trans diene was observed and 90 could be trapped with perchlorocyclopentadiene. Since the geometry required for the concerted g2s + o2a reaction pathway is so prohibitive, the hypothesis adopted to explain these observations provides for initial homolytic cleavage generating 91 or 93 followed by conformational ring inversion to 92 which then delivers 90.

Diradical 93 was also generated by Allred upon gas phase thermolysis of azo compound 9**^2.

He observed formation of 85 and 8*1 with the amount of 8*4 decreasing with reaction time. However, no 90 was pro- “NT duced. 69

The syn (95) and anti (96 ) dimers of cyclobutadiene

react similarly to give cyclooctatetraene, Dewar^3

invoked a diradlcal mechanism for this transformation

and on the basis of MINDO/3 calculations hypothesized that intersystem crossing to a triplet species may be involved but that this species is so close to the ground state in energy that its detection might prove difficult.

In a closely related example, however, heating of syn-

benzotricycloC^I. 2 . 0 . *5]oct-3-ene (98) in the presence 70 of 9fl0-dibromoanthracene (97) was found to give benzo-

£c3-l*3,5-cyclooctatriene (99) and visible light correspond ing to fluorescence from 97* A triplet species is thus implicated during the thermolysis of 98. Based on the observed positive entropy of activation for the conversion of 95 and 96 to cyclooctatetraene , Frey**1* believes that a major triplet pathway is not likely even though minor amounts of singlet to triplet inter system crossing may be occurring.

Despite the general acceptance of the boat-chair diradical interchange for explaining the stereoselectivity of tricyclooctane cleavage reactions, the intercession of conformationally distinct biradical intermediates remained but a deductive assumption. Accordingly, further insight into this mechanistic question seemed highly* desirable. This section reports the results of a study in which both the 02s + a2a mechanism and the boat to chair inversion process are blockaded by microcyclic annulation of the syn- and anti-trieyclooctadlene frames.

Restriction of the energy surface in this manner was expected to alter the steric course from antarafacial to suprafacial regardless of the mechanism prevailing 71 in the lesser restricted cases. The problem can be restated briefly as an examination of whether stereo- specific (Z,Z) ring opening of pentacyclic molecules such as 100 and 101 might now be realized. The three- membered rings not only impose a suitable level of conformational constraint, but also introduce the requisite degree of stereochemical labeling. Because the kinetics of these cycloreversions also gain considerable importance, the rates of thermal reorganization of the stereoisomeric tetracycloC^.3•0.0^*5,o?»9]non-3-enes and their 8-oxa analogs have been measured too. This last group of

100 101

compounds represents a systematic link between the tricyclooctadienes 95 and 96 and the pentacyclic series. RESULTS

Synthesis. The starting syn-(95) and anti-tricyclooctadienes (96) can be prepared conveniently by slight modification of Nenitzsecu's method^®. Under his conditions, treatment of cls-3,^-dlchlorocyclobutene with lithium amalgam in ether resulted in ready conversion to 9 6 . In contrast, we found'the effectiveness of sodium amalgam under entirely comparable circumstances to be greatly reduced and to require prohibitively long reaction periods (ca 2 weeks). Through the use of 25% HMPA in ether as solvent, however, conversion to 95 was complete in 7 hr. When exposed to ethylzinc iodide and diiodomethane®^, the preferred cyclopropanating agent in this work, both 95 and 96 afforded mixtures of mono and bis • m • m adducts which could be separated by VPC techniques.

Greater control of product formation was realized with m-chloroperbenzoic acid, one equivalent of reagent providing chiefly 102b and 103b. Diepoxides 100b and 101b were obtained through use of excess peracid. These latter materials are not especially amenable to gas chromato-

72 73 graphy, but can be purified by standard column methods.

2 1 fP

v b ,X -0 V 96 103a,bv 101a b

Stereochemical assignments to the various products follow from simple steric and spectral considerations.

Because the inner (endo) faces of these olefins are hindered by a cyclobutene ring in 95 and bridgehead hydrogens in 96, attack at an outer (exo) face is most likely. The fact that further chemical elaboration of the tetracyclic molecules affords highly symmetrical pentacylic structures also implicates initial bonding to the less congested outer face.

Given the numerous NMR studies on bicyclo[2,1,0]- pentanes^S some involving detailed computer simulation^, much detailed information on long range coupling and 7*4

shielding effects is available. For example, it is

now recognized that the endo cyclopropyl proton gen

erally experiences shielding relative to its exo counter

part as a consequence of its proximity to the cyclo-

butane ring^Sb,66,67. Marked shielding of the endo

cyclobutyl hydrogens by the three-membered ring has

also been demonstrated in these systems^5b,66,68.

Vicinal coupling constants in bicyclopentanes follow

the Karplus relationship and have in general been accepted

as diagnostic of stereochemistry‘ 69 ^.

In the present work, it Is especially noteworthy that the (Tables V and VI) and NMR spectra

(Table VII) of bisadducts 100 and 101 are compatible ^ ^ ^ only with structures having symmetrically disposed X groups. The noise-decoupled carbon spectra of 100a ^ ^ and 101a show three lines, while 101b and 100b exhibit two-line patterns Indicating two symmetry elements for all pentacyclics. The proton coupled spectra provide corroborative evidence for the gross structures.

Since the multiplicities of and H g ^ o in 100a and 101a make it clear that very low levels of spin-spin coupling (<0.5 Hz) to bridge protons H2(6 and are operative, these pairs of hydrogens must Table V, 1h NMR Shift Data (6, CDCl^ solution) for Tetracyclononanes3,

Compd Olefinic Bicyclo[2.2.0]bridge Bicyclo[2.1.0]bridges Methylene Groups Allylic Saturated Cyclopropyl Oxiranyl Anti Syn

102a 6.42 3.32 2.31 1.43 (m, 2H) 0.68 (d of t , 0.39 (d of t ^ ^ ^ (m, 2H) (m, 2H) (m, 2H) J - 5 and 4.5 J = 4.5 and Hz, 1H) 1.5 Hz, 1H)

103a 6.30 3.21 2.19 1.63 (m, 2H) 0.87 (d of t, 0.42 (d of t (m, 2H) Cm, 2H) (m, 2H) J - 5 and 4.5 J = 4.5 and Hz, 1H) 1.5 Hz, 1H) 102b 6.39

103b 6.31 3.20 2.69 3.89 Cd, 4* H t (m, 2H) (m, 2H) Cm, 2H) J = 3.5 Kz» 2H) a See Appendix Table VI. NMR Chemical Shift Data (6 > CDCl^ solution) for Fentacyclodecanesb

Compd Bicyclo[2.2.0]bridges Bicyclo[2.1 . 0]bridges Methylene Groups Saturated Cyclopropyl Oxiranyl Anti Syn

100a 2.50 2.00 (d, 0.60 (d of t, 0.35 (d of t, % ^ % V (m, 2H) J = 5 Hz, J = 5 and 4.5 J = 4.5 and 4H) Hz, 2H) 1.5 Hz, 2H)

101a 2.42 1.58 (d, 0.73 (d of t, 0.41 (d of t, ** ^ (m, 4H) J » 5 Hz, J = 5 and 4.5 J * 4.5 and 4h Hz, 2H 1.5 Hz, 2H) 100b 2.80 4.12 ^ ^ ^ (m, 4H) Cm, 4H)

101b 2.78 (d, “m 3*94 (d, J - 3-5 Hz, £ - 3.5 Hz, 4h ) 4n)a

J Additional fine coupling is apparent, to See Appendix -4 O'! Table VII. l^C NMR Chemical Shift Data (ppm from TMS, CDCI3 solution, 22.6 MHz).

Bicyclo[2.2. Ojbridges Bicyclo[2.1.0]bridges Compd Olefinic Allylic Saturated Cyclopropyl Oxiranyl Cyclopropyl

102a 139.85 *12,73 38.0H 15.65 10.41 "m -m — (167.2 ) (158.9) (1*18.9) (180.7) (159*1)

103a 1*10.71 50.07 44.57 17.97 13.16 ^ ** *0 ^ (166.2 ) (152.6 ) (156.3) (174.6) (136.7)

100a 38.95 16.24 12.09 ^ ** (1*17.7) (175.8) (163.6) r 101a ne.no 17.70 11.28 ^ ^ +*■ (inn.o) (174.6) (156.3)

100b 54.84 ^ ^ ^ ^ ni.73 (153.8) (197.8)

101b 47.32 55.68 - m *+ -m (151.1) (205.1) be geometrically fixed in an anti relationship. The

substantial deshielding experienced by the saturated bicyclo[ 2.2.0]hexyl bridge protons upon replacement at

X of CH2 by 0 ■ 0 .30-0 .3 8 , Table V) likewise accords

fully with exo orientation of the heteroatoms?0 .

This effect is also reflected in the 13c-H coupling

constants exhibited by the central carbons of 100 and 102 ^ ^ ^ relative to those of 101 and 103, in agreement with the

greater inherent s-orbital character of these bonds in

the more highly strained syn series?1 .

Product Studies. Thermal activation of all the tetracyclic

molecules gave rise to products having the bicyclc[6.1.0]-

nonane skeleton. Epoxides 101b and 103b at 80° gave * ^ « «r « cyclooctatetraene oxide 104 whose spectral parameters " m matched those of an authentic sample obtained by m-

chloroperbenzoic acid oxidation of cyclooctatetraene.

Attempts at GLC purification of this epoxide resulted in

rearrangement to cycloheptatrienyl aldehyde 105 and ^ P-phenylacetaldehyde 106, the amount of 106 increasing « % ^ ^ ^ with injector temperature. This process is known to be cata lyzed by'various raetal complexes?^ with the seven-membered ring aldehyde serving as intermediate in the reaction. 79

Heating 102a at 100° for 6 hr resulted in initial ^ ^ isomerization to c_is-bicyclo[6 .1. 0]nona-2, ^ ,6-triene which, because of its thermal lability, underwent subsequent rearrangement to an ca 85:15 mixture of cis- and trans-8,9-dihydroindene73. At the higher temperature required for pyrolysis of 103a (130°), exclusive conversion m m * * m H * to cis-8,9-dihydroindene was realised.

The thermal chemistry of 100a is characterized by ^ ^ ready conversion to diene 107 whose structural assignment follows from its temperature dependent ^-H and NMR spectra (vide infra). Ring inversion of 107 returns the

107 identical molecule but with interchange of the chemical environments of all twelve protons and all ten carbon atoms. When the exchange process becomes rapid on the NMR time scale signal averaging results. This behavior

differs intrinsically from that expected of cls-1,5-

bis-homocyclooctatetraene where conformational inversion now generates a non-identical structure. Additionally,

as its two methylene groups are transposed from the

extended to the folded form, high order steric congestion

develops representing a much less favorable and more energy demanding state of affairs than found in 107* ■ w ^ Diene 107 was also isolated from the pyrolysis of ^ 101a, but much polymer was produced as well due to the « % ^ higher temperature required for rearrangement (170°).

This is almost certainly not the direct result of product decomposition since pure 107 undergoes smooth first-order isomerization under these conditions to a product believed to be 108 on the basis of spectral data. This vinylcyclo- ^ ^ ^ propane ■+ cyclopentene bond reorganization operates to the exclusion of epimerlzation in The NMR spectrum (in 0f this 2-norcarene derivative consists of a low-field multiplet at 6 6.0-6.2 and a doublet of doublets (J *» 9.5 and 2 Hz) centered at 5.22 due to the a and 6 protons, respectively, and two multiplets at 0.6 and 1.1 arising from the gemlnal cyclopropyl hydrogens. The cyclopentene oleflnic pair Is seen as a pseudosinglet at 6 5*8 and the remaining sp3-bound protons give rise to multiplets in the 1.2-3.7 region.

The features of these signals were not sufficiently first-order to permit unequivocal assignment of stereo chemistry to the ring juncture. The 1^C spectrum was % supportive, showing three cyclopropyl, three allylic, and four vinylic carbons.

The structurally related diepoxide 101b was more ^ ^ ^ well behaved, being converted exclusively to luy at 150c .

Rearrangement of 100b proceeded at a somewhat lower temperature (130°) to give chiefly 109 (“* 95?) with -V slight contamination by 110 ( - 5?)- These stereoisomers ^ * * r** are readily distinguished by their -*-H NMR spectra76.

Since neither lOOa/lOla, lOOb/lOlb, nor 109/110 interconvert under the reaction conditions it would appear that both sets of pentacyclic molecules may be reacting through a common intermediate.

Of mechanistic importance, too, is the fact that conversion of 100a to 107 and 100b to 109 requires 82

109 110

conformational inversion of two bicyclo[2.1.0]pentane

bridgehead positions, a phenomenon not observed for

101a and 101b where configuration is retained.

Kinetic Results. The isomerizations of 101a,b and 100a, cr i were studied kinetlcally and the results appear in summary

form in Table VIII. The pyrolyses of the hydrocarbons

were conducted in benzene solution in sealed Pyrex

ampoules and reactions monitored by VPC analysis. [NMR

techniques were used for the VPC unstable diepoxides].

Good first-order dependence through two half-lives was

obtained for all compounds and Arrhenius parameters and

standard deviations were derived by the method of least

squares. Examination of the data reveals two striking

facts. First, the compounds with anti ring structure

react significantly more sluggishly than their syn counter-

'^5 Table VIII. Rate Constants and Activation Parameters for the Thermal Rearrangement of the Pentacyclo[5.3.0.02»o.o3»5.o8,10]decanes

Compd T, °C kxlO^, sec*1 AH*, kcal/mol AS*, eu Ea, kcal/mol log A

100a 99.8 3.00±0.07 30.68±0.99 2 ,71±2.55 31.38±0.93 13.9410.53 110.2 10.43±0.19 130.2 74.81+1.12

100b 110.9 2.88+0.05 29.22±0.63 -3.71±1.6l 29.78±0.50 12.46+0.28 120.0 6.92+0.44 129-7 18.05±0.40

101a 159.6 4.43±0.07 35.6 36.5212.96 14,1211.47 ^ 3.0 , 170.0 10.43+0.19 180.3 30.61+0.99

101b 129.7 1.31±0.04 35.14 5 .8l±2.12 36.01+1.02 14.71±0.54 140.0 3.90+0.10 149.9 ll.39iO.37

CD U> 84 parts (ca. 5 kcal/mol). Secondly, within any given stereoisomeric subset, replacement of the cyclopropyl methylene groups by oxygen is seen to exert no discernible effect on AH^ and Ea. The difference in Ea observed between 100a,b and 101a,b would seem to be the result of greater steric strain in the syn isomers which Increases their ground state energy.

These same trends are further corroborated by the results of the tetracyclic derivatives 102a and 103a,b % ^ ^ ^ ^ (Table IX). These reactions were also carried out in benzene solution. The difference in Ea between 102a ^ ^ « and 103a is very similar (4,7 kcal/mol) to that obtained ^ N V for the pentacyclic counterparts and epoxide 103b and - m ^ hydrocarbon 103a react with similar energetics. The “ m “ • “ w unsubstituted tricyclooctadienes also feature activation parameters which are larger In the anti example (AEa =

2.1 kcal/mol*’1*). in the fully saturated series, the energy gap appears to be significantly smaller, although kinetic measurements were only carried out at two temp eratures. It Is possible, however, that saturation of these tricyclic ring systems minimizes the differences in their strain energies. Since the cyclopropyl and oxirane derivatives isomerize at similar rates and since Table IX. Rate Constants and Activation Parameters for the Thermal Rearrangement of the Tetracyclo[4.3.0.02 >5.o7,9]non-3-enes and Related Compounds

Compd T, °C kxlO^, sec-1 AH*, kcal/mol. AS*, eu Ea, kcal/mol log A

102a 90.4 4.37±O.Q7 28.97t0.76 0.8l±2.04 29.64±0.59 13.51±0.35 « « « ^ 100.0 12.01± 0.52 110.2 36.59±1.22

103a 119.8 3.40±0.09 33.64±0.22 6.1U0.56 34.33±0.34 I4.69t0.02 m* 4* 130.2 10.48±0.26 140.1 29.72+0.79

103b 120.3 7.71±0.10 31.18+0.80 1.3^±1-99 32.09t0.69 13.76t0.38 •V 130.1 19.98+0.19 l4o .1 54.83±0.71

83® 30.9 3.4 31.7 13.6

95b 30.49t0.l6 14.22±Q.09

84a 31.2 4.1 32.1 13.3

96b 32.59t0.17 14.01±0.09

a Reference 77. ^Reference 64, 86

existing precedent would suggest that rate-determining cleavage of the central C-C bond in these three-membered rings would be characterized by a higher activation energy

for the oxygenated derivative, it would seem reasonable to

conclude that one of the bonds in the central portion of the cyclobutane network is being cleaved ir the initial step. The activation parameters in Tables VIII and IX fall into a consistent pattern based on this rationale but further discussion is deferred to the Discussion Section.

Variable Temperative NMR Studies. It Is apparent that 107 and 109 are capable of degenerate ring inversion and that the environments of all carbons and attached protons have been reciprocally interchanged. Passage through the coalescence temperature is tantamount to changing the effective shape of the molecules from q^ to C2h*

- 107, X - X'* CH2

109, X « X"- 0 • m 4* 111, X * CHo, x' « 0 4* ^ “ In addition, adduct 111 containing one cyclopropane and

p * ^ one oxirane ring was synthesized by peracid oxidation of

102a giving 100c followed by thermolysis at 125° (see

Experimental Section). In the case of 111, competition between epoxide oxygen and the cyclopropyl methylene group for "Inside" and "outside" positions gains impor tance. At -45° and below, the off-resonance decoupled

13c NMR spectrum of 107 (CDClo solution) consists of 6 I V J lines: 128.35, 126.90, 22.71, 17.80, 16.51, and 14.02 ppm. As the temperature is raised, these signals broaden and then sharpen to generate above +41° a triad of singlets at 127.62, 20.55, and 15.26 ppm. This mutual averaging of peaks shows that rapid equilibration between 107a * * **> and 107b leads to total environmental exchange between ^ ^ the four methine, four olefinic, and two methylene carbons (Figure 1). The equilibration of each set of methine, methylene, and vinylic carbons was viewed as a simple two-site exchange. By Inserting the chemical shifts and line widths of the signals in the absence of exchange into Binsch's^® DNMR 2 program, a series of calculated spectra were obtained correspond ing to first order rates of exchange (k). The calculated 88

308

266

263

233

-3 2 5,00 - 450.00 - 325.00 EXPERIMENTAL CALCULATED

Figure 1. Variable temperature NMR of the vinylic region of 107 in CDCI3 89 and experimental spectra were matched in the following manner. The line width at half-height (vjy2) for the signals in the experimental spectra were measured and plotted against observed temperature. Likewise v l/2 vs k was plotted for the calculated spectra. The observed v^/2 values were then matched with the cor responding value on the calculated plot and a value for k interpolated thus giving k vs T. This allowed for calculation of the values in Table IV in the usual way.

The room temperature NMR spectrum of 109 (o-dichlorobenzene solution, 90 MHz) consists of a downfield AB quartet (5^7.3 and *499.3 Hz), an upfield pseudodoublet

(338.8), and a broadened pseudosinglet (325*5). As judged from Dreiding models, the exo orientation of an oxirane ring causes the >CHO-hydrogens to adopt a dihedral angle of **90° with the adjacent olefinic protons and consequent low spin-spin interaction. Inner folding of the epoxide moiety decreases this angle to ~*40° and is therefore expected to measurably enhance coupling.

By standard LA0C00N techniques for AA'BB1 systems, the relevant coupling constants were determined to be:

jIaa' = -dd' * £ab “ °*50, 1 3 0 .2 9 , J3B t m

J^ct » 0.01, and Jcq ** 2.7*4 Hz. Upon gradual heating of 109, 4 2 6 '

395

378

3 6 3

313

- 350.30 310.30 - 350.30 310 . 30 EXPERIMENTAL CALCULATED

Figure 2. Variable temperature 1H MNR of the oxiranyl region of 109 in o- dichlorobenzene 91

* w the oxlrane protons were seen to become fully Isochronous

only at 150° (Figure 2). Under these limiting conditions,

the chemical shift of the >CHO-singlet (332.1 Hz) ap

peared midpoint between those in the static structure.

The observed coupling constants were then used in a

complete line shape analysis according to the method

of Binsch79. The calculated spectra were matched to the

experimental by simple superpositioning.

The 1h NMR spectrum of static 111 (-20°, ClpC15

CClp solution) is very similar to that of 109 in the ^ ^ oxiranyl proton region except that the singlet (304.3

Hz) is not diminished in intensity. The pseudodoublet

Table X. Thermodynamic Data and Activation Parameters for the Ring Inversions in 107, 109, and 111

111 107^ v ** 109^ ^

Ea, kcal/mol l6.4l±0.72 16.77±0.75 l8.8l±0.85 log A 14.08±0.56 13 .11±0.53 12.03± 0.49

AH+, kcal/mol 16.17±0.58 16.40+0.73 l8.17±0.76

AS+, eu -5.21±2.05 -0.05±2.00 - 5 .60±2.00

appears at 324.4 Hz. Upon slow heating to 71°, these signals broadened and then coalesced at a chemical shift 92

(318.8 Hz) (Figure 3) which proved to be a linear combination of the shifts of conformers 111a and 111b. The 111a:111b V 4 4 ^ ^ 4 ratio (1:2.58) therefore gives no indication of meaningful alteration with temperature, the concentration imbalance re presenting a AG° value of 0.56 kcal/mol. Kean life- "298 times as a function of temperature were determined by computer simulation of the oxiranyl proton signals as before; use of a relaxation time of 0.250 sec afforded optimal results (Table X).

The findings fully corroborate the structural assign ments of 107 and 109 as anti-bishomocyclooctatetraenes. ^ ^ ^ ^ ^ ^ In addition they disclose that diepoxide 109 undergoes 4 >V ^ tub-to-tub interconversion with less facility than hydro carbon 107. Since the AS+ components of these processes “ v ^ “ w “ are so similar, the 2.5 kcal/mol difference in Ea must be due to entropy factors. The behavior of 111 provides - m m v ^ Important clues. First, replacement of one cyclopropyl methylene by oxygen lowers the free energy of 111b with the « ^ ^ ^ oxygen "Inside" relative to that of 111a with the cyclopropyl ^ ^ Inside by oa 0.5 kcal/mol, probably as a result of the smaller steric bulk of the oxygen lone pair relative to a cyclopropyl hydrogen. This added ground state stabili- 344

313

307

300

£54

i j - 3 4 0 .3 0 - 290.30 - 340.30 290.30 EXPERIMENTAL CALCULATED

Figure 3. Variable temperature 1H NMR of the oxiranyl region of 111 in Cl2C»CCl2 zatlon could account for the reduced conformational

mobility of 111 relative to 107 (AA G*= 0.2 kcal/mol)

as represented in Figure 4 , It is not known to what An o extent the shorter C-C bond of an oxirane (1.472A

for ethylene oxide) as compared to cyclopropane®1

(1.510X) affects the attainment of planarity necessary

for 1 ,5-cyclooctadiene ring inversion.

In any case, bond length factors alone cannot account for the AAH^ of 2,0ikcal/mol between 107 and ^ ^ «

16.17kcalh mol A ,

18.17

109a -109a

Figure ^. Possible free energy profiles for ring inversion in 107* 109, and 111

^109. ^ ^ However, as shown schematically, dipolar interactions can gain Importance in the folded conformation

of the diepoxide (109c) internal solvation of the "inner

oxygen lone pair by the transannular oxiranyl carbons

leading to enhanced ground state stabilization. In

progressing to the ring inversion transition state, such

interactions must not only be decoupled, but an adverse

dipole-dipole situation develops which is perhaps ac

centuated by the interconnective double bonds (see 109d) % *** **+ ***

r o

109c 109d

The abnormally high AH* for 109b is therefore viewed to be the result of combined steric and dipolar electronic ground state stabilization In tandem with some dipole- induced transition state destabilization (Figure 4).

The differential response of 107, 109, and 111 to ^ ^ ^ the Input of thermal energy nicely illustrates the 96 sensitivity of medium ring inversion energetics to subtle structural factors. DISCUSSION

The mechanistic discussion entered into here will

concern intself with the behavior of pentacyclic molecules

100a,b and 101a ,b since the tetracyclic cases do not

possess the requisite number of stereochemical markers

to permit differentiation between the various mechanistic possibilities. Product and kinetic studies bring to

light the following facts: (1) syn and anti geometries

generate the same product, a process which is tantamount to configurational inversion for the syn isomers and retention for the anti isomers; (2) activation energies

(and enthalpies) are consistently lower for the syn cases; (3) the differences in energy between isomers in any given stereoisomerlc subset are minimal. These factors must be rationalized by the final mechanism.

Product-determining cleavage of the two central cyclobutane bonds seems rigorously excludable. If this process were occurring in a stepwise manner, both structural types would likely form els,cis-cyclooctadlenes. Since the geometries of the three-membered appendages would

97 98 not be affected by this process, 100a,b would serve ^ ^ as precursors to cls-1,5-blshomocyclooctatetraenes while

101a,b would generate the trans counterparts. This does m ^ ^ ^ not coincide with the observed fact that trans-1,5- bishomocyclooctatetraenes are generated in both cases under conditions where the cis compounds are stable.^1*

Concerted cleavage would require a severly twisted and strained geometry and would of necessity produce a

1 ,5-bishomocyclooctatetraene^containing one cis and one trans olefinic bond. Clearly, the energetic demands of this situation are extreme and not at all consistent with the observed parameters:

Assuming the behavior of 112 (AAH* - 3-3) and 113 « ^ « “ % n

112a,« « ^ ^ X - 0; AH* * 22.7 kcal/mol b, X - CH2 ; AH* « 19-^ kcal/mol

113a, X * 0; AH* - 28.6 kcal/mol

b, X - CH2i AH* - 25.0 kcal/mol 99

C AAH* = 3.6) to provide serviceable a n a l o g y ^ ( then the enthalpies of activation for the oxygen heterocycles should be of appreciably greater magnitude than those of structurally related hydrocarbons if internal bicyclo-

[2.1.0]pentane bond cleavage is kinetically important.

In actual fact, the activation parameters within each stereoisomeric subset are less disparate than the above: for 103a/103b, AAH* = 2.46; lOOa/lOOb, 1.46; and lOla/lOlb, - v -W 1^1 ^ “V V ^ V 0.46. Given the error limits of these measurements, the deviations are seen to be insufficiently meaningful to permit firm mechanistic conclusions.

Since the members of the syn series undergo ring opening more readily than do their anti counterparts greater intrinsic strain in the syn counterparts is implicated. This effect is best accommodated by homolysis of an internal blcyclo[2.2.0]hexyl bond or C2 g) with generation of biradical species such as 114 and 115 « 4 f % because of its close parallelism with the behavior of the parent tricyclic systems (Table IX). In contrast to the capability of biradicals 79, 91, and 93 for facile conformational inversion to their chair forms, these boat-like cyclohexanediyls are rigid Intermediates wherein bond rotation is precluded from operating because of the structural constraints imposed by the two three- 100

membered rings. In mechanistic terms, this blockade

of the normal ring flip can provide for subsequent dis-

rotatory opening of the neighboring bicyclo[2.1.0]-

pentane bond and formation of all-cis dienes 116 and 117.

Although 116 and 117 are generated in distinctive con- * W ~W "W “W

116

115 X 117

formations, these representations nevertheless comprise a single chemical species. Consequently, if the syn and anti isomers do not experience prior structural inter conversion, it would be at this stage that their identities would be lost. Simple ring inversion in extended con formation 117 provides the somewhat more sterically demanding colled form 116 necessary83 for the Cope re- 101

arrangement which would deliver the products ultimately

observed.

The combined weight of the above kinetic and stereo

chemical data forms the basis of our conclusion that

diradical energy surfaces are thermally accessible.

The formation of cis,trans-dlenes thought to characterize

such transformations is blocked due to inhibition of the

boat to chair interconversion and thus the apparent

ff2s + 02a fragmentation lead§, instead, to all supra-

faciality. The biradical pathway, then, offers the most

self-consistent mechanistic rationale for the cleavage of the strained cyclobutanes in these systems. EXPERIMENTAL

The ^-H NMR spectra were obtained with Varian T-60,

Varian A-60A, and Jeolco MH-100 spectrometers and apparent splittings are given in all cases. A Bruker 90 spectro meter was employed for the recording of spectra.

Mass spectral measurements were made on an AEI-MS9 spectrometer at an Ionizing potential of 70 eV. Pre parative vpc work was done on a Varian Aerograph A90-

P3 instrument equipped with a thermal conductivity detector. Microanalyses were performed by the Scandinavian

Microanalytical Laboratory, Herlev, Denmark. s^n-Tricyclo[4.2.0.02»5]octa-3,7-diene (95). A solution

containing 12.3 g (0.10

mol) of cis-3,4-dlchlorocyclo- 84 butene in 200 ml of dry 95 ether and 60 ml of anhydrous hexamethylphosphoramide was added under nitrogen to 650 g of 0.5Jt sodium amalgam contained In a 500 ml three-necked

Morton flask equipped with a mechanical stirrer, nitrogen inlet, and addition funnel. The mixture was stirred

102 vigorously for 7 hr, the ethereal solution decanted, and the amalgam washed with ether (2 x 100 ml). The combined ether layers were filtered through Celite and the major portion of solvent removed by distillation through a Vigreux column. The remaining ether was removed at 0° (40 mm) and the hydrocarbon obtained by raising the temperature to 40°. There was isolated 1.8 g (34.6#) of 95 having spectral properties identical to those reported,

Ip,26,56,66,7a,9a-Tetracyclo[4.3.0.0^»5o.7 >9]non-3-ene

(102a) and IB ,26 ,3ct ,5oi,6(3,73,83 ,10a-Pentacyelo[5 ■ 3 • 0 .

»6,o3»5, » 1 0 ]decane (100a). To a magnetically stirred

solution of ethylzinc iodide**^

(50 ml, 1M, 50 mmol) contained

in a 250 ml three-necked flask

equipped with condenser, ad

dition funnel, and gas inlet tube

was added dropwise under nitro

gen 5.35 g (20 mmol) of diiodo-

methane dissolved in 10 ml dry

ether. This solution was heated

at reflux for 1 hr and cooled while 1.04 g (10 mmol) of 95 and 5-35 g of diiodomethane dissolved in 15 nil of ether was added dropwise. After an additional 9 hr at the reflux temperature, the reaction mixture was cooled and poured into ice-cold saturated ammonium chloride solution. The organic phase was dried and concentrated in vacuo to leave a pale yellow oil consisting of two compounds in a 55:45 ratio. These were separated by preparative VPC (5J£ SE-30 on Chromosorb G,

70°), The more rapidly eluted component was identified as 102a; m/e calcd 118.0782, found 118.0783.

Anal. Calcd for 91.47; H, 8.53. Found:

C, 91.25; H, 8.73.

The product of longer retention time proved to be

100a; m/e calcd 132.0939, found 132.0935. ^ ^ ^ ^ — “

Anal. Calcd for Ci q H ^ : c , 90.85; H, 9.15. Found:

C, 90:91; H, 9.32.

4 ,9-Dioxa-lB,28,3ct,5a,6e,70 »8a,10a-pentacyclo[5■3 * 0^0^ ,

03,5,0® »1®]decane (100b). To a magnetically stirred

solution of 95 (900 mg, 9.0 «■» 0 mmol) in ether (10 ml) cooled to

100b -20® under nitrogen was added dropwise *15 ml of 0,5 N monoperphthalic acid in ether

(32 mmol). The mixture was allowed to warm to room temperature where stirring was continued for 17 hr.

The phthalic acid was separated by filtration and washed 105 with ether. The combined organic layers were processed as above to leave a white solid. Additional product was isolated by continuous dichloromethane extraction of the first bicarbonate wash. The combined solids were sublimed (60°, 0.1 mm) to furnish 610 mg of 100b • a M mp 141-1^2°, after sublimation.

Anal. Calcd for CqHq 02: C, 70.57; H, 5*92. Found:

C, 70.71*; H, 6.09.

8-0xa-lfi ,28 ,50 > 66 , 7a,9<*-tetracyclo[4 .3.0*02 * 5 7 * 9 ] - non-3-ene (102b). A mechanically stirred solution of 1.0

g (9*6 mmol) of 95 in 20 ml of

dichlornmethane was treated drop-

wise with 1.82 g (10.6 mmol)

102b of m-chloroperbenzoic acid ^ dissolved in 75 ml of the same solvent. After 1 hr the reaction mixture was poured into

200 ml of water and the dichloromethane layer brought to equal volume. The organic phase was separated, washed with saturated sodium bicarbonate solution (2 x 50 ml), water

(50 ml), and brine (50 ml), dried, and evaporated. The residual oil was chromatographed on Florisll (elution with 106

15? ether In pentane) to give 700 mg (47%) of 102b as

a waxy white solid, mp 31-32° (after sublimation); m/e

calcd 120.0575, found 120.0577.

IB,2a,5a,60,7a,9a-Tetracyclo[4.3.0.0^»5.q 7 >9]non-3-ene

(103a) and IB ,2a,33,56,6a,76,8a,10a-Pentacyclo[5^3^0,0^*6^

o3»5,o®>^®]decane (101a). Treatment of 96 (2.5 g, 24

mmol) and dilodomethane (13*5

g, 48 mmol) In 50 ml of ether

with a reagent prepared from 100

103a ml of 1 M ethereal ethyl-

zinc Iodide and 13-5 g of dl-

iodometbane (reflux 1 hr)

according to the predescribed

101a procedure afforded a mixture + * + * comprised of unreacted 96*

103a and 101a. The components were separated by preparative « « « « v ^ ^ VPC (5? XF-1150 on Chromosorb G, 50°). Monoolefin

101a was obtained as a colorless oil, 470 mg (23%); m/e calcd 118.0782, found 118.0783.

Anal. Calcd for CgH10: C, 91.47; H, 8.53. Found:

C, 91.19; H, 8.32.

There was also Isolated 900 mg (28?) of 101a : m/e * * " m calcd 132.0939, found 132.0942. 107

Anal. Calcd for C-^qH^: c* 90.85; H, 9.15- Found:

C, 90.42; H, 9.47.

8-Oxa-lfJ ,2a, 5 a ,68 ,7a,9a-tetracyclo[4 .3 .0.0^*5,o?»9]-

non-3-ene (103b) and 4,9-Dioxa-16,2a,33,5B,6a,78,8a,

10a-pentacyclo[5•3.0.0^*^.o3»5.o®»^®3 decane (101b ).

To a magnetically stirred

solution of 1.04 g (10 mmol)

of 96 in 10 ml of ether cooled

to -20° was added dropwise 15.4 103b ^ ^ ml of ethereal 0.65 N mono-

perphthalic acid. The solution

was allowed to warm to room •SSOCJ;. temperature where after 8 hr 101b m > the phthalic acid was separated

by filtration and washed with ether. Workup in the predescribed manner left a white paste, chromatography of which on Florisil (elution with pentane-ether 4:1) furnished 300 mg (25%) of 103b and 380 mg (28*) of 101b. M V *>i The monoepoxide was an oil; m/e_ calcd 120.0575, found 120.1577.

The diepoxide was obtained as colorless crystals, mp 111-112.5° after sublimation; m/e calcd 136.0524, 108

found 136.0526.

Anal. Calcd for CgHg02: C, 70.57; H, 5*92.

Pound: C, 70.^7; H, 5.89.

Kinetic Measurements. A. NMR Method. A standard solution of the dlepoxide (300 mg) in 1.5 ml of dry benzene-dg was prepared and 250 pi aliquots were introduced into NMR tubes which had been pretreated in turn with 10$ nitric acid, 10$ ammonium hydroxide, water, and acetone prior to oven drying at 70°. The tubes were sealed at -780 and 15 mm before immersion in an oil bath preheated to the appropriate temperature. The tubes were removed individually at appropriately timed Intervals, cooled to 0°, and the ^-H spectra recorded at ambient temp erature. The amount of residual starting material at any given time was determined by suitable integration of the

>CHO-signals in the original epoxide and product.

Plots of In fa/a-x yvs. time (sec) were linear through two half-lives. Least-squares treatment of these plots pro vided the data in Table VIII.

B. VPC Method. Standard solutions were prepared by ^ ^ ^ ^ ■ f m dissolving the substrate (50 mg) and internal standard

(50 mg) in 5 ml of dry benzene. For the tetracyclononenes, 109 the internal standard was n-decane, while cyclododecane and cyclodecane were utilized for the lOOa/lOla pair and ^ 4 ^ 4 103b, respectively. Into prewashed (see above) Pyrex tubes were introduced 25 pi aliquots of the solutions.

The tubes were sealed and manipulated as above except that the progress of reaction was monitored by flame ionization VPC.

Preparative Scale Thermolysis of the Tetracyclononenes.

A. Syn Isomer 102a. A 30 mg sample of 102a was dis- “w ^ iw % ^ v 4 4 * * " w ^ ^ solved in benzene-dg and its

signals integrated with respect

to the residual benzene protons. 102a After heating at 100° for

15,000 sec (10 tjyg)* ^MR spectrum was shown to be superimposable upon that of the mixture obtained upon heating cls-blcyclo[6.1.0]nonatriene under the same conditions. The conversions were both essentially quantitative. Vpc analysis of the two product mixtures revealed the composition to consist of an 85:15 ratio of cis- and trans-8,9-dihydroindene.

During the kinetic runs which were monitored by vpc, a peak was seen to gain prominence during the first half 110

life of 102a, reach a maximum at and gradually

transmute to the dlhydrolndenes. The retention time of

this material was identical to that of cis-bicyclo-

[6.1.0]nonatriene on several columns.

Anti Isomer 103a. The runs were carried out as above

but at a somewhat higher

temperature (150°). No cis-

bicyclo[6.1.0]nonatriene could + now be observed because of the 103a ~ necessarily higher temperatures.

Preparative Scale Thermolysis of 100a. A sample of 100a

(300 mg, 2.67 mmol) was dis

solved in 3 ml of benzene and

this solution was sealed into

100a a thick-walled Pyrex ampoule ^ at -78° and 15 mm before im mersion into an oil bath preheated to 150°. After 6 hr, the tube was returned to -78° and opened. The single volatile component was Isolated by preparative VPC on a 12 ft x 0.25 In. 15? XF-1150 column (Chromosorb W,

90°, 100 ml/min). There was obtained 250 mg (83?) of Ill

107 as a colorless oil; 6CDC13 5.57 (br s, 4h ), 1.56 TMS (m, 5H), and 0.78-1.21 (m, 3H); m/e calcd 132.0938, found 132 .09*12.

Anal. Calcd for C^qH ^ : c > 90.85; H, 9.15*

Found: C, 90.75; H, 9.31-

At -45° and below, the off-resonance decoupled

13C NMR spectrum of 107 (CDClo solution) consists of + * ^ ■ ' 6 lines: 128.35, 126.90, 22.71, 17.80, 16.51, and 14,02 ppm. As the temperature Is raised, these signals broaden and then sharpen to generate above +41° a triad of singlets at 127.62, 20.55, and 15.26 ppm. This mutual averaging of peaks shows that rapid equilibration between 107a M V and 107b leads to total environmental exchange between the four methine, four olefinic, and two methylene car bons. Matching of calculated (DNMR 2) and observed methylene and olefinic segments of these spectra provided a series of rates from which the following thermodynamic data were obtained: Ea , 16.41 kcal/mol; log A,14.08;

AH*, 16.17 kcal/mol; and AS*, -5.2 eu.

Preparative Scale Thermolysis^of^100b ,^101bi__103bt and 101a.

Because of limited quantities of materials, the remaining

"preparative" scale experiments were conducted in the 112 following way. In the first, 30 mg of the compound to be studied was dissolved in benzene-d^ and this solution was sealed in a precleaned (see above) NMR tube. The spectrum was recorded and the various signals carefully Integrated with respect to the residual benzene peak. The tube was then immersed in an oil bath pre heated to the temperature utilized in the most rapid kinetic run and heating was continued for at least ten half-lives. The spectrum was recorded again and inte grated as before to establish the extent of conversion.

For 100b and 101b,>91% conversion to a single product was noted. The spectra were superimposable upon those of authentic products. For 103b, the conversion was quantitative; in the case of 101a, the yield of 107 was determined to be 21%t Much polymer formation was in evidence.

Thermal Rearrangement of 107. A solution of 107 (55 mg,

0.41 mmol) in 1 ml of benzene

A was sealed in a Pyrex tube as V described above and heated for

107 12 hr at 170°. After cooling 113

4 the contents were removed and the single product isolated by preparative VPC on the XF-1150 column. There was obtained 48 mg (87?) of 108 as a colorless liquid; « W * r *H» 5CDCI3 6 .0-6.2 Cm, 1H), 5.8 (br s, 2H), 5.22 (dd, J =* TMS 9.5 and 2 Hz, 1H), 3.45 (m, 1H), 2.0-3.1 (br m, 3H),

1.4 (m, 2H), 1.1 (m, 1H), and 0.6 Cm, 1H); 13C NMR

(CDCI3) 135.64, 128.41, 127.98, 126.79, 42.14, 41.11,

34.58, 15.16, 14.19, and 9.55; m/e calcd 132.0939, found 132.0941.

Anal. Calcd for C10Hi2: C * H, 9.15.

Found: C, 90.57; H, 9.18.

4-Oxa-18,2S,3^T»5rt.68>7P,8a,l 0m-pentanyr.lo[ 5 . 3 • 0 . 0^ ,6 #

O3,5.o®»10]decane. To a magnetically stirred solution of

278 mg (2.4 mmol) cyclopropyl

hydrocarbon in 10 ml dry ether

at -20° under argon was added 100c dropwise 20 ml O.685N monoperphthalic acid in ether and the solution slowly allowed to warm to room temperature. After 12 hr the phthalic acid was removed by filtration and the total ether volume brought to 60 ml. The ether solution was washed with 50 ml portions of 10% sodium carbonate, 10? sodium bisulfite,

103E sodium carbonate, and saturated brine solutions, 11*1

dried and evaporated to leave 244 mg (77%) of a pale

yellow oil: 6TMS 4.05 (d, J = 3.0 Hz, 2H, oxiranyl), CDCI3 " 2.95 (m, 2H), 2.39 (m, 2H), I .65 (m, 2H, cyclopropyl methine), O .85 (dt, J = 5.0 and 4.5 Hz, anti cyclopropyl),

0.30 (dt, J = 4.5 and 1.5 Hz, 1H, syn cyclopropyl). ant1-4,5-EpoxybicycloC6.1.0]nona-2,6-diene (111). A

200 mg (1.49 mmol) sample of

100c was dissolved in 3 nil of

dry benzene and sealed under

111 house vacuum at -78°C in a thick- walled Pyrex ampoule. The ampoule was Immersed for 12 hr in an oil bath preheated to

125°, cooled to -78°, and opened. The product was separated by preparative GLC (2', 5% SE-30 on Chrom G 60/80 at 50°) to give 114 mg (57%) of a clear oil: 0DC13 5.12 - 5.86 (m, 4h , vinyl), 3-51 (m, 2.5H, epoxide),

3.30 (s, 0.5H, epoxide), 1.08 - 1.89 (m, 2H, cyclopro- pylmethine), 0.74 (dt, J « 5.0 and 4.5 Hz, 1H, methylene), and -0.07 (dt, J * 4.5 and 1.5 Hz, 1H, methylene); calcd for C9H100 m/e 134.0732, found 134.0733. V

SOLVOLYTIC STUDIES OP TRICYCL0[H.2.0.02

OCTA-3 17-DIENE DERIVATIVES INTRODUCTION

Transmission of electronic effects in organic molecules can take place either "through bonds" or

"through space". Such interactions are dictated by molecu lar geometries which orient either it or o bonds in a suit able position for efficient overlap. The intent of the pre sent work is to examine whether the solvolytic behavior of syn- and ant_I-tricyclo[4 .2.0. 0^» 5]octa-3*7-diene derivatives might serve as an assay of the extent to which through bond and through space n-bond interaction with developing cationic centers operates in these highly condensed molecules.

If two isolated olefins are brought face to face, the two degenerate it orbitals may split apart energeti- or cally. In norbornadiene (118) , a classic example, the in-phase linear combination of the olefins is placed at lower energy than the out-of-phase combination.

118

116 117

The antibonding levels likewise split with the result that the energy gap between the highest occupied and lowest unoccupied molecular orbitals is minimized and the ioni

zation potential is lowered with respect to that of

norbornene^. Closure to quadricyclane, a common photo- 8 8 chemical reaction of norbornadiene derivatives is thereby facilitated.

A similar mixing and splitting is predicted for the

two-electron 7-norbornenyl cation®^t the net stabili zation being reflected in an extremely large solvolysis rate enhancement for 120®9 as compared to 7-norbornyl

^OTs rOTs

J b

^119 m m m m

^rel 1

tosylate (119). That the interacting filled orbital « ^ ^ need only be rich in ^-orbital character is evidenced by the solvolytic behavior of 1 2 1 $°, the record holder for 118 rate enhancement due to anchlmerlc assistance.

Through bond coupling is experienced by diazabicyclo- octane 122. Extended Hiickel calculations^i predict ^ ^ ^

N N < T )

122 A

that the symmetric combination (S) of the lone pairs should be 1.6 eV higher in energy than the anti-symmetric

(A), in opposition to expectation. Photoelectron spectro scopy^ has confirmed these predictions. The phenomenon is explained®^ by the fact that the lone pairs couple through the parallel a bonds thus lowering A with respect to S. (See Figure 5).

A similar phenomenon is observed solvolytically in comparing the behavior of the quinuclidine derivative 119

Figure 5. Diagram of orbital splitting due to through bond Interaction >

123 to l-bromobicyclo[2.2.2]octane (124). Heterocycle 120

A CH2 m : + 11 ^ p * Br CH2 123

1 *rel 5x10^ Products

123^3 solvolyses faster than 12^1 by a factor of 5x10^ ^ ^ ^ ^ ^ and the products are those of fragmentation.

The syn and anti dimers of cyclobutadiene 95 and

96 offer a unique orbital topology in terms of the above discussion. Both isomers possess large degrees

95 96 125

of symmetry and have o bonds intervening between 121

the olefins appropriately disposed to through bond

coupling. The syn geometry, however, orients the

n-bonds face-on and proximal enough for through space

interactions to operate. These effects have been examined 94 qk both calculationally and experimentally^ via

photoelectron spectroscopy. Because of the different

configurations of 95 and 9 6 it is possible to separate the different types of interactions.

Through bond influences ^re found to dominate in both cases, with the first ionization potential of 96 appearing “ m m m 0.15 eV lower than the model 125. The syn isomer had % ^ ^ a still somewhat lower ionization potential (by 0.14 eV), an effect due to the destabilization of the tt- orbitals arising from through space interaction. MINDO calculations also show that the cyclobutene-cyclobutane angle In 95 is 3° larger than in 96 probably due to repulsive interactions of the ethylenic units.

Very little of the chemical consequences of these orbital effects is known although the failure to obtain cubane from 95 by direct or sensitized irradiation has been explained^>95 as a result of through-bond coupling.

The retro reaction catalyzed by Rh1 has been reported^. 122

RhI 95

It was the intent of the present work to synthesize alcohols 126-129 as precursors to derivatives for solvolysis. Kinetic and product studies of these reactions might then serve as further cljemical probes into the nature and magnitude of the orbital interactions, if any, with the developing cationic centers in these systems.

(Rr (P 1 2 6 127 + * « % ^

128 129 RESULTS

Synthesis. Alcohols 126 and 128 were synthesized from dienes 95 and 96 by selective hydroboration with 9- borabicyclo[3.3* 1]nonane in tetrahydrofuran in the custo mary manner^. The realized yields were fair due to the difficulty inherent in selective functlonalization of two equivalent olefinic centers. 'Attempted hydride reduction of expoxides 102b and 103b proved non-effective. The ^ ^ ^ ^ "W assignment of exo stereochemistry was based on simple steric considerations and NMR data. The respective signals for the proton a to the hydroxyl appear as a broadened triplet due to splitting by the adjacent cyclobutyl methylene. Spin-spin interaction with the adjacent bridgehead proton is minimal due to the near perpendicular orientation of the two substituents

(see Part II). Oxidation of 127 with N-chlorosuccini- ^ mide and dimethyl sulfide^ yielded a ketone whose in stability precluded its characterization. However, its direct reduction with lithium aluminum hydride produced a new alcohol (130) different from 126. Now the proton 12H

„ k HN-NH y] K 9-BBN ^ q y » _ | « f “

95 126 127

9-BBN OH

96 128

LIAIH4

| W ^ - H

HO 129 130

a to the hydroxyl group was seen as a broadened multiplet because of added coupling to the bridgehead hydrogen.

Additionally, one of the vinylic protons, formerly a 12k

9-BBN

95 126 127

9-BBN OH

128

LlAlHjj

HN=NH

HO 129 130

a to the hydroxyl group was seen as a broadened multiplet because of added coupling to the bridgehead hydrogen.

Additionally, one of the vinylic protons, formerly a Table XI. Rates of Acetolysis in Acetic Acid 0.0510 N in NaOAc

65.2W compd k(xlo5)( sec”-1- AH* T, °C AS* -rel

79.0 3.75*0.10 ^ O T s 90.0 10.75*0.27 27.16±1.99 -2.12+5.67 3.65a ^131 4 % 68.2 0.91*0.02

79.0 0.95*0,04

iR r 0Ts 90.0 3.05*0.14 26.10*0.28 -7.40*0.76 la

132 100.6 8.97*0.23

68.2 0.47*0.02

^ p O T s 85.2 3 .04±0.09 26.22+0.21 -6.33*0.59 1.62

133 100.6 14.32±0.47

125 M CttyOTs f\J 85.2 3.23*0.30 1.72 ui 134 ^ a Interpolated values. 126

finely spaced triplet in 126, is now shifted downfield

(6 6.62 as opposed to 6.50 in 126) as a result of de-

shielding by the oxygen lone pairs. The saturated analogs

127 and 129 could be obtained from 126 and 128 by reduction

with excess diimide generated in situ from dipotassium

azodicarboxylate-*-1^ . The corresponding tosylates101, in? 3,5-dinitrobenzoates > and acetates were prepared in

the classical manner.

The solvolytlc rate constants for these tosylates,

determined in buffered acetic acid by titration of the

unconsumed sodium acetate buffer, are listed in Table XI.

Surprisingly, they fall within a rather narrow range.

In each case slightly less than the theoretical amount

of acid was liberated. The rate constants were determined using the "infinity titer" observed after 10 half-lives

and represent the average of two independent runs. The

solvolyses were followed through 1.5-2 half-lives and

good first order plots were obtained in each instance.

The major products formed upon acetolysis of 131 v ^ (95? yield) proved to be acetates 135 (85?) and 136

(6.5?). Two unknown compounds (3? and 6 %) eluted with

135 and could not be obtained in a pure state for ■ » s ^ 127

identification. Structural assignment of 135 follows

OAc 'Ac {KT 131 135 136

from comparison of its NMR features with those pub

lished by Moriarity and Yeh^-^3 for the product of acid-

catalyzed HOAc addition to semibullvalene. These

workers also demonstrated the stability of this

acetate to the solvolysis conditions. The structure of

136 is based solely on its NMR, and particularly + m + W on a direct comparison with that of the corresponding endo

OAc

1) NaBH/i 2) AC20

137

isomer 137. This acetate was synthesized from the readily 4 ^ 128

available bicyclo[ 4 .2 .0 ]octa-^ t7-dien-2-one10i* by

sequential sodium borohydride reduction and acetylation.

Prom steric considerations known to operate in hydride

reductions and the analogous behavior of bicyclo[4.2.0]-

oct-iJ-en-2-one (vide infra), the endo configuration in

137 is deemed conclusive. The spectral patterns of the

vinylic regions of both 136 and 137 are very similar with “ v “ V “n ^ two cyclobutenyl protons being observed farther downfleld than the two proton signal from the cyclohexenyl hydro gens. The absorption of the proton u to the acetate in

* 136 is a narrow multiplet indicating little spin interaction with proximate hydrogens, while the high level of spin interaction operating in 137 produces a >^r doublet of triplets (J = 7.0 and 10.0 Hz).

Comparable buffered acetolysis of 132 led to a + * - m spectrum of products (68% yield) consisting of 138 * +

OAc

+ + 132 138 (21%) 139 (11%) ^ ^ ^

OAc OAc

+ +

m o (11%) m i (5%) OAc 1^2 (52%) 129

(21%), 139 <11*), 140 (11*), 141 (5*), and 142 (52*).

The 138/139, and 140/141 acetate pairs were isolated as mixtures by preparative vpc. The correspondaqce of the observed *H NMR absorptions of 142 with those published ^ ^ for the indicated structure**^ are striking, particularly that of the proton a to the acetate group which appears as a doublet of triplets centered at 6 4.97 (J - 9-00 and 6.46 Hz). The structure was further elucidated by

•*

n V OAc OH 143

repetition of Pincock’s scheme of reduction to the alcohol followed by oxidation to ketone 143 which had the ap propriate carbonyl stretching frequency (1740 cm"*).

Both bicyclo[3.3.0]octenyl acetates 138 and 139 had been previously prepared in these laboratories*0^ and their pmr spectra and vpc retention times proved identical to those of the authentic samples. Subsequent reexposure of 130

each acetate to the reaction conditions resulted in no further structural change. A known mixture of the acetates 140 and l4l10^ (predominantly endo) was available by reduction of bicyclo[4 ,2 . 0]oct-4-en*-2-one with sodium borohydride followed by acetylation. Direct comparison of the NMR data (particularly the chemical shifts of the acetate protons) and vpc retention times of this mixture with that obtained from the solvolysis attested to the identity of the carbon.framework of 140 and 141.

Anti isomer 133 solvolyzed to give four products

(69% yield) as determined by vpc analysis. The two more rapidly eluted components could not be identified due to limited accessibility in the first case (3?) and coincidental elution with 144 in the case of the second (12#). The ■V -w + * two remaining compounds were identified as the acetate of retained structure (144, 3350 and 7-acetoxy-l,3,5-cyclo- ^

145 131

octatriene (145, 52?). Acetate 144 was shown to have the

identical NMR spectrum and vpc retention time as the

authentic sample derived directly from 128, while 145 **» ^ ^ had NMR parameters identical to those described in the

literature^? . Both Kroner and Huisgen10^ have observed

that 145 is in equilibrium with its bicyclic tautomer

(145"). In fact, Huisgen reported that 53? of the bicyclic

is present at 60°, Acetate 145 likewise exhibits the ^ ^ ^

expected two acetate methyl singlets ("-1:1) in its 1h

NMR spectrum.

Relevantly, 145 was Isolated as the only volatile ^ product (5?) from reduction of cyclooctatrienone (146) ^ with 9-borabicyclononane10^ and subsequent acetylation.

Submission of 144 to the reaction conditions also pro- ^ ** vided 145 exclusively, although vpc monitoring of the pro- « ^ duct mixture from the solvolysis of 133 with time showed % ^ 144:145 to be constant at 1,2,5, and 10 half-lives.

Syn Isomer 147 also yielded 145 as the sole product of rearranged structure (45?) upon submission to the acetolysis conditions. These ring openings may be simple thermal processes as the temperatures at which the solvolyses were conducted are sufficiently elevated to promote such transformations (see Part II). 132

1) 9-BBN » 2) amino - ethanol 147 3) Ac 20 146

The product mixtures derived from 134 (58? yield) **r -w ^ consisted simply of 140 (66?) and l4l (34?), both of ^ ^ A * ^ ^ which had been previously identified.

OTs

134 140 (66?) 141 (34?)

Finally, the saturated acetates 148 and 149 were 4# 4t ^ ^ found to be inert to acetolysis conditions.

OAc 148 DISCUSSION

The acetolysis rate constants for all four tosylates are seen (Table XI) to fall within a very narrow range, the difference between the fastest and slowest being merely a factor of 3*65. Bicyclo[2.2.0]hexyl tosylate

(150)^-10 actually solvolyses five times faster than 1 3 2 . — -ta *** This comparison demonstrates that the solvolysis rates for 131-134 are close to those expected for ionization of strained cyclobutyl tosylates.

It might be argued that the similarity between'the behavior of the saturated and unsaturated compounds in a given stereoisomeric subset may be due to compensatory rate enhancing and retarding factors (perhaps by an inductive mechanism) in the unsaturated cases. Indeed, in norbornyl systems possessing vinyl groups

OTs

150 132

5.5 1

133 134

CH2

OTs OTs

151

trel 1 22 prohibited from through space interaction with the reaction center as in 7-fliethylene-endo-2-norbornyl tosylate

(151), a >20-fold rate retardation is seen to be the result of adverse inductive contributions^-. A similar effect, albeit greatly leveled, obtains for the 2-norbornyl

£-nitrobenzoates 1§2 - 155'L1^* P°lar rate retarding

anisyl nisyl

OPNB OPNB

152 153 4 « V *0 +0 iSrel 1 5

PNB PNB

anisyl anisyl 154 155 ^ ^ ^ ^ ^ ^ ^rel 1 4.5 135 effect of the double bond, leveled by the anisyl group, is now five.

The addition of another intervening carbon atom mediates this effect still more. Bly11^ found that the acetolysis rates of brosylates 156 - 158 also fell within ^ ^ ^ ^ ^ a very narrow range. Anti isomer 157 was observed to ^ ^ ^ solvolyze 2.4 times slower than its saturated analog 158 “ w " r due to Inductive destabilization while 156 was 2.3 times ^ ^ ^ slower. The inductive effect is, therefore, small and through space interaction in 156 is minimal. The major + * ^ product in each case was the acetate of retained structure.

OBs OBs

156 157 158

^rei^°l.1 1 2.4

If one assumes a destabilizing mechanism of similar magnitude to be operative in the tricyclo[4.2.0.0^ octyl skeleton (rate retardation by a factor of ca. 2) then slight rate enhancements due to through bond and through space Interactions are observed. The fact that 133 solvolyses at the same rate as its saturated counterpart 136

13^ (Table XI) implicates a through bond rate enhancing m m m m factor of approximately 2, operative in a direction inverse

to the inductive mechanism. Hypothesizing that the same

factor obtains for through bond interaction in 131» a + m through space factor for rate enhancement relative to 132 m m ** ca. 3.6 can be deduced.

Product studies reveal no participation of the ole-

finic or peripheral cyclobutane bonds with the reaction

center. The formation of thetbicyclo[4.2.Ojoctyl framework

(from 131. 132, and 13*0 can be readily rationalized in * * * * * * ** -* — terms of participation of the adjoining internal cyclo

butane bond as in 159. The resulting homallylic cyclo- ^ ^ ^

OAc + HOAc

hexenyl cation (160) can be trapped from either of two m m directions depending upon prevailing steric conditions.

Exo orientation expectedly predominates in all cases.

Bicyclo[2.2.0]hexyl tosylate ISO1-0 exhibits similar ^ r n m rnm behavior, yielding 4-cyclohexenyl acetate as 51Z of the product mixture. 137

OAc

Construction of the bicyclo[3.3•0]octane skeleton can result from either concommitant or subsequent 1,2- shifting of the second central bond in 159 or 160, respectively. Note that this process occurs only where syn geometry exists, and is likely related to the suit able orientation of the cyclobutane Walsh orbitals^!** in the second internal bond such that maximum bonding can be achieved at all times. Although the above fact may argue for synchronous migration of both internal cyclo butane bonds to give an ion such as l6l, it is not pos- -V sible to rule out a more circuitous route involving ring

161 162

opening to the 2 ,5-cyclooctadienyl cation 162 and subse quent transannular cyclization to give 161. ^ ^ 138

The seemingly most anomalous product observed is

142, but its formation from 132 can be rationalized by *“■ —'*** well precendented processes (Scheme VI). A Wagner-Meerwein

SCHEME VI

163 164

TsO

121 166

OTs

OAc 168 AcO 142

shift subsequent to formation of 163 or sychrononus with departure of the leaving group in 132 can lead directly to ^ 139

cation l6*f. A comparable 1,2 carbon shift has been -w »w previously observed in the case of 150^10 where exo- “v “m “ v bicyclo[2.2.l]hexyl tosylate has been isolated in *115S

yield as the product of internal return. Cyclobutyl

150 ^ ^ participation, a process known to be facile in lOS-1-1^. " m +0 can then give rise to cation 1§5* "Bridge-flipping"'1-^ could yield 166. The square-pyramidal ion 167, which has 4 V ^ “V *— —•* been directly observed by Masamune^^ from 121-Cl or ■* ^ l42-ci> might serve as an all-encompassing representation. A* ^ Ions 166 or 167 are known to trap acetic acid to give 1^2 as the exclusive product.

The only product which argues for through bond stabilization is the acetate of retained structure (144) ^ ^ from solvolysis of 133. If the cation resulting from 133 - w " w (i.e., 169) is stabilized by a through bond interaction ^ ^ (169a), a capture of solvent to return the starting frame- N “w ' r work might very well be expected. Participation of the two central cyclobutane bonds in 169 could account for the formation of ring opened product 1^5 * as the internal sigma orbitals are poorly aligned for 1 ,2-shifting to deliver the bicyclo[3 .3 .0 ]octyl skeleton (vide supra).

Thermal ring opening of 144, a process which was shown to occur under the reaction conditions, could account for * the isolation of 145, but this is viewed as less tenable 4 ^ ^ than direct formation from 169 since the 144:145 ratio was not observed to change significantly with time when the progress of reaction was monitored by vpc.

The apparent lack of through space interaction in

131 need not be attributed to disymmetric orientation of ^ ^ « the n-bond with the developing cationic center. Winstein has found11® that octahydrodimethanonaphthyl system

170 is 10 times more reactive than the anti-7-norbornenyl

OCOCF3

170 171 1*41

i derivatives in spite of such dissymmetry. In addition, the product observed (171) results from Ti-participation.

The mitigating factor in cation 172 is more likely the distance between the m-bond and the empty £-orbital.

Bly113 attributed the lack of anchimeric assistance in o 173 to a distance of 2.8A between the carbon bearing the empty orbital and those of the Tr-bond is cation 173-

OTs

172 173

MINDO calculations for hydrocarbon 95 show the distance between the two closest corresponding carbons to be 2.93&, probably very close to that in cation 172.

The inefficiency of through bond coupling in a frame work in which photoelectron spectroscopy suggests facile o-bond relay of electronic effects^^ ±s demonstrated by the anti-tricyclo[*t.2.1.02 »5]nona-3»7-diene system. The parent diene 17*4 exhibits through bond coupling in its photoelectron ^ ^ ^ spectrum in a manner comparable to 96. A chemical consequence of this effect^^ is the ready photolytic cleavage of <^-02 giving bisally1 radical 175* The 9-toluenesul- fonyl derivative (176)^^, however, solvolizes eight times

slower than syn-7-norbornenyl tosylate indicating no ac

celerating effect due to coupling of the reaction center

to the cyclobutenyl olefin vialC^_C2 > C^-Cg.

The experimental findings show that the interactions

of the ir-bonds in 131 and 133 with the developing cationic ^ ^ ^ ^ centers are much smaller than what might have been expected

on the basis of the large interactions between the v-

bonds in 95 and 96 as shown by photoelectron spectroscopy.

In the absence of pronounced effects through space and

through the a framework, development of a consistent

rationale of the facts becomes difficult. Nevertheless,

arguments using the (7-norbornyl)methyl (156 - 158) and

anti-trlcyclo[fr.2.1.0^*^Jnonadiene (17^ - 176) systems as models, provide precedent for the results observed and close analogy to the tricyclooctadiene system in terms of inter

atomic distances and suitability of high energy sigma bonds

for efficient relay of electronic effects. EXPERIMENTAL

The NMR spectra were obtained with Varian T-60,

Varian A-60A, and Bruker 90 (FT) spectrometers and ap

parent splittings are given in all cases. The Bruker 90

spectrometer was also employed for the recording of

13c spectra. Mass spectral measurements were made on an

AEI-MS9 spectrometer at an ionizing potential of 70 eV.

Preparative vpc work was done on a Varian Aerograph A90-

P3 instrument equipped with a thermal conductivity detector.

A 6 ft x 1/4 in. column of 1035 OV-11 on 60/80 mesh

Chromosorb G at 115° was used unless otherwise stated.

Microanalyses were performed by the Scandinavian

MIcroanalytical Laboratory, Herlev, Denmark.

syn-Trlcyclo[4.2.0.0^ ]oct-7-en-exo-3-ol (126). To a

magnetically stirred solution of

2.08 g (20 mmol) of syn-

tricyclo[4.2.0.0^'^Jocta-

126 3,7-diene in 50 ml of tetrahydro-

furan under nitrogen at 0° was added dropwise 45 ml of 0.5 N 9-borabicyclononane in

143 144 tetrahydrofuran. Stirring was continued for 2 hr, whereupon 15 ml of 1556 aqueous sodium hydroxide solution

followed by 15 ml of 30? aqueous hydrogen peroxide were added dropwise and the solution was allowed to warm to room temperature. The tetrahydrofuran was evaporated and the aqueous residue was extracted with ether (3 x 50 ml).

The combined ether layers were washed with water (75 ml) and brine (75 ml), dried, and concentrated to leave a yellow oil. Chromatography on silica gel (elution with 15? v ether-pentane) yielded a yellow solid which was recry stallized from pentane giving 0.67 g (27?) of white needles, mp 52-5°: 6.50 (m, 2H, vinyl), 4.18 (t, CDCI3 J « 6.0 i h , a to hydroxyl), 3-40 (m, 2H, allyl),

2.75 (m, 2H, methine), 2.39 (br s, exch. DjO, hydroxyl), and 2.00 - 2.40 (m, 2H, methylene); calcd for m/e

122.0732, found 122.0733.

The tosylate was prepared in the usual way, mp

61 - 62°, as was the 3 ,5-dinitrobenzoate, mp 127*5 -

128.5°.

Anal: Calcd for ci5Hi2N2°6: C* 56.96; H, 3*83;

N, 8 .8 6 . Pound: C, 57.10; H, 4.07; N, 8.64. 145

anti-Trlcyclo[4.2.0.02 joct-7-en-exo-3-ol (128). To

a magnetically stirred

solution of 2.08 g (20 mmol)

of anti-trlcyclo[4.2.Q.02 »5]~

octa-3,7-diene In 40 ml of dry tetrahydrofuran at 0° under nitrogen was added dropwlse

40 ml (20 mmol) of 0.5 N 9-borabicyclononane in tetrahydro furan over 1 hr. Stirring was continued for an additional hour, whereupon 15 ml of 15? aqueous sodium hydroxide solution followed by 15 ml of 30? aqueous hydrogen per oxide were added dropwise and the mixture was allowed to warm to room temperature. The tetrahydrofuran was evap orated and the aqueous residue was extracted with ether

(3 x 50 ml). The combined ether layers were washed with water (2 x 100 ml) and brine (100 ml), dried, and concen trated to leave a yellow oil. Chromatography on silica gel (elution with 15? ether-pentane) yielded 500 mg (25?) of the alcohol as a clear oil: 6.32 (nm, 2H, CDCI3 * vinyl), 4.21 (t, J=6.2 Hz, 1H, a to hydroxyl), 3.13 (nm,

2H, allyl), 2.40 (s, 1H, exch. DgO, hydroxyl), and 2.36

(m, 4h, methine and methylene); calcd for m/e 122.0732, found 122.0733. 146

The tosylate was prepared in the usual way, mp 62-

63°» as was the 3>5-dinitrobenzoate, mp 151.5 - 153°.

Anal. Calcd for 56.96; H, 3*82;

N, 8.86. Found: C, 57.00; H, 3*90; N, 8.88. syn-Trlcyclo[4 .2.0.02 * ^]octan-exo-3-ol (127). To a

magnetically stirred solution

OH of 1.8 g (14.8 mmol) of exo-

alcohol 126 in 150 ml methanol 127 under nitrogen at 0°C was added

19 g (100 mmol) of dipotassium azodicarboxylate in one portion. To the stirred slurry was added 15 ml of acetic acid via syringe and stirring was continued until the yellow color of the potassium salt was discharged.

The methanol was extracted with ether (3 x 50 ml) and the combined organic layers were washed with saturated sodium bicarbonate solution (2 x 100 ml), water (100 ml), and brine (100 ml), before drying over magnesium sulfate.

The solution was filtered and the solvent evaporated to leave 1.35 g (85%) of the saturated alcohol as a clear oil: ^£^5^3 4.98 (t, J=6.0 Hz, 1H, a to hydroxyl),

2.93 (m, 4H, methylene), 2.66 (s, 1H, hydroxyl), and

2.36 (m, 6h , methine); calcd for C g H ^ O m/e 124,0888, 147

found 124.0890.

The tosylate was prepared in the usual way, mp 44-

45°, as was the 3 ,5-dinitrobenzoate, mp 144.0 - 144.5°C.

Anal. Calcd for C15H12N 206 : C, 56.60; H, 4.43;

N, 8.86, found C, 56.57; H, 4.50; N, 8.71.

antl-Trlcyclo[4. 2 . 0 . 0 ^ 3octan-exo-3-ol (129). To a

magnetically stirred solution

of 1.1 g (9.0 mmol) of 128 in ^ . V 100 ml of methanol at 0° under

nitrogen was added 13-9 g

(72 mmol) of dipotassium azo- 129 dicarboxylate In one portion.

Subsequently, 9.7 ml of acetic acid was added slowly over 1 hr and stirring was con tinued until the yellow color was discharged. The methanol was evaporated, the residue was dissolved in

100 ml water, and the aqueous phase was extracted with ether

(3 x 50 ml). The combined ether layers were washed with sodium bicarbonate solution (2 x 75 ml) and brine (75 ml), dried, and concentrated to leave 590 mg (53%) of the alcohol as a clear oil: 6CD^l3 4.26 (t, J=6.0 Hz, a to hydroxyl), TMS “ 2.88 - 1.83 (m, 4H, methine), 2.65 (m, 6h , methylene), and 2.01 (s, 1H, exch. D20, hydroxyl); calcd for m/e 12*1.0888, found 124.0890.

The tosylate was prepared In the usual way and used directly as a clear oil. syn-Trlcyclo[4.2.0.0^ 3oct-7-en-endo-3-ol (130). To

a magnetically stirred solution

of 218 mg (1.64 mmol) of N-chloro-

succinimide in 10 ml of methylene

chloride cooled to 0° under

nitrogen was added 102 mg (1.68

mmol) of dimethyl sulfide via syringe. A white precipitate formed and the mixture was cooled to -23°. The alcohol 126 (100 mg, 0.82 mmol) in 3 ml of methylene chloride was added in one portion and stirring was continued for 2 hr, whereupon 84 mg

(0.84 mmol) of triethylamine was introduced. The mixture was allowed to warm to room temperature and poured into ether (15 ml) and water (50 ml). The aqueous phase was separated and the organic layer was washed with water

(2 x 25 ml), dried, concentrated and transferred to a

25 ml round bottomed flask equipped with a stirring bar, condenser, and nitrogen inlet tube. Lithium aluminum hydride (50 mg, 1.3 mmol) was introduced and the mixture stirred at room temperature for 1 hr. Saturated sodium sulfate solution was added dropwise until the supernatant

liquid became clear and the precipitated aluminum salts were filtered and washed amply with ether. The filtrate was dried and concentrated to leave 30 mg (30JS) of alcohol

as a clear oil: 6.62 (nm, 1H, vinyl), 6.40 (nm, TMS 1H, vinyl), 4.16 (m, 1H, a to hydroxyl), 6.52 (br d,

2H, allyl), 2.56 (m, 2H), and 2.17 (m, 2H).

Acetolysis Kinetics

A. Reagents. Anhydrous acetic acid was prepared by re- —* -*v ■“» •• fluxing a solution of acetic anhydride in glacial acetic acid for 24 hrs and subsequent fractional distillation in

« dr-y atmosphere. Perchloric acid (ca. 7050 was standardi

zed against aqueous sodium hydroxide solution using phenolph- thalein as indicator. Standard perchloric acid (ca. 0.021

M) used in titrating acetolysis aliquots was prepared by the addition of an accurately weighed amount of 7056 per chloric acid to a known volume of anhydrous acetic acid.

Standard sodium acetate in acetic acid was prepared by dissolving anhydrous sodium carbonate in acetic acid; the water of neutralization was not removed. The result ing solution was standardized against perchloric acid using bromphenol blue as the indicator; the color change is from yellow to colorless. The automatic pipettes were calibrated with glacial acetic acid. 150

B, Kinetics Procedure. A ca. 0.02 M solution of tosylate

(accurately weighed) in 0,0510 N sodium acetate in acetic acid was prepared in a 25 ml volumetric flask. Aliquots of this solution were removed and sealed in glass ampoules which had been washed sequentially with 10? hydrochloric acid, 10? ammonium hydroxide, and water, and dried at

70° overnight. The ampoules were placed in a constant temperature bath and after 10 min the first ampoule was re moved and quenched in ice water. At this point an accurate timer was started. The ampoule then was allowed to warm to room temperature for ca. 5 min and exactly 1.985 ml of solution was removed via the automatic pipette and titrated with standard perchloric acid in acetic acid. Three drops of bromphenol blue was used as indicator and the endpoint was considered to be reached when the yellow solution turned clear. The remaining ampoules were removed at approp riately timed intervals and treated as above. Points were taken through 1-2 half-lives, and an infinity titer after 10 half-lives.

C. Determination of a Solvolytic Rate Constant. The raw data from a kinetics experiment consisted of volumes of perchloric acid (Vt ) used to titrate the residual sodium acetate in a given ampoule and the elapsed times at which those ampoules were quenched, including a reading at 151

zero-time (V0) and one at 10 half-lives (V„). The fraction

of unreacted tosylate at any given time (Ft) is given by: Ft = V* --V» VD - Vw

For the first order reactions (i.e., solvolysis) the

natural logarithm of Ft is a linear function of time t^

with slope -k (the first order rate constant) and inter

cept zero. The method of least squares was used to obtain

the best slope (-k) of In F^ vs t^ The error included with

the rate constants is an average deviation for two runs.

The half life is determined from the relationship

t^/2 = (0.693/k).

Rate constants for all tosylates obtained in this way ap pear in Table XI. A sample run Is shown in Table XII.

D. Calculation of Activation Parameters. The familiar expression for the first-order rate constant from transition state theory is

k = RTeAS*/Re-AH*/RT (1) Nh where k = first order rate constant at temperature T(k) r = gas constant N ° Avogadro's constant h = Plank's constant AS*= entropy of activation enthalphy of activation 152

Table XII. Example of a Solvolytic Rate Calculation

(1 ) (2 ) (3) (4) Time(secxlO-3) HCIO4 (ml) Ft -in Ft

0 4.730 - —

7.77 4.659 .9582 .0426

19.^9 4 .550 .8941 .1119

27.63 4 .550 .8588 .1522

43.47 4.373 .7900 .2357

83.22 4.125 .6441 .4398

68.23 4 .229 .7053 .3491

126.2 3.942 .5365 .6227

177.9 3.700 .3941 .9311

CO 3.03 — -

For 135 at 68.2° In 0.0510 N Na02CCH3. The correlation co- efficient was 0.999 and k = 0 .509±0.01x10“5 via the method of least squares.

\pOTs H0Ac )

133 ^ V v 153

Eqn. (1) rearranges to

R[ln(k/T)-ln (R/Nh) = AH# (1/T) + AS* (2) which Is a linear relationship between the left side of the equation and (1/T) with slope -AH# and intercept AS#.

A program (ACTIPAR-CAL (X)) written for the Wang 360 calculator by Dr. Ian R. Dunkin converts reaction temp eratures In °C and rate constants Into pairs of dummy variables XT and Yrp:

XT - R (k/T) - lp (R/Nh)

YT “ 1000/T

The Xrp and YT values for several temperatures are treated by a least-squares sub-program to give AH* and AS# directly. All activation parameters obtained in this work result from this program. 154

Preparative Scale Solvolysis of 131* A magnetically stirred

solution of 1.0 g (3*63 mmol) of

131 and 530 mg (5*0 mmol) of ^ “ m sodium carbonate in 25 ml of 131 acetic acid was heated at 90° for 18 hr (10 half-lives). The solution was cooled, poured into 100 ml of water and extracted with ether (3 x

30 ml). The combined ether layers were washed with 50 ml portions of 103! aqueous sodium hydroxide solution (2x), V water, and saturated aqueous sodium chloride solution, be fore decolorization with charcoal, filtration through

Celite, drying and evaporation to leave 560 mg (95%) of a clear oil. Analysis by vpc showed four components, two of which could be preparatively separated.

The first compound to be eluted was identified as

136 (6.5%): 6CDC13 6.34 (nm, 1H, cyclobutenyl), 6.14 --- t m s (nm, 1H, cyclobutenyl), 6.03 (d, J=4.5 Hz, 2H, cyclohexenyl),

5.55 (nm, 1H, a to acetate), 3 .00 (AA'BB", 2H, bridgehead),

2.38 (t with fine splitting, J=8.0 Hz, 1H, allyl), and

2.06 (s, 3H, acetate).

The third component was 135 (85%) as shown by comparison of its pmr data to those published^-03: ^ M S * ^ ^.05

(dd, J=2.0 and 5.5 Hz, 1H, vinyl), 5.50-5.80 (m, 3H, vinyl), 5.40 (m, 1H, a to acetate), 3.65 (m, 1H, bridge- 155

head), 3.30 (m, 1H, bridgehead), and 2.00 (s, 3H,

acetate); calcd for CiQ^i2^>2 1^4.0837, found 164.

0839.

Two other components (3% and 6jE) could not be separated

in a pure state due to coincidental elution with 135.

Preparative Scale Acetolysis of 132. A solution of 1.25

g (*1.5 mmol) of 132 and 500 mg ^ ^ ^ (4.7 mmol) of sodium carbonate

132 in 25 ml of acetic acid under

argon was heated at 90° for

63 hr (10 half-lives). The usual workup yielded 500 mg

(685E) of a clear oil exhibiting 5 peaks on vpc which were preparatively separated. The first two compounds to elute were collected together and shown to be 138

(21 JO and 139 (11?). The third and fourth components + * were likewise collected together and identified as 140 “ m ^ (11%) and 141 (5?).

The final compound was 142 (52%): XH NMR 6CDCI3 ■*" TMS 4.97 (d of t, 1H, J=6.46 and 9.00 Hz,a to acetate),

2.67-1.12 (series of m, 10H, methine and methylene), and 2.00 (s, 3H, acetate); 13C NMR (CDCI3 ) 1 7 0 .3 6 , 79.82,

44.23, 31.29, 27.08, 26.69, 26.17, 23.71, 21.06, and

19.31 ppm. 156

Preparative Scale Acetolysis of 133- A magnetically

stirred solution of 600 mg (2.18

■Ts mmol) of 133 and 500 mg (4.73 mmol) of sodium carbonate in

133 25 ml of acetic acid was stirred

at 90° under argon for 26 hr (10 half-lives). The usual

workup yielded 250 mg (69%) of a clear oil containing

four components by vpc. The two major compounds could

be separated. The first was identified as 144 (23%) t by comparison of -*-H NMR and vpc retention times with those

of an authentic sample. The second component was shown

to be 145 (65%) by comparison of its NMR data with ^ * * those published10^: 5.95 (m, 2H, vinyl), 5-tJU TMS (m, 3H, vinyl), 5.15 (m, 1H, a to acetate), 3.15-3*00

(m, 1H, bridgehead), and 2,50 (m, 2H methylene); calcd

for C10H12°2 m/e 164.0837, found 164.0839.

The two remaining components (5% and 7%) eluted simultaneously and could not be obtained in a state pure enough for identification.

Preparative Scale Acetolysis of 134. To a solution of

1.1 g (4.0 mmol) of 134 in

10 ml of acetic acid under argon

was added 424 mg (4,0 mmol) of

sodium carbonate and the 157 solution was stirred at 90° for 30 hr (10 half-lives).

The reaction mixture was processed as before to leave

500 mg <58JE) of a clear oil which was purified by pre parative vpc. There was isolated a mixture of 140 ^ (602) and 141 (34JS). ^ ^ endo-2-Acetoxybicyclo[4.2.0]octa-4,7-diene (137). To a

solution of 200 mg (1.6*1 mmol)

of bicyclo[4 .2 .0]octa-4 ,7-

dlerl-2-one in 25 ml of.methanol 137 at -20° under argon was added

2148 mg (6.5 mmol) of sodium borohydride in two 124 mg portions at a 10 min interval. The solution was allowed to warm to room temperature, stirred for 5 hr, and poured into 150 ml water. The aqueous solution was extracted with ether (3 x 50 ml) and the combined ether layers were washed with water (2 x 75 ml) and saturated sodium chloride solution (75 ml) before drying and evaporation. The resulting yellow oil was dissolved in 5 ml of pyridine

1.4 g (13-8 mmol) of acetic anhydride was introduced, and the solution was stirred at room temperature for

24 hr. The solution was poured into 50 ml water and the water layer was extracted with ether (3 x 30 ml). The combined organic layers were washed with 50 ml portions of water, 10J5 aqueous hydrochloric acid (2x), saturated 158

aqueous sodium bicarbonate solution, and saturated aqueous

sodium chloride solution prior to decolonization with charcoal, filtration through Celite, drying, and con centration. There remained 238 mg (852) of the acetate as a clear oil: 6CDCl3 6.23 (d, J=3.0 Hz, 1H, cyclo- TMS - butenyl), 5.70 Cm, 2H, cyclohexenyl), 4.86 (d of t, J=

7.0 and 10.0 Hz, 1H, a to acetate), 3.58 (m, 2H, bridge head), 2.82 (m, 2H, allyl), and 2.07 (s, 3H, acetate); calcd for C^oHi2°2 !!l/® 164.0837, found 164.0839*

Control Experiments Concerned with the Stability of

Acetates 138, 139, 140 and l4i, A 50 mg (0.09 mmol) sample containing 66% l4n pnd 342 ^4l was d1s — solved in 2 ml of acetic acid and the solution heated at 90° under nitrogen for 48 hr. The solution was poured into 25 ml of water and extracted with ether (3 x 10 ml).

The combined ether layers were washed with 15 ml portions of 152 aqueous sodium hydroxide solution (2x), water, and hrine. Drying and evaporation left 30 mg (66j£) of a mixture of 140 and l4l( the 1H NMR of which was « « « « « ^ identical to that of the starting sample.

The identical treatment of 30 mg (0.18 mmol) of a mixture of 138 and 140 (66:34) yielded after workup

25 mg (832) of the unchanged starting mixture. 159

Control Experiments Concerned with the Stability of Ace- tates 144 and 147. To a solution of 82 mg (0.2 mmol)

of 147 in 2 ml of acetic acid ^ under argon was added 2.2 mg

147 (0.2 mmol) of sodium carbonate.

This solution was stirred at

Ac 90° for 18 hr (10 half-lives).

The usual workup afforded 80 mg

C 9856 ) of a mixture of starting acetate (5556 ) and 145 t ** * * " v (45?) by integration of the vinylic signals in the pmr.

Identical treatment of 144 (100 mg, 0.55 mmol) * * * - * * # yielded 70 mg (70?) of 145. -*r *»«

Control Experiments Concerned with the Stability of

Acetates 148 and 1*19. To 100 mg (0.6 mmol) of 148 in

5 ml of acetic acid was added OAc 63.6 mg (0.6 mmol) of sodium

carbonate and this solution was

heated for 48 hr at 90° under OAc argon. Upon workup, 85 mg (85?)

of starting acetate was obtained.

Identical treatment of 149 (200 mg, 1.2 mmol) yielded 160 mg (80?) of unchanged acetate. APPENDIX

1H NMR spectra of 100a, 101a, 102a,b, and 103a, i cr i « t > 102a

* MO I I*' *'

103a ^ ^

> V H

8 7 6 5 H 3 2 1 0 162

o o mX) f * xi t oj I o i iH I

t—

oo 09

CT\

ft-* .■ o I o fp>

rv»

E9T REFERENCES

1. A, Kekul£, Bull. Soc. Chim. France, 3* 98 (1865).

2. E. Hiickel, Z. Phyzlk, 70, 204 (1931); 78, 628 (1932).

3. R. Wilstatter and E. Waser, Chem. Ber. , 44, 3423 (1911).

4. T.L. Burkoth and E.E. van Taraelen in "Nonbenzenold Aromatics," Vol. I, p. 63ff (1969).

5. S. Masamune and N. Darby, Accounts Chem. Res., 5, 272 (1972).

6. K. Mislow, J. Chem. Phys. , 20, 1489 (1952).

7. N.L. Alllnger, T.J. Walter, and M.G. Newton, J. Amer. Chem. Soc., 96, 4588 (1974).

8. E. Vogel, M c Btsknp. W. Pretzer, and W.A. Boll, Angew. Chem. Intern. Ed. Engl., 3» 642 (1964).

9. F. Sondheimer and A. Shani, J. Amer. Chem. Soc., 86, 3168 (1964); 89, 6310 (195TT

10. E. Vogel, Special Fubl. No. 21, p. 122, The Chemical Society, London (19675 -

11. H. Gunther, H. Schmickler, H. Koningshosen, K. Recker, and E. Vogel, Angew. Chem. Intern. Ed. Engl., 12, 243 (1973).

12. T.J. Katz, J. Amer. Chem. Soc., 82, 3784 (i960).

13. E.A. LaLancette and R.E. Benson, J. Amer. Chem. Soc., 85, 2853 (1963).

14. R.B. Woodward and R. Hoffmann, "The Conservation of Orbital Symmetry", Verlag Chemie, Academic Press (1970).

15. E.E. Van Tamelen and B.C.T. Pappas, J. Amer. Chem Soc., 85, 3296 (1963).

164 165

16. W. von E. Doering and J.W. Rosenthal, J. Amer. Chem. Soc., 4534 (1967).

17. M. Jones, Jr., Ibid. , 8$, 4236 (1969).

18. T.L. Burkoth, Ph.D. Thesis, Stanford University, 1967.

19. J. Masamune, C.G. Chin, R-Hojo, and R.T. Seldner, J. Amer. Chem. Soc., §9, 4804 (1967).

20. S. Masamune, K. Hojo, Kiyomi Hojo, G. Bigam, and D.L. Rabenstein, J. Amer. Chem. Soc., 93* 4966 (1971).

21. E. Vogel, W. Meckel, and W. Grimme, Angew. Chem., 7.6. 786 (1964); Angew. Chem. Intern. Ed. Engl., 3, 643 (1964).

22. H.E. Zimmerman and G.L. Grunewald, J. Amer. Chem. Soc., §6, 1434 (1964).

23. L.A. Paquette, J. Amer. Chem. Soc., 93. 7110 (1971) .

24. (a) K. Alder and K.H. Backendorf, Chem. Ber. , 2199 (1938); (b) L.A. Paquette, R.E. Wingard, Jr., J.C. Philips, G.L. Thompson, L.K. Read, and J. Clardy, J. Amer. Chem. Soc., 4508 (1971).

25. L. Friedman, D.L. Fishel, and H. Shechter, J. Org. Chem., $0, 1453 (1965).

26. W. Meckel, Ph.D. Dissertation, University of Koln.

27. S. Masamune, R.T. Seidner, H. Zenda, M. Wiesel, N. Nakatsuka, and G. Bigam, J. Amer. Chem. Soc., 90, 5286 (1968).

28. We are aware of one previous assignment of stereo chemistry to a diester in the 9,10-dihydronaphthalene series solely on the basis of electronic spectra [H.-D Martin and M. Hekman, Angew. Chem. , 8jj, 615 (1973); Angew. Chem. Intern. Ed. Engl., 12, 572 (1973)3. These" researchers took the uv maximum 166

of 1 [285nm (e 2600)in acetonitrile] to define the XO '"2”3 H 3C02C trans-fusea character of the two six-membered rings.

29. A generous gift of this compound was made to us by Prof. M.S. Newman of this Department.

30. These diesters were prepared by esterification of the structurally related anhydrides [T.M. Lyssy, J. Org. Chem., 27, 5 (1962)].

31. R. Cookson, J. Hudec, and M. Marsden, Chem. Ind., (London), 21 (1961).

32. Prepared by methoxide-promoted isomerization of J2. This product was, however, not fully characterized.

33. W.B. Scott, Ph.D. Thesis, University of British,Col umbia, 1965.

3*1, (a) E. Vogel, W. Maier, and J. Eimer, Tetrahedron Lett., 655 (1966); (b) J.J. Bloomfield and J.R.S. Irelan, ibid. , 2971 (1966).

35. H.M. Frey and R. Walsh, Chem. Rev., §£, 103 (1969).

36. H.-D Martin and M. Hekman, Chimia, 28, 12 (197*0.

37. T. Miyashi, M. Nitta, and T. Mukai, J. Amer. Chem. Soc., 93, 3****1 (1971); J .E. Baldwin and M.S. Kaplan, ibid., 93, 3969 (1971). A p A / 38. J.J. Gajewski, Mech. Molecular Migrations, *}, 1 (1971).

39. M.J. Kukla, Ph.D. Thesis, The Ohio State University, 197**. A Cl,7] vinyl shift has recently been reported: E.E. Waali and W.M. Jones, J. Org. Chem.. 38, 2573 (1973).

*10. P. Schiess and P. Fiinfschllling, Tetrahedron Lett., 5191 (1972). 167

41-, L.L. Miller, R. Greisinger, and R.P. Boyer, J. Amer. Chem. Soc., 91* 1578 (1969). “ “— 1 " 42. R. Huisgen, F, Mietzsch, G. Boche, and H. Seidle, Chem. Soc. Spec. Publ. , No. 19, 3 (1965).

43. G. Scholes, C.R. Graham, and M. Brookhart, J. Amer. Chem. Soc., 9 6 , 5665 (1974).

44. C.W. Spangler, T.P. Jondahl, and B. Spangler, J . Org. Chem., 38,* 2478 (1973)-

45. J.M. Photis, Ph.D. Thesis, The Ohio State University, 1975. 46. L.A. Paquette, M.R. Short, and J.F. Kelly, J. Amer. Chem. Soc., 93, 7179 (1971). ■ ■■ ■— ,VM # 47. E.F. Bradbrook and R.P. Linstead, J . Chem. Soc., 1739 (1936).

48. J.C. Philips, Ph.D. Thesis, The Ohio State University, 196 8.

49. At 10 0 Mb t the splitting pattern exhibited by the olefinic protons is identical to that of a-pyronene

' (ii, 220-MHz spectrum): S.G. Traynor, Ph.D. Thesis, Trinity College, Dublin, 1974.

50. R.B. Woodward and R. Hoffmann, Angew. Chem., 8l, 781 (1969).

51. R. Srinivasan and J.N.C. Hsu, Chem. Comm., 1213 (1972).

52. L.A. Paquette and M.J. Kukla, Tetrahedron Lett., 1241 (1973).

53. For reviews, (a) H.E. O ’Neal and S.W. Benson, Int. J. Chem. Klnet. , J, 221 (1969); 2, 423 (1970); (b) S.W. Benson and H.E. O ’Neal, Nat. Stand. Ref. Ser., Nat. Bur. Stand., No. 21, 14 TWfOT*

54. C. Steel, R. Zand, P. Hurwitz, and S.G. Cohen, J. Amer. Chem. Soc., 89, 679 (1964). 1 ™ 1 'V 168

55*. (a) M.J. Goldstein and M.S. Benzon, J. Amer. Chem. Soc., 9Ji, 5119 (1972); (b) A. Sinnema, F. Van Rantwijk, A.J. deKoning, A.M. Van Wijk, and H. van Bekkum, JCS Chem. Commun., 364 (1973).

56. L.A. Paquette and J.A. Schwartz, J. Amer. Chem. Soc., 12, 3215 (1970).

57. W.R. Roth and M. Martin, Tetrahedron Lett., 3865 (1967).

58. M. Avram, I.G. Dinulescu, E. Marica, G. Mateescu, E. Sliam. and C.D. Nenitzescu, Chem. Ber. , 97, 382 (1964).

59. E. Vogel, 0. Roos, and K.-H. Disch, Justus Liebigs Ann. Chem., 653* 63 (1962). I ■ * S ' - HV# 60. D. Bellus, H.-C Mez, G. Riks, and H. Sauter, J. Amer. Chem. Soc. , 96, 5007 (1974).

61. H.-D. Martin and E. Eisenmann, Tetrahedron Lett.. 661 (1975). 62. E.L. Allred and J.C. Hinshaw, Tetrahedron Lett., 387 (1972).

6 3 . R.S. Case, M.J.S. Dewar, S. Kirschner, R. Pettit, and W. Sleglr , J. Amer. Chem. Soc., 96 , 7581 (1974).

64. H.M. Frey, H.-D. Martin, and M. Hekman, J.C.S. Chem. Comm., 204 (1975).

6 5 . (a) J.P. Chesick, J. Amer. Chem. Soc., &4,» 3250 (1962); (b) W.R. Roth and M. Martin, Justus Liebigs Ann. Chem., J02, 1 (1967); (c) M.G. Waite, G.A. Sim, C.R. 01ander,~R.J. Warnet, and D.M.S. Wheeler, J. Amer. Chem. Soc. , %l, 7763 (1969); (d) J.J. Tufariello and A.C. Bayer, Tetrahedron Lett., 3551 (1972).

66. (a) K.B. Wiberg and D.E. Barth, J. Amer. Chem. Soc., 91, 5124 (1969); (b) E. Block, H.W. Orf, and R.E.K. Winter, Tetrahedron, 2§, 4483 (1972).

6 7 . J.G. Traynham, J.S. Dehn. and E.E. Green, J. Org. Chem., 33, 2587 (1968).

68. E.L. Allred and R.L. Smith, J. Amer. Chem. Soc., 91, 6766 (1969). 169

6 9 . P.A. Bovey, "Nuclear Magnetic Resonance/' Academic Press, New York, N.Y. 1969.

70. L.A. Paquette, A.A. Youssef, and M.L. Wise, J . Amer. Chem. Soc., 94, 997 (1972); (b) K.B. Wiberg, Adv. Alicyclic Chem., 2, 185 (1968); (c) M. Christl, Chem. Ber., 108,, 2781 (1975).

71. For leading references, consult: (a) R.D. Bertrand, D.M. Grant, E.L. Allred, J.C. Hinshaw, and A.B. Strong, J. Amer. Chem. Soc., 9*j, 997 (1972); (b) K.B. Wiberg, Acfv\ Alicyclic Chem., 2, 185 (1968); (c) M. Christl, Chem. Ber., 108,' 2 T 8 l Tl975).

72. G. Strukul, P. Vigline, R. Ros, and M. Graziani, J. Organometal. Chem., 7fj, 307 (197*0.

73. This ratio has previously been realized in the direct thermal rearrangement of cls-bicyclo[6.1.Ojnonatrlene at 90°: E. Vogel, W. Wiedemann, H. Kiefer, and W.P. Harrison, Tetrahedron Lett., 673 (1973).

7*0 cis-1,5-Bishomocyclooctatetraene has been prepared by the direct ^ycT opropanatl nn of ovol ooctatetraene with diazomethane and cuprous chloride: W.W. Frey, Dissertation, Universitat zu Koln, 1966. When pyrolyzed in a sealed tube at the much higher temperature of 250° for 50-75 min, this tricyclic diene was partially converted to iii, iv (probable stereochemistry indicated), and two unidentified products. We are grateful to Professor Vogel for making a copy of the Frey dis-

sertation available to us

75. L.A. Paquette and S.E. Wilson, J. Org. Chem., 37, 38*19 (1972).

76. (a) A.G. Anastassiou and E. Reichmanis, J. Qrg. Chem., 37, 38*19 (1972).

77. H. Tanida, T. Teratake, Y. Harata and M. Watanabe, Tetrahedron Lett.. 53*11 (1969). 170

7 8 ’. G. Blnsch and D.A. Kleier, Program 1*40, QCPE, Indiana University (1969),

79. D.A. Kleier and G. Binsch, Program 165, QCPE, Indiana University (1969).

80. C.L. Cunningham, Jr., A.W. Boyd, R.J. Meyers, W.D. Gurln. and W.I. Levan, J. Chem. Phys., 19, 676 (1951). ------

81. 0. Bastiansen, F.N. Fritsch, and K. Hedberg, Acta Cryst., 17, 538 (1964); R.M. Schwendeman, G.D. Jacobs, and T.M.“Krigas, J . Chem. Phys. , 40,, 1022 (1964).

82. J.C. Pommelet and J. Chuche, Tetrahedron Lett.. 3897 (1974); W, Grimme and K. Seel, Angew. Chem. Intern. Ed. Engl., 12, 507 (1973).

8 3 . W. von E. Doering and W.R. Roth, Angew. Chem. Intern. Ed. Engl., 12, 507 (1973).

84. R. Pettit and J. Henery, Org. Synth., 50, 36 (1970).

8 5 . S. Sawada and Y. Inoue, Bull. Chem. Soc. Japan, 42, 2b6y (1969).

86. R. Hoffmann, Acc. Chem. Res. , 4,, 1 (1971).

87. P. Bishof, J.A. Hashmall, E. Heilbronner, and V. Hornung, Helv. Chim. Acta., 52, 1745 (1969).

88. S.J. Cristol and R.L. Snell, J. Amer. Chem. Soc., 76, 5000 (1954).

89. S. Wlnstein, Quart. Rev., 23, 141 (1969) -

90. H. Tanida, T. Tsuji, and T. Irie, J. Amer. Chem. Soc., 82, 1953 (1967), M.A. Battiste, C.L. Deyrup, R.E. Pincock, and J. Haywood-Farmer, ibid., 8 9 , 1954 (1967). "

91. R. Hoffmann, A. Imamura, and W.J. Hehre, J. Amer. Chem. Soc. , 90, 1499 (1968). * v *v # 92. P. Bishof, J.A. Hashmall, E. Heilbronner, and V. Hornung, Tetrahedron Lett., 4025 (1969). 171

93* P. Brenneisen, C.A. Grob, R.A. Jackson, M. Ohta, Helv. Chlm. Acta., 48, 146 (1965).

94. H, Iwamura, H. Klhara, K. Morio, T.L. Kunli, Bull. Chem. Soc. Jap. 46, 3248 (1973).

95. R. Gleiter, E. Hielbronner, M. Hekman, and H.-D. Martin, Chem. Ber. , 10,6,, 28 (1973).

96. R. Criegee, Angew. Chem., 74,, 703 (1962).

97. L. Cassar, P.G. Eaton, and J. Halpern, J. Amer. Chem. Soc. 92, 3515 (1970).

98. E.F. Knights and H.C. Brown, J. Amer. Chem. Soc., 90, 5280 (1968).

99. E.J. Corey and C.U. Kim, J. Org. Chem., 3 8 , 1233 (1973).

100. E.E. van Tamelen, R.S. Dewey, and R.J. Timmons, J. Amer. Chem. Soc,, 8 3 , 3725 (1961).

101. L.F. Fieser and M. Fieser, "Reagents for Organic Synthesis", Vol. 1 , p. 118, Wiley (1967).

102. D.J. Pasto and C.R. Johnson, "Organic Structure Determination", p.'36l, Prentice-Hall (1969).

103. R.M. Moriarity and C.-L. Yeh, Tetrahedron Lett., 385 (1972).

104. G. Buchi and E.M. Burgess, J. Amer. Chem. Soc., 84, 3104 (1962).

105. J.S. Haywood-Farmer and R.E. Pincock, J. Amer. Chem. Soc. , gl, 3020 (1969).

106. L.A. Paquette, 0. Cox, M. Oku and R.P. Henzel, J. Amer. Chem. Soc., §6, 4892 (1974).

107. M. Kroner, Chem. Ber., 10,0, 3172 (1967) -

108. R. Huisgen, G. Boche, A. Dahmen, and W. Hechtle, Tetrahedron Lett., 5215 (1968). We thank Professor Huisgen for a copy of the pmr spectrum. 172

109. H.C. Brown, S. Krishnamurthy, and N.M. Yoon, J. Org. Chem., 41, 1778 (1976).

110. R.N. McDonald and C.G. Reineke, J. Org. Chem., 32 1878 (1967).

111. C.F. Wilcox, Jr. and H.D. Banks, J. Amer. Chem. Soc., £4, 8231 (1972).

112. E.N. Peters and H.C. Brown, J. Amer. Chem. Soc., 94, 5899 (1972).

113. R.K. Bly and R.S. Bly, J . Org. Chem. , 31^, 1577 (1966).

114. C.A. Coulson and T.H. Goodwin, J . Chem. Soc. , 3161 (1963).

115. M. Sakai, A. Diaz, and s: Winstein, J. Amer. Chem. Soc. . 92, 4452 (1970). The rate of 168 is enhanced 10**'3 Bimes relative to 7-norbornyl tosylate.

116. S. Masamune, M. Sakai, A.V. Kemp-Jones, H. Ona, A. Venot, and T. Nakashima, Angew. Chem. Intern. Ed . Engl. , 12, 769 (1973).

117. S. Masamune, M. Sakai, and A.V. Kemp-Jones, Can. J. Chem., 52, 858 (1974).

118. P. Bruck, D. Thompson, and S. Winstein, Chem. Ind. (London), 590 (I960).

119. E. Haselbach and W. Eberbach, Helv. Chim. Acta, 5.6 1944 (1973); H.-D. Martin, S. Kagabu, and R. Schweisinger, Chem. Ber., 107, 3130 (1975).

120. R.M. Coates and K. Yano, J. Amer. Chem. Soc., 95, 2203 (1973).