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 will 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.

University Microfilms International 300 North Zeeb Road Ann Arbor, Michigan 48106 USA St. John's Road. Tyler's Green High Wycombe, Bucks, England HP10 8HR I I 78-5830

DETTY, Michael Ray, 1951- CON JUGAT I VE AND HOMOCONJUGATIVE INTERACTIONS OF CYCLOPROPANE RINGS IN NEUTRAL AND CATIONIC SYSTEMS.

The Ohio State University, Ph.D., 1977 Chemistry, organic

University Microfilms International, Ann Arbor, Michigan 48106 CONJUGATIVE AND HOMOCONJUGATIVE INTERACTIONS OF

CYCLOPROPANE RINGS IN NEUTRAL AND CATIONIC SYSTEMS

DISSERTATION

Presented in Partial Fulfillment of the Reauirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Michael Ray Detty, B. Sc.

* * * * *

The Ohio State University

1977

Reading Committee: Approved by

Prof. Leo A. Paguette Prof. John S. Swenton Prof. Gary G. Christoph ACKNOWLEDGEMENTS

It has been an honor to have Professor Leo A. Paquette as an adviser. His enthusiasm and knowledge of chemistry have been inspirational. His guidance has been active when needed, but more subdued when the progression of research has not required it.

Importantly, his day to day interest in my research was unceasing and was greatly appreciated. VITA

January 23, 1951 ...... Born - Springfield, Ohio

1972 ...... B. Sc., Bowling Green State

University, Bowling Green, Ohio

1972-1973 ...... Chemist, General Latex and Chemical

Corporation, Ashland, Ohio

1973-1974 ...... University Graduate Fellow, The Ohio

State University, Columbus, Ohio

1974-1976 ...... Research Assistant, The Ohio State

University, Columbus, Ohio

1976-1977 ...... University Dissertation Fellow, The

Ohio State University, Columbus, Ohio

PUBLICATIONS

Michael R. Detty and Leo A. Paquette,"Stereocontrolled Synthesis, Con­

formational Features, and Response to Thermal Activation of the

Seven Possible Bis- and Trishomocycloheptatrienes," J. Am. Chem.

Soc., 99, 821 (1977).

Leo A. Paquette and Michael R. Detty, ” Stereoisomeric Bishomo-3,5-

cycloheptadienyl £-Toluenesulfonates as Probes of the Geometric

and Conformational Dependence of Long-Range Cyclopropyl Interaction

during Acetolysis," J. Am. Chem. Soc., 99, 828 (1977). iii Michael R. Detty and Leo A- Paquette, "The Fate of Bishomocyclohepta-

dienyl Cations Generated by Deamination," J. Am. Chem. Soc., 99,

834 (1977).

Michael R. Detty and Leo A. Paquette, "The Question of Hexahomobenzene 3 Automerization. New Synthetic Approaches to cis -1,4,7-Cyclo-

nonatriene," Tetrahedron Lett., 347 (1977).

FIELDS OF STUDY

Major Field: Organic Chemistry

iv TABLE OF CONTENTS

Page ACKNOWLEDGEMENTS ...... ii

VITA ...... iii

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

Chapter I. Synthesis and Properties of the Bis­

and Trishomocycloheptatrienes...... 1

INTRODUCTION...... 2

RESULTS AND DISCUSSION ...... 9

Chapter II. The Question of Hexahomobenzene

Automerization...... 38

INTRODUCTION...... 39

RESULTS AND DISCUSSION ...... 42

Chapter III. Studies of Long-Range Cyclopropane

Interactions with Carbonium Ion Centers. Con­

trasting Behavior of Cations Generated by Sol-

volysis and Deamination in the 3,5-Bishomocyclo-

heptadienyl System...... 47

INTRODUCTION ...... 48

RESULTS AND DISCUSSION ...... 52

Chapter IV. Functionalization of the Bishomocyclo-

heptatrienes. Studies of Cyclopropyl Halide

Solvolysis and 1,3-Bishomotropylium Cations...... 84

v Page INTRODUCTION ...... 85

RESULTS AND DISCUSSION ...... 94

EXPERIMENTAL ...... 127

REFERENCES ...... 201

vi LIST OF TABLES

Table 1 O Page 1 C Chemical Shift Data of the Trishomocycloheptatrienes.... 18

2 Kinetic and Thermodynamic Data for the Thermal Rearrangement of 34...... 22

3 Kinetic and Thermodynamic Data for the Thermal Rearrangement of 88...... 26

4 Kinetic Data for Acetolysis in Buffered Acetic Acid...... 58

5 NMR Spectra of the Isomeric Bicyclo[6.1.0]non-3- en-6-yl Acetates and Their Deuterated Derivatives...... 60

6 NMR Data for the 1,3-Bishomotropylium Cation...... 112

7 NMR Data for Unprotonated and Protonated 256...... 115

8 NMR Data for Unprotonated and Protonated J251...... 116

9 Kinetic Data for Solvolysis in 80% Aqueous Acetone...... 120

vii LIST OF FIGURES

Figure Page 1 1,5-Dienyl and Homo-1,5-Dienyl Hydrogen Shifts in Bicyclo[6.1.0]nonadienyl Systems...... 23

viii Chapter I.

Synthesis and Properties of the

Bis- and Trishpmocycloheptatrienes.

1 INTRODUCTION

Hie concepts of homoconjugation and homoaromaticity have elicited much interest in recent times. ^ Of particular interest have been cyclic polyene systems in which the conjugated rr ribbon has been interrupted one or more times by a methylene unit or by substitution of a cyclopro­ pane ring for a double bond. 1,3,5-Cycloheptatriene (1) is a represent­ ative of such a molecular type and has often been referred to as homo­ benzeneo . 2 1 2 sw (V 3

The question of whether cycloheptatriene and its derivatives are neutral homoaromatic molecules is one of long standing. A central issue has been whether or not cycloheptatriene (1) is in equilibrium with its norcaradiene valence isomer (2), or whether cycloheptatriene is best thought of as a resonance hybrid of the two as in 3. From electron diffraction and NMR studies, it is clear that cycloheptatriene adopts a boat shaped conformation as in j4 with a barrier to degenerate ring 3—5 inversion of about 6 kcal/mole. Although the existence of an

H.

AH* = 6 kcal/mole

4a 4b equilibrium between cycloheptatriene and norcaradiene cannot be detected experimentally, ® many of its reactions imply the existence of fL Its thermal conversion to toluene and capture by dienophiles in

Diels-Alder reactions are best explained as arising from the norcaradiene tautomer. ^-13 More convincing evidence for this equilibrium comes from

7.7-disubstituted cycloheptatrienes where 5^ and 6 have been isolated as separate entities and £ has been shown to be the preferred form of the

7.7-dicyano derivative. *4-16 Direct analysis of the NMR spectrum of

M £• * M *1 Q cycloheptatriene and diamagnetic susceptibility data provide evidence that cycloheptatriene can support an induced diamagnetic ring current. Finally, theoretical calculations combined with the photo-

CN CN CF: CN 6 7 /w ^ electron spectrum of 1 argue for the existence of a small homoaromatic interaction between the two carbons at the termini of the triene unit. 19'20

The theoretical calculations mentioned above are based on the structural parameters of cycloheptatriene determined by electron diffraction studies. 3 The hydrocarbon adopts a boat conformation (4) with the methylene group and the opposite carbon-carbon double bond bent out of the basal plane. The intemuclear distance between and Cfc was

O determined to be 2.511 A and the p orbitals on these carbons were canted toward each other at an angle of 67.42° relative to the basal plane. 3 Through the use of vector analysis and Slater orbitals, the extent of interpenetration of these p orbitals as given by the overlap integral

S was found to be 0.042, which is indicative of a low level of homo- aromatic interaction between and Cg. In particular, the magnitude of

S is considerably smaller than that calculated for 1,4,7-cyclononatriene 19 (8), triquinacene (9), and 6,7-benzoelassovalene (10).

■ A .

8 10

Cycloheptatriene also undergoes some interesting thermal and photochemical reactions. For the 7-deuterio derivative, thermal 1,5- hydrogen shifts as in the conversion of 11 to 12 have been observed 21-23 to proceed with an energy of activation of 32 kcal/mole.

hv

13 11 12

Photochemically induced 1,2-hydrogen shifts as in the isomerization to

A J AP 13, have also been observed. ' The photoclosure of cycloheptatriene 26 to bicyclo[3.2.0]hepta-2,6-diene (14) has been reported as well. CD A logical extension of the chemistry of 1 is that of the 1,2-

(1£) and 3,4-homocycloheptatrienes (16). These hydrocarbons were initially isolated from a mixture of products obtained by the reaction of carbene (prepared from diazomethane and cuprous chloride) with cycloheptatriene. 2^ The tetramethyl derivative 17^has been prepared in more elaborate fashion. 28 Conformational analysis by NMR of 16 and

17 reveals the existence of an equilibrium between the two different 27 28 conformers ^18a and 18b. ' Although the barrier to interconversion of 18a to 18b has not been determined, an estimate has been made that transoid conformation 18a is approximately 4 kcal/mole more stable than the cis form (18b). 28a The lesser stability of 18b is presumably due

CH.

CH.

CH. 15 16 IT to the nonbonded repulsions which develop on the interior of the structure because of the interpenetration of the two endo protons. The

1 ,2-homoderivative (15) is also capable of adopting two conformations represented by ,19a and 19b, although no data is available concerning 27 their relative stabilities.

18a R = H, CH; l8b 19a 19b The thermal chemistry of 3,4-homotropilidine and its derivatives has been of considerable interest. Structures of type 20 and 21 are capable of rapid, degenerate [3,3] sigmatropic shifts provided they adopt the cisoid conformations shown. 27-29 The transition state for such reactions is well represented by the bishomobenzene formulation

22. 27 »2® For the parent hydrocarbon, the energy of activation (Ea) and enthalpy of activation (AH+) for the Cope rearrangement were

20 21 22 determined to be 13.0 kcal/mole and 12.3 kcal/mole, respectively, by 13 28d variable termperature C NMR analysis. Similar values were obtained for the 1 ,2 ,3,4 ,5 ,6 ,.7 ,7-octadeuterio derivative (12.6 kcal/mole for Ea and 12.0 kcal/mole for AH^> by variable temperature NMR analysis. 28c For 17, the free energy of activation (AG?) of the Cope rearrangement has been estimated to have a lower limit of 22.8 kcal/ mole. The 3,4-homotropilidme nucleus forms the backbone for the degenerate rearrangements of semibullvalene (23) and bullvalene 30,31 (24)-. When 1(5 is heated to 305°, a vinylcyclopropane 7 27 32 rearrangement occurs to give bicyclo(3.3.0] octa-2,6-diene (25). '

The thermal chemistry of 1£ proceeds along different pathways.

When heated to 225°, homo-1,5-hydrogen shifting occurs leading to 1,3,6- 27 29 cyclooctatriene initially. '

When homotropilidine 15, is irradiated, the major product isolated is 33 the tricyclic olefin £6. This product presumably arises due to

26 disrotatory closure which minimizes nonbonded interactions between the two vinyl protons at the termini of the diene system and the endo cyclopropyl hydrogen.

While a wealth of chemistry has been derived from cycloheptatriene and its derivatives, and particularly the homocycloheptatrienes, extensions to more highly cyclopropanated analogs of ^ had not been reported. Although Doering alluded to the presence of bis- and tris- cyclopropanated cycloheptatrienes in his original preparation of the homotropilidines ,15 and ,16, these molecules were not characterized nor were their properties discussed. Sasaki and co-workers obtained one bis-dichlorocarbene adduct and one tris-dichlorocarbene adduct when cycloheptatriene was treated with dichlorocarbene under conditions

•j a of phase transfer catalysis. However, no structural assignments were attempted- During the course of the present study, Prinzbach reported the synthesis of some epoxy analogs of the bis- and trishomo- 1C cycloheptatrienes.

The primary goal of this research program was to devise unequivocal synthetic routes to the seven possible bis- and tris-cyclopropanated analogs of cycloheptatriene, such that all possible stereoisomers were fully characterized. Ancillary objectives were examination of the conformational characteristics of the hydrocarbons, determination of their susceptibility to thermal bond reorganization, determination of reaction pathways for acid catalyzed rearrangements, and determination of their susceptibility to transition metal catalyzed rearrangements. RESULTS AND DISCUSSION

Synthetic Considerations. The consequences of adding dichlorocarb- ene to cycloheptatriene under conditions of phase transfer catalysis were first examined. Although Sasaki and coworkers reported the production of single bis- and tris-dichlorocarbene adducts, ^4 there is actually formed a mixture of 27-29 which can be readily separated into the pure components by a combination of column chromatography on silica gel and fractional crystallization. Rigorous stereochemical assignment to these adducts could not be made on the basis of their NMR spectra.

However, because the cyclopropyl protons in 27 appear as a narrow signal at 61.80 and the pair of methylene protons gives rise to multi- plets centered at 2.36 and 1.18, a symmetrical structure was seemingly implicated for this product. The level of anisotropic shielding operat­ ing on the latter of these hydrogens is particularly noteworthy. The

.Cl Cl V CI Cl Cl Cl

Cl Cl Cl Cl Cl

27 28 29 30 31 spectra of ^28 and 29^ also reveal symmetrical patterns consistent with molecular frameworks having either C or Co symmetry. For 28, the #n 5 ^ r w w olefinic protons appear as a singlet at 55.67. The pair of methylene protons appear as two widely separated multiplets at 2.51 and 1.10, while the cyclopropyl protons appear as a multiplet between 2 .3 and 2 .0 .

Por J29, the olefinic protons appear as a narrow multiplet at 5.85. The pair of methylene protons appear as a triplet centered at 2.58, while the remaining cyclopropyl protons appear as a multiplet between 2.25 and

2.05.

Prom the demonstrated absence of 20, and 31 (vide infra), the /v w ' preferred initial site of dichlorocarbene attack on cycloheptatriene is inferred to be the 0^,02 double bond. Subsequent reaction occurs at

CgfCg with little stereochemical discrimination. This point was estab­ lished by resubjecting both 28^ and 29^ to the original experimental conditions. While 28 entered readily into further reaction, 29 proved rwv stable to further cyclopropanation. Since 28^ was converted uniquely to

27r the two flanking cyclopropane rings in the tris adduct must be stereodisposed as in the precursor molecule. Furthermore, the combined isolated yield of this pair of products (29%) is seen to be roughly comparable to that realized for 29^ (23%).

Reductive dechlorination of 29-31 with either sodium or lithium and tert-butanol in ammonia/ether led smoothly to the structurally 13 related hydrocarbons 32-34. The C NMR spectrum of 32 consists of six ■ signals and is therefore that of a symmetrical molecule. Although the most direct proof for the anti,anti arrangement of the three-membered rings in 3j2 comes from its independent synthesis (vide infra) , the rather complex NMR spectrum also attests to this stereochemistry. Thus, in addition to the cycloheptyl methylene protons which are characterized by widely divergent chemical shifts (A6 ~ 2 ppm), the twelve cyclopropyl protons appear as three quite differently weighted multiplets centered at 5 0.70 (9H), 0.10 (1H), and -0.05 (2H) . This pattern conforms 3-1 to that expected for (a) the sura of the six peripheral and three ''exo' cyclopropyl hydrogens, (b) the ''endo*' proton of the central cyclo­ propane ring, and (c) the remaining two shielded protons, respectively. no Assignment of structure to 33 rests upon its definitive C (five peaks) and NMR spectra. In CDCl^, the olefinic protons appear as a sharp singlet (55.55) and the cycloheptyl methylene protons as two

Na, t_-Bu0H

$ Br Br 2t & & widely separated multiplets (2.40 and 0.10). A rather broad multiplet in the 1.53-1.03 region due to the four remaining peripheral hydrogens is accompanied by two other complex absorptions attributable to the

"exo*' (0.78) and mutually shielded ''endo'* protons (0.00). These 13 data are to be compared to those recorded for ^4 whose C spectrum features four signals in the intensity ratio 2:1:4:2 and whose NMR spectrum is characterized by an AA'BB' multiplet centered at 5 5.56

(2H), a broadened triplet at 1.92 (2H), and an intricate pattern rang­ ing from 1.40 to 0.5 (8H). Relevantly, the cycloheptyl methylene protons have now gained equivalence due to axial symmetry and no anisotropic shielding is in evidence. Although 3J3 and J34 cannot be assigned their respective structures unequivocally by spectroscopic means, chemical methods can readily distinguish the two.. Thus, carbenoid attack on 33 from either face would yield a tris-cyclopropanated molecule possessing a mirror- plane of symmetry, while carbenoid attack on either face of 34^would lead to only one tris-cyclopropanated product which must be unsymmet- rical.

Dibromocarbene addition to 34^in pentane solution (KOtBu, HCBr^) or by means of phase transfer catalysis in a benzene-water mixture

(NaOH, HCBr-^) gave one tetracyclic adduct 3^ whose ^-3C NMR spectrum consisted of ten lines. Reductive debromination with sodium and tert- butanol in tetrahydrofuran gave rise to anti, syn trishomo derivative

26/ whose 13C NMR spectrum consisted of ten lines.

Exposure of 33^ to ethylzinc iodide 36 and methylene iodide proceeded stereospecifically to provide an alternate synthesis of 32.

No other products could be detected.

^NaH, PhCH„8r

OH

hi Cl Cl b2 UO

Preparation of the stereoisomeric 1 ,3-bishomocycloheptatrienes began by sodium borohydride reduction of tropone to form 3,5-cyclo- 37 heptadienol (,32). That Simroons-Smith cyclopropanation of 37 leads to cis-bicyclo[5.1.0]oct-5-en-3-ol (38) has been reported previously by Lambert. Exposure of 38 to an excess of zinc-copper couple and methylene iodide in ether resulted in very little further reaction even after prolonged periods of reflux (up to 6 days). Although 39 do recourse to zinc-silver couple and diethylzinc * proved somewhat more efficacious (47-48% conversion to bis-cyclopropanated products), ethylzinc iodide ^ emerged as the reagent of choice. After a reaction period of 23 hr, 3J3 was transformed quantitatively into a mixture of bisadducts 39 and 40 (70:30 to 90:10). The two alcohols were readily separated by column chromatography on silica gel or by preparative scale VPC on SE-30. The same product mixture of bisadducts was obtained by treating 3j7 with 5 mole equivalents of methylene iodide and ethyl zinc iodide for 48 hr.

In agreement with the established directing effect of the hydroxyl group, ^ the major product has all syn geometry. The Cg symmetry of 39 is clearly revealed by its ^3C (five signals) and NMR spectra supported by appropriate Eu(fod)3 shift studies. ^ The stereochemical reversal which results in the formation of 40 appears to be the result of steric congestion on the syn face of 3S3 and not to unusual conform­ ational factors such as those encountered in medium-ring allylie alcohols. If this hypothesis is correct, the problem of reversing the directionality of this reaction so as to obtain 4£ with high stereochemical control reduces to substitution of the hydroxyl function.

In actuality, reaction of benzyl ether 41^, which is available from 38^ in

97.5% yield, with dichlorocarbene under phase transfer conditions followed by reduction with sodium in ethereal liquid ammonia furnished aa,in 80% overall yield. The *3C spectrum of this alcohol shows nine signals indicative of its unsymmetrical nature. The NMR spectrum

consists of a multiplet at 5 4.10 (1H) , a broad one-proton singlet at 1.08, multiplets of area 2 centered at 2.13, 1.27, and 0.00, and a

four-proton multiplet at 0.72. C l C l

cr ^5

An alternative synthesis of was devised. 3,5-Cycloheptadienol

can be protected as its dimethyl t-butyl silyl ether 43, which is formed

in quantitative yield. ^4 Reaction of 4^, with dichlorocarbene under

conditions of phase transfer catalysis yielded tetrachloride 44, in 85% yield. Reduction with lithium and tert-butanol in tetrahydrofuran gave silyl ether 4Jj, in 95% yield. The silyl protecting group was readily

removed with tetra-n-butylammonium fluoride to give 4& as a white

crystalline solid in 89% yield. 44 This route is the method of choice

for preparing 40 based on yield and ease of running the requisite

reactions.

Li,N H3

k6 &

Li.NH ether 45 15 Collins oxidation of 39 and 40 led essentially quantit­ atively to ketones ,46 and 49. Conversion of ,46 to enol phosphate ,47 and subsequent treatment with lithium in ammonia gave syn bishomocyclo- heptatriene 48 in greater than 80% overall yield. Comparable application 46 of the Ireland procedure to 49 likewise resulted in smooth conversion to stereoisoraeric olefin 51 by way of the intermediacy of enol phosphate

50.

It should be noted that attempts to proceed from the alcohols 39, and 40 directly to hydrocarbons 48 and 51 were not nearly as successful.

Treatment of 39, and 40 with the Burgess reagent^ gave the desired hydro­ carbons in only 40 to 54% yield. Treatment of 39 and 40 with thionyl 48 chloride and pyridine led to deep seated skeletal rearrangements, which will be discussed later. Attempted elimination of either methane- sulfonic acid or j>-toluenesulfonic acid from the corresponding mesylates or tosylates with potassium tert-butoxide in dimethylsulfoxide or 49 benzene gave the desired hydrocarbons in 40 to 50% yield. However, a significant side reaction occurred which led to products that detonated upon heating during distillation.

Cyclopropanation of 48 with a large excess of methylene iodide and zinc-silver couple gave a 65:35 mixture of 36 and 52, the third and final trishomocycloheptatriene. The final step in the synthesis of 52

$

52 involves a reaction so transparent that structural assignment 13 hardly requires more elaborate justification. The C NMR spectrum of

six lines is indicative of a symmetrical molecule. The NMR spectrum

shows the cycloheptyl methylene protons to be two widely separated

multiplets at 62.37 and 0.28. Furthermore, the anisotropic shielding

effects prevailing upon all three ''endo'' cyclopropyl methylene hydrog­

ens [6-0.05 (m, 2H) and -0.38 (m, 1H)1 conforms uniquely with the syn,

syn orientation of the three-membered rings.

Conformational Features. Detailed NMR studies have established 4 5 that cycloheptatriene ’ and a number of its derivatives exist as

interconverting pairs of nonplanar conformers. ^ For the parent hydrocarbon, the two conformers are isoenergetic with an energy barrier

separating them of approximately 6 kcal/mole. In terms of exper­

imental observation, the methylene protons are seen to remain isochron­

ous down to -140° at 60 MHz. Below this temperature, the ''axial'' and

''equatorial'' H-j protons gain individual identity and appear as two bands with the 11 axial'' proton experiencing the greater shielding.

As substituents are introduced onto the ring, the barrier to conform­

ational inversion may be decreased or enhanced depending upon their

nature and position. For example, the trifluoromethyl groups in 7,7- bis(trifluoromethyl)cycloheptatriene remain isochronous down to -185®,

presumably because the ring in this instance is rendered more planar

and can invert more readily. In contrast, 7-tert-butyl-l-methyl-

cycloheptatriene undergoes slow inversion at room temperature and

prefers to exist in that conformation where the tert-butyl group is

axially disposed. 17 Clearly, the conformational populations of such systems are delicately balanced by a number of factors. Nonbonded repulsions which 27 28 may occur such as those described for the 3,4-homocycloheptatrienes '

are certain to play a significant role in controlling the equilibrium

dynamics of the bis- and trishomocycloheptatrienes described herein.

As concerns the anti.anti trishomo derivative 32, conformation 53a

is seemingly free of serious nonbonded interactions since all three

cyclopropyl substituents adopt an ’1 equatorial11 relationship relative

to the central seven-membered ring. Because conformational inversion

leading to 53b^generates a structure having closely proximate pairs of

hydrogens on both of its surfaces (and particularly the underside as

drawn), this conformer is forced at best to approach a more energy

demanding planar form to decrease such interactions. Distortions of

this type should decisively favor a conclusion which is fully

supported by the NMR spectra. Its inflexible conformational nature at

room temperature is signaled not only by the widely separated chemical

shifts of the two cycloheptyl methylene protons (5 2.38 and ca 1.0) , but

also by the appearance of two appreciably shielded cyclopropyl hydrogens

(6 -0.05). This observation requires that long range anisotropic shield­

ing by one or more of the cyclopropane rings^1 operate on a pair of chemic­

ally equivalent protons. This criterion is uniquely satisfied by the "endo"

methylene protons of the 1 ,2- and 5,6-bridges which find themselves la mutually projected into positions above the plane of the opposite cyclopropyl group. The methylene protons on the central cyclopropane ring are oriented away from the inner core and consequently do not experience these effects. Additionally/ not one of the three-membered rings in §3$, is aligned geometrically to be capable of shielding the sp^-hybridized cycloheptyl carbon. ^ As a consequence, the chem­ ical shift of this atom (32.47 ppm) is lowest of all three isomers (Table 1.).

Table 1. chemical shift Data of the Trishomocycloheptatrienes.a

Compd Cycloheptyl methylene Methines Cyclopropyl methylenes

32 32.47 17.23, 15.32, 12.19 14.82, 14.22 36 26.98 16.91 15.54, 14.95 14.65, 12.41, 5.26 13.54, 12.19, 11.60

/W52 29.46 18.29, 16.62, 10.36 12.63, 7.55 av ■"__

St 5a

Using similar lines of reasoning, the preferred conformations of

and 5 £ are deduced to be ,53 and £4, respectively. That ring flipping is again inconsequential to these molecules on the NMR time scale at

30° is evident from the appearance of the '1 equatorial'1 proton at

6 2.37 (note constancy of this signal) and its ,,axial,, counterpart above 1.55. In £4 this particular cyclopropyl triad arrangement serves to orient the endo proton of the 3,4- and 5,6-bridges into the shielding 19 region below the opposite three-membered ring. These protons are assumed to be the pair which gives rise to the high field multiplet at 5 0.1 to -0.01. In 55 the severe nonbonded interaction generated between the cycloheptyl methylene group and the carbon of the 3,4- bridge is expected to be relieved by attainment of a more planar arrange­ ment. Notwithstanding, at least four protons should be subject to effective anisotropic shielding. In agreement with this interpretation, two sets of resonances do appear above TMS at

(2H).

Furthermore, the rigid three-dimensional features of 55, are such that Cg can only experience moderate shielding by the somewhat remote central cyclopropane ring. For 54, one can discern that C7 should experience an anisotropic effect from the more proximal axially disposed 13 1,2-bridge. The actual C chemical shifts of these carbons are indeed shifted to higher field than that in 53, and the ordering (29.46 and

26.98 ppm, respectively) is commensurate with the above estimate of long range shielding.

The most reasonable conformations of 33, and 5j^ are those (£6, and

in which both cyclopropyl groups are positioned in extended ’'exo'' fashion. In hydrocarbon axial and equatorial now exhibit nearly identical resonances (ca 2-13). This phenpmenon is not attributable to a rapid conformational inversion process. Rather, this is a particularly good example of the differing spatial projections of Ti-olefinic and cyclopropyl anisotropic influences in rigid systems. 20

5&' 5 1 £

Prom a consideration of Dreiding models, the more thermodynamically favored conformation of 48 is deemed to be 38 where the 3,4- and 5,6- bridges reside in ''equatorial'’ and ''axial'' environments, respectively.

In the ring-flip form, considerable interpenetration of ''axial'' with the endo proton of the 3,4-bridge occurs with resultant energetic disfavor. Although the room temperature NMR spectrum of 4^ exhibits two well defined H7 proton multiplets centered at 6 2.80 and 2.42, insufficient evidence is currently available to permit full consideration of the conformational mobility of this structure.

On the other hand, bishomo derivative 34, is seen to undergo rapid ring inversion at room temperature. The dynamic process is degenerate

(59a. ^ 52fe) an(5 gives rise to an identical molecule in which the chemical environments of all six types of protons (labeled a-f) have been interchanged. Therefore, sufficiently rapid exchange of the

He

% *'Ha

52* 5 2 £ environments of these six pairs of protons generates a spectrum determined by the time averaged environments of the exchanging nuclei. As the process is gradually slowed by external cooling, the spectrum passes through an intermediate stage showing broadened lines and ultim­ ately appears as a superposition of the resonances of all individual

nuclei.

Upon cooling chlorodifluoromethane solutions of 34, the original

cycloheptane methylene triplet first coalesced and then separated at

-120° into two distinct signals due to Hf' and (partially hidden,

see 59) at 6 2.54 and 1.34 (J = 13.5 Hz), respectively. No further

changes were noted below this temperature. Simultaneously, three protons originally centered at 1.00 were shifted upfield to 0.88 (Hc),

0.76 (Hc i), and -0.03. The experimental spectra were insufficiently

resolved to provide all necessary coupling constants of the four protons

in question. However, trishomocycloheptatriene 36 serves as a good model for conformationally frozen 34^ since the only structural permut­

ation involves replacement of the olefinic linkage by a cyclopropane

ring, the effect of which is to lock the tetracyclic structure into

the closely related conformation depicted by 54^. The following exper­

imentally determined coupling constants were utilized in the computer

simulation: JH . = 9 , „ = 0.5, J . „ = 6 , J,, , = 6 , -Hc',Hf —He,Hf —Hc',Hf» “c *Hf and J = < 0.1 Hz. From data acquired at 8 temperatures, the -Hc,Hc* relevant thermodynamic and activation parameters were determined to be

Ea = 8.13 1 0.25 kcal/mole, AH* = 7.74 * 0.27 kcal/mole, AS_* = -4.5 -

1.4 eu, and AG* = 9.07 kcal/mole. — 298 Since the free energy of activation at 298° for ring inversion 4,5 in cycloheptatriene is on the order of 6.75 kcal/mole, 34^

exhibits a measurably greater hindrance (ca 2.3 kcal/mole) to

attainment of a presumably planar central seven-membered core. Interestingly, this decrease in conformational flexibility parallels closely in magnitude the barrier found for several 1,5- 53 bishomocyclooctatetraenes relative to cyclooctatetraene.

Attempts to treat tetrachloride 29 in a similar manner met with failure. The compound was insoluble in all solvents suitable.for low temperature NMR work.

Thermochemical Studies. When a benzene solution of 34 was heated 1 1111 "■ 1 “ ■■ ■" '■*■ i i — /s/V at ISO0 for 5 hr, essentially quantitative conversion to a single isomeric hydrocarbon (86% isolated) occurred. On the basis of direct spectral comparisons with an authentic sample, this product was identified as cis -1,3,6-cyclononatriene (60). The kinetics of this re­ arrangement as monitored by VPC analysis of reaction mixtures obtained at 151-174° and the resulting thermodynamic data are summarized in

Ta^lg^j^ Kinetic and Thermodynamic Data for the Thermal Rearrangement

of /34. w

T, °C k x 105, sec 1 Activation parameters

151.0 2.44 AH* = 32.7 + 0.2 kcal/mole 2.53 AS* = -3.1 + 0.5 eu 162.6 7.16 AG^298 = 33.7 kcal/mole 7.20 EA = 33.6+0.2 kcal/mole 173.8 18.9 19.1

Table 2. It was immediately evident that 34^ undergoes reaction with an energy and enthalpy of activation very similar to those found previously 54 by Winstein for the and 6^ §4^ interconversions.

Considering the geometry of (as given by formula 59) , we see that the endo cycloheptyl hydrogen is both ideally aligned with, and in adequate proximity to, the tt bond for concerted homo-1,5-dienyl shift 23

AH*=30.5 AS*= -5

AH*=32.5 AS* = -5

AH*= 29.3 H AS* = — 7

AH* =3 2 .1 AS* = - 6 &

55 as illustrated in Scheme I.

Scheme I

* H’ A

3 H P H & &II H H

This process gives rise to bicyclo[6.1.0]nona-2,5-diene 65, the 56 conversion of which to 60, has previously been studied. Since the homodienyl proton shift which transforms 65^ to 60, is complete within

1 hr at 130°, whereas the half-life for the conversion of 34 to 60 at 24

150° is approximately 10 hr, the rate-determining step would have to be the first hydrogen migration. In accord with this analysis, the presence of 6!§, was not detected in the various reaction mixtures.

To gain further insight into this rearrangement, the dj derivative

6& was prepared by reduction of 2^, with sodium in tert-butanol-0-d and tetrahydrofuran. Thermolysis of 66, gave a tetradeuterio labeled cis^-

1,3,6-cyclononatriene whose NMR spectrum shows a 3:1 ratio of olefinic to allylic protons. Accordingly, deuterium was present exclusively at allylic sites. The doubly allylic methylene triplet seen at 6 2.67 in

6JJ, was now clearly absent, and the original four-proton allylic multiplet

__ No, t-BuOD £2 = - THF

66 §L centered at 2.10 was now a somewhat broadened two-proton doublet.

Importantly, the multiplicity of the 1,3-dienyl olefinic protons was greatly simplified while that of the isolated vinylics was appreciably less so. On this basis, the labeling scheme must be tnat given by 67,, indicating that the original cyclopropyl methylene groups retain their integrity.

These data do not rule out the possible operation of a radical process involving initial homolytic cleavage of an internal cyclopropane bond followed by opening of the second three-membered ring and ultimate

1,2-hydrogen shift to deliver 67. This distinction was made by specifically labeling 34 with two deuterium atoms at the cycloheptyl 25 methylene group. The labeled hydrocarbon 68 was obtained by synthetic methods totally analogous to those employed for the preparation of 34.

68 69

Cycloheptatriene-7,7-d2 (69) was prepared by the addition of diazo- 57 58 methane-c^ to a refluxing slurry of benzene and cuprous chloride.

Dichlorocarbene addition to 69 under conditions of phase transfer / v v

catalysis“ yielded a mixture of a70-72 o i i~> i from which r72 n i was readily 34 isolated. Reduction of 72 with lithium in tert-butanol and tetra- Cl Cl Cl

Cl Cl 71 hydrofuran yielded 68. The H NMR spectrum of £8 showed a complete absence of the cycloheptyl methylene protons. The cyclopropyl protons appeared as a multiplet between 1.40-0.50. The splitting pattern of these protons was simplified relative to that of ^34.

When a benzene solution of 68 was heated at 185° for 10 hr, conversion to an isomeric hydrocarbon occurred whose ^-H NMR spectrum conforms to that of 1,3,6-cyclononatriene-3,7-d? C73) . The ratio of olefinic to allylic protons was 4:6 indicating deuterium substitution for two olefinic protons. Because the doubly allylic methylene protons

appear as a triplet (J = 8.0 Hz) at 6 2.63, no deuterium 26

73 substitution has occurred at the adjacent olefinic carbons. Two triplets having identical coupling constants of 8.0 Hz were observed at 6 5.45 and 5.57. These protons spin interact with and correspond to HB and HE , respectively. Since no further splitting is observed, Hfi and HE must be flanked by deuterium. Proton Hq appears as a broadened doublet (J = 10.5 Hz) centered at 5.85, while HD is seen as a multiplet centered at 5.48. The remaining allylic protons Hp resonate between 2.20-1.70 as a multiplet whose multiplicity is greatly reduced relative to that observed for these protons in £0. This is consistent with deuterium incorporation at an olefinic carbon flanking two of the protons Hp .

Table 3^ ^ Kinetic and Thermodynamic Data for the Thermal Rearrangement of 6£.

T, °C k x 105 , sec"* Activation parameters

183 7.37 7.37 AH^- = 38.2 kcal/mole

173 2.81 Ea = 39.1 kcal/mole 2.80 27 The kinetics of the rearrangement of 6<3 to 73^ were monitored by

VPC analysis at 173° and 183°. The resulting rates and thermodynamic

data are summarized in Table 3. At 173°, kn/kn is 6.7 where k^ is

the rate of rearrangement of 34 and kr> relates to the rate of rearrange-

ment of 68. The magnitude of the deuterium isotope effect is indicative

of C-H bond rupture in 34^ or C-D bond rupture in 68^ being the rate

determining step.

The data above are consistent with only one mechanism, that being

the twofold 1,5-homodienyl rearrangement outlined in Scheme I, where * the fate of the isotopic labels is represented by H . In this particular 59 pathway, the molecule makes recourse to "hydrogen rebounding" to

avoid otherwise nonconcerted reactions. The first rate-determining

deuterium transfer which involves migration from C7 to sets the

stage (presumably because of exothermicity) for subsequent transfer

from C4 back to the original Cg site. Because of the usual stereo- 54 60 selectivity demands on 1,5-homodienyl shifts, ' conformational ring

inversion to the "saddle" form must precede the "rebound." This flexing

of the ring simultaneously exchanges the relative positioning of

hydrogen and deuterium such that only hydrogen is now stereodisposed

for migration. The final cyclononatriene is therefore labeled as shown

in 73. The biradical mode of rearrangement would require instead that

the nine-membered ring be isotopically substituted at vicinal carbon

atoms.

When a benzene solution of 60 was heated at 210° for 3 hr, smooth

rearrangement to a single isomeric hydrocarbon was observed. From

NMR spectral data, the rearrangement product was identified as 28

11 76 cis-3,4-divinylcyclopentene (ZA)• The olefinic protons appear as two four proton multiplets centered at 6 5.63 and 4.96. The doubly allylic methine proton resonates at 3.20 (overlapping d x d, J = 8 , 8 Hz), while the other methine proton appears at 2.85 (d x d x d, J = 16, 8 , 8 Hz).

The allylic methylene protons appear as a multiplet between 2.30-2.08.

Further corroboration of structure was achieved through catalytic hydroqenation of 74 with 10% palladium on carbon in hexane to give cis-

1,2-diethylcyclopentane (75). An authentic sample of 75 was prepared from 2-ethylcyclopentanone by the procedure of Untch and Martin. ^

77MM 178 .

A [3,3] sigmatropic shift to give 78, has been reported upon heating •

77. ***■ A similar Cope rearrangement of 60, would lead to 74, directly.

In order to gain further insight into this rearrangement, a benzene solution of 73 was heated as described above. Analysis of the product by NMR was consistent with cis-3-vinyl-4-(vinyl-l-d)~cyclopentene-2-d

(7£). Two different olefinic protons appear between 6 5.92-5.50 which is indicative of deuterium substitution for one cyclopentenyl olefinic proton and for one of the vinyl group protons adjacent to the ring. The multiplicity of the methine proton at 6 2.85 has been reduced to a doublet of doublets (J = 16, 8 Hz) indicating deuterium substitution on an adjacent carbon. The remainder of the spectrum is similar to that of

7k. The labeling pattern is consistent with a Cope rearrangement of 73 to give 76^. From the labeling pattern in 76^ and the lack of other products, the [3,3] sigmatropic shift occurs much more rapidly than

1,5-dienyl hydrogen shifting.

The remaining six bis- and trishomocycloheptatrienes do not undergo thermal rearrangement with facility. At temperatures up to 500°C, flash vacuum pyrolysis conditions produced no structural changes. At 550°C, carbonization was observed in all cases. Under these conditions, only

52 afforded recoverable volatile materials (< 20%). However, the complexity of these reaction mixtures (VPC analysis) discouraged more detailed examination. Based upon the conformational considerations presented earlier, we conclude that the cyclopropyl groups and olefinic bonds in these hydrocarbons are not suitably oriented for interaction.

ACld Catalyzed Rearrangements. When a chloroform solution of syn-

1,3-bishomocycloheptatriene (48^) was treated at the reflux temperature with a catalytic amount of 70% perchloric acid for 0.5 hr, an isomeric mixture of hydrocarbons was obtained that was dominated by one product to the extent of 70%. This compound was identified as cis^-1,4,7- cyclononatriene (79) by comparison of its ^H NMR spectrum with that of an authentic sample. 54'56,62 a plausible mechanism for the 48^ to 79 30 transformation is outlined in Scheme II. Initial protonation of the bishomotropilidine forms a cyclopropylcarbinyl cation which delivers

Scheme II.

H+ -H

79, ® = ch2

80 , • = CD' 8 1 , © = cd2

the cyclononatriene product following twofold cyclopropane ring cleavage and deprotonation.

In order to test the above mechanism, the tetradeuterio bishomo­ tropilidine 80 was prepared and subjected to a catalytic quantity of

m

82 83 8U perchloric acid as before. Our approach to 80 began with the cyclopro- A/V panation of 3 ,5-cycloheptadienol (37) using zinc-silver couple and methylene iodide-^. There resulted a mixture of products from which 82

45 was readily isolated. Collins oxidation of §2, gave in quantitative yield the ketone §.3 whose tosylhydrazone J84 was readily prepared.

Treatment of 8^, with lithium 2,2,6 ,6-tetramethylpiperidide ^ gave the 31 tetradeuterio hydrocarbon 80^ in good yield. When treated with perchloric acid as predescribed, this labeled bishomotropilidene underwent rearrange­ ment to l,4,7-cyclononatriene-3 ,3 ,6 ,6-d^4 (81)* The NMR spectrum of

81 at -30° is characterized by a series of three multiplets centered at A/V

6 5.5 (6 , H), 3.7 (1H), and 2.1 (1H). the formation of 81^is entirely consistent with the mechanism outlined in Scheme II.

When bishomotropilidine 51^ was treated with perchloric acid at reflux in chloroform solution, a rapid disappearance of starting material was noted. However, the product mixture was determined to be mostly polymeric. VPC purification of the volatile products gave a 5% yield of 79^ as the major component. Apparently, conformational features or the anti orientation of the cyclopropane rings divert the major rearrangement manifold away from the twofold cyclopropane ring cleavage leading to 1,4,7-cyclononatriene (79}.

When hydrocarbons 33 and 34 were exposed to a catalytic amount of perchloric acid as described, no rearrangement was observed. Higher concentrations of acid and longer periods of reflux gave only polymeric materials.

The trishomocycloheptatrienes 32, 36, and 52, when treated with either perchloric acid or £-toluenesulfonic acid, gave very complex product mixtures containing more than 50 products in each case. From

30-50% of these product mixtures could be identified as a mixture of all six possible dimethylethylbenzenes.

Transition Metal Catalyzed Rearrangements. The ability of rhodium

(I) complexes to initiate skeletal rearrangements in systems containing strained o bonds is well documented. ^ The skeletal rearrangements 32 appear to be net four electron processes involving an oxidative addition of rhodium (I) to the strained a bonds of the substrate to give first a rhodium (III) complex, then rearrangement and finally reductive

64g loss of rhodium (I) from the rearranged complex. The presence of the catalyst allows rearrangements to occur that are thermally forbidden in a concerted manner such as the quadricyclane (85) to norbornadiene 64c d (86) or homonorbomadiene (87) to homoquadricyclane rearrangements. '

Rh(l) *

85

88

The preference for net four electron processes appears to dominate even in systems ideally suited for six electron rearrangement. Thus, although diademane (89) readily opens to triquinacene (9) in the presence of acid or silver (I) or with thermal activation, in the presence of rhodium (I), the net four electron process leading to 64d. snoutene (90) is the only rearrangement pathway observed. 33 When a benzene solution of syn-1 ,3-bishomocycloheptatriene (£8) and

8 mole percent rhodium dicarbonyl chloride dimer was heated at 60° for

42 hr, smooth rearrangement occurred to give a 59:41 mixture of bicyclo-

[6.1.0]nona-2,5-diene (j65) and cis 3-l,3,6-cyclononatriene (60) as attested to by spectral comparisons of the products with authentic samples. ^ Upon heating a benzene solution of j55 to 60°, slow thermal conversion to 60 was observed. ^ Resubmission of 60 to the conditions of reaction showed no rearrangement, while j4jQ was found to be thermally stable to refluxing benzene. This data is consistent with 6!5 being a primary rearrangement product while the formation of (50 arises, at least in part, from a thermal rearrangement of 6j5.

The overall bond reorganization in converting 48 to 6^5 is equivalent to a net homo[3,3]sigmatropic shift (Scheme III). A homodienyl hydrogen shift in-65 would lead to 60. Alternatively, 6jD

Scheme III.

Bishomo Homo [1,5] P,3] H-shift shift

Bishomo U,7] Homo H-shift [1,5] H-shift 34 might arise directly from 48 through rhodium catalyzed bishomo (1,5] or bishomo [1,7] hydrogen shifts. It is also possible that more complex skeletal rearrangements leading to both 65 and 60 may occur. In order to address the mechanistic questions raised, the series of deuterium labeled compounds 80, 91, and 92 was subjected to rhodium catalyzed rearrangement.

D D

D D 92 80 91

Treatment of 80 with rhodium dicarbonyl chloride dimer in benzene at 78° for 12 hr gave a 59:41 mixture of two products whose NMR spectra are consistent with structures 93 and 94. The spectrum of 93

D 93 9^ displays no allylic protons, 4 olefinic protons, and 4 cyclopropyl protons. The NMR spectrum of 94 indicates a 5;3 ratio of olefinic to allylic protons with the doubly allylic methylene triplet still intact plus one other allylic proton. Treatment of 80^ and the catalyst in benzene at 85° for 30 hr gave a 7:93 ratio of the same two products.

The NMR spectra of these products were identical to those of the products isolated from the first run. 35 The monodeuterio hydrocarbon 91 was readily prepared by sodium borodeuteride reduction of ketone 46 to give an epimeric mixture of a-deuterio alcohols 95, reaction of these alcohols with sulfene to give the corresponding mesylates, and subsequent elimination of methanesulfonic acid with potassium tert-butoxide in refluxing benzene to deliver 91.

When 91 was treated with 8 mole percent of rhodium dicarbonyl chloride dimer in benzene at 78° for 22 hr, two rearrangement products were isolated in a 30:70 ratio whose NMR spectra are consistent with 96^ and 97, respectively. The spectrum of 96 is nearly identical to that of

96 97

65 except that the allylic bridgehead cyclopropyl proton at 5 1.47 is absent. The NMR spectrum of 97^displays a 5:6 ratio of olefinic to allylic protons. The doubly allylic protons appear as a doublet (J =

7 Hz) centered at 6 2.68. This is indicative of deuterium substitution at one of the adjacent olefinic centers.

The trideuterio bishomotropilidine 92^ was next prepared. Exchange of the protons a to the carbonyl for deuterium in 46 was accomplished

NNHTs

D 98 99 36 with a catalytic amount of sodium carbonate in deuterium oxide-tetra- hydrofuran. Two exchanges were sufficient to give 90% deuterium incorporation by 1h NMR. The final exchange was made in methanol-0-c3 with a catalytic amount of £-toluenesulfonic acid. Tosylhydrazine was added to the methanolic solution, and the tosylhydrazone 99 was isolated as a white crystalline solid. Decomposition of 99 with lithium

2,2,6,6-tetramethylpiperidide ^ gave 92 with greater than 95% deuterium incorporation by NMR.

When a benzene solution of 92 was treated with 7 mole percent of rhodium dicarbonyl chloride dimer at 70° for 21 hr, a 52:48 ratio of two rearranged products was observed whose 1h NMR spectra are consistent with 100 and 101, respectively. The spectrum of 100 shows the absence of the cyclopropyl methylene protons at 6 0.68 and -0.11 and the absence of one olefinic proton. The NMR spectrum of 101 shows a complete absence of the doubly allylic methylene protons at 6 2.68 and a 5:4 ratio of olefinic to allylic protons. The data from the rhodium catalyzed rearrangements of 80, 91, and 92 is consistent with only one mechanistic pathway leading to products. The observed distribution of the deuterium labels is summarized in Scheme XV and is conclusive

100 101, of a net homo[3,3Isigmatropic shift leading to bicyclo(6.1.0]nona-2,5- diene products followed by 1,5-homodienyl hydrogen shifting to give 37 Scheme IV.

1/3,6-cyclononatriene products. The rearrangement is novel in the sense that a 6 electron process is being catalyzed by rhodium (I) instead of 64 the usual 4 electron bond reorganizations.

The remaining bis- and trishomocycloheptatrienes were found to be stable to rhodium dicarbonyl chloride dimer. Similarly, all seven of the hydrocarbons were found to be inert to silver perchlorate in benzene . Prolonged heating of the hydrocarbons in silver perchlorate- benzene solution allowed decomposition of the catalyst to occur with concomitant onset of acid catalyzed rearrangements. CHAPTER II.

The Question of

Hexahomobenzene Automerization.

38 INTRODUCTION

In recent years, much interest has been given to molecular systems

capable of undergoing degenerate thermal rearrangements. Systems

involving bond reorganizations of tt electrons or a blend of ir and a

electrons have been most extensively studied. The bond shifting in the

automerizations of cyclooctatetraene (102) ^ and methano-1,7-[12]

annulene (103) ^ are representative examples of degenerate rearrange- ments involving only it electrons. The observed energetics for such

102

103

rearrangements cover a wide range as exemplified by the 5 kcal/mole 66 barrier to interconversion in 103 and the 27 kcal/mole barrier, as a

lower limit, to automerization for 1,2,3- trimethylcyclooctatetraene

(104). 67 The degenerate thermal rearrangements of 3,4-homocyclohepta- r w w

triene (16), ^ semibullvalene (23), bullvalene (24), ^1 and

39 40

CH. c h 3" 10U

68 hypostrophene (105) are representative examples of systems involving a mixture of ir and a electrons.

105

Despite the high level of interest accorded to the types of isomerizations described above, automerization reactions of molecules constructed entirely of o bonds have been given little attention. transformations based upon the prototypical 4 electron process 106 107 69 have been briefly sought, but gone undetected; however, orbital 29 symmetry can be expected to exert an untoward influence in these cases.

Winstein prepared the hexahomobenzene 108 by eyelopropanation of 1,4,7- cyclononatriene (79) with the hope that the molecule might exist as the 69 i delocalized 110. However, the H NMR spectrum of 108 showed it to 67 possess distinct cyclopropyl and "ordinary" alicyclic methylene protons.

Because the 108 109 isomerization (a 6 electron change) is in principle 29 a thermally allowed reaction, it appeared conceivable that sufficient thermal activation might now result in observable degenerate rearrangement. O

jsfe AAA 108

110

Relevant to the question of hexahomobenzene automerization was

Prinzbach's report, which appeared toward the conclusion of this work, of the noninterconvertibility• of the two "trioxa-hexa-a/ir-homobenzenes" 70 111 and 112. This related pair of nondegenerate valence isomers did not undergo interconversion at temperatures up to 450°.

Ill 112

In order to study the thermal behavior of 10J3, under similar

conditions, it was necessary to label the two types of methylene protons

in the molecule if interconversions within the twofold degenerate system vere to be observed. To this end, synthetic routes to the hexadeuterio

hexahomobenzenes 113 and 114 were devised, and the response of these 42 molecules to thermal activation was determined.

Hi).

RESULTS AND DISCUSSION

This study first required access to cis -1,4,7-cyclononatriene (79).

Because the preexisting methods for gaining entry to the triene suffer from low overall yields and difficulties in purification, efficient alternative synthetic approaches were sought. Winstein prepared 79 in

79

30% yield from 1 ,3,6-cyclooctatriene by cyclopropanation, pyrolysis, 71 and isolation of 79^ from a mixture of products. Untch prepared 79^

in more elaborate fashion from indane in 7 steps in less than 3% overall yield. ^ Our approach utilized 1 ,5-cyclooctadiene (115) as starting material. Dichlorocarbene addition to 115^ followed by reduction of the monoadduct 116 with lithium in tert-butanol and tetrahydrofuran gave i i i 73 74 cis-bicyclo[6.1.0]non-4-ene (117) in 66% yield. ' Bromination of Scheme V.

Li, Cl 3C 02N a Cl (c h 3)3coh ME o THF 115 116 117 Br2 , iCCI4 LiF, Br 79 LigCOa C6 Ha HMPA Br

119 123

117 with bromine in carbon tetrachloride gave dibromide 118 in quantit- ative yield. Dehydrobromination with lithium fluoride and lithium carbonate in hexamethyl phosphoramide at 100° in the presence of powdered glass gave diene 119 in 72% yield. 66 Thermolysis of 119 at 175° in benzene solution gave 79^ in 87% yield. The overall transformation as

shown in Scheme V involves 5 steps and proceeds in 42% overall yield. 56 The pyrolysis of 119 delivers 7j3 after two rearrangements.

Initial 1,5-dienyl hydrogen shifting delivers 120 which undergoes a 44 56 homodienyl hydrogen shift to give 1 ,4 ,7-cyclononatriene (79).

The synthesis of 114, required the preparation of 1,4,7-cyclonona-

triene-3,3,6,6,9,9-dg (121). None of the synthetic approaches in the

literature are suitable for such extensive deuterium incorporation.

Furthermore, attempts to promote H/D allylic exchange in 79 with a

121 122

strong base have seemingly failed since dihydroindane 122^ is the major product formed upon treating cyclononatriene 79, with potassium tert- 7c butoxide. In principle, more extensive specific deuterium substit-

ution of 75, should be realized if two of the three constituent methylene

groups were present in a precursor molecule where control of H/D

labeling was possible and elaboration of the third -CD2- moiety was

reserved for the final step. Guided by this rationale and the previously

explored acid-catalyzed rearrangement of 80^, the pentadeuterio hydrocarbon

123 was viewed as the precursor of choice. Subsequent deuteration

should give 121 directly. D D. JD 45 The desired molecule was readily prepared by sodium borodeuteride

reduction of ketone to give an epimeric mixture of pentadeuterio

alcohols Reaction of this mixture with sulfene followed by treat­

ment with potassium tert-butoxide gave bishomocycloheptatriene 123^in

73% overall yield. Cyclononatriene 121^was obtained by treating 123^in

chloroform solution with a catalytic quantity of perchloric acid in a

100-fold excess of deuterium oxide. As anticipated, the NMR spectrum

of 121 (-30°) produced in this manner displays only a slightly broadened

singlet at 6 5.48 as well as greater than 90% deuterium incorporation.

Furthermore, the elaboration of the third CD2 unit is consistent with

the mechanism proposed earlier.

Treatment of 79^with methylene iodide-d^ and zinc-silver couple in

ether followed by preparative VPC purification afforded 113^ as a white

crystalline solid. In agreement with the structural assignment, the

NMR spectrum of this tetracyclic hydrocarbon lacks the upfield

cyclopropyl methylene absorptions which are characteristic of the 69 unlabeled molecule. When 121^ was exhaustively cyclopropanated as before but with methylene iodide, the white crystalline 114^was isolated.

The NMR spectrum of 114^ exhibited a 9-proton multiplet centered at

6 0.70 and a 3-proton multiplet at -0.27.

Zn - Ag

m 1 1 U 46

Pure samples of 113^and llj^were next individually subjected to conditions of flash vacuum pyrolysis in a quartz reactor packed with quartz chips (nitrogen entrainment). At temperatures up to and

including 485° (contact time ~3 sec), the two hexahomobenzenes were recovered totally unchanged. Progression to 500° caused the onset of low level decomposition, but no automerization- Clearly, there exists insufficient interaction among the three constituent rings in 113 and 1 14^ to allow for bond shifting.

The demonstrated lack of interconversion of 113 and 114 does not ( W v r w w appear attributable to adverse electronic effects if the facile 6 electron reversion of diademane (89) to triquinacene (9) serves as suitable 64a analogy. Rather, it is more likely that the geometric conditions prevailing in these hexahomobenzenes preclude the possibility of attaining interatomic distances sufficiently proximate to permit interaction. Given the well characterized three-dimensional structure 76 2 2 ° ^ 29, we see the crucial nonbonded 1 ,3sp C-sp C distances (2.46A) to be such as to deny the possibility for meaningful homoaromatic delocal- ization. 7 7 Since cyclopropanation of the cis- 3-1,4,7-cyclononatriene 69 system inflexibly fixes the 9-membered ring while simultaneously compounding compression energies within the molecular crown, we view an improved juxtapositioning of the key reactive centers as unlikely. Chapter III.

Studies of Long-Range Cyclopropane Interactions with Carbonium

Ion Centers. Contrasting Behavior of Cations Generated by Solvolysis

and Deamination in the 3 ,5-Bishomocycloheptadienyl System.

47 INTRODUCTION

One of the classic properties of carbon-carbon double bonds has been their ability to interact with an adjacent carbonium ion center to 78 form allylic cations. More recently, the ability of the double bond

to interact with a carbonium ion center two or more bonds away has been 79 realized such as the homoallylic cation. The cyclopropane ring, which in many ways is more similar to the carbon-carbon double bond than

to its larger cycloalkane counterparts, has also been found to be

capable of interaction with an adjacent positive charge to give 80 resonance-stabilized cyclopropylcarbinyl cations. In the last 15 years, the ability of the cyclopropane ring to interact with distant 81 carbonium ion centers has been realized.

Progression from the cyclopropylcarbinyl cation (125) to homologous

0-cyclopropylethyl systems (126) is accompanied by a dramatic alteration

in the geometry adopted to achieve maximum conjugative overlap. The

spatial requirement in 125 is a bisected structure which necessitates

h , h

H i u H H 126

that the remote carbon-carbon bond of the three membered ring be 80b,81,82 aligned "anti periplanar" to the leaving group. In rigid

molecules where such an arrangement is unattainable, direct conjugative

assistance is absent and pronounced kinetic deceleration is seen because

48 83 49 of sizable inductive contributions by the proximate three-membered ring. When homoconjugated transition states such as 1J26 are involved, kinetic effects are frequently not as pronounced and their detection has relied heavily upon product analysis.

In actuality, rate comparisons have been of little value in this area since formation of a delocalized intermediate need not necessarily 84 give rise to an enhanced rate. One can envision the formation of transition states such as 126 occurring along two different pathways.

If the cyclopropane ring anchimerically assists in the ionization of the leaving group, then an enhanced rate of solvolysis should be observed.

Alternatively, if the cyclopropane ring interacts with the cationic center only after it is formed, then no kinetic acceleration should be expected. Tosylates 12J7 and 128 serve as useful contrasts. Thus, 12J7

H OTs

2SL H — OTs

OBs

128

exhibits little rate acceleration relative to its epimeric tosylate, presumably because the conformation necessary for participation is not oc preferred, while the conformationally rigid and strained 128 is 86 . 50 highly reactive toward ionization. The early controversy concerning the extent of cyclopropyl participation during ionization of parent 87 brosylate 129 further accentuates the difficulty in establishing definitively whether the cyclopropyl group anchimerically assists in formation of the carbonium ion or interacts with the cationic center 81 only after it is formed. As noted by Haywood-Farmer studies of long-range interactions in B-cyclopropylethyl systems have been surprisingly few in number despite the inherent fascination of the subject. The difficulties mentioned above have apparently comprised a major deterrent to many potential investigators.

Herein we report details of the solvolytic behavior of the isomeric bishomo-3,5-cycloheptadienyl tosylates 13(3-1^32, three molecules which incorporate structural features particularly well suited for elucidation

V

H H

H H

OTs SSL

of long-range cyclopropyl participation. The special characteristics unique to each of these molecules is detailed on an individual basis below. It is emphasized here that the thrust of this research was not the possible generation of tetrahomocyclopentadienyl cations of type

133. Because such intermediates clearly should suffer from thermodynamic

destabilization, even when geometry is favorable, participation by the

pair of cyclopropyl groups in this fashion was considered highly

unlikely from the outset. Rather, our objective was to gain extra

kinetic evidence for the existence or absence of homoconjugative cyclo­ propyl participation during ionization by the simple expedient of

symmetry-based structural design.

The solvolysis of 130-132 may involve significant neighboring

group participation and consequent distortion of reactant geometry in progressing to products. Alternatively, unperturbed cations such as

those represented by 134 and 135^ might offer an interesting contrast,

since subsequent long-range cyclopropyl interactions would be with a

235, £ £ 12L

fully developed carbonium ion center.

At this point, it appears useful to define "vertical" and "non­

vertical" ionization. During vertical ionization, a carbonium ion is

formed via the loss of some leaving group free from any intramolecular

neighboring group involvement. In contrast, nonvertical processes are

characterized by direct neighboring group interactions at the stage of

rate-determining ionization. 52 Solvolytic studies on B-cyclopropylethy1 systems are recognized to proceed by both pathways in many cases, and often the line dividing them is frustratingly obscured. Frequently, widely varying product ratios are found for diazonium ion decompositions when compared with solvolysis mixtures. This has been attributed to much lower activation energies resulting from the excellent leaving group characteristics of the nitrogen molecule. The initial generation of carbonium ions in a 88 vertical ionization manifold is thought to be involved. In order to understand better the relationship between the stereodisposition of cyclopropyl groups and their possible involvement with anchimeric assistance in bishomocycloheptadienyl cations, a study of the fate of

and 12J3 as generated by deamination was undertaken. Solvolysis of

JJJO, and i-n a common solvent (methanol) was also studied. To complete the mechanistic picture, we have examined as well the behavior of 2 ,5-bishomocycloheptadienyl cations 133, and 1JX when generated under comparable conditions.

RESULTS AND DISCUSSION

The svn.anti-3,5-Bishomocycloheptadienyl System. Stereochemically pure the synthesis of which was described earlier, and its crystall­ ine tosylate derivative 130, can exist in the two conformationally distinctive forms anc^ UiSfe,. In saddle geometry JJ33& the proximity of ''axial" H2 to the endo proton of the 5,6-bridge gives rise to transannular nonbonded repulsions and, as a consequence, extended form

138b is considered to be thermodynamically favored. Furthermore, the

OT*

H

two cyclopropane rings in 138b are projected in the equatorial plane in contrast to their axial disposition in 138a. The possibility that

138a and 138b interconvert is, of course, not ruled out. However, neither cyclopropyl group in 138a is properly aligned for long-range anchimeric assistance to ionization (see 126). In 138b, only the syn cyclopropane ring has its internal bond suitably oriented to provide neighboring group assistance. Accordingly, should cyclopropyl partici­ pation gain importance during the ionization of this system, interaction is necessarily restricted to the syn three-membered ring for geometric reasons. On the other hand, if such direct involvement is lacking at the solvolytic transition state, secondary carbocation 139 could intervene. Because 139 possesses axial symmetry and is constructed such that either cyclopropane ring can now participate with equal probability in charge delocalization, the earlier restriction of unique syn cyclo­ propyl involvement is no longer strictly applicable. On this basis, the extent of cyclopropyl assistance during and after ionization could be

indicated by the extent to which the anti three-membered ring has experienced appropriate perturbation while going to product(s). The 89 possible operation of ''memory effects'' (now considered to be the 90 result of unsymmetrical ion pairing ) has not been taken into con­ sideration in the above analysis (vide infra).

Titrimetric solvolysis rates for 130 were measured in sodium acetate buffered (0.0510 M) acetic acid by the aliquot method and good first-order behavior was exhibited throughout. The kinetic data and activation parameters are summarized in Table 4. since the infinity titers

invariably agreed well with the calculated values, internal return to a less reactive tosylate was not an issue. The product mixture was comprised of a single acetate to the extent of > 98-5%. The remaining

1 .5% appeared to be a mixture of hydrocarbons and was not further

investigated. The structure of the nearly exclusive product was established as 140 by sequential lithium aluminum hydride reduction to

141 and oxidation of this alcohol to bicyclic enone 142. The independent

synthesis of 142 took advantage of the facile thermal rearrangement of 91 143 to 3,6-cyclooctadienone, which afforded 144^upon treatment with 91 92 sodium borohydride. Controlled Simmons-Smith cyclopropanation of

144 led to 145, the epimer of 141. which was likewise transformed to 142 when treated with the chromium trioxide dipyridine complex (Scheme VI).

A mechanistically plausible rationalization of the conversion of

130 to 140 considers initial homoconjugative interaction of one of the

cyclopropane rings to generate a cyclopropylcarbinyl cation intermediate

which is trapped by solvent as its homoallylic equivalent. Due to the Scheme VI. 55

OAc 'OH

JJ& 1^1 | C r03-2py

OH CpHgZnl

2.Na BH 4 CH2I 2 % - a

m . unsymmetrical nature of 140, it becomes possible by suitable deuterium labeling to establish not only which cyclopropane ring enters into which phase of this reaction, but also the extent to which these cyclo­ propyl interactions are specific. The pursuit of this objective was made possible by the synthetic sequence outlined in Scheme VII. Treat­ ment of 3,5-cycloheptadienol with diiodomethane-d^ anc^ zinc-silver

Scheme.VII. 'V

,0H NaH ,0CH„Ph OH

1U6 1U8

TsCI, py

NoOAc OTs

OAc HOA c

i5° 56 couple gave 146 admixed with lesser quantities of the two biscyclo- propanated tetradeuterio alcohols. Conversion to benzyl ether 14^7 followed by reaction with dichlorocarbene under phase transfer conditions and sodium/ammonia reduction provided 148. The assignment of .stereo­ chemistry to 148 is based reliably on the synthetic method which is known to introduce the first cyclopropane ring onto 3 ,5-cyclohepta- dienol stereoselectively from the syn direction. Therefore, there exists no question that the syn cyclopropane ring in tosylate 149, carries the pair of deuterium substituents. When subjected to acetolysis as before, 149 was converted to acetate 150.

The labeling pattern assigned to 150 follows principally from comparisons of its ^"H NMR spectrum with that of 140, (Table V) . In the case of 140 the noteworthy spectral features are: (1) the doublet of doublet of doublet multiplicity of each olefinic proton arising from mutual coupling and spin interaction with the adjoining methylene groups; (2) the deshielded nature of anti H2 and (<5 3.16-2.50) caused by their spatial proximity to the acetate substituent; and (3) the rather similar chemical shifts of syn H2 , syn H5 , and anti H7 which give rise to a series of overlapping multiplets at 2.25-1.92. In dideuterio derivative ^50, the complex signals at 3.16-2.50 and 2.25-1.92 are both decreased in intensity by the relative area 1. While this feature alone does not constitute proof of the loci of deuterium substitution, the emergent patterns of the remainder of the spectrum suggest the structure to be 150.. Thus, while the 5.63 absorption due to H3 experiences simplification to a doublet (J_ = 11 Hz) and the multiplet at 1.1-0.5 is somewhat altered in appearance, the remainder of the signals do not lose 57 their original character. Consequently, acetate 150 must be substituted

with two deuteriums at C2 .

Further confirmation of this assignment was derived from an exam­

ination of the solvolytic behavior of 152 under comparable conditions.

Prepared from alcohol 151,» a product of the cyclopropanation of 3,5-

cycloheptadienol with CD2I2 1 this tosylate underwent ready conversion to

153. The revealing aspects of its spectrum includes the collapse of

0H TsCt Na0Ac D --- py hoac -r>s=B^%''0Ac D D D

^ 152. both olefin signals to doublets, the total absence of the 6 3.16-2.50 multiplet, and reduction in relative intensity of the next upfield

absorption (2.15) to area 1 combined with its simplification. These

data are uniquely compatible with the indicated 2,2,5,5-tetradeuterio

substitution plan.

Once attention is called to the overlapping nature of the and

signals, it is clear that a direct measure of the ’’leakage1' of

deuterium from H2a to Hga during acetolysis is not feasible. Fortunately,

an indirect assessment of this question is possible. When the multi­

plicities of the a-acetoxy (Hg) and olefinic protons (H^ and H^) in well

resolved spectra of 150 are examined closely, two points emerge: (a)

relative to their appearance in unlabeled acetate 140 the Hg and

patterns are not permuted; (b) no vestiges of the original multiplicity

of the 5 5.63 signal (due to H3) are visible, it being replaced by a Tgkle_4. Kinetic Data for Acetolysis in Buffered Acetic Acid

40° Compd Temp^ k, sec-1 AH , kcal/mol AS*, eu ^rel

+ MQ 79.7 3.50 X 10“4 24,8 + 0.6 -4,5 1.8 1 63.4 5.43 X 10"fi 40.0 3.46 X 10’6 m , 51.1 2.63 X 10"4 23.9 + 0.5 -1.3 ± 1.5 18.7 40.0 6.48 X 10"5 28.4 1.49 X 10~5 132 79.7 6.38 X 10-4jk 24.1 ± 1.2 -5.0 ± 3,6 2.1 66.4 2.09 X 10-4 50.0 3.06 X 10-5 40.0 7.41 X 10~6

OTs 40.0 4.43 X io_5a'b 21.9 ± 0.5 -8 ± 4 12.8

40.0 1.95 X io-5C#d 23.3 -5.7 5.6

interpolated value based on the activation parameters. B. Lambert, A. P. Jovanovich, J. W. Hamersma, F. R. Koeng, and S. S. Oliver, J. Am. Chem. Soc., £§, 1570 (1973). cExtrapolated value based on the activation parameters. ^H. C. Brown and G. Ham, ibid., 78, 2735 (1956).

U1 oo 59 doublet showing additional narrow long range coupling (J ~ 1 Hz). Based on these observations, the deuterium would appear to be localized very heavily, and perhaps exclusively, at C2 - Consequently, to the level of

accuracy accorded by this analysis, the question of cyclopropyl participation is resolved in favor of the syn three-membered ring.

The svn.svn-3,5-Bishomocycloheptadienyl System. Conformational analysis of 131^reveals that the molecule can exist as the two inter­ convertible forms 154a^and 154b. In either structure, one axial and one equatorial cyclopropyl ring is projected from the tub-shaped cycloheptane frame. From models, the steric interaction prevailing between the endo axial cyclopropane and transannular '’axial'' proton pair in these conformers seemingly causes a greater degree of flattening of the medium ring than found, for example,, in 138b. Although the axial three-membered

H

OTs

154a rings cannot provide anchimeric assistance to ionization, this restrict- • ion does not apply to their equatorial counterparts. But since both

and l§4b, possess a cyclopropyl group properly oriented for partici­ pation and because the activation energy for ionization should exceed the barrier to their interconversion, solvolysis will likely occur from both conformers.

Determination of the acetolysis rate constant revealed 131^to be well behaved kinetically and to be almost 19 times more reactive than 1h NMR Spectra of the Isomeric Bicyclo[6.1.0]non-3-en-6-yl Acetates and their Deuterated Derivatives (6 , CDCl^', 90 MHz).

Compd h4 H2a'H5aa H2s ^ 5s>H7aa Hi,Hq,Hq h3 h6 CH3 H7s 1 0 yexo H9endo

140 5.63 5.47 4.95 3.16-2.50 2.25-1.92 2.04 1.20 1.1-0.5 -0.1 (d of d of d) (d of d of d) (d of t) (m) (m) (s) (m) (m) (m)

150 5.63 5.47 4.95 2.92 2.25-1.92 2.04 1.20 1 .1-0 .5 -0.1 (d) (d of d of d) {d of t) (m, 1H) (m, 2H) (s) (m) (m) (m)

153 5.63 5.47 4.98 2.15 2.05 1.20 1 .1-0 .5 -0.1 (d)

Compd H, H H3 »4 «6 H2a'H2s'H5a'H5s'H7aa CH3 7s l'H8'H9exo **9endo

5.84 5.58 4.73 2.67-2.08 2.00 1.40 1 .0-0 .5 0.05 (d of t) (d of d of d) (q of t) (m) (s) (q of t) (m) (m)

157 5.84 5.58 4.74 2.34 2.01 1.40 1 .0-0 .5 0.05 (d) (d) (d of d) (d of t, 1H) (s) (q of t) (m) (m) aThe subscripts a and s are stereochemical descriptors and refer to the syn or anti orientation of the proton relative to the cyclopropane ring.

cn o ^22, at 40° (Table 4). When conducted on a preparative scale, the solvolysis led to an oil shown to be composed of a single acetate

(> 98.5%) identified as 155 by its reduction to alcohol 145. Therefore, the ionizations of 13£ and 131 proceed with like skeletal rearrangement but with high levels of contrasting selectivity. The absence of crossover between intermediate ions is thereby demonstrated. Although an intermixing of rearrangement pathways is difficult to imagine as long as the reactive species are tricyclic, these results establish that merging of the rearrangement manifolds does not occur at a later stage during the evolution of bicyclic character. This aspect of the problem gains importance in view of the independent demonstration that

NaOAc k0 Ac OTs

As revealed previously by the conversion of 152 to 153, the syn,syn and syn,anti series share this common reactivity pattern. The comparative spectral features of 155 and 157 are detailed in Table 5.

The anti.anti-3,5-Bishomocycloheptadienyl System. Heating of 39 with aluminum isopropoxide in a 2-propanol-acetone solvent system resulted in epimerization and gave 132—OH as the major component (58%).

Isomer separation was achieved by chromatography on silica gel and the composition of the mixture (58:42) was established as the equilibrium 62 distribution by similar reaction of the pure, anti,anti alcohol. In contrast to 3d where the quasi-axial functional group is predisposed for anchimeric assistance by the equatorial cyclopropane ring, similar neighboring group participation appears less likely for 132. If the prevailing conformations are indeed 158a and 158b, then neither internal cyclopropane bond attains a geometric relationship relative to the tosylate group which permits realization of that orbital orientation

(cf 126) necessary to aid departure of the leaving group. However, long-range cyclopropyl participation in 132 would not be altogether precluded if a more shallow tub conformation were adopted. On this

OTs OTs

basis, operation of the presumed low energy pathways open to 13(3 and 131 would not necessarily be fully deterred, but the possibility exists that solvolytic pathways not encountered in the earlier examples could now become operational. One of these may, of course, merely be preliminary ionization to the unrearranged secondary carbocation.

The acetolysis of 132 proceeded as smoothly as that of its isomers, the rate constants denoting a reactivity level at 40° approximately double that of 130 (Table 4) . Four products were generated in high yield, two of which were hydrocarbons and the others acetates. The spectra of the acetates require that they be lTL-OAc (9%) and 15!5 (55%).

The hydrocarbons were similarly readily identified as syn-3,5-bishomocycloheptatriene (48/ 25%) and cis -1,4,7-cyclononatriene

(2S.t 11%) • This product distribution clearly discounts the possibility that tosylate 132 ionizes to exactly the cationic intermediate formed from 131.

The solvolytic conversions of the syn,anti- and syn,syn-3,5-bishomo- cycloheptadienyl tosylates 130 and 131 to the epimeric acetates 140 and

155 are, within the limits of our analytical method, completely stereo- specific. As little as 0.3% intercontamination could have been detected but was not. In both instances, complete stereochemical control is maintained during twofold cyclopropane ring cleavage with solvent capture occurring from the opposite and same surfaces, respectively, as that utilized by the departing tosylate ion. These findings are consistent with, but do not by themselves require, utilization of the secondary

C3-C4 bonding cyclopropane electrons in synchronous backside partici­ pation. What is strictly demanded is that no rotation or other conform­ ational perturbation occur about C4 prior to rupture of the c5-cg bond and charge annihilation at C^. This would certainly be reasonable if molecular rigidity were gained by long-range interaction to C^, to be followed by stereospecific rearrangement of the rather unique cyclo- propylcarbinyl cations 159_ and 160. Scheme VIII illustrates simplified versions of this analysis. 64

5

HOAc

OTs

HOAc H 160

From the nature of the acetate obtained from solvolysis of deuterium labeled tosylate 149,/ it is clear that the syn,anti-3,5- bishomocycloheptadienyl system does not undergo rate-controlling ioniz­ ation to secondary cation 139. Rather, the stereospecific formation of

150 supports intervention of bishomoallyl cation 159^ rather than evolution of this symmetrical intermediate. A not unreasonable assumption is that contributions from unsymmetrical ion pairing may also retard full equilibration to 139. Since 130 ionizes at about the same rate as

132 whose geometry is not particularly conducive to anchimeric assistance, rigorous analysis of the kinetic data is made somewhat complicated.

There is little doubt that the ground states of both 131^and 132^ are thermodynamically less stable than that of 130^because of the structurally enforced requirement that one of their cyclopropyl appendages be project­ ed axially. This factor may be chiefly responsible for the somewhat

reduced AH* values of 131 ^ W »K/ and 132 AAAJ as compared to 130. But if 130 ~ — — - and 131 do solvolyze with anchimeric assistance (permissible on the basis

of product formation), how is the ^acetolysis for 13Jj to be assessed?

The ionization of 132 seemingly does lead in part (55% of 155) to

a cation similar to 160 but with anti tosylate departure. This reaction

channel could reflect cyclopropyl participation from a rather shallow

tub conformation as previously discussed. The nature of the other

three products indicates that delocalized intermediates of comparable

type are unimportant in their formation and that a classical electron-

deficient species (161) probably intervenes. Direct solvent capture

from the less hindered backside surface now provides 131, although the

possibility that this product arises by direct SN 2 displacement cannot

be summarily dismissed, while 3 proton loss leads to olefin 48. The

formation of 1,4,7-cyclononatriene C79) may be attributable to a 1,2-

hydride shift process generating 162 which by well precedented electronic

reorganization can release strain with simultaneous introduction of

alternating single and double bonds. Cyclononatriene ^79 may also arise

from subsequent protonation of ^48 during acetolysis. However, this

question was decided in the negative when ^48 was recovered unchanged

after resubmission to the original conditions. That 161 is capable of

hydride shifting provides important insight into the level of electron

deficiency which builds up at C^. In particular, this specific trans­ formation demonstrates by its absence during acetolysis of 130 and 131

that structural factors present in the latter tricyclic systems deny

comparable development of cationic character at C]_. On pragmatic

grounds, direct bishomoallyl cation intervention concisely rationalizes

this dichotomy. 66

H

^i6i 12

If the above analysis is valid, then the acetolysis rate constant for becomes a composite of k^ and kg terms. The extent to which

15 j> is produced (55%) can be taken to indicate that k^ is the dominant component. On this basis, the relative kinetic ordering of 132^ (Table IV) becomes comprehensible. Since differences in activation energies comprise the sum of ground state and transition state inequalities, the for

132 should parallel rather closely in magnitude the value experimentally determined for 131. This state of affairs arises because 131, and 132^ are merely epimeric at and the anchimerically assisted transition states (cf lf>0) are closely similar. Accordingly, one is tempted to conclude that cyclopropyl participation operates in all three isomeric systems, although to a reduced extent in 132^because of adverse orbital geometry factors. It is interesting to note in this context that both epimers of cis-bicyclo[5.1.0]oct-3-yl brosylate undergo acetolysis to give mixtures of products, the nature of which clearly reveals inter- 94 action of the cyclopropane ring with the cationic center. Importantly, the presence of two non-interacting cyclopropane rings in 130jl32^ would be expected to reduce their rates of ionization significantly below that of cycloheptyl tosylate due to adverse inductive effects, but this is not observed (Table IV). 67 These findings cause the converse viewpoint which would argue that all three tosylates solvolyze without anchimeric assistance and that cyclopropyl participation, if operative, occur only during rate-limiting ionization to be less than convincing. Under these terms, the single feature which would preclude 130, from becoming fully symmetric (cf 139), and the reaction manifolds of 131 and 132 from fully merging, is oriented ion pairing with solvent relaxation occurring more slowly than possible molecular skeletal vibrations. Although ion pairing is known to gain 90 importance in amine deamination processes, it has not acquired 95 comparable prominence in acetolysis reactions. But because the importance of metastable unsymmetrically solvated intermediates cannot be directly assessed, their role cannot be dismissed. We note, however, that because capture of acetate ion occurs at Cg and not in the immediate vicinity of tosylate departure, counter ion control of stereospecificity becomes rather unlikely. This conclusion is supported by the stereo­ chemical crossover observed with 130 and 131 (compare ^159 and ,160) • A further relevant issue concerns whether the counter ion is responsible for the inability of ,139 to become symmetric and of ,161 (as generated from the two epimeric tosylates) to relax fully to a common geometry.

In the latter instance, it is also obligatory that the level of charge buildup at also be controlled by solvation factors. Although the completely symmetric conformations of ,139 and ,161 may not coincide with the minima in the vibrational potential curves, there can be no doubt that these structures would be attained during a given molecular vibration. It therefore becomes necessary for solvent to trap these

— 1 1 nascent cations in less than ^ 1 0 sec. To the extent that this 68 criterion is unlikely, the hypothesis of direct cyclopropyl participation gains credence. The problem reduces to one of how the symmetry propert­ ies of nonisolable intermediates are to be construed in mechanistic studies.

There remains the need to consider the possibility that both cyclo­ propane rings open simultaneously leading to cations best represented by A&3. and it is instructive to consider first the syn,anti derivative 130.Molecular models (see 138) indicate that the prevailing geometry is rather unfavorable for full cyclopropane delocalization because the relevant orbitals are canted to opposite surfaces of the molecule. Such geometry has previously been recognized not to be 96 conducive to efficient electronic interaction. In the 131 and 132^ examples, an almost coplanar arrangement of the inner cycloheptane core is necessitated. Such conformations can indeed be attained and extended cyclopropane - cyclopropane interaction seemingly gains meaningful importance during the ionizations of 131 (but not 132) if the results obtained by deamination provide serviceable analogy.

When tosylhydrazone salts are irradiated in alkaline aqueous or alcoholic solutions, the diazoalkanes which result experience ready in situ protonation, the procedure virtually ensuring kinetic control by the attacking solvent nucleophile. Such an experimental approach was deemed 69 ideally suited to the stated problem and accordingly was utilized in 97 this study.

The known ketones 49, and 4£, readily afforded their respective cry­ stalline tosvlhydrazones 165 and 166. For the preparation of anti-2,5- bishomocycloheptadienone (168), olefin 3^ was successively hydroborated and oxidized with Collins reagent. The propensity of £4 for rapid conformational ring inversion at room temperature causes both surfaces of the ir bond to be equally accessible to the electrophilic reagent.

For steric reasons, however, exo attack on 34a and 34b is assumedly favored. Conveniently, the inherent £2 symmetry of this hydrocarbon renders both pathways fully equivalent. Electronic effects, in contrast, contribute substantially to the regioselectivity of addition. Thus, development of positive charge during hydroboration should favor carbon- boron bond formation adjacent to the axially disposed three-membered ring because of more favorable orbital interaction with the internal cyclopropane o bond in this alignment. That 167 is produced in > 95% purity is testimony to the control which conformationally distinguishable cyclopropane rings can exert on such processes.

Ketone 168a is also conformationally mobile, both faces of its carbonyl group being available for attack by nucleophilic reagents.

However, since 167 is formed preferentially {> 95%) upon lithium aluminum hydride reduction of 168a, attack on conformer B from the exo direction

is seen to be kinetically favored. This finding agrees with independent

assessments of prevailing steric factors made with molecular models. 70

«b*. V !&' R = o 1^6 , R =0 165, , R = NNHTs 166 , R = NNHTs

C?47 == isA

3ba / 3kb

i & & s. 168a, R = 0 l68b , R = NNHTs 71 Comparable hydroboration of the more conformationally rigid bishomocycloheptatriene 33 occurred with > 95% exo stereoselectivity to provide 169, oxidation of which gave 170a. Upon treatment with sodium borohydride, this ketone led expectedly to 169 and 171 in a ratio of 1:4.

H HO

HO

H

m. lJ0a,R = 0 b , R = NNHTs

When irradiated (450W Hanovia lamp, Pyrex) in anhydrous methanol

0.2 N in sodium hydroxide, 165 was converted to a three-component • /w srv mixture comprised of olefin 51 (3%) and ethers 172 (19%) and 173 (78%). A/Si rf’WWs, (WV

The structural assignments to 172 and 173 follow from their revealing

hv OCH 3 -h c h 3o h (N oO H )

a 172 173

^■H NMR spectra and independent synthesis from the known alcohols.

Photodeamination of 165 in aqueous base gave a reaction mixture containing

40 (16%) and 167 (84%), but no detectable amounts of 51. 72 Comparable photolysis of 166 in methanol was significantly more complex and provided nine detectable products. Preparative VPC separation and detailed spectral analysis revealed several of the constituents to have retained both cyclopropane rings. These were iden­ tified as 173 (3%), 48 (12%), 174 (19%), 175 (11%), and 176 (22%). Three of the remaining ethers contained but one three-membered ring and these

hv 16 6 ------173 + CH^OH ^ (N aO H )

,och 3 OCH

'OCH H' m m m

were characterized as 177 (6%) , 178 (22%) , and 179 (2%) . The ninth component (3%) remains unidentified. Although alternative access to 177 could readily be gained by suitable chemical modification of 141 and 145, respectively, the isomeric bicyclic ether 178 required more extensive elaboration. For this purpose, 166 was photodeaminated in aqueous base and the product mixture directly oxidized. Product analysis established that ketones 4£ (17%), 168a (25%), 170a (27%), 180 (5%), and 181^ (26%) had been formed. The identity of 180 was revealed by its independent synthesis from 141 and 145, for the unequivocal preparation of 181, a stream of oxygen was bubbled through a warm sample of 1,5-cyclooctadiene 98 (115) and the intermediates directly reduced with sodium borohydride. r v w

By preparative VPC methods, pure alcohol 182 was obtained and 73 ^7°

39 cyclopropanated under Simmons-Smith conditions. The major component was separated and assigned the indicated anti stereochemistry (183a) in 99 line with existing precedent. Its methyl ether (183b) was prepared and found to be different from 178. Collins oxidation furnished 181. identical in all respects to the previously isolated ketone. Sodium borohydride reduction and methyl ether formation yielded 178 and 183b in a 62:38 ratio. The stereochemical disposition of the cyclopropane ring and methoxyl groups in this pair of ethers follows not only from the

l)02»A t CHoI OH Zn-Cu

m 182 , R = H b , R = CH, C r 0 3- 2py

1) NaBH 4 181 2 ) NaH, CH^I

method of synthesis, but from the relative shielding effects of the neighboring cyclopropane ring on the oxygen substituted methine C-H groups as well. 74 For relevant mechanistic reasons, it was important to eliminate

the chance possibility that bicyclo[6 .1 .0 ]nonenyl alcohols and ethers

structurally related to ketones 188 and 192 had been formed in small

amounts but not detected in the product analyses because of peak

overlapping, etc. It was instructive, therefore, to effect epoxidation

of bicyclo[6 .1.0]non-4-ene (llj) with m-chloroperbenzoic acid. The resulting pair of epoxides were separated under preparative VPC

conditions and individually subjected to ring opening with lithium diethylamide in ether. Molecular models show only 184 to have a reasonable capacity for conformational folding with resultant trans

diaxial orientation of the epoxide ring and a vicinal C-H bond. The

epoxide isomer which was returned unchanged under these conditions was

accordingly assigned structure 185. The allylic alcohol to result from

* &

HI

0 'OR 186a , R — H 187a, R= H 188 ~ T , R - C H 3 R = ch 3

184 is therefore considered to be trans isomer 186a, a conclusion

supported by its isolation as the major photooxygenation product of ITT

Both 186 and the second photooxygenation alcohol 187a underwent smooth

oxidation to 188 and raethylation to give 186b and 187b, respectively. 75

Using suitable VPC methods, it could be demonstrated that none of these three compounds had been produced during the deamination of 166 in either

solvent studied.

Treatment of alcohol 141 with thionvl chloride gave chloride 189, /■WWW reduction of which with sodium and tert-butyl alcohol in tetrahydrofuran gave hydrocarbon 190^in 80% overall yield. It should be noted that chloride mixture 189 could also be prepared by treating the tricyclic alcohols 89 and 40 with thionyl chloride at 0°. This is analogous to the behavior of ^130 and 131 during acetolysis. Singlet photooxygenation of 190 resulted in 90% conversion chiefly (> 90%) to one allylic alcohol whose NMR spectrum showed a combination of six allylic and cyclo- propylcarbinyl protons exclusive of the >CH0H signal. These features

require the substance to be 191 and rule out alternative formulation 193.

Collins oxidation of 191 led cleanly to ketone 192 which also could be

dismissed as a product of photodeamination-oxidation of 166 in aqueous base.

v V S0CI2 -- 'y^,C\ Na 0H J_-Bu0H ' l4l m I) hv, Og sens . 2) NaBH.

Cr03-2py

OH

m . 76 Light-induced deamination of 168b^in methanolic sodium hydroxide afforded hydrocarbon 34 (11%) in addition to the three ethers 173 (24%), IN/V

179 (11%), and (54%). In aqueous potassium hydroxide solution, the alcohols 167 (20%), 141 (10%), and 1 9 4 (70%) were formed, but hydro­ carbon 34 was not observed. To establish the carbon framework in 194a and demonstrate its structural relationship to 194b, the alcohol was both

RO

lgfca ,R = H b , R = CH3 oxidized and methylated. Ketone 168b and ether 194b^ were produced in high yield.

Finally, irradiation of 170b^ in alkaline methanol afforded trace quantities of olefin 33^ together with ether 176^ (99%). In water, 171 proved to be the exclusive product.

In the case of 168b,, the product distributions which result in methanol and water solution suggest that a mixture of epimeric diazonium ion-pair intermediates are formed, with anti protonation of the > 0 = ^ group occurring twice as frequently as the syn alternative. This partitioning, which is likely related directly to the anticipated conformational mobility of this diazo intermediate (see 168), assumes that tight ion pairs are formed which collapse sufficiently rapidly so that solvent capture occurs with retention of configuration. Were solvent-separated ion pairs involved, a much greater predominance of 173 would be expected as in the case of 165/ Significantly, the cyclopropyl- carbinyl carbonium ion resulting from loss of nitrogen, i.e. 136, is not particularly prone to structural rearrangement. Although low levels

{ ~10%) of conversion to cation 195 are operative (note stereospecificity

in the capture of this ion), no products were detected having structures

+ +

125.

corresponding to 134. We view this noninterconvertibility to be a reflection of the enhanced thermodynamic stability of 136 relative to

134, the energy gap separating them being adequate to insulate 136 from

134,even when vertical stabilization is likely operative. However, 136 is not stable enough to deter rearrangement to its homoallylic counterpart

195. It is of interest that the extent to which 179 (11%) and 141 (10%) are formed falls far short of the true thermodynamic position of equilibrium. Independent treatment of 167 with 10% hydrochloric acid

(50°, 20 min), for example, resulted in ready conversion to bicyclic alcohol 141 (90%). Thus, the deamination product distribution is kinetically controlled.

Because diazo compound 196 which is formed from 170b is conform- ationally rigid, highly directed solvent protonation from the anti direction is anticipated. Its decomposition through ion-pair intermediate

197 with capture by the geminate methoxide or hydroxide ion shortly 78

agg. i21

after formation of cation JJtXwill lead to 176 and 171. respectively.

Since no other products (except for barely detectable levels of deproto­ nation) are formed, the stability of 137^ at least prior to relaxation is sufficient to preclude electronic reorganization. This result is once again an obvious frustration of thermodynamics, since alcohol 171 is rapidly isomerized to 1^45 (95%) in 10% hydrochloric acid (50°, 20 min) .

The absence of syn-3,5-bishomocycloheptadien-l-yl derivatives provides insight into the disinclination of 137 to transform to 135.

The behavior of 165^ under photochemical deamination conditions is such that solvent capture without skeletal rearrangement was prominent in both media (16-19%) as reflected in the isolation of 172 and 40.

Clearly, the most important process (78-84%) consists of skeletal rearrangement and solvent trapping with full stereoselectronic control.

The lack of involvement of the second cyclopropane ring is particularly noteworthy. This behavior is a dramatic departure from that observed during tosylate acetolysis, and could speak to the decreased level of molecular distortion required to achieve stabilization of the positive charge under vertical ionization conditions. To conclusively establish this point, however, it becomes imperative to recognize possible differences between tosylate solvolysis in acetic acid and deamination in 79 methanol or water. The far higher nucleophilicity of the latter solvents can conceivably lead to interception of a carbonium ion sequence at an earlier stage. More specifically, reaction of solvent with a carbonium ion intermediate in which delocalization to one cyclo- propane ring has occurred can be faster in methanol or water than the conformational changes required for delocalization to the second ring.

For these reasons, the solvolysis of 130 in methanol was invesigated and observed to give rise to olefin 51 (1%) and ethers 172 (19%), 173 (69%), and 179 (11%). The similarity of this product distribution with that derived from the deamination studies is most striking; the contrast with the acetolysis results is equally so. The behavior of 130 can again be best explained by principal ionization to cation 159^with simultaneous participation by the syn cyclopropyl group. Although there is some apparent leakage to 163 or 195 (compare the fate of 136), one can arrive at the plausible conclusion that cyclopropane - cyclopropane interactions do not gain importance in this particular system. Since this behavior is not seen in acetolysis, the increased nucleophilicity of methanol serves to intercept the sequential series of carbonium ion rearrangements after opening of the first cyclopropane ring (rate-determining).

Given the reactivity patterns of 166, the results of the photo- deamination of 165 can be best rationalized in terms of preliminary

conversion to cation (WV134 and its further reaction along three different pathways as noted above. The cyclopropyl participation (one of these groups only) which occurs subsequently in part leads to a new increase in thermodynamic stabilization (bisected cyclopropylcarbonyl cation generation) and is therefore favored. Should the structural features of 134 not be complicated by ion pairing and the like (rather unlikely - see above), then its axial symmetry would lead to participation by either three-membered ring. A cation of lesser symmetry would favor neighboring group assistance only by the syn cyclopropyl group. However, because the first-formed diazo compound also shares this symmetry pro­ perty, its protonation by solvent from top and bottom faces then becomes statistically equivalent and redistribution of any label would occur prior to loss of nitrogen. There exists, therefore, no convenient way to address this stereochemical question; fortunately, this point is of low relevance to the present study.

Product mixtures from the photodeamination experiments involving 166 were substantially more complex. The combined yields of 118, 174, and 175

(42%) reveal that interconversion of 135^with isomeric cations is less prevalent in methanol than in water (17% of 46). In line with prevailing conformational features, syn protonation of the diazo intermediate dominates by a factor of approximately 2 (assuming negligible incursion of Sjj2-type reactions) . The formation of 176^ (22%) and 170a (27%) denotes that 135 is capable of B-cyclopropyethyl ->■ cyclopropylcarbinyl interconversion as is 134. Under normal circumstances, conventional stereocontrol would be expected within both stereoisomeric subsets (134,

136, 195/, 135 , 137, 198) such that leakage from one series to the other becomes unlikely. This conclusion is generally valid for most acetolysis reactions, but vertical ionization manifolds can sometimes provide surprises as in the present example. Thus, the isolation of 173^ (3%) and

168a (25%) corresponds to leakage to the anti-2,5-bishomocycloheptadien-

1-yl cation (136). From molecular models, it is seen that syn homoallylic 81

m

■ H" • H' »b 162 200

m cation 198^experiences noteworthy nonbonded interactions not present in its anti conformer 195. If one makes the reasonable assumption that

198 as generated by twofold rearrangement of 135^is capable of relaxation by ring inversion and generation of 19J5, then reclosure of this last intermediate to give 137^ becomes a likely event prior to solvent trapping.

In fact, the ratio of 173, to 168a, which favors 168a^ by a factor of 8 can be accounted for in terms of the enhanced ability of solvent water to permit attainment of more optimal stabilization of positive charge prior to its annihilation by covalent bonding. Additionally, isolation of the epimeric ethers 177 and 179 from the methanol photodeaminations conforms to the presence of both 195, and 198, respectively, in view of the expectancy that these two cations should react stereospecifically with solvent. 82 The formation of 178 and 181 denotes that this portion of the product mixtures must arise from a hydride-shift pathway. Since neither

19JL, 198, nor 200^ can be reasonably implicated, this phenomenon which is unique to 166 is thought to arise by rearrangement of 135 to 198, a cation which is geometrically well disposed for ring opening to 199.

This intermediate is both cyclopropylcarbinyl and homoallylic in nature and could thus be thermodynamically attractive. In actuality, the

sequence■ 135 — «- — » -> 198 & -*■ 199 - — — - has previously been invoked as the primary stages •a in the transformation of cation 135 to cis -1,4,7-cyclononatriene under acetolysis conditions.

The solvolysis of 131 in methanol led to a markedly different ratio •* /w w v of products: 48 (4.4%), 174 (8.0%), 175 (0.6%), 177 (85.3%), and 179

(1.7%). The high percentage (87%) of compounds resulting from scission of both cyclopropane rings contrasts markedly with those results realized upon deamination of 166 (8% in MeOH; 5% in f^O). Additionally, there was found no evidence for the formation of 3/76 a principal product of diazonium ion decomposition in methanol. Nor was hydride shifting operational. Furthermore, the stereoselectivity attending production of the epimeric ethers 177 and 179 (50:1) substantially exceeds that realized from 166 (3:1). These composite observations provide convincing evidence for the direct generation of 139 upon solvolysis of 1J/L.

One general conclusion to be drawn from the present experiments is that tosylhydrazones can provide ready access to intermediate carbonium ions not accessible by solvolysis. When the relative orientation between the developing cationic center and the proximate cyclopropane rings can be varied substantially as it can in the bishomocycloheptadienyl series. 83 a most critical dependence upon their mutual geometry is clearly evident.

In solvolysis, the level of neighboring group participation reaches a maximum, with the result that high stereospecificity during conversion to extensively rearranged products is observed. When an all-cis arrangement is present as in 131,, ionization by solvolysis results in all-out participation by the pair of three-membered rings. Inversion of stereochemistry of one cyclopropane ring (e.g., as in 130) restricts delocalization only to the syn oriented group. In the vertical ionization manifolds reached by deamination, structurally less encumbered cations are first formed. Greater complexity in product formation then develops but generalizations cannot be transferred from system without adequate consideration of geometric considerations. Thus, whether neighboring cyclopropane rings experience direct involvement with a developing cation center or not can be controlled by conformational factors as well as by the method of generation. As shown above, the specific reaction type employed can lead to interception of a series of carbonium ion or ior pair intermediates at different points in the reaction manifold and shed light on the sequential or simultaneous participation of cyclopropyl groups. Chapter IV.

Functionalization of the Bishomocycloheptatrienes. Studies of

Cyclopropyl Halide Solvolysis and 1,3-Bishomotropylium Cations.

84 85

INTRODUCTION

The concept of homoaromaticity became an experimental reality 20 years ago with Winstein's study of the solvolytic behavior of the bicyclo-

[3.1.0]hex-3-yl cation. Since this pioneering work, homoaromatic

systems have been the object of a great number of experimental and

theoretical investigations, the results of which are summarized in 102 several excellent reviews. A homoaromatic system can be defined as

a cyclic array of 4ii + 2 electrons in which the tt ribbon has been

fractured one or more times to accommodate at least one sp -hybridized

center. In order to maintain electron delocalization, the two tt

orbitals flanking the homobridge must become canted toward one another.

As a result, overlap becomes restricted to that surface of the molecule

opposite the bridging atom(s) and involves only single lobes of the

canted ir orbitals.

Homoderivatives of the tropylium cation have been systems of

continual interest in organic chemistry. Homotropylium ion chemistry

was first developed with Pettit's early observation of an unusual NMR

spectrum for the cation ^201, generated by protonation of cycloocta-

tetraene. Subsequently, the stereochemistry of protonation of

201 104 13 105 86 cyclooctatetraene, C NMR analysis of J201» and various metal „ _ 106,107 carbonyl complexes of 201 have been reported.

Other homotropylium species have been generated in a variety of ways. Protonation of suitable acceptors such as the keto compounds

fl*

202 OH

H

OH

205

H+

H

SI

108 109 110 111 202, 203, 204, and 205 gave the corresponding tropylium

ions which were observable by NMR. Electrophilic attack on cycloocta­

tetraene and its derivatives gives rise to products best explained by the 87 intermediacy of homotropylium species as illustrated for the addition of chlorine to cyclooctatetraene and the addition of chlorosulfonyl 113 isocyanate (CSI) to various monosubstituted cyclooctatetraenes.

More recently, electrophilic attack upon a cyclic precursor such as 209

_ / f y ci2

206

CSI +

207 208

114 has generated a homoaromatic cation. Solvolysis of an appropriate substrate such as 210 yields products best explained by a homotropylium . 115 ion.

FsSbOs o=s SbFc

S 0;

HOAc 05 210 .88 For those homotropylium species that have been observed by NMR, several features emerge that are shared in common. All of the ring protons in the charged species are highly deshielded relative to the uncharged precursors. The exo and endo protons of the methylene bridge show widely separated chemical shifts with the endo proton being more , . , , . 102 shielded. The chemical shifts of the protons in 201 are an illust-

rative example. 107b 13c analysis shows all of the basal protons 105 are involved in delocalizing the positive charge.

(ft 0.73) H (6 *0-7) H (6 5-13)

(6 122.2) (6 8.2T)H H (6 6 .U8 ) 144.7)

(6 8.57) H H (6 8.39) (6 lU3>2) 155*T)

The introduction of a second sp^-hybridized center into the tropylium system leads to the bishomotropylium cations. Three general

types of bishomotropylium ions are possible: the 1,2- (£11), 1,3- (£12),

and 1,4-bishomotropylium (£13) ions. ^-02c To no examples of

213 211 212

116 211 are known although they have been sought unsuccessfully. 102 Examples of 212 and 213 are both known. 89 The bridged 1,4-bishomotropylium cation 214a has been generated in superacidic media from a variety of alcohol precursors such as 215a, 117-119

*216a, n fi ri i • 217, — ^—» * ' and 221 Protonation and dehydration of 215b, and 119 216b similarly gave the monomethyl cation 214b, which was also 120 obtained by protonation of 218-220. NMR observation of these HO,

(6 7.38)

b, R = CH; H (6 8.23)

'H (6 6.40) OH H (6 7.52)

214a, R = H b, R = CH3

%^H + 216a, R = H

b, R = CH. p+ OH

221

219 220 cations has provided the evidence for homoaromatic character. Thus, the ring protons experience considerable deshielding consistent with 13 the delocalization of a positive charge. The C NMR spectrum of 214b is indicative of extensive charge delocalization, although a significant amount of the charge appears to be localized on the allyl cation portion 13 when comparison is made to the C NMR of 201, which shows a more 102 . . equitable deshielding of the basal carbon atoms. It is interesting 119 to note that solvolytic studies of exo- and endo-221-OPNB, 119 121 217, and 222 provide no apparent rate enhancement leading to

the intermediacy of 214.1, although 1,4-bishomotropylium intervention after rate-determining ionization cannot be ruled out.

A second bridged 1,4-bishomotropylium species was generated by 119. protonation of 22J3, to give 224 in a superacidic medium below -100°. 122 123 1 The pronounced deshielding of the ring protons as well as the 91 sizable shielding of the bridging methylene protons are entirely consistent with structure 224. The charge dispersal appears to be more uniform in this cation than in 214. Furthermore, the substantial

226

OH

4 5 225/22,6 rate ratio (10 ) signals that anchimeric acceleration can 119 operate in precursors to ,£24. Other electrophilic additions to 223 and its derivatives support the mechanistic premise that initial electrophilic attack at one of the ethylene carbons followed by a 1,2- 124-126 shift of the butadiene bridge leads to the delocalized species

The only example of an unbridged 1,4-bishomotropylium derivative in the literature is the cation generated by protonation of the cisoid 127 bishomotropone 227. The exo and endo methylene protons in 228 display widely different chemical shifts with the endo proton being highly shielded. In contrast, protonation of the transoid bishomotropone

229 does not give a homoaromatic species presumably because of adverse 127 electronic interactions from the trans cyclopropane rings.

1,3-Homotropyliurn species have been generated by protonation of 3 128 ci£ -bicyclo[6.1.0]nonatriene (230) and its anti-9-methyl 105 derivative to give cations of structure 231. Stereoselective electrophilic attack from the exo side of 230 was demonstrated by examining the reactivity of the syn- and anti-9-methyl derivatives.

The former is prevented by obvious steric reasons from attaining a 92

.+ H H

(5 183.4)

£ 2 1 231 folded conformation and in fact proved to be unreactive under conditions where the anti isomer was rapidly transformed to the bishomotropylium species. The NMR spectrum of 2^31 shows that although there is a pronounced ring current present, it is clearly weaker than in the monohomotropylium case.

Higher order homotropylium species have been sought, but have so far escaped detection. Solvolysis of £32-OPNB (short-lived conditions) 129 and exposure of ,232-OH, ^232-Cl, or 2^32-00*12 to superacidic conditions conditions 129,130 ^ave faiied to give the tricyclot5.3.1.0^,11]undeca-

2,5,8-trienylium cation (,233) . Only simple allylic character has been attributed to the charge delocalization observed. Winstein attempted to generated the perhomotropylium cation 234 by solvolysis of 235 and

H

131 its epimer. The corresponding rates of solvolysis, products, and lack of deuterium scrambling indicate that 234 is not formed and that delocalized intermediates are not important. 93 k3tX’

sat s »

The primary goals of this work were the synthesis of functionalized bis- and trishomocycloheptatrienes and the study of cations generated from suitable derivatives under solvolytic (short-lifetime) and super- acidic (long-lifetime) conditions. 94

RESULTS AND DISCUSSION

Synthetic Considerations. The bis- and trishomocycloheptatrienes prepared in Chapter I can be readily classified into two categories for subsequent functionalization. The three trishomocycloheptatrienes

32,, 36,, and 52,, as well as the two 1,5-bishomocycloheptatrienes 33, and

3£, possess cycloheptyl methylene positions that are biscyclopropyl- carbinyl in nature. Such activation should be amenable to functional­ ization by free radical methods. The two 1,3-bishomocycloheptatrienes

48 and 5jL, on the other hand, possess cycloheptyl methylene hydrogens that are allylic. Standard methods of allylic functionalization should be useful for these systems.

The hydrocarbons 3J3, and 34, incorporate an interesting spectrum of unique structural features. For example, the conformationally rigid

33^ where both cyclopropyl groups adopt the thermodynamically favored extended "equatorial" arrangement, constrains H2 and Hg to lie in plane with the pn orbitals located on C3 and C4 and projects H7endo into a

endo. 7 endo 7 exo

bisected relationship with both three-membered rings (see . In anti isomer 34,, Hj- remains well aligned with the C3-C4 tt bond, but H2 is now almost orthogonal to that plane (see 237a). Also, while H?endo bisects the 5,6- fused cyclopropane ring, Hjexo is oriented properly 95 for maximum interaction with the "axial" three-membered ring positioned at Cj_,C2 - However, since 237a is in mobile degenerate equilibrium with

237b at room temperature (Chapter I) and these pairs of hydrogens experience rapid time-averaging under such conditions, all four can presumably attain a chemical environment capable of substantially enhancing their chemical reactivity.

Those forms of possible conjugative interaction available to £3 and

34 are therefore: (a) allylic, but with requisite abstraction of a cyclopropyl hydrogen (Hj or in £36, in 237a or H2 in 237b) to achieve delocalization; (b) biscyclopropylcarbinyl (Hyendo in 33 only); and (c) cyclopropylcarbinyl (H7encj0 or H7exo • In 311 effort to determine the extent to which these extrasymmetric factors affect the regioselectivity of free radical attack, £3 and £4 have been subjected to chlorination with tert-butyl hypochlorite. Strikingly specific abstraction of H2 or has been observed in both examples with formation of structurally unrearranged chlorides. This work concerns an analysis of this unprecedented selectivity and contrasts the structural integrity of the intermediate bishomocycloheptatrienyl free radicals to the lability of their carbocationic counterparts.

Upon reaction of simple hydrocarbons with elemental chlorine in the gas phase, a reactivity order for hydrogen abstraction of 3° > 2° > 1° is generally seen, but the quite low selectivity (maximum range of 5-6) 132 usually leads to production of complex mixtures. Low level C-H bond 134 elongation at the transition state has consequently been implicated, although various degrees of looseness in the activated complex seemingly 135 can become operational as hydrocarbon structure is altered. The 96 incorporation of cyclopropane rings into the molecular framework contributes to enhanced selectivity, especially favoring the cyclopropyl- carbinyl site. Thus, while Walling and Fredricks established that cyclopropane itself has l/20th the reactivity of a normal secondary 136—137 hydrogen toward Cl' at 0°, Roberts and Mazur showed that methy1- cyclopropane undergoes photochlorination predominantly via methyl hydrogen abstraction to give chloromethylcyclopropane and 4-chloro-l- 138-139 butene. 140 With more highly strained molecules such as bicyclo[2.1.0]pentane 141 and nortricyclene, chlorine prefers to enter into polar addition reactions. Olefinic substrates share this common problem. For abstraction of allylic hydrogens, the reagent of choice is tert-butyl hypochlorite, since tert-butoxy radical addition to double bonds does not prevail. Walling's extensive examination of the reactions of tert-

BuOCl with a variety of olefins has shown that an allylic hydrogen is 142 activated with respect to a corresponding saturated hydrogen atom.

More recent work by de Meijere's group has established that the cyclo- propylcarbinyl sites in trishomobarrelene (^38) and trishomobullvalene 143 (£3$) are specifically halogenated by this reagent. Furthermore,

Schallner has demonstrated that utility of fluorotrichloromethane 144 (Freon 11} as a low temperature solvent for such photochlorinations.

Irradiation of a cold (-63°) solution of 33 and tert-butyl hypochlorite in Freon 11 for 45 min provided a single monochloride (240) and a dichloride (241) as major products (85%) , together with several more extensively halogenated materials which were not further investigated.

Preparative VPC purification gave 240 and 241 in 46% and 6% isolated 97

tert -BuOCI

tert - BuOCI

CCL ,-2 0 °

2 ^ yields, respectively. When 240, was resubmitted to the conditions of chlorination, efficient conversion to 24^ (71% isolated) was realized. 13 Since the C NMR spectrum of 240, consists of 9 lines arising from one quaternary, five tertiary (including the two olefinic centers), and three secondary carbon atoms, substitution clearly had not occurred at a biscyclopropylcarbinyl site (Cj) , a cyclopropyl methylene position, or the double bond (C-j,C^) . Comparable analysis of 24^ revealed this dichloride to be a symmetrical molecule. Off-resonance decoupling established the five signals to arise from two quaternary, four tertiary and three secondary sites. These data are uniquely consistent with radical substitution at C.i. ,C, b or C_,C ^ 5 in the bishomotropilidene.

ci 98 A clean distinction between these possibilities can be made by analysis of the NMR spectra. In CDCl^, H-^exo and H7endo ^y^ro” carbon 33, appear at 6 2.47 and 0.32 as a doublet of multiplets (J = 14 Hz) and a multiplet, respectively. This rather characteristic pattern is unaltered in 240, (6 2.50 and 0.30) and somewhat more sharply structured in 241, [6 2.62 (d of t, J = 14 and 5.5 Hz) and 0.07 (d of t, J = 14 and

11 Hz)]. Therefore, only long-range or virtual H-H coupling is affected upon chlorination. The protons adjacent to C-j must consequently have remained intact. That chlorination had occurred at C2 and is further revealed by the reduced level of spin-spin interaction operating on the cyclopropyl methylene protons relative to 33.

With these assignments of structure, it is immediately obvious that the allylic cyclopropane hydrogens in 33^are most reactive. Furthermore, halogenative substitution has occurred without structural rearrangement, in line with the preestablished reluctance of cyclopropyl radicals to relieve strain through ring opening. 145

Submission of 34, to analogous reaction with tert-butyl hypochlorite gave 62% of a colorless oil composed predominantly (85%) of monochloride

242. Although attempts to purify 242^by VPC and column chromatographic 13 techniques usually led to rearrangement (see below) , C NMR analysis of the unrectified sample showed clearly that substitution had again taken place at a tertiary allylic cyclopropyl site (see Experimental

Section). The *H NMR spectrum shows the protons bonded to to be widely separated (multiplets at 6 2.40 and 1.62) but does not provide indication whether the favored conformer is 242a, or 242b. The olefinic protons of the chloride appear as a multiplet centered at 6 5.67, at slightly (

99

01 y y

242a 242b

lower field than in 34. The two cyclopropyl methylene protons which

are projected over the seven-membered ring have moved to higher field

with that presumed to be on the chlorinated ring appearing at 1.02 and

the other at 0.72. The remaining hydrogens resonate in the region of

1.52-1.08.

Molecule is so structured that the chlorine atom occupies a

pseudoequatorial position, while in 242b the halogen is projected in a

pseudoaxial direction and normally should be less favored for the usual

energetic reasons. For mechanistic reasons, 242b must be the initial

product of free radical substitution. Since 242 is not subject to

further chlorination as is 240, this may mean that conformer 242a (where

is now properly stereoaligned with the tt bond) is not present in

reasonable quantities at low temperatures. The inference would be that

242b is highly populated under these conditions. However, since the

argument rests on negative evidence, this conclusion is obviously

tenuous.

An interesting feature of the free radical chlorination of bishomo-

cycloheptatrienes 3^ and 34 is the regioselectivity of abstraction favoring

an allylic cyclopropyl hydrogen. The observed reactivity is viewed as

the combined result of structurally enforced geometry which fixes H2 or

1 100 Hc in plane with the pn orbitals and the resonance stabilization available 5 to free radicals 243 and 244. The customary bond strength and higher

H

2kk

electronegativity of a cyclopropyl C-H bond is thereby effectively lowered to the point where other possible reactions are hardly compet­ itive. This reasoning assumes that the transition states leading to

243 and 244 resemble products more than starting materials. In this way, the activated complexes can profit energetically from their proximity to the potential minima on the reaction profile.

The formation of tertiary chlorides 240 and 242 without structural rearrangement parallels earlier discoveries which have shown that cyclopropyl radicals normally fail to ring open. The closest analogy known to us is the vapor-phase chlorination of bicyclopropyl (245) which proceeds via cyclopropylcarbinyl stabilized radical 246^ to 247 with retained structural integrity.

Cl* Cl.

2h6 2kJ

Although the reaction of tert-butyl hypochlorite with 33, and 3jl gave fairly clean product mixtures, attempted chlorination of the 101 trishomocycloheptatrienes in an analogous manner gave complex mixtures of products. The NMR spectra of these product mixtures displayed a large percentage of olefinic protons which might arise from opening of the cyclopropane rings in a radical rearrangement. Mass spectral analysis was indicative of tert-butyl hypochlorite addition products as well as polyhalogenated compounds. Attempted halogenation of 32, 33, 34, 36, /s/v-* A/s/ A/V and 52 with N-bromosuccinimide or N-chlorosuccinimide gave similar results. Attempted photochemical bromination of the trishomocyclo­ heptatrienes with bromine gave only cyclopropane addition products.

The ability of singlet oxygen to react with olefins possessing at least one allylic hydrogen to give allylic hydroperoxides with allylic rearrangement has been well documented. Subsequent reduction of the hydroperoxides gives the corresponding allylic alcohols. We felt that the reaction of singlet oxygen with the hydrocarbons 48 and 51 might lead to allylic alcohols suitable for solvolytic work and studies in superacidic media.

When a solution of 51 in 9:1 methylene chloride-methanol was /SA/ irradiated at 0° in the presence of Rose Bengal as well as a steady stream of oxygen, complete consumption of starting material was observed by TLC after 11 hr. Sodium borohydride reduction of the intermediate hydroperoxides gave an 89:11 mixture of two allylic alcohols, assigned structures 248 and 249. The major component was obtained in pure form as a white crystalline solid by fractional crystallization from petroleum ether. The NMR chemical shifts of the protons alpha to the to the hydroxyl group are 5 4.03 for the major component and 5 4.62 for the minor component. According to shielding parameters for epimeric 148 cyclopropylcarbinyl alcohols described by Winstein, the alcohol with the more shielded alpha proton has that proton syn to the adjacent cyclo­ propane ring. On this basis, the major allylic alcohol from photo­ oxygenation of jy. is assigned structure 248. Examination of molecular models shows that the C-H cr bond of the endo hydrogen in /51^as

endo

250 represented by 250 is nearly parallel to the ir orbitals of the adjacent olefinic linkage. Preferred attack of singlet oxygen on the face of the molecule such that this hydrogen is eventually abstracted with an overall 147 allylic rearrangement would lead to 248 after reduction.

Oxidation of the epimeric alcohol mixture with manganese dioxide in refluxing cyclohexane gave a single product identified as enone 251.

The two olefinic protons of 2%X appear in the ^H NMR spectrum as a doublet of doublets of doublets (.J = 12, 7, 1.5 Hz) at 6 6.48 and a broadened doublet at 5.38 (J = 12 Hz). The carbonyl stretching frequency of 1635 cm”^ in the infrared and the ultraviolet absorption 103

251 at 263 nm (ethanol, e 4500) are consistent with the enone formulation

The conformation shown is assumed to be preferred since both cyclopropanes are in pseudoequatorial environments.

Reduction of 2J5Jl with diisobutylaluminum hydride in ether at -78° 149 gave an 84:16 mixture of 249, and 248,, respectively. The major component of this mixture was obtained in pure form as a white crystall­ ine solid by fractional crystallization. This crystalline alcohol was identical to the minor component of photooxygenation. Examination of molecular models indicated that attack on the carbonyl of 251 from the pseudoequatorial direction ought to be preferred sterically. Thus, hydride delivery to this surface would project the hydroxyl group and the flanking cyclopropane ring into the same face of the molecule as in

£43-

Photooxygenation of 48 for 20 hr and reduction of the intermediate hydroperoxides in a manner totally analogous to the photooxygenation of

Jil gave a more complex mixture of alcohols. Careful high pressure liquid chromatography of the alcohol mixture on Florisil (elution with 65:35 hexane-ether)gave two product fractions. The more rapidly eluting

fraction contained two allylic alcohols assigned structures 252 and

253 in a 75:25 ratio. The NMR chemical shifts of the protons alpha to the hydroxyl group are <5 4.35 for the major component and 4.82 for 104

OH H OH

253 the minor component. On the basis of these chemical shifts, the major 148 component is assigned structure 252 and the minor component, 253

Further substantiation of this structural assignment was obtained by endo exo (✓* 1) 1o2 ------a = 252 2) NaBH„ HO

25^ examination of molecular models. The C-H a bond of the endo hydrogen is aligned nearly parallel to the prr orbitals of the adjacent olefin.

Attack of singlet oxygen on this face of the molecule, allylic rearrangement, and reduction of the intermediate hydroperoxide would be expected to give 252 as the major product, which is consistent with the 147 structural assignment.

The allylic alcohol mixture was not separable by standard chemical methods. Attempted fractional crystallization returned only oily products, while chromatography on silica gel or alumina gave only rearranged products. Similarly, VPC techniques yielded only rearranged alcohols. A successful separation of the two structures was achieved by converting the alcohols to their corresponding jD-nitrobenzoates and separation of 105 the resulting esters by fractional crystallization. In this manner, a pure sample of 252-OPNB was obtained.

The allylic alcohol mixture accounted for approximately 80% of the products. The remaining 20% appeared to be a mixture of epoxy alcohols

255. Mass spectral analysis indicated a molecular weight of 152 which

corresponds to a structural formula of CgH^O^. NMR ana^Ys^s this mixture indicated multiplets between 6 3.0 and 4.5 which are consistent with oxirane protons. The protons alpha to the hydroxyl group appeared at 6 4.00 and 3.75. The nature of these reaction products was not investigated further.

Oxidation of the mixture of 25^2 and 2J53 with manganese dioxide in refluxing cyclohexane gave a single product identified as enone 256^

The *H NMR spectrum of 256 displayed the olefinic protons as a doublet

256 257 of doublets at 5 6,33 (J = 12, 3 Hz) and a doublet of triplets at 5.61

(J 12, 0.8 Hz) . The carbonyl stretching frequency at 1650 cm * in the infrared and the ultraviolet absorption at 225 nm (ethanol, e 5200) 106 are consistent with enone 25j5.

Reduction of ,£§£, in ether at -78° with diisobutylaluminum hydride gave an 86:14 mixture of £53 and £52, respectively. The major component was obtained as a white crystalline solid by fractional crystallization.

This alcohol appeared to be identical to the minor component of photo­ oxygenation of 48 by NMR.

.In contrast to the preferred conformation of auti-bishomotropone

251 in which both cyclopropane rings can adopt pseudoequatorial positions,

256 must necessarily have one pseudoaxial and one pseudoequatorial cyclopropane ring. Molecular models suggest that nonbonded repulsions between the cyclopropyl methylene hydrogens should be greater in 2ET7 than in 256. Thus, on this basis, 256 should be the preferred conformation. Attack of the reducing agent from the pseudoequatorial direction in 256 would deliver 253 as the major allylic alcohol.

Further support for significantly different conformations of the two bishomotropones comes from their NMR and ultraviolet spectra. The difference in the chemical shifts of the two olefinic protons in 2J5JL is

1.10 ppm while the difference in 256^ is only 0.72 ppm. Furthermore,

Xmax is 263 nm in 251,, but is 225 nm in 25(5. These data indicate a much 150 greater degree of planarity in the enone segment of 251, than in 256,,

since conjugative overlap improves with increasing planarity.

Rearrangement Reactions of £40, and 242,. To achieve purposeful

rearrangement of 242, under purely thermal conditions, the neat chloride

was sealed in a glass ampoule and immersed in an oil bath heated at 150°

for 4 hr, NMR analysis of the product indicated that complete

rearrangement to 258, had occurred. Similar results could be achieved upon passing 242 through VPC columns heated to 180°, or by elution through Florisil with hexane (100% yield). Structural assignment to

242 follows from mass spectral, combustion, ultraviolet, and NMR evidence. For example, its electronic spectrum (in cyclohexane) is characterized by a single maximum at 232 nm (e 3100) as expected for 27 . • such a homotropilidene. In the NMR spectrum, the olefinic protons appear as a doublet of doublets (.J = 5 and 11 Hz) at 5 6.22 (1H) and a multiplet centered at 5.40 (2H) . Additionally, the -C^Cl protons are seen as a broadened singlet at 3.87, downfield of the allylic cycloheptyl methylene multiplet at 2.43-1.97. This pattern, as well as that exhib­ ited by the cyclopropyl hydrogens, is generally characteristic of this class of dienes and particularly 259.

OCH

c h 2o c h 3 OCHi 261 262

The greater solvolytic reactivity of 242 as compared to ^240 was made evident through experiments conducted in anhydrous methanol. Upon heating 242 in this solvent at 100° (sealed tube) for 4 hr, conversion to 108 a mixture of 260 (94%), 261 (4%), and 262 (2%) was observed; 240 remained unresponsive and could be recovered. However, at 150° 240 was transformed to ,260 (95%), 261 (4%), and 262 (1%) after 4 hr.

Treatment of either chloride with a slight excess of silver trifluoroacetate in methanol at room temperature for brief periods of time led to the same three methyl ethers, although in somewhat different ratios (for ,240, 73.5, 7, and 19.5%; for 2£2, 67.5, 15, and 17.5%).

In the case of ,260, its structure follows directly from its ultraviolet and NMR spectra. With the exception of the methyl signal shown by ,259 and the methoxylmethyl peaks exhibited by 260, their spectra are otherwise essentially superimposable. The proton spectra of 261 and 262 clearly define their gross structural features. To distinguish between the epimers, recourse to molecular models was made.

For 263, the preferred conformation appears to be that which places the methoxyl substituent telow the plane defined by the exo methylene

group. In the geometry given by A, the magnetic environments of the

terminal olefinic protons become rather similar and could reasonably

correspond to the broadened singlet absorption (2H) shown by 261 at

6 5.03. For 262, the same ring conformation appears again to be favored,,

thereby necessitating that the methoxyl group approach syn proton

rather closely (see B.). Operation of a through-space shielding effect would be expected to cause to appear at lower field than H& . In

fact, ether 262 shows and Ha to be widely divergent in their

chemical shifts. These protons appear as multiplets centered at <5 5.27

and 4.97, respectively. Although this basis for stereochemical

assignment is not totally unequivocal, it does appear entirely

reasonable. Independent treatment of 258 with silver trifluoreoacetate in

methanol afforded 260, 261, and 262 in the ratio 55: 20.5: 24.5. Since

260 proved to be stable to these reaction conditions, ethers 261 and

262 must be primary products in all the rearrangements examined in this

study.

In contrast to the stability of 243 and 244, the cyclopropane

ring comprising the seat of reaction during carbocation generation does

not remain structurally intact. Since the product compositions obtained

from direct methanolysis of these chlorides are identical, the inter­

vention of a common intermediate satisfactorily represented by allyl

cation 248 is implicated. The differences in the ratios of ethers 260-

^262 obtained in the Ag+ - assisted reactions are considered to arise

because of somewhat earlier transition states where less structural

reorganization has had time to occur. The isomeric differences between

243 and 244^ consequently become more apparent.

Qualitatively speaking, the solvolytic behavior of 240 and 242 is

somewhat enhanced relative to that of simpler cyclopropyl

halides. ^-^'152,153 Chloride 242 which necessarily must have one of

its cyclopropane rings axially disposed, is the more reactive as a

consequence of its higher ground state energy. The products of 110 methanolysis are consistent with electrocyclic ring opening concerted with ionization, as found previously for 1-vinyl-cyclopropyl 155 tosylate. Significantly, disrotatory ring fission synchronous with departure of the leaving group would necessitate that only low level positive charge density actually develop on the originally functionalized carbon, such charge distribution precluding the possibility of 157 efficient vertical stabilization by the adjoining vinyl group as m

Evidently, any stabilization which might accrue to 264 is inadequate to impede the facile ring opening giving rise to 263 which releases inherent strain while likely occurring with little or no energy 156,158 barrier.

In summary, the preceding results reveal that the tt bond of a vinylcyclopropane can activate an allylic cyclopropyl hydrogen to free 159 radical substitution but cannot reduce the barrier to (conrotatory?) ring opening of the three-membered ring despite the possible release of 160 ca 30 kcal/mol of energy. When cations are involved, essentially the reverse chemical response is seen. Cyclopropyl ring fission concurrent with ionization presumably obtains while delocalization of positive charge by comparable vertical stabilization is not of primary importance.

Studies of 1 ,3-Bishomotropylium Cations under Conditions of Long-

Lifetime. Protonation and dehydration of alcohol 252 in a solution of fluorosulfuric acid - sulfuryl chloride fluoride (FSO-jH-SC^ClF) and methylene chloride-d? at -140° (liquid nitrogen - iso-pentane slush bath) gives rise to a yellow solution which displays fairly clean Ill and 13C NMR spectra at -100° These spectral parameters are tabulated in 13 Table 6. The C NMR spectrum was indicative of a symmetrical species possessing 7 olefinic-like carbons and 2 saturated carbons- The NMR spectrum displayed 7 olefinic-like protons and 4 protons attached to saturated carbons. In fact, the NMR spectra are virtually identical to those of the cation generated by protonation of cis^-bicyclo[6.1.0]non- atriene. ^^,128 slight differences are attributable to small differences in solvent composition as well as differences between internal and external reference signals. The structure of this cation is best represented by formulation 231, although the Mobius-like transoid

231 265 105,128 configuration represented by cannot be strictly discounted, if analogy to the bridge 1 ,4-bishomotropylium species 214 and 224 holds 117-120,122,123 true. In the latter two homoaromatic species, the homobridges are constrained to the same face of the molecule such that transoid cations are precluded.

We had hoped that protonation and dehydration of the allylic alcohols 248 and 249 might allow the direct observation of 26E3.

Interestingly, when 248 was treated with FSOoH-SOCIF at -140°, a pale yellow solution was formed whose ^"H and NMR spectra at -100° were identical in all respects to the cation generated from 252 under super- acidic conditions. This somewhat surprising result further clouds the 112

Table 6. NMR Data for the 1,3-Bishomotropylium Cation.

5 s' f?F % V He / \ 4 4 / \ -WA- h d , «D -/ ' ' \ j t' ' \ » ' + * / H HC 0\ ^ y •jf / c ' 3 V + ' / 3 % HB' 1 h a

Chemical Chemical Proton Shifta,b Carbon Shift9 »c

9.08 1 181.4 h a 1 h b ,h b * 7.98 2,2' 165.9

Hc,Hc ' 7.18 3,3' 139.2d

7.00 4,4* 135.2d W 3.84 29.8 W 5'5 ' 1.89 W a6,ppm from TMS. ^Internal CHDCI2 as standard. cExternal CHDCI2 as standard. ^Interchangeable values. 113 issue of the exact structure of the 1,3-bishomotropylium cation; that is, whether structure 231 or 265 best describes the species.

It is interesting to note that the thermodynamic driving force for attainment of the 1,3-bishomotropylium cation is large enough to allow rupture of a cyclopropyl o bond followed by inversion of the two carbons.

Such would be the case with £48 generating £31, or 252 generating 265.

Hie 1,3-bishomotropylium cation can be viewed as the interaction of a pentadienyl cation with an ethylenic unit as shown for £66 and 267.

266 267

However, the species has thus far been generated by combination of an allyl cation with two a bonds as in the protonation and dehydration of

248 and 253 or by combination of a pentadienyl cation with one o bond 105 128 as in the protonation of £30. ' We felt that a good model system to show the direct interaction of a pentadienyl cation with a remote “5 ethylenic moiety might be the protonation and dehydration of cisJ-

2,4,7-cyclononatrienol (268). The alcohol was readily prepared by the 114 procedure of Winstein. Treatment of the alcohol with FSO^H-SOCIF at -140° gave a pale yellow solution whose NMR spectra at -100° were

identical to those described in Table VI . Theory predicts that most of the charge should be contained in the pentadienyl segment of the delocal- 119 ized cation. Unfortunately, all attempts to quench the cation with sodium methoxide-methanol at -78° returned only polymeric materials.

It seemed feasible to distinguish cations of the type ,233. and ,265 by attenuation of the amount of charge delocalized into the ring. As the thermodynamic driving force leading to the homoaromatic species is

decreased, the bond rupture and ring inversion observed leading to the parent species should cease. To this end, the NMR spectra of 253. and

256 and those of their conjugate acids generated in FSO3H-SOCIF were

compared. The positive charge created by protonation of the carbonyl

oxygens should by localized fairly extensively on the enone segment of

the molecule.

The NMR spectral data for unprotonated and protonated 256, is

compiled in Table 7 and that for protonated and unprotonated 251,/ in

Table 8 . It should be noted that these NMR spectral parameters are

constant over the temperature range of -100° to -65°, and that quenching

of these cations with sodium acetate returns the starting enones 256,

and 251. Although the complexity of the NMR spectra makes assignment

of the protons H^,, HQ , HE , and Hp somewhat nebulous, the ene protons

and the cyclopropyl methylene protons are readily assigned. The olefinic

protons of are separated by 0.72 ppm. Protonation increases this

separation to 1.55 ppm. In 256, the cyclopropyl methylene pairs, Hq and

H-r and H„ and H-, are separated by 0.77 and 0.32 ppm, respectively. In Ir H J 115

Table 7, NMR Data for Unprotonated and Protonated 256.

OH

Chemical Shift a

Proton Unprotonated ^ Protonated

Ha 6.33 8.39 Hfi 5.61 6.84 HcfHp 2.08 3.39 HE ,Hp 1.53 2.31 Hq j Hh 1.02 2.31,2.87 Hj 0.70 1.70 Hj 0.25 0.76

Chemical Shift a

Carbon Unprotonated Protonated A5_

Cx 202.3 212.5 10.2 C 2 127.6 117.8 -9.8 C 3 136.6 168.4 31.8 C4 19.8d 35.8d 16.0 C5 14.4d 27.2d 12.8 C6 17.ld 19.9d 2.8 C7 33.0d 35.8d 2.8 C8 11.8e 12.3^ 0.5 C9 15.7e 12.86 -2.9 a5, ppm,from TMS. ^Internal TMS as standard. cExternal TMS as standard. d 'e Interchangeable values. 116

Table 8. nmr DATA for Unprotonated and Protonated 251.

— H:

OH

Chemical Shift a

Proton Unprotonated ^ Protonated

6.48 8.32 HA 5.38 6.29 % HC »HD 1.92 2.98,3.35 h e »h f 1.47,1.70 2.98 H[g 'h h 1.33 2.57 Ht 1.33 2.24 H. 0.6S 2.24

Chemical Shift a

Carbon Unprotonated Protonated A6

202.2 213.3 11.1 125.0 114.8 - 10.2 143.9 180.3 36.4 28.4 41.3 12.9 23.0 29.0 6.0 17.3 22.9 5.6 25.8 35.6 9.8 8 23.1 36.9 13.8 17.9 33.8 15.9 a6 , ppm from TMS. bInternal TMS as s t a n d a r d . cExternal TMS as standard. ^ ,e Interchangeable values. 117 the conjugate acid, this separation actually increases to 1.55 and 1.07 ppm, respectively, with Hj and being fairly shielded (5 1.70 and 0.76).

In bishomotropone 2151, the separation of the ene protons is 1.10 ppm.

Protonation increases this value to 2.03 ppm. The cyclopropyl. methylene protons are quite similar in chemical shift with the separation in the conjugate acid being only 0.33 ppm. From the A6 's observed for the ene protons, the positive charge appears to be more highly localized on the enone segment in protonated 251 than in protonated £56- Furthermore, the difference in chemical shift of the "internal" and "external" cyciopropyl methylene protons (1.02-1.55 ppm) in protonated £56 more resembles the separation observed for Hg and Hp in the parent bishomo- tropylium species (1.95 ppm) than does this difference in protonated

251 (0.33 ppm). This data is indicative of a modest ring current in protonated 256_, but not in protonated 251.

The *3C NMR data is also indicative of more extensive charge delocalization in protonated £56 than in protonated 251. Although conformational changes might affect the magnitude of the AS's observed between protonated and unprotonated forms, several differences between

251^ and 256 are apparent. The separation in chemical shift between C2 and C3 is 50.5 ppm in protonated £56, an increased separation of 41.6 ppm relative to the unprotonated form. For protonated 251, the C2~C3 separation is 65.5 ppm, an increase of 46.6 ppm relative to unprotonated

251. Although the charge is significantly localized on the enone in both protonated forms, it is obviously more so in 251^. The cyclopropyl methylene carbons of 251 are deshielded by a total of 30 ppm upon protonation. For 256, these carbons are actually shielded slightly upon 118 protonatiorx, which is indicative of greater charge through the 7 basal carbons. Unfortunately, unambiguous structural assignments cannot be made by NMR to C4 , C5 , , and which would address the magnitude 1 13 of a ring current directly. However, the combined H and C NMR spectral data are indicative that protonated 256 is best represented by 269, while

OH OH OH

269 270 271

protonated 251 is more aptly depicted by 270, or 27^.

Although the ring current in 269^ is greatly diminished relative to the parent cation, the data infer 231, to be the most reasonable representation of the 1 ,3-bishomotropylium cation.

Solvolytic Studies of 1 ,3-Bishomotropylium Cations (Short-Lifetime).

The j>-nitrobenzoate esters 272,-275, contain a variety of structural features for studying 1 ,3-bishomotropylium cations under solvolytic

OPNB OPNB OPNB OPNB conditions. These esters were readily prepared from the corresponding alcohols with £-nitrobenzoyl chloride in pyridine and were individually solvolyzed in both 80% aqueous acetone and 75:25 methanol-tetrahydro- furan. The structural features and results of solvolysis are detailed on an individual basis below. The abbreviated nomenclature utilized in this section first describes the stereochemical relationship of £-nitro- benzoate group with the flanking cyclopropane ring followed by the stereodisposition of the two cyclopropane rings.

The syn,anti-2,4-Bishomotropylium System. The stereodisposition of the cyclopropane rings in restricts possible participation during solvolysis to the three-membered ring flanking the £-nitrobenzoate group.

The remote cyclopropane ring is not favorably disposed for participation during solvolysis or for conjugative interaction with the second three- membered ring.

Titrimetric solvolysis rates for 272 were measured in 80% aqueous acetone by the aliquot method and good first-order behavior was exhibited throughout. The kinetic data and activation parameters are sunanarized in Table 9. Since the infinity titers invariably agreed well with the calculated values, internal return to a less reactive

£-nitrobenzoate was not an issue. The product mixture was comprised of alcohols 248 and 249 in a ratio of 88:12, respectively. Solvolysis of

272^in 75:25 methanol-tetrahydrofuran gave a product mixture comprised of a single methyl ether identified as 276. The independent synthesis of 276 was realized by treating alcohol 248 with sodium hydride in tetrahydrofuran followed by quenching of the alkoxide with methyl iodide. Table 9. Kinetic Data for Solvolysis in 80% Aqueous Acetone.

. 48.90 Compd Temp, °C k , a sec”^ AH^, kcal/mole AS+,eu ii rel

m 33.9 1.95 X 10-5e 40.6 4.49 X 10”5 48.6 9.91 X 10"5 21.2 ± 0.7 - 11 ± 2 1.6 48.9 1.09 X 104b m . 34.1 5.09 X 10~5_ A 41.6 1.07 X 10 4 18.8 ± 0.3 - 17 ± 1 3.2 49.9 2.42 X 10"4 48.9 2.18 X l0-4b m 34.9 3.03 X 10"5 22.1 ± 0.4 - 7 ± 1 2.3 48.9 1.53 X 10-4

34.9 1.32 X 10-5 42.4 3.34 X 10“| 22.6 ± 0.6 - 8 ± 2 1 48.9 6.80 X 10-5 aAverage of duplicate runs."Interpolated value based on the activation parameters.

H M O 121

OCH

. . 276

The absence of alcoholic products in the methanolysis is indicative of alkyl oxygen cleavage in the solvolysis process.

The anti,anti-2,4-Bishomocycloheptatrienyl System. The mutual stereodisposition of the cyclopropane rings in 273 is identical to that existing in 272. However, the opposite orientation of the leaving group allows only the possibility of the remote cyclopropane ring becoming involved during solvolysis.

Determination of the solvolytic rate constants in 80% aqueous acetone revealed 273 to be well behaved kinetically and to be twice as reactive as 272^ at 48.9° (Table 9). When conducted on a preparative scale, the solvolysis gave an 89:11 mixture of 24J3 and 24J3, respectively.

Methanolysis gave 27J5 as the only product.

The anti,syn-2 ,4-Bishomocycloheptatrienyl System. Neither cyclo­ propane ring in 274 is favorably aligned to participate during solvolysis /WWW of the leaving group. However, the syn orientation of the cyclopropane rings is favorable for conjugative interaction.

The solvolysis of 274 in 80% aqueous acetone was as well behaved as those of ^272 and 273the rate constants denote a slightly greater reactiv­ ity than 272 at 48.9°(Table 9). The product mixture, when solvolysis was conducted on a preparative scale, was composed chiefly (> 95%) of a single alcohol identified as 2,4,7-cyclononatrienol (2^68). A second minor component was not investigated. Methanolysis of 274 gave only methyl ethers as products with 277 predominating (> 95%). A second minor component was again not investigated The independent synthesis

277 of 277 was achieved by treating 268 with sodium hydride in tetrahydro- furan followed by quenching with methyl iodide.

The syn,syn-2,4-Bishomocycloheptatrienyl System. The cyclopropane rings in 275^are favorably aligned for direct participation by both rings during solvolysis as well as for conjugative interaction between the cyclopropanes. This orientation was felt to be favorable for possible anchimeric assistance during solvolysis.

The solvolysis of 275,in 80% aqueous acetone proceeded as smoothly as its isomers. However, 275 was the least reactive of all (Table 9).

The product mixtures from 275^ in aqueous acetone and methanolic tetra- hydrofuran were identical in all respects to those from 274_,

In order to contrast the products of kinetic control with those of thermodynamic control, the alcohols 248, 249^, 252^ and 2jjJJ, were treated with hydrochloric acid in aqueous tetrahydrofuran. In each case, the predominant product (> 95%) was 268. A second minor component was not investigated. 123 The kinetic data and thermodynamic parameters in Table IX imply

that cyclopropyl participation during solvolysis is not important in any

of the isomers. One might expect that the thermodynamic driving force

to generate an allyl cation is sufficient to "level" the relative rates of all four £-nitrobenzoate esters. Subsequent cyclopropyl participation

after an initial rate-determining ionization could then occur. Wiberg

has shown that solvolysis of £-nitrobenzoate or 3 ,5-dinitrobenzoate

esters of optically active medium-ring allylic alcohols need not proceed 161 to a symmetrical allyl cation. He also advances the rather

interesting idea that olefinic participation may be unimportant in the

solvolysis of these allylic alcohol derivatives. The thermodynamic

parameters contained in Table IX are quite similar to those determined

by Wiberg for the solvolysis of the £-nitrobenzoate of 2-cycloheptenol.

The nearly identical nature of the product mixtures from the

solvolyses of 272^ and 273^argues for a common intermediate leading to

products. One might implicate a symmetrical allyl cation such as 278^

2 7 9 278 280

although an examination of molecular models suggests that this may not

be the case. A symmetrical allyl cation would require a coplanar

arrangement of the seven basal carbons of 278^ which would introduce a

large amount of strain. The twisted allyl cations 279^and 280^on the

other hand, appear to incorporate a minimal amount of strain into the 124 molecular framework. Interconversion of 279 and 280 would pass through

2^8 as a transition state.

In order to address the question of the exact nature of the inter­ mediate cation leading to products from 272, and 273, the monodeuterio

£-nitrobenzoate 281, was prepared. Sodium borodeuteride reduction of

OPNB D(^)

OCH

281 282

251 and treatment of the resulting alcohol mixture with £-nitrobenzoyl

chloride in pyridine gave 281,. Solvolysis of 281, to give a static,

unsymmetrical allyl cation such as 279, should give products in which

the deuterium label is highly localized at the site of solvent capture.

Alternatively, the intervention of a symmetrical allyl cation (278)

or a rapidly equilibrating pair of unsymmetrical cations (279, 280)

Ought to give products in which the deuterium is evenly distributed

between the site of solvent capture and the remote carbon of the double

bond, leading to a 3:1 ratio of protons for olefinic to solvent capture

sites as in 282,. Methanolysis of 231, gave methyl ether products in which

the ratio of interest was 2.7:1. This result implies that a symmetrical

allyl cation or a rapidly interconverting pair of unsymmetrical cations

is the intermediate species. The observed products from 2JJi and 2JJ^

are consistent with either species. In a planar species such as 2JSb

solvent capture should be preferred at the terminus of the allyl cation

further removed from the cyclopropane projecting into the same face as 125 solvent approach. In a species such as 279, solvent capture from the pseudoequatorial direction should be sterically less demanding than pseudoaxial attack. Solvent attack in these ways would lead to 248 and

276 as the major products.

The similarity of the product mixtures from 274 and 275 again

implies a common intermediate. However, the difference in the nature of the product mixtures of 272,273 and 274,275 suggests that quite different

intermediates are involved in the two pairs. An initial allyl cation

28k 283 285

ia probably formed which may be symmetrical (283) or unsymmetrical

(284,285). Molecular models suggest that 284 and 285 minimize ring

strain while the planar 283 is highly strained. At this point, a

fundamentally important difference is observed between the planar,

symmetrical cations 278 and 283. In 278, the bent a bonds in the basal plane of the species project into opposite faces of this plane. Normal

conjugative interaction is not possible and a more highly delocalized

intermediate would have a Mobius-like nature. This is not observed.

In 283, the bent a bonds project into the same face of the basal plane.

A more extensively delocalized intermediate not only becomes possible

but highly probable in the form of the 1 ,3-bishomotropylium cation

(231), Solvent capture at either terminus of the pentadienyl segment PLEASE NOTE: This page not included in material received from the Graduate School. Filmed as received. UNIVERSITY MICROFILMS 127

EXPERIMENTAL

Infrared spectra were recorded on a Perkin-Elmer Model 467

spectrophotometer. The NMR spectra were determined with Varian

A-60A, Varian T-60, Varian HA-100, and Broker HX-90 instruments, and 13 apparent splittings are given in all cases. The C NMR spectra were

recorded on Bruker HX-90 and Bruker WP-80 instruments. Mass spectra were measured on an AEI-MS9 spectrometer at an ionizing energy of 70 eV.

Preparative scale VPC separations were performed on a Varian Aerograph

Model A-90-P3 instrument equipped with thermal conductivity detectors.

Analytical VPC separations were performed on a Hewlett-Packard Model

5750 instrument fitted with electronic integrator and flame ionization

detector. Microanalytical determinations were performed at the

Scandinavian Microanalytical Laboratory/ Herlev, Denmark.

Diehlorocarbene ^Addition _ to jCyc^heptatriene. Into a 1-S. three-necked flask equipped with a mechanical stirrer was placed 16 g (0.17 mole) of

cycloheptatriene, benzene (30 ml), 50% aqueous sodium hydroxide solution

(300 ml), and triethylbenzylammonium chloride (0.5 g). With efficient

stirring of this mixture, 70 ml of chloroform (104 g, 0.87 mole) was

introduced dropwise during 10 hr. After completion of the addition,

the mixture was poured onto 1 i of ice water, and the products were

extracted into chloroform (3 x 200 ml). The combined organic layers

were washed with water, dried, and evaporated to leave a residue which

was purified by chromatography on silica gel. Elution with hexane 128 followed by recrystallization of the properly combined fractions from methanol afforded 8.35 g (14.3%) of 27, mp 134-135° (lit mp 139-142°);

6.47 g (14.5%) of 28, mp 100-104°; and 10.2 g (22.9%) of £ , mp 53-56° 34 (lit mp 48-51°). .

For 27: NMR (CDC1,) 6 2.36 (m, lH), 1.80 (m, 6H), and 1.18

(m, 1H); u... (CHC1,) 2935, 1448, 1303, 1220, 1135, 1042, 1020, 898, U l o A J and 770 cm"'*'; m/e 338.

For ,28: 1H NMR, (CDCI3) 6 5.67 (s, 2H), 2.51 (m, 1H), 2.3-2.0

(m, 4H), and 1.10 (m, 1H); vmax (CHC13) 2995, 2940, 2875, 1650, 1448,

1210, 1130, 1086, 1025, 897, 840, 608, and 633 cm-1; m/e 256.

For ^9: 1H NMR (CDCI3 ) 6 5.85 (m, 2H), 2.58 (t, J = 5 Hz, 2H), and

2.25-2.05 (m, 4H) ; vmit (CHCI3) 2925, 1720, 1090, 1023, 851, and 750 cm”*-; m/e 256.

anti,anti-Trishomocycloheptatriene Q 2 ) . To a stirred solution of

4.6 g (0.20 g-at) of sodium in 150 ml of anhydrous liquid ammonia cooled to -78° was slowly added 5.20 g (15.4 mmole) of 27. and 7.40 g

(0.10 mole) of tert-butanol in 150 ml of ether. After 2 hr at this temperature, solid ammonium chloride was introduced in small portions until the blue color faded. The ammonia was allowed to evaporate overnight, water (100 ml) was added, and the organic phase was dried and evaporated. Distillation afforded 1.33 g (65%) of £2, as a colorless oil, bp 85-90° (30 mm); 1H NMR (CDCl3) 6 2.38 (m, lH), 1.0-0.3 (m, 10H),

0.10 (m, 1H), and -0.05 (m, 2H); 13C NMR (CDCI3) 32.47, 17.23, 15.32,

14.82, 14.22, and 12.19 ppm; V j ^ (neat) 3050, 2980, 2900, 1450, 1060,

846, 818, and 790 cm"^; m/<5 134. 129 Anal. Calcd for ®^,49; H, 10.51. Found: C, 89.32;

H, 10.56.

^syn-l,5-Bishomocycloheptatriene (33). a 384 mg (1.50 mmole) sample of 28 and 444 mg (6.0 mmole) of tert-butanol in 150 ml of dry — — ether was added dropwise to 345 mg (15.0 rag-at) of sodium in 15 ml of ammonia as described above. Molecular distillation (100°, 38 mm) afforded 115 mg (64%) of 33 as a colorless liquid; NMR (CDCI3) <5 5.55

(s, 2H), 2.40 (m, 1H) , 1.53-1.03 (m, 4H), 0.78 (m, 2H), 0.10 (m, 1H) , and 0.00 (m, 2H); 13C NMR (CDCI3) 128.9, 33.6, 17.4, 14.9, and 13.0 ppm; (neat) 3045, 2920, 2860, 1730, 1643, 1451, 1440, 1022, and max 708 cm~^; m/e^ 120.

Anal. Calcd for CgH^2 : C, 89.94; H, 10.06. Found: C, 89.71;

H, 10.19.

anti-1^5-Bishor^ Reduction of 2.56 g

(0.010 mole) of 29 with 3.00 g (40.6 mmole) of tert-butanol in ether

(60 ml) and liquid ammonia (60 ml) containing 2.30 g (0.10 g-at) of sodium metal provided after distillation 830 mg (69%) of 34^as a colorless oil, bp 86-88° (38 mm); NMR (CDCI3) 5.56 (m, 2H), 1.92

(t, J 4.5 Hz, 2H), and 1.40-0.50 (m, 8H) ; 13C NMR (CDCI3) 127.1,

27.5, 14.3, and 12.7 ppm; vmav (neat) 3050, 2990, 2820, 1650, 1450,

1440, 1020, 782, 708, and 696 cm-1; m/e 120.

Anal. Calcd for CgH12: C, 89.94; H, 10.06. Found: C, 89.95;

H, 10.23. composed of jg, (800 mg, 6.7 mmole), potassium tert-butoxide (2.0 g,

19.2 mmole), and 15 ml of pentane cooled in an ice bath was added dropwise a solution of bromoform (2.0 g, 7.91 mmole) in pentane (5 ml).

Upon completion of the addition, the mixture was stirred at room temperature for 4 hr before pouring into water (50 ml) and extraction into pentane (3 x 25 ml). The combined organic layers were washed with water, dried, and concentrated to leave after filtration through silica gel 1.68 g (83%) of 35 as a colorless oil; NMR (CDCI3) 6 2.47 (m, 1H),

1.86 (narrow m, 2H), 1.5-0.3 (m, 7H), and 0.16 (m, 1H); 13C NMR (CDC13)

38.6, 33.9, 29.4, 26.5, 16.6, 14.7, 13.1, 10.9, 9.9, and 8.0 ppm; vmax

(neat) 3000, 2916, 2848, 1448, 1030, 981, 940, 875, 852, 831, 789, and 727 cm”l; m/e_ 290.

anti,syn-Trishomocycloheptatriene Q 6) . A 1.60 g (5.48 mmole) sample of 35 was treated with sodium shot (2.3 g, 0.10 g-at) and tert-butanol (2.0 g, 27 mmole) in tetrahydrofuran (25 ml) at reflux

for 2.0 hr. The excess sodium was quenched with methanol. The mixture was poured into water and extracted with pentane (3 x 25 ml). The combined organic layers were washed with water, dried, and concentrated.

Preparative vpc purification (105°, 15 ft x 0.37 in. 5% SE-30 on

Chromosorb G) afforded 412 mg (56%) of ^6, as a colorless oil; NMR

(CDCI3) 6 2.37 (d of t, J = 14 and 6 Hz, 1H) , 1.55 (d of d, J_ = 14 and

9 Hz, 1H), 1.3-0.1 (m, 10H), and 0.1 to -0.1 (m, 2H); 13C NMR (CDCl)

26.98, 16.91, 15.54, 14.95, 14.65, 13.54, 12.41, 12.19, 11.60, and 5.26

PPm »* vmax (neat> 3068* 2999» 2913» 2855, 1451, 1021, 846, 832, 819, and 131 805 cm m/£ 134.

Anal. Calcd for C1(JH : C, 89.49; H, 10.51. Found: C, 89.30;

H, 10.62.

Cyclopropanation^of^3^ Into a 15 ml flask equipped with a condenser, rubber septum, and magnetic stirring bar was placed 5 ml of a 1 M solution of ethylzinc iodide in ether (5.0 mmole). Methylene iodide (1.33 g, 4.96 mmole) was introduced via syringe and the resulting solution was refluxed for 0.5 hr. A solution of 33^ (120 mg, 1.0 mmole) in 2 ml of ether was added in a comparable manner and the mixture was heated at reflux for 12 hr before being poured into 20 ml of cold saturated ammonium chloride solution. The ether layer and the single extract (10 ml) of the organic phase were combined, washed with saturated ammonium chloride solution, dried, and carefully concentrated.

Final purification was achieved by VPC methods (140°, 12 ft x 0.25 in.

10% XF-1150 on Chromosorb P). There was isolated 63 mg (47%) of 32 and

15 mg (13%) of unreacted

syn,syn- and syn,anti-3 ,5-Bishomocycloheptadienol (39^ and 40).

Methylene iodide (67 g, 0.25 mole) was added dropwise to 250 ml of 1 M ethylzinc iodide under nitrogen. The resulting solution was heated to reflux for 1 hr whereupon 6.0 g (54 mmole) of dissolved in 10 ml of -V- ether was added dropwise. After a 23 hr reflux period, the predescribed workup was employed to give a colorless oil which was analyzed by VPC methods (10 ft x 0.35 in. 10% SF-96 on Chromosorb G). Present was

6% of anti-bicyclo[5.1.0]oct-5-en-3-ol, 74% of 38, 8% of 40 and 14% of 132 The entire procedure was repeated to give 5.06 g (67%) of a colorless oil, bp 70-84° (0.75 mm). The isomeric alcohols were separated by chromatography on silica gel using pentane-ether (20:1)

elution. The ratio of / '♦39 v ' to 40 varied from 70:30 to 90:10.

For ,29: % NMR (CDCl-j) 6 4.05 (m, lH) , 2.36 (m, 2H) , 1.97 (s, lH)

1.50 (m, 2H), 1.08 (m, 2H), 0.73 (m, 4H), and 0.13 (m, 2H) ; 13C NMR

(CDCl-j) 71.4, 35.2, 13.5, 12.0, and 11.2 ppm; (neat) 3340, 3065,

1450, 1067, 1032, and 942 cm-1; m/e 138.1047.

Anal. Calcd for C9H140: C, 78.21; H, 10.21. Found: C, 77.92;

H, 10.46.

For 1H NMR (CDCI3) <5 4.10 (m, 1H), 2.13 (m, 2H), 1.08 (s, 1H)

I.27 (m, 2H), 0.72 (m, 4H), 0.00 (m, 2H); 13C NMR (CDCl^ 70.4, 37.3

37.2, 14.9, 13.9, 13.5, 13.1, 12.8, and 9.0 ppm; vmax (neat) 3340, 3064

2995, 2920, 1449, 1070, 1039, 1025, 994, and 730 cm-1; m/e 138.1047

Anal. Calcd for CgH140: C, 78.21; H, 10.21. Found: C, 78.04;

H, 10.25.

syn-3-Benzyloxybicyclo[5.1.0]oct-5-ene (£L). To a suspension of sodium hydride (750 mg of a 57% mineral oil dispersion freshly washed with pentane and tetrahydrofuran, 18 mmole) in 30 ml of anhydrous tetrahydrofuran was slowly added 2.00 g (16 mmole) of 38, dissolved in

8 ml of the same solvent. This mixture was heated at reflux for 6 hr before dropwise addition of benzyl bromide (2.8 g, 16 mmole) dissolved in 5 ml of tetrahydrofuran. Heating was continued overnight before addition of methanol (5 ml) and water (100 ml). The product was extracted with ether (3 x 30 ml) and the combined organic layers were washed with water, dried, and evaporated. Distillation yielded 3.36 g

(97.5%) of 41 as a colorless oil, bp 98-100° (0.10 mm); NMR (CC14)

6 7.32 (s, 5H), 5.76 (m, 1H), 5.42 (m, 1H), 4.52 (s, 2H), 3.72 (m, 1H),

2.32 (m, 1H) , 1.67 (m, 1H) , 1.4-0.5 (m, 4H) , and 0.20 (m, 2H) ;

(neat) 3062, 2998, 2920, 2858, 1660, 1496, 1455, 1360, 1091, 1070,

1028, 732, and 695 cm-1; m/e 214.

Anal. Calcd for ci5hiq0 : c » 84.07; H, 8.47. Found: C, 84.21;

H, 8.69.

Dichlorocarb^ne^Ad^ition^to^Jij-^ Chloroform (8.1 ml, 0.10 mole) was added dropwise to a stirred slurry of £1 (3.36 g, 15 mmole), benzyltriethylammonium chloride (0.10 g), benzene (5 ml), and 50% aqueous sodium hydroxide solution (30 ml). After being stirred overnight, the reaction mixture was added to water (100 ml) and the product was extracted with chloroform (2 x 30 ml). The combined organic layers were washed with brine, dried, and evaporated.

Chromatography on silica gel (elution with 15% ether in pentane) afforded 4.05 g (87.5%) of 42 as a pale yellow viscous oil; NMR

(CDCl3)

(m, 2H), 1.26 (m, 2H), 0.84 (m, 3H), and 0.18 (m, 1H); vmax (neat)

3070, 3000, 2930, 2860, 1450, 1343, 1070, 1027, 732, and 695 cm~^; m/e 296.

^syn^^ti-^^-Bi^honrocyj:lo h ^ t adienol (40). A solution of 42

(4.05 g, 13.6 mmole) dissolved in anhydrous ether (25 ml) was added dropwise to a magnetically stirred solution of sodium (2.6 g, 0.11 g-at) in liquid ammonia (25 ml) cooled to -78°. After 1 hr, the mixture was

allowed to warm to room temperature. After 4 hr, methanol (25 ml) was

added followed by water (100 ml). Ether extraction (3 x 30 ml), washing

and drying of the combined organic layers, solvent removal in vacuo, and

distillation gave 1.73 g (92%) of 40, bp 56-60° (0.3 mm). The spectral

properties of this colorless liquid were identical to those of the

material isolated earlier.

°1_ _ijmethy-1, tej^t-B u t y l ^ S i J ^

Dimethyl tert-butyl silyl chloride (5.44 g, 36 mmole) and imidazole

(4.90 g, 72 mmole) were added in rapid succession to a solution of 37

(3.30 g, 30 mmole) in 20 ml of dry dimethyl formamide under a nitrogen

atmosphere. After stirring for 24 hr at room temperature, the contents of the flask were poured into 250 ml of water, and the product was extracted with ether (3 x 50 ml). The combined ether extracts were washed with 5% hydrochloric acid, dried over magnesium sulfate, and concentrated. Distillation yielded 6.66 g (99%) of a colorless oil, bp 52-55° (0.02 torr); NMR (CDC13) 6 5.72 (m, 4H), 4.08 (quintet,

1H, J = 6.5 Hz), 2.43 (m, 4H), 1.07 (s, 9H), and 0.07 (s, 6H); UI c I a

(neat) 3015, 2960, 2860, 1608 (w) , 1255, 1080, 835, and 773 cm-1; m/e

224.1602.

Anal. Calcd for C13H240Si: C, 69.58; H, 10.78. Found: C, 69.70;

10.8 6 .

Dichlorocarbene Addition to 43.^ Chloroform (180 g, 1.51 mole) was added dropwise over an 8 hr period to a mechanically stirred slurry 135 of 43 (33.8 g, 0.151 mole), 80 ml of benzene, 250 ml of 50% sodium hydroxide, and 1.5 g of benzyltriethylammonium chloride. The resulting mixture was allowed to stir for an additional 8 hr at room temperature.

The reaction mixture was diluted with 600 ml of water. The product was extracted with methylene chloride (300 ml, 2 x 200 ml). The combined organic extracts were washed with water, dried over magnesium sulfate, and concentrated to yield a pale yellow solid. Recrystallization from methanol yielded 50.0 g (85%) of 44 as white needles, mp 80-81°;

^■H NMR (CC14) 6 4.15 (m, 1H) , 2.43-1.30 (m, 8H) , 0.95 (s, 9H) , and 0.12

(s, 6H) ; Vn,av (KBr) 2965, 2940, 2870, 1475, 1468, 1366, 1260, 1201,

1085, 1051, 1020, 978, 950, and 901 cm"1; m/e_ 398.

Anal. Calcd for C^gl^/jCl^jOSi: C, 46.17; H, 6.20. Found: C, 46.08;

H, 6.18.

syn^anti^,^-Bishoinocyclohepta^enol Dimethyl tert-Butyl Silyl

Ether (45). A solution of 44 (37.6 g, 0.094 mole) and tert-butanol

(30 g, 0.40 mole) in 200 ml of tetrahydrofuran was added dropwise to a rapidly stirred slurry of small pieces of lithium wire (6.9 g, 1.0 g-at) in 100 ml of tetrahydrofuran at a rate sufficient to maintain gentle reflux. After addition was complete, reflux was continued for 4 hr.

The reaction mixture was filtered through a pad of glass wool. The inorganic solids were washed with 600 ml of low boiling petroleum ether. The combined organics were washed with water (4 x 500 ml), dried over magnesium sulfate, and concentrated. Distillation yielded

22.7 g (95%) of 45 as a colorless oil, bp 110-115° (0.10 torr); NMR

1080, 833, and 771 cm-1; m/e 252.1912.

Anal. Calcd for C15H2gOSi: C, 71.37; H, 11.18. Found: C, 71.26;

H, 10.72.

syn,anti-3,5-Bishomocycloheptadienol (4£J. Tetra-n-butyl ammonium fluoride (65.3 g, 0.25 mole) was added to a stirred solution of

45. (20.4 g, 0.081 mole) in 150 ml of tetrahydrofuran under a nitrogen atmosphere. The resulting solution was allowed to reflux for 12 hr.

The reaction mixture was concentrated, and the residual oil was taken up in 300 ml of water. The aqueous mixture was extracted with ether

(3 x 75 ml). The combined ether extracts were washed with water, dried over magnesium sulfate, and concentrated. Distillation yielded 8.07 g

(75%) of dimethyl tert-butyl silyl fluoride, bp 35-40° (0.35 torr), and a second, viscous oil, bp 68-71° (0.35 torr). Recrystallization of the higher boiling product from pentane yielded 9.65 g (86%) of 40 as a white solid, mp 38-40°. The spectral properties of this product were identical to those described earlier.

syn-3,5-Bishomocycloheptadienone (46). Chromium trioxide (3.48 g,

34.8 mmole) was added in three portions to 5.50 g (69.6 mmole) of pyridine in 35 ml of distilled methylene chloride. After 30 min, 480 mg (3.48 mmole) of 39 in 5 ml of methylene chloride was added and stirring was continued for 0.5 hr. The solution was decanted and evaporated. The insoluble inorganic solids were washed with ether

(2 x 25 ml) and these washings were added to the methylene chloride 137 residue. A chalky precipitate was filtered and the filtrate was washed with 10% hydrochloric acid (2 x 50 ml) and saturated sodium bicarbonate solutions, dried, and evaporated. Molecular distillation afforded 450 mg (95%) of 46 as. a colorless oil; 1H NMR (CDC1,) 2.93 (dd, J = 16 and

5 Hz, 2H), 2.23 (dd, J = 17 and 7 Hz, 7H), 1.60-0.4 (m, 6H), and -0.10

(dd, J = 10 and 4.5 Hz, 2H); (neat) 3070, 3000, 2920, 1695, 1440,

1415, 1350, 1290, 1258, 1163, 1028, 830, and 810 cm"1 ; m/e 136.0890.

Anal. Calcd for CgH^O: C, 79.37; H, 8 .88. Found: C, 79.38;

H, 8.96.

J>yn--3^ 3-Bishomocj^^ (j^7) ' T o a lithium diisopropylamide solution (prepared from 0.42 ml (1.0 mmole) of njbutyllithium in hexane and 101 mg (1.0 mmole) of diisopropylamine in

5 ml of dry tetrahydrofuran) cooled to -78° was added dropwise a solution of 46, (136 mg, 1.0 mmole) in 1 ml of tetrahydrofuran. After

10 min, 173 mg (1 mmole) of chloro diethyl phosphate and 0.1 ml of

TMEDA were slowly introduced. The cooling bath was removed and stirring at room temperature was maintained for 2 hr. One ml of saturated ammonium chloride was added and the reaction mixture was poured into

25 ml of water. The product was extracted with methylene chloride

(3 x 8 ml) and the combined organic layers were washed with 10% hydro­ chloric acid (2 x 20 ml) and saturated sodium bicarbonate solution before drying and solvent removal. The residual pale yellow oil was filtered through a short silica gel column (elution with 40% ether in hexane) to give 240 mg (88%) of 4^7; 1H NMR (CDCl-j) 6 5.50 (br s, 1H),

4.13 (overlapping q, iT = 7 Hz, 4H) , 2.57 (m, 1H) , 2.35 (m, 1H) , 1.35 138 (t, J = 7 Hz, 6H) , 1.6-0.6 (m, 6H) , and 0.2 (m, 2H) ; m/£ 272.

syn-3,5-Bishomocycloheptatriene (48). To a solution of lithium metal (235 mg, 35. mg-at) in 20 ml of liquid ammonia cooled to. -78° was added dropwise a solution of 41 (1.09 g, 4.0 mmole) in 20 ml of ether.

After 2 hr at this temperature, solid ammonium chloride was added in portions until the blue color faded and the ammonia was left to evaporate overnight. Ether (20 ml) was added and this mixture was poured into water (50 ml). The ether layer was dried and carefully concentrated.

The residue was subjected to molecular distillation thereby yielding

410 mg (85%) of 48,was a colorless oil; XH NMR (CDCl3) 5.70 (br d, J =

10 Hz, 1H), 5.32 (m, 1H), 2.80 (m, lH), 2.42 (m, 1H), 1.6-0.5 (m, 6H), and -0.2-0.5 (m, 2H); (neat) 3070, 3000, 2920, 2850, 1657, 1450,

1260, 1026, and 830 cm-1; m/e 120.0941.

Anal. Calcd for CgH-j^: *-•' 89.94; H, 10.06. Found: C, 89.91;

H, 10.26.

B±shomocadienone^ Oxidation of 40 according to the procedure described above produced 49 in yields of 90%; -^-H NMR

(CDCI3) 2.50 (m, 2H), 1.98 (m, 2H), 0.82 (m, 6H), and 0.08 (m, 2H); vmax (neat) 3060, 2970, 2950, 2900, 1708, 1448, 1410, 1282, 1182, 1029,

840, and 803 cm-'*-; m/e^ 136.0890.

Anal. Calcd for CgH-^C^ C , 79.37; H, 8 .88. Found: C, 79.27;

H, 9.01. anti-3,5-Bishomocycloheptatriene (51). Conversion of 338 mg

(2.5 mmole) of <49, to its enol phosphate in the predescribed manner afforded 540 mg (80%) of ^50 as a pale yellow oil. Reduction of a 540 mg sample (2.0 mmole) of 5.0, as outlined above furnished 220 mg (92%) of J51 as a colorless oil after molecular distillation; ^-H NMR (CDCI3)

5.72 (m, 2H), 2.13 (m, 2H) , 1.4-0.6 (m, 6H), 0.42 (m, 1H) , and 0.17

(m, 1H); Vnjax (neat) 3070, 3000, 2930, 2860, 1645, 1450, 1025, 830,

804, and 693 cm-^; m/f2 120.0941.

Anal. Calcd for CgH-^: C, 89.94; H, 10.06. Found: C, 89.92;

H, 10.17.

syn,syn-Trishomocycloheptatriene (52). Methylene iodide (6.0 g,

22 traoole) was added to 3.0 g (46 mmole) of zinc-silver couple in 20 ml of anhydrous ether. This mixture was heated at reflux for 30 min, cooled, and treated dropwise with 200 mg (1.67 mole) of 51^ in 5 ml of the same solvent. After 24 hr at the reflux temperature, the ether solution was decanted into 40 ml of cold, saturated ammonium chloride solution. The organic layer was washed with brine, dried, and carefully concentrated. Preparative VPC purification (108°, 12 ft x

0.25 in. 10% XF-1150 on Chromosorb P) gave 50 mg (25%) of unreacted

51, 24 mg (11%) of 36, and 12 mg (6%) of 52, !h NMR (CDCl.) 2.37 (d of m, J = 14 Hz, 1H), 1.33 (m, 2H), 1.03 (m, 4H), 0.63 (m, 2H), 0.28

(m, 2H), -0.05 (m, 2H), and -0.38 (q, J = 4 Hz, lH); 13C NMR (CDCI3)

29.46, 18.29, 16.62, 12.63, 10.36, and 7.55 ppm; vmax (neat) 3065, 3000,

2920, 1475, 1020, 840, and 690 cm-1; m/e 134.1099.

Anal. Calcd for C, 89.49; H, 10.51. Found: C, 89.43;

H, 10.51. 140 Kinetics Procedure. A standard solution of 34, in benzene was prepared by dissolving 45 mg of the hydrocarbon and 25 mg of cyclooctane

(internal standard) in 6.0 ml of benzene (distilled from calcium hydride).

Aliquots (15 yl) were sealed in each of nine ampoules (constructed from

2 mm glass tubing and washed with hydrochloric acid and ammonium hydroxide before use) under a slight vacuum for individual runs. The sealed ampoules were introduced simultaneously into a constant temperature oil bath. After 5 min, the first ampoule was removed and immediately cooled while a timer was started. The kinetic data were determined by monitoring the disappearance of 34 with respect to the internal standard by electronic integration of vpc traces (Hewlett-

Packard Model 5750 instrument fitted with a flame ionization detector;

1/8 in. x 8 ft column packed with 15% SE-30 on Chromosorb G, 50°).

Preparative^ Scale ^ e rmolyais^^f ^ 4 ^ A 120 mg (1.00 mmole) sample of 34 in 1.0 ml of benzene was sealed in a thick walled ampoule and heated at 180° for 5 hr. Solvent was removed by distillation. Analysis of the residual oil by vpc showed greater than 99% of a single component. Molecular distillation yielded 103 mg (86%) of a colorless oil whose spectral properties (ir, nmr) were identical to those of 3 authentic cis -1 ,3,6-cyclononatriene (60).

anti-1,5-Bishomocycloheptadiene-d4 (67). Sodium sand was prepared by shaking 2.30 g (0.10 g-at) of molten sodium in 20 ml of hot xylene.

The sand was allowed to cool and was subsequently washed with dry ether

(3 x 25 ml) and dry tetrahydrofuran (2 x 20 ml). Tetrachloro compound 141 29, (768 mg, 3.0 mmole) and tert-butanol-0-d (1.00 g, 13.3 mmole) were — dissolved in 20 ml of dry tetrahydrofuran and added to the sodium sand.

The resulting mixture was allowed to exotherm and then cool with stirring for 0.5 hr. The reaction mixture was filtered through a pad of glass wool into 100 ml of water, the inorganic solids were washed with 20 ml of hexane, and the aqueous solution was extracted with hexane (3 x 25 ml). The combined organic extracts were dried and concentrated, and the residual oil was subjected to molecular distillation (80°, 30 torr).

There was isolated 205 mg (55%) of 66, NMR analysis of which indicated

80% d^4 incorporation; NMR (CCI4 ) 5 5.50 (br m, 2H) , 1.93 (t, iJ = 5 Hz,

2H), and 1.43-0.70 (m, 4H); m/e 124.1192.

• Thermolysis of 66. A 30 mg sample of g6 was vaporized into a quartz tube packed with quartz chips maintained at 330° and 50 mm (slow nitrogen entrainment). The volatile components were collected in a trap cooled in a Dry Ice-acetone bath. NMR analysis denoted complete conversion to 67; (CCI4) 5 6 .0-5.2 (m, 6H) and 2.42-1.92 (br d, 2H); m/e 124.1192.

1,3,5-Cycloheptatriene-7,7-d2 (69J. The apparatus was free of ground glass joints and all glass pieces were carefully fire polished.

The entire apparatus was contained behind a plexiglass shield.

Dideuterio diazomethane was added to a rapidly stirred, refluxing slurry of benzene (312 g, 4.00 mole) and cuprous chloride (1.5 g, washed with saturated sodium bisulfite and dried under vacuum at 50°) by bubbling a stream of nitrogen through 200 ml of a 0.05 M (0.10 mole) 142 ethereal solution of d iazome thane-clj. The escaping gases were passed

through a potassium hydroxide drying tube before entering the benzene

slurry. The nitrogen was bubbled through the ethereal diazomethane-d2

until the yellow color disappeared. The reaction mixture was filtered

through Celite and concentrated to 10 ml by distillation through a 40 cm

Vigreux column. VPC (75°, 6 ft x 0.25 in 5% SE-30 on Chromosorb G) and

^■H NMR analysis showed 40% of the concentrate to be 69^ (3.7 g, 40%);

^■H NMR (Benzene) 5 6.42 (m, 2H) , 6.00 (m, 2H) , 5.12 (broadened d, 2H,

iJ = 10 Hz). Integration of this spectrum showed 96% incorporation of

two deuterium atoms; m/e 94.

CalculationDeuterium^ Isotope Incorporationi inThe mass spectrum of cycloheptatriene was determined at 70eV and gave the

following ion ratios. [For comparison, see S. Meyerson, J. Am. Chem.

Soc. , ££, 3340 (1963)):

M-4 M—3 M—1 M M-4-1 M+2 M = 9 0.002 0.094 2.10 1.00 0.059 0.016

The mass spectrum of cycloheptatriene-7 ,7-d2 (6^) was also determined

at 70eV and gave the following ion ratios:

M-3 M M+l M+2 M+3 M+4 M = 92 0.014 0.73 1.85 1.00 0.077 0.0035

It was assumed that the M-3 peak in the labeled compound arose strictly

from the presence of monodeuterio and unlabeled cycloheptatriene. From

the ratio of the M-3 peak in cycloheptatriene and the same ratio in the

labeled compound, the lower limit of cycloheptatriene-d2 is 85.2%;

cycloheptatriene-d^, 14.6%; and unlabeled cycloheptatriene, 0.2%. Dichlorocarbene Addition to Cyc^c^eptatriene-7,7-d^ ^Into a

100 ml three-necked flask equipped with a mechanical stirrer was placed

1.0 g (0.011 mole) of cycloheptatriene, benzene (5 ml), 50% aqueous sodium hydroxide (20 ml), and benzyltriethylammonium chloride (0.10 g).

With efficient stirring of this mixture, 9 ml of chloroform (13.5 g,

0.11 mole) was introduced dropwise over a period of 4 hr. After completion of the addition, the mixture was poured onto 250 ml of ice water, and the products were extracted into methylene chloride (3 x 25 ml). The combined organic layers were washed with water, dried over magnesium sulfate, and concentrated to leave a residue which was purified by chromatography on silica gel. Elution with hexane followed by recrystallization of the properly combined fractions from methanol afforded 0.40 g (11%) of 70, mp 136°; 0.53 g (19%) of /KV71, mp 103-104°; and 0.89 g (32%) of 72, mp 54-55°.

For Ay70: ^-H NMR (CCl*) 5 1.95 (narrow multiplet, 6H) ; m/e — — 340. For 71: XH NMR (CC1.) 6 5.70 (s, 2H), 2.13 (s, 4H); m/e 258. H --- For 7£: 1H NMR (CC14) 5 5.87 (m, 2H) , 2.10 (m, 4H) ; m/e_ 258.

atriene-^7^7-^^ (,&.§)_• A 320 mg (1.24 mmole) sample of 72, and tert-butanol (0.74 g, 10 mmole) in 15 ml of tetrahydrofuran was added dropwise to lithium wire (0.35 g, 50 mg-at) cut in small pieces. The initially exothermic reaction was allowed to cool to room temperature where stirring was maintained for 4 hr. The reaction mixture was filtered through a pad of glass wool. The inorganic solids were washed with 30 ml of petroleum ether. The combined organic filtrates were washed with water (4x 20 ml) , dried 144 over magnesium sulfate, and concentrated to give a pale yellow oil.

Preparative scale VPC (90°, 6 ft x 0.25 in. 5% SE-30 on Chromosorb G) yielded 34 mg (23%) of 6*3,as a colorless oil; NMR (CCI4 ) 6 5.63

(m, 2H), 1.4-0.4 (m, 8H); m/e 122.1067. Calcd for C9H10D2: 122.1065.

Kinetics Procedure. A standard solution of 68, in benzene was prepared by dissolving 8 mg of 68 and 5 mg of cis-decalin (internal standard) in 1.0 ml of benzene (distilled from calcium hydride).

Aliquots were sealed and kinetic data were determined as described before.

Preparat i ve ^ Seal e _Th o rmoly si so f A 12 mg (0.10 mmole) sample of 68 in 0.25 ml of benzene-dc was sealed under reduced pressure in a n e y — 10 glass ampoule. The ampoule was immersed in a 185° oil bath for 8 hr.

VPC analysis of the reaction mixture showed greater than 95% of a single component. Preparative VPC (90°, 6 ft x 0.25 in. 5% SE-30 on

Chrome G) yielded 9 mg (75%) of 73j NMR (benzene-dg) 6 5.85 (d, 1H,

J = 10.5 Hz), 5.57 (t, 1H, J = 8 Hz), 5.48 (m, 1), 5.45 (t, 1H, J = 8

Hz), 2.63 (t, 2H, J = 8 Hz), 2.20-1.70 (m, 4); m/e 122.1067. Calcd

for C9HiqD2: 122.1065.

Thermolysis of 60. A 24 mg sample of 6£, in 0.25 ml of benzene was sealed under vacuum in a thick-walled glass ampoule. The sealed ampoule was immersed in an oil bath heated to 210°. After 3 hr, the ampoule was cooled to room temperature and opened. The product was purified by preparative scale VPC (6 ft x 0.25 in. 5% SE-30 on 145 Chromosorb P, 85°) to yield 20 mg (84%) of cis-3,4-divinylcyclopentene

(74J; NMR (C6D6) 6 5.89-5.50 (m, 2H), 5.63 (broadened S, 2H), 4.98

(broadened d, 2H, J = 12 Hz) , 3.20 (d x d, 1H, £ = 8, 8 Hz) , 2.85

(d x d x d, 1H, £ = 16, 8, 8 Kz), 2.30-2.08 (m, 2A); m/£ 120.0941.

Hydrogenation of 74,. A 20 mg sample of 7^, in 3 ml of hexane was placed under a hydrogen atmosphere and 20 mg of 10% palladium on carbon was added. After 12 hr of vigorous stirring under a hydrogen atmosphere, the reaction mixture was filtered through a pad of Celite and concent­ rated. Preparative VPC purification (6 ft x 0.25 in. 5% ?E-30 on

Chromosorb P, 85°) yielded 13 mg (60%) of cis-1 ,2-diethylcyclopentane 61 identical in all respects to an authentic sample; ■‘■H NMR (CCl^)

6 2.0-0.8 (series of multiplets); vm_„ (neat) 2860, 2870, 1463, and IQaX

1380 cm"1.

Thermolysis of 73. A 12 mg sample of 73, was thermolyzed as described above to give a single rearrangement product identified as

76; 2H NMR (C^-Dc)O O

Acid Catalyzed Rearrangement of j48. A 20 mg sample of 70% perchloric acid was added to a stirred solution of 150 mg (1.25 mmole) of 48. in 5 ml of chloroform. The resulting mixture was brought to reflux for 20 min. The reaction mixture was diluted with 10 ml of methylene chloride, was washed with saturated sodium bicarbonate, dried over magnesium sulfate, and concentrated. Preparative scale VPC 146 purification (10 ft x 0.25 in. 15% XF-1150 on Chromosorb P, 115°) yielded

43 mg (29%) of 79, as a white, crystalline solid, mp 55-57°.

syn-3,5-Bishomocycloheptadienone-d_4 (J33) . Chromium trioxide

(1.20 g, 12.0 mmole) was added in three portions to 1.90 (24.0 mmole) of pyridine in 40 ml of methylene chloride. After 30 min, 426 (3.00 mmole) of £2, in 10 ml of methylene chloride was added and stirring was continued

for 0.5 hr. The solution was decanted and evaporated. The insoluble inorganic solids were washed with ether <2 x 25 ml) and these washings were added to the methylene chloride residue. A chalky precipitate was filtered, and the filtrate was washed with 10% hydrochloric acid (2 x 50 ml) and saturated sodium bicarbonate solutions, dried, and concentrated to yield 401 mg (95%) of J33, as a colorless oil; NMR (CCl^) 6 2.73

(d x d, 2H, J = 16, 5 Hz), 2.07 (d x d, 2H, J = 16, 7 Hz), 1.4-0.5 (m,

4H); m/e 140.1141.

syn- 3,5-Bishomocj^ lol^ptadienone-d^4 JTosy Ihydrazone (

Tosylhydrazine (186 mg, 1.00 mmole) was added to 132 mg (0.945 mmole) of Jg3, in 5 ml of absolute ethanol. A single crystal of £-toluenesulfonic acid was added, and the resulting solution was refluxed for 50 hours.

The reaction mixture was allowed to cool slowly. The crude crystalline product was recrystalli2ed from ethanol to yield 139 mg (48%) of 84 as a white solid, mp 150-153°; m/e 308.1500.

syn-1,3-Bishomocycloheptatriene-d^ (80). Lithium tetramethylpiper- idide was prepared by the addition of 0.80 ml of 2.5 M n-butyllithium 147 {2.0 mmole) in hexane to 282 mg £2.00 mmole) of 2 ,2 ,6 ,6-tetramethylpip- eridine in 1 ml of hexane. This solution was added dropwise to a stirred slurry of 139 mg (0.452 mmole) of 84 in 5 ml of anhydrous ether at 0°.

The resulting orange solution was warmed to room temperature for 1.0 hr followed by reflux for 1.0 hr. The reaction mixture was quenched by the addition of 1.0 ml of water. The organic phase was separated, washed with 10% hydrochloric acid and saturated sodium bicarbonate solutions, dried, and concentrated. Preparative scale VPC purification (10 ft x

0.25 in. 15% QF-1 on Chromosorb G, 115°) yielded 40.4 mg (72%) of J30 as a colorless oil; NMR (CDCI3) 66.00—5.00 (m, 2H), 3.10-2.00 (m, 2H) ,

1.6-0.8 (m, 4H); m/e 124.1192.

Acid Catalyzed Rearrangement of 80. A mixture of 10 mg of 70% perchloric acid and 40 mg (0.32 mmole) of 80 in 2 ml of chloroform was treated as described above to give 6 rag (15%) of a white solid, mp

54-57°; 1H NMR (CDC13 , -30°) 6 5.5 (m, 6H), 3.70 (m, 1H), 2.10 (m, 1H); m/e_ 124.1192

General^Proce^ure _for Rhodium (I) Rearrangement. Into an argon flushed small glass ampoule or nmr tube, 7-8 mole percent of rhodium dicarbonyl chloride dimer was weighed. The hydrocarbon in 0.4 ml of benzene-d^ was added. The vessel was again flushed with argon, cooled to -78°, and sealed under vacuum. The sealed vessel was then immersed in an oil bath at the appropriate temperature. Workup consisted of diluting the reaction mixture with 20 ml of ether, washing the organic solution with 10 ml of a saturated potassium cyanide solution, drying 148 over magnesium sulfate, and concentrating. Product analyses and separations were by VPC (10 ft x 0.25 in. 15% XF-1150 on Chromosorb P,

85°) techniques.

^ o d i u m jI^) Rearranjgement^ o f ^ A 120 mg sample of 48 (1.00 mmole) and 12.5 mg (0.0350 mmole) of rhodium dicarbonyl chloride dimer heated as described for 32 hr at 78°. Product analysis showed 20% unreacted

48 and a 15:25 mixture of 6j5 and 60.

For 65,: XH NMR (CDC13) 6 5.65 (m, 4H) , 3.38 (m, 1H) , 2.40 (m, 2H) ,

1.47 (m, 1H) , 1.00 (m, 3H) , -0.11 (m, 1H) ; v , , (neat) 3070, 3020, 2920, ulaX 1645, 1445, 1030, and 833 cm-1.

For 60: ^-H NMR (CcD,-) 6 6 .0-5.0 (m, 6H) , 2.68 (t, 2H, J = 7 Hz), / V W Q O — 2.03 (m, 4H); vmax (neat) 3000, 2930, 2850, 1630, 1455, 773, 720, and

665 cm 1.

Rhodium^(I| Jtearr a r^gemen t of A 25 mg sample (0.20 mmole) of

80 and 2.5 mg (0.0070 mmole) of rhodium dicarbonyl chloride dimer were heated as described for 30 hr at 85°. Product analysis showed a 7:93 mixture of 93^ and 94,.

For 1h nmr (CDCl-j) 6 5.67 and 5.53 (2 broadened s, 4H) , 1.47

(m, 1), 1.08 (m, 1), 0.68 (m, 1), -0.12 (m, 1); m/e 124.1192.

For 1H NMR (CDCI3) 6 6 .0-5.2 (m, 5H), 2.67 (t, 2H, J = 7 Hz),

2.05 (m, 1H).

A second run was made with 90 mg (0.73 mmole) of 80 and 10 mg

(0.026 mmole) of catalyst at 80° for 12 hr. Product analysis indicated

a 59:41 mixture of 93 and 94 as well as some unreacted starting

material. 149 Sodium^^orpdeuteride Reduction of syn~3^5-Bish^o^qlohe^tad^enone^

(46)^. Sodium borodeuteriae (42 mg, 4.0 meq) was added to 136 mg (1.00 mmole) of 4(5. in 3 ml of methanol-0-d. The resulting solution was allowed to stir at room temperature for 12 hr. The reaction mixture was concentrated, and the residue was taken up in 5 ml of ether and 5 ml of water. The aqueous layer was extracted with an additional 5 ml of ether. The combined organic extracts were dried over magnesium sulfate and concentrated to give 120 mg (88%) of an epimeric mixture of alcohols 95.

syn-1,3-Bishomocycloheptatriene-6-d (91). Mesyl chloride

(200 mg, 1.75 mmole) in 5 ml of methylene chloride was added dropwise to a 0° solution of 120 mg (0.86 mmole) of _r95 and 290 mg (2.9 mmole) in 20 ml of methylene chloride. The resulting solution was allowed to stir at 0° for 1.0 hr. The methylene chloride solution was washed with cold 10% hydrochloric acid (2 x 20 ml) and saturated sodium bicarbonate solution, dried over magnesium sulfate, and concentrated. The pale yellow oil was taken up in 15 ml of benzene. Potassium tert-butoxide

(220 mg, 2.0 mmole) was added, and the resulting slurry was brought to reflux for 4 hr. The organic solution was washed with water, dried over magnesium sulfate, and concentrated. Molecular distillation (80°,

40 mm Hg) yielded 43 mg (51%) of J9J; NMR (CCl^) <5 5.57 (broadened s, 1H), 2.72-1.78 (m, 2H), 1.40-0.20 (m, 6H), -0.15 (m, 2H); m/e

121.1004. 9^ and 5.0 mg (0.013 mmole) of rhodium dicarbonyl chloride dimer were heated as described for 22 hr at 78°. Product analysis showed 5% of unreacted starting material as well as a 30:70 mixture of J96, and 91

For 96: NMR (CDCI3) 5 5.83-5.43 (m, 4H), 3.40 (m, 1H), 2.45 (m,

2H)/ 1.48 (m, 1H), 1.03 (m, 1H), 0.73 (m, 1H), -0.10 (m, 1H); m/e

121.1004.

For %i'. 2H NMR (CDC13) 6 6 .0-5.0 (m, 5H) , 2.68 (d, 2Hf J = 7 Hz),

2.03 (m, 4H); m/e 121.1004.

Sodium carbonate (50 mg) was added to a stirred mixture of 300 mg (2.20 mmole) of £6 in 3.5 ml of dry tetrahydrofuran and 3.5 ml of deuterium. The resulting mixture was allowed to stir for 30 hr at room temperature.

The reaction mixture was diluted with 15 ml of deuterium oxide, and the product was extracted with methylene chloride (3 x 10 ml). The combined organic extracts were dried over magnesium sulfate and concentrated. The described procedure and workup were repeated to yield a pale yellow oil (280 mg, 91%). NMR analysis showed greater than

90% deuterium incorporation; NMR (CCl^) 6 1.98 (m, residual H from exchange), 1.3-0.0 (m, 6H), -0.27 (m, 2H); m/e^ 140.1141.

syn-3,5-Bishomocycloheptadienone-2^2/7/7-d^ Tos y lhy drazone

A single crystal of £-toluenesulfonic acid was added to a stirred solution of 280 mg (2.00 mmole) of 98, in 5 ml of methanol-0-d. After stirring for 0.5 hr at room temperature, tosylhydrazine (372 mg, 2.00 151 nmole) was added, and the resulting solution was brought to reflux for

3 hr. The reaction mixture was diluted with 25 ml of water. The product was extracted with methylene chloride (3 x 10 ml). The combined organic extracts were washed with saturated sodium bicarbonate, dried over magnesium sulfate, and concentrated. The crude product was recrystallized twice from ethanol to give 260 mg (42%) of a white solid, mp 153.5-155°; NMR (CDC13) 5 8.35 (d, 2H, J = 8 Hz), 7.78 (d, 2H, J

= 8 Hz), 3.00 (broadened s, 1H), 2.75 (s, 3H), 1.80-0.20 (m, 6H), -0.20

(m, 2H); m/£ 308.1502.

syn-1 ,3-Bishomocycloheptatriene-5,7,7-d-^ (J|2J . Lithium tetramethylpiperidide was prepared from 0.835 ml (2.00 mmole) of 2.40

ri-butyllithium in hexane and 1.28 g (8.00 mmole) of 2 ,2 ,6 ,6-tetra- methylpiperidine in 3 ml of hexane. The resulting solution was added dropwise to a 0° slurry of in 5 ml of ether. The resulting solution was allowed to stir at room temperature for 12 hr. The contents of the flask were diluted with 20 ml of ether. The organic solution was washed with cold 10% hydrochloric acid (3 x 20 ml) and saturated sodium bicarbonate solutions, dried over magnesium sulfate, and concentrated.

Molecular distillation (80°, 40 mm Hg) yielded 81 mg (78%) of 92^

NMR (CC14) 6 5.10 (broadened s, 1H), 1.20-0.20 (m, 6H), -0.30 (m, 2H) m/e 123.1129.

Rhod iurn^ (I)_ Re arrange me n t of 92^. A 65 mg (0.53 mmole) sample of

95|^and 7 mg (0.018 mmole) of rhodium dicarbonyl chloride dimer were heated as described at 70° for 21 hr. Product analysis indicated 30% 152 unreacted starting^ material and a 52:48 mixture of a100 m / and 101.

For 100; XH NMR (CDCI3) 5 5.65 (m, 3H)t 3.39 (m, 1H) , 2.44 (m, 2H) ,

1.49 (m, 2H), 1.02 (m, 1H); m/e 123.1129.

For 10J: NMR (CDCl.^ 5 6 .0-5.0 (m, 5H) , 2.03 (m, 3H) m/e_

123.1129.

cis-9^9-D^h3^rc^j.cyclo [6 ^1.0]non-4 - e n e (lj.^) . A 1-1, three­ necked Morton flask equipped with an efficient mechanical stirrer,

Friedrichs condenser, thermometer, and nitrogen inlet tube is charged with 325 g (2.98 mmole) of cis^-1 ,5-cyclooctadiene (115), 111 g (0.599 osnole) of sodium trichloroacetate, and 175 ml of 1,2-dimethoxyethane.

The resulting suspension is slowly heated to 100° with a thermostated oil bath. During the first 2 hours, considerable carbon dioxide is evolved and a dark brown color develops. After 20 hr, the mixture is

allowed to cool and filtered through a pad of diatomaceous earth. The residue is washed with 500 ml of ether and the combined filtrates are distilled through a 20 cm Vigreux column. The pressure within the distillation apparatus is gradually reduced to 10 mm and the product is collected at 95-110°. There is obtained 74-82 g (66-71%) of 116^ as a colorless liquid of sufficient purity for use in the next step.

ci^-Bic^^oJ6 0] n^n-4-ene (11,7) ^ A 2-1 three-necked Morton

flask equipped with a mechanical stirrer, Friedrichs condenser, 500-ml pressure-equalizing addition funnel, and nitrogen inlet tube is charged with 500 ml of dry tetrahydrofuran. Lithium wire (16.2 g, 2.33 g -at)

is cut into small pieces, washed free of oil with pentane, and introduced 153 into the reaction vessel. A solution containing 67.9 g (0.354 mmole) of

116/ 65.7 g (0.89 mole) of tert-butyl alcohol, and 400 ml of tetrahydro- furan was slowly added dropwise with intermittent cooling in an ice-bath to maintain a gentle reflux rate. Upon completion of the addition, the mixture is refluxed for 2 hours, cooled, and filtered through a glass wool plug. To the filtrate is added 2 1 of pentane and this solution is washed with water (4 x 500 ml) and saturated brine (300 ml) prior to drying over anhydrous magnesium sulfate. The solvents are removed by distillation through a 20 cm Vigreux column at atmospheric pressure.

The residue is distilled under reduced pressure to yield 36-40 g

(84-93%) of hydrocarbon as a colorless liquid b.p. 78-81° (25 mm).

^is-Bicyclo[6 .1.O^i^ n ^ - ^ ^ ^ iene^ (ll^j A solution of 30.9 g

(0.253 mmole) of cis-bicyclo[6.1.0]non-4-ene in 250 ml of carbon tetra­ chloride is stirred magnetically at 0° (ice-bath cooling) while a 25%

(v/v) solution of bromine in carbon tetrachloride is added dropwise until the red color just persists. The reaction mixture is concentrated on a rotary evaporator to give a pale yellow oil which partially crystallizes upon standing.

To a solution of this dibromide in 300 ml of hexamethyl phosphor- amide is added 116.3 g (1.74 moles) of anhydrous lithium carbonate,

43.0 g (1.66 moles) of anhydrous lithium fluoride, and powdered soft glass. With magnetic stirring, this mixture is heated at 100° for 12 hours in a thermostated oil bath, then cooled and poured into 2 1 of ice and water. The product is extracted into hexane (5 x 400 ml.), washed with water (4 x 300 ml) and brine, dried with anhydrous magnes­ ium sulfate, and concentrated at atmospheric pressure by distillation 154 of solvent through a 20 cm Vigreux column. The resulting dark red oil

is distilled under reduced pressure to yield 21.4-22.9 g (70-75%) of the

diene as a colorless liquid, b.p. 85-88° (45 mm). XH NMR (CC14) <5 5.8

(m, 4H), 2.4 (m, 4H), 0.8 (m, 3H), -0.1 (m, 1H); m/e 120.0941.

^ Z ^L!dlZ3^5i:^5I°-atriene Into a walled glass ampoule is placed 7.0 g (0.058 mole) of cis-bicyclo[6.1.01 nona-3.5-d-ip u p and an '

equal weight of benzene. The ampoule is cooled to -70° and sealed under

vacuum prior to being immersed in an oil bath heated to 175° for 1 hour.

The ampoule is allowed to cool to room temperature before opening. The

benzene is removed by distillation through a 12 cm Vigreux column and

the residue is purified by sublimation at 80° and 25 mm in an apparatus

fitted with ice-cooled condenser. The white crystalline sublimate

is collected to give 5.9-6.1 g (84-87%) of cis3-l,4,7-cyclononatriene, mp 35-46°. Although this material is of adequate purity for most purposes, further purification can be achieved by sublimation at 25° and 25 mm and/or recrystallization from methanol, mp 50-51°.

Sodium Borodeuteride 5-Bishomo- cjJclo^^tadi^none^83K Sodium borodeuteride (168 mg, 16.0 meq) was added to a solution of 400 mg (2.86 mmole) of 83, in 7 ml of methanl-0-d.

After stirring for 2 hr at room temperature, the methanolic solution was concentrated. The residue was taken up in 10 ml of ether and 5 ml of water. The organic phase was washed with brine, dried over magnesium sulfate, and concentrated. Molecular distillation (0.1 mm,

85°) yielded 389 mg (95%) of 124, that was a 60:40 mixture of the two epimers by VPC (12 ft x 0.25 in. 15% DEGS on Chromosorb G, 140°);

NMR (CCL.) 62.7-0.4 (series of multiplets); m/e 143.1361. 4 ---

syn-1^3^j^shqmocyclohepta.triejie-la, la^, 3aj 3a, 6-dg (1,23^) . Methane-

sulfonyl chloride (406 mg, 3.50 mmole) in 2 ml of methylene chloride was added dropwise to a 0° solution of 389 mg (2.72 mmole) of 124. and

455 mmole) of triethylamine in 35 ml of methylene chloride. The

resulting solution was allowed to stir for 0.5 hr at 0°. The organic

solution was washed with cold water (50 ml), cold 10% hydrochloric acid

(50 ml) and saturated sodium bicarbonate solutions (50 ml), dried, and

concentrated. The residual oil was taken up in 30 ml of dry tetra- hydrofuran. Potassium tert-butoxide (884 mg, 7.00 mmole) was added

and the resulting mixture was brought to reflux for 1.0 hr. The

reaction mixture was quenched with 2 ml of water and was concentrated.

The residue was taken up in 50 ml of hexane. The organic solution was washed with water (3 x 25 ml), dried, and concentrated. Molecular distillation (87°, 40 mm) yielded 262 mg (77%) of 124 as a colorless oil; 1H NMR (CDC13) 6 5.62 (broadened s, 1H), 2.72 (d, 1H, J = 14 Hz),

2.27 (d x d, 1H, J = 14, 5 Hz), 1.5-0.6 (m, 4H); m/£ 125.1255.

cis-^-l,4,7-Cyclononatriene-3,3,6,6,9,9-d^ ()Q ) • A 150 mg

sample of 70% perchloric acid in 2 ml of deuterium oxide was added to

125 mg (1.00 mmole) of 3J23, in 2 ml of chloroform-d. The resulting

mixture was refluxed for 6.0 hr. Workup as before yielded 24 mg

(19%) of 121 as a white solid, mp 47-50°; XH NMR (CDCI3) <5 5.48

(broadened s); m/e 126.1317. (3.25 g, 0.050 g-at) was added with stirring to 0.1 g of silver acetate

in 50 ml of hot acetic acid. The acetic acid was decanted, and the

zinc-silver couple was washed with several portions of ether.- The

flask was now charged with 25 ml of anhydrous ether, and diiodomethane-

dj (8.1 g» 30 mmole) was introduced dropwise. 1,4,7-Cyclononatriene

(120 mg, 1.00 mmole) in 2 ml of ether was added, and the reaction

mixture was allowed to reflux for 12 hr. The clear ether solution was

decanted into 50 ml of cold 10% hydrochloric acid solution. The organic

phase was washed with 10% hydrochloric acid (3 x 25 ml) and saturated

sodium bicarbonate solutions (25 ml), dried over magnesium sulfate,

and concentrated. Preparative VPC (10 ft x 0.25 in. 15% XF-1150 on

Chromosorb P, 140°) yielded 44 mg (25%) of 113 as a white solid, mp

55-60°; 1H NMR (CCl^ 6 2.20 (d, 3H, J = 13 Hz), 1.20-0.40 (m, 9H) ; m/e 168.1788.

^ c l o ^ opanation_of_ 3^21^w^th_Methy1 ene_Iodide^ A 20 mg (0.16 mmole)

sample of 121. was treated with 1.34 g (5.00 mmole) of diiodomethane and 0.65 g (10 mg-at) of zinc-silver couple as described above. Pre­ parative VPC yielded 5 mg (19%) of 114 as a white solid, mp 55-60°;

*H NMR (CC14) 6 0.70 (m, 9H), -0.27 (m, 3H); m/e 168.1788.

Thermolysis^of ^ d ^ ! 1 4 ^ Each isomer was vaporized individually into a quartz chip filled hot tube at 500° that was under reduced pressure and continuously purged with nitrogen. The products were trapped in a collector immersed in a Dry Ice-acetone bath. NMR 157 analysis was indicative that both isomers were recovered without detect­ able automerization.

A solution of syn,anti-3,5-bishomocycloheptadienol (360 mg, 2.6 mmol) and £-toluene- sulfonyl chloride (1.0 g, 5.2 mmol) in pyridine was kept as -20° for 24 hr, poured into ice water, and extracted with ether (3 x 25 ml). The combined ether extracts were washed with cold 10% hydrochloric acid and saturated sodium bicarbonate solutions before drying, evaporation, and recrystallization of the residue from pentane. There was obtained 470 mg (62%) of 131 as a white solid, mp 70.5-71.5°; -^H NMR (CDC1_) 5 7.70 / V W V J (d, J = 8 Hz, 2), 7.42 (d, J = 8_Hz, 2), 4.62 (m, 1), 2.40 (s, 3), 2.22

(m, 2), 1.30 (m, 2), 0.68 (m, 6), and -0.05 (m, 2); m/e 292.1138

(292.1133).

Anal. Calcd for C, 65.72: H, 6.89; S, 10.97. Found:

C, 65.73; H, 7.06; S, 11.11.

syn^syn-3^5-Jii^omocyc^^ J131) . Reaction of

500 mg (3.6 mmol) of the alcohol with 1.50 g (7.8 mmol) of £-toluene- sulfonyl chloride in pyridine (7 ml) as above afforded 675 mg (64%) of

131 as colorless crystals, mp 53-54°, from pentane; NMR (CC14) 5 7.73

(d, J = 8.Hz, 2), 7.27 (d, J = 8_Hz, 2), 4.73 (m, 1), 2.43 (s, 3), 2.33

(d of t ) , J = 16 and 5 Hz, 2), 1.65 (m, 2), 1.05 (m, 2), 0.70 (m, 4), and 0.13 (m, 2).

Anal. Calcd for C16H20°3S: C' 65*72'* H * 6 *89J s r 10.97. Found:

C, 65.48; H, 6.97; H, 11.02. PLEASE NOTE: This page not included in material received from the Graduate School. Filmed as received. UNIVERSITY MICROFILMS 159

anti,anti-3,5-Bishomocycloheptadienol. A mixture of syn,syn-3,5-bishomo-

cycloheptadienol (690 mg, 5.0 mmol), aluminum isopropoxide (4.1 g, 20

mmol), and acetone (0.1 ml) in 10 ml of 2-propanol was sealed in a heavy

walled glass ampoule and heated at 130° for 72 hr. The cooled contents

were poured into cold 10% hydrochloric acid and the product was extracted with ether (3 x 20 ml). The combined organic extracts were washed with

saturated sodium bicarbonate and sodium chloride solutions, dried, and

concentrated. Nmr analysis showed the anti,anti isomer to dominate by

a factor of 58:42. Careful silica gel chromatography gave pure anti,anti

alcohol (220 mg, 32%) as a white solid, mp 54-56°; 1H NMR (CDC13) 6 3.58

(quint, J_ = 6 Hz, 1), 1.97 (m, 4), 1.50-0.30 (m, 6), and 0.0 (m, 2);

3340 cm"1 ; m/e 138.1047 (138.1045). IuaX “ Anal. Calcd for CgH140: -C, 78.21; H, 10.21. Found: C, 78.01; H,

10.47.

anti^nti-JI^^Bistomo^c^ol^eptadJ.enyl JTosylate (^132j . From 186 mg

(1.4 mmol) of the alcohol and 344 mg (1.8 mmol) of £-toluenesulfonyl

chloride in 4 ml of pyridine, there was obtained 240 mg (61%) of 132^as

a white solid, mp 58-59°, from pentane; 1H NMR (CDClg) 6 7.70 (d, J. = 8

Hz, 2), 7.25 (d, J = 8 Hz, 2), 4.37 (m, 1). 2.40 (s, 3), 2.40-1.40 (m, 4),

1.40-0.25 (m, 6), and -0.08 (m, 2).

Anal. Calcd for Cl6H2003S: C, 65.72; H, 6.89; S, 10.97. Found:

C, 65.55; H, 7.07; S, 11.10.

Preparative Scale Acetolysis of 130. A 900 mg (3.1 mmol) sample

of 130 was dissolved in 5 ml of 0.0510 M sodium acetate in acetic acid / V w — ” 160 and sealed in a glass ampoule. The ampoule was heated at 65° for 6 hr

and cooled before pouring the contents into 60 ml of ice water. The

aqueous solution was extracted with ether (3 x 20 ml) and the combined

organic layers were washed with saturated sodium bicarbonate solution

prior to drying and concentration in vacuo. Molecular distillation of

the residue yielded 390 mg (71%) of actetate 140 as a colorless oil

shown by VPC analysis to be 99% pure. See text for ^-H NMR data.

This material was dissolved in dry ether (5 ml) and this solution

was added dropwise to a stirred slurry of lithium aluminum hydride (63 mg)

in dry ether (7 ml). The resulting mixture was refluxed gentlv for 1 hr

and hydrolyzed by addition of 63 yl of water, 50 yl of 10% sodium hydro­

xide solution, and 63 yl of water. The ether solution was decanted from

the white solids which were then washed with ether. The combined organic

phases were dried and evaporated to give 200 mg (86%) of 141 as a colorless

oil; NMR (CDCI3) 6 5.63 (m, 2), 3.90 (m, 1), 3.17-0.60 (m, 6), 2.10 (s,

1, -OH), and -0.03 (m, 1); m/e 138.1047 (138.1044).

Anal. Calcd for C9H140: C, 78.21; H, 10.21. Found: C, 77.95; H, 10.38.

cis-Bicyclo[6 .1.0]non-3-en-6-one (142). Chromium trioxide (90 mg,

0.9 mmol) was introduced under nitrogen to a magnetically stirred

solution of pyridine (144 mg, 1.8 mmol) in 3 ml of methylene chloride.

After 30 min, 12.5 mg (0.09 mmol) of 141 was introduced via syringe.

After 15 min, the solution was decanted and evaporated. The solid residue was washed with ether (2 x 8 ml) and the combined organic extracts were shaken with 10% hydrochloric acid and saturated sodium bicarbonate solutions, dried, and concentrated. There was obtained 6 mg (60%) of 1 „ 161 ketone 142. as a colorless oil; H NMR (CDCl^) 6 5.67 (m, 2), 3.5-1.3

(series of m, 6), 1.2-0.5 (m, 3), and 0.10 (m, 1); m/e^ 136.0890

(136.0888).

Anal. Calcd for C^H^O: C, 79.37; H, 8 .88. Found: C, 78.85; H,

9.33.

Diiodomethane

(440 mg, 1.64 mmol) was added under nitrogen to 10 mmol of ethylzinc iodide in ether. The resulting solution was heated at reflux for 30 min at which point 100 mg (0.81 mmol) of 144 was added via syringe. After an additional 3 hr of heating, the solution was poured into 20 ml of cold 10% hydrochloric acid. The ether layer was washed with brine, dried and evaporated. Preparative scale VPC isolation of the major component

(6 ft x 0.25 in. 5% SE-30 on Chromosorb G 150°) furnished 33 mg (30%) of

145; *H NMR (CDC13) 6 5.65 (m, 2), 3.70 (m, 1), 2.8-0.3 (series of m, 9),

2.48 (s, 1, -OH) , and -0.02 (m, 1); m/e^ 138.1047 (138.1046).

Anal. Calcd for CgHj^O: C, 78,21; H, 10.21. Found: C, 78.11;

H, 10.39.

Oxidation of this alcohol in the predescribed manner led exclusively to formation of 142.

Diiodomethane-d^• Under a nitrogen atmosphere, sodium (2.3 g,

0.01 g-at) was added in small pieces to deuterium oxide (50 g, 2.5 mol) with vigorous stirring. Diiodomethane (48 g, 0.18 mol) was added and the mixture was heated overnight at the reflux temperature. The diiodomethane was separated and added to fresh NaOD-D20 (24 hr). The 162 recovered halide was dried and distilled to give 36.4 g (76%) of colorless liquid, bp 81-85° (10 mm) . ^-H NMR analysis indicated the material to be

99% dideuterated.

Cyclopropanation of J3 ,_5-^cj.oheptadienol-j?ith^M^ty 1^ene Iodide-dgj

The zinc-silver couple was prepared by addition of zinc powder (6.2 g,

91 mg-at) to 0.5 g of silver acetate in hot acetic acid with vigorous stirring. The acetic acid was decanted and the couple was washed several times with ether. Dry ether (50 ml) was added followed by 18.0 g (67 mmol) of CD2I2 after 30 min of heating 3,5-cycloheptadienol (1.20 g, 11.0 mmol) dissolved in 10 ml of dry ether was added dropwise. Heating was continued for 48 hr at which time the solution was decanted into 100 ml of cold 10% hydrochloric acid. The organic phase was washed further with the acid (2 x 50 ml) and then with saturated sodium bicarbonate solution before drying and evaporation. The product mixture was separated into its three components by preparative VPC at 150° (12 ft x 0.25 in. 5% DEGS on Chromosorb G) : 327 mg (21%) of §2^, 160 mg (10%) of 151, and 211 mg

(15%) of 146. 38 AAA/

For jj)2: 3H NMR (CDC1 ) 5 4.05 (m, 1) , 2.37 (d of t, J = 15 and 5.5

Hz, 2H), 1.48 (m, 2), 1.3-0.4 (m, 4), and 1.08 (s, 1); m/h 142.1298

(142.1295). For 151: "^H NMR {CDl3) 6 3.97 (m, 1), 2.25 (m, 2), 1.58 (s, 1), and 1.5-0.3 (m, 6).

syn^anti-3^,5-Bi^5 ^ 9 5 ^ 5 ^Z. J I Reaction of 100 mg (0.794 mmol) of 146 with 41.3 mg (1.0 mmol) of 57% sodium hydride and 96 yl (0.80 mmol) of benzyl bromide as described previously furnished 165 mg (96%) of ether 147. This material was added to 3 ml of 50% sodium hydroxide solution, 300 yl of benzene, and 0.03 g of triethylbenzylammonium chloride with vigorous stirring. Chloroform

(400 yl, 5 mmol) was introduced dropwise via syringe and the mixture was stirred at room temperature for 2 hr before pouring it into water

(15 ml). Workup led to isolation of 219 mg (96%) of the dichlorocarbene adduct which was reduced as predescribed with 460 mg of sodium in 5 ml of ammonia containing 370 mg of tert-butyl alcohol. The usual process­ ing gave 65 mg (64%) of pale yellow oil. The crude alcohol was further purified by VPC (150°, 5% SE-30 column) ; 30 mg (29%) of pure JL48^ was thereby obtained.

Reaction of this material (21 mmol) with 100 mg (0.52 mmol) of

£-toluenesulfonyl chloride in 1.5 ml of dry pyridine as above afforded

20 mg (33%) of JL49,, mp 70-72°, after three recrystallizations from pentane.

Acetolysis of 149. The tosylate (20 mg, 0.068 mmol) was dissolved in 1.5 mi of 0.0510 M sodium acetate in acetic acid, heated at 100° for 1 hr, and poured into cold water. Customary processing provided 12 mg of a colorless oil that was comprised almost solely (> 98%) of 150■, NMR

(CDC13) 6 5.63 (d, J = 9 Hz, 1), 5.47 (d of d, 1), 4.95 (d of t, 1), 2.92

(m, 1), 2.25-1.92 (m, 2), 2.04 (s, 3), 1.20 (m, 1), 1.1-0.5 (m, 3), and

-0.10 (m, 1); m/e^ 182.1279 (182.1276).

syn,anti-3,5-Bishomocycloheptadienyl^Tosylate-d^ (152K

£-Toluenesulfonyl chloride (130 mg, 0.68 mmol) was added to a cooled solution of 151, (60 mg, 0.42 mmol) in 2 ml of dry pyridine. After 48 hr at -20°, there was obtained a solid which was recrystallized three times 164 from pentane: 25 mg (20%), mp 69-71°.

Acetolysis of 152. The above tosylate (25 mg, 0.084 mmol) was dissolved in 1.7 ml of 0.0510 M sodium acetate in acetic acid, heated to

100° for 1 hr, and worked up as above to give 14 mg (90%) of a colorless oil containing > 98% of a single product identified as NMR

(CDC13) 6 5.63 (d, J = 12 Hz, 1), 5.47 (d, J = 12 Hz, 1), 4.98 (br s, 1),

2.15 (m, 1), 2.05 (s, 3), 1.33-0.50 (m, 4), and -0.10 (m, 1); m/e

184.1403 (184.1401).

Preparative Scale Acetolysis of JL31. An 840 mg (2.88 mmol) sample of 131 dissolved in 12 ml of 0.0510 M sodium acetate in acetic acid was sealed in a glass ampoule and heated at 85° for 2 hr. Workup in the predescribed fashion and molecular distillation yielded 340 mg (66%) of a colorless oil that contained > 99% of a single product. This substance

(jJJ5§_) was dissolved in dry ether (5 ml) and added dropwise to a stirred slurry of lithium aluminum hydride (50 mg) in dry ether (7 ml). The resulting mixture was gently refluxed for 1 hr prior to hydrolysis. There was isolated 200 mg (87%) of 145.

Acetolysis^of^!56^ Reaction of 82^ (115 mg, 0.81 mmol) with j>-toluenesulfonyl chloride (230 mg, 1.2 mmol) in dry pyridine (-20°, 24 hr) furnished 105 mg (44%) of pure ,156/ mp 55-56°. When solvolyzed as before (0.0510 M NaOAc in HOAc, 100°, 1 hr), there was isolated 46 mg

(70%) of a colorless oil containing > 98% of acetate 157: NMR (CDCl^)

6 5.84 (d, J = 11 Hz, 1), 5.58 (d, J = 11 Hz, 1), 4.74 (br d of d, J = 1.0-0.5 (m, 3), and 0.08 (m, 1); m/e 184.1403 (184.1401).

Preparative^Scale^cetolysis of ]L3g^^ A 100 mg (0.31 mmol) sample of 132 dissolved in 10 ml of 0.0510 M sodium acetate in acetic acid / U t A 1 was sealed in a glass ampoule and heated at 80° for 1 hr. Product analysis on a 12 ft x 0.25 in. 10% XF-1150 on Chromosorb P column (155°) indicated four products to be present in the ratio of 25:11:55:9. The first hydrocarbon isolated proved identical in all respects to syn-3,5- bishomocycloheptatriene (48.) * The second hydrocarbon displayed spectral properties identical to those reported for 1,4,7-cyclononatriene (79) .

The major acetate proved to be 155 while the minor acetate was identical to 131-OAc.

A solution containing 40 mg (0.29 mmol) of 39^ in 2 ml of dry pyridine was added to

300 mg (2.9 mmol) and acetic anhydride and allowed to stand overnight.

The mixture was poured into water (20 ml) and the product extracted with ether. Molecular distillation of the oil so obtained provided 38 mg neat l (73%) of 131-OAc as a colorless oil: v 1730 cm -1-; H NMR (CDCl,) ✓w'- max j 6 5.07 (m, 1), 2.37 (d of t, J = 15 and 5 Hz, 2), 2.00 (s, 3), 1.55

(d of t, J = 15 and 7 Hz, 2), 1.4-0.4 (m, 6), and 0.20 (m, 2); m/e^

180.1154 (180.1150).

Anal. Calcd for C1-H.J,0„: C, 73.30; H, 8.95. Found: C, 73.41; ~ XX Xo Z prepared by distillation of glacial acetic acid from acetic anhydride.

Perchloric acid (70%) was standardized against sodium hydroxide with

phenolphthalein indicator. Standard perchloric acid in acetic acid was

prepared by weighing the standard perchloric acid into a volumetric flask

and filling to the mark with dry acetic acid.

Standard sodium acetate in acetic acid was prepared by weighing dry

sodium carbonate (flame dried and allowed to cool in a desiccator) into

a volumetric flask and filling to the mark with dry acetic acid. The water of neutralization was not removed. After standing for one week,

the sodium acetate in acetic acid solution was standardized against the perchloric acid in acetic acid solution using bromophenol blue indicator.

B ^ D ^ e C T ^ ation_ of_ Data^ Solutions of tosylates 3d0, 131 and 132

in buffered acetic acid were prepared by weighing the appropriate

tosylate into a 10.0 ml volumetric flask and filling to the mark with

0.0510 M sodium acetate in acetic acid. The concentration of tosylate varied from 0.0114 M to 0.0224 M over all runs. The resulting solution was divided into 9 glass empoules which were sealed under partial vacuum.

All ampoules were simultaneously immersed into a constant temperature

bath. After 5 min, one ampoule was removed from the rate bath and

placed in an ice water mixture. A timer was started upon removal of

the first ampoule. The remaining ampoules were removed and cooled at

appropriate intervals covering two to three half lives. The final

ampoule was removed after at least 10 half-lives to give an infinity

point. The individual ampoules were allowed to warm to room temperature, 167 at which point a standard aliquot was removed, diluted with 1 ml of dioxane, and titrated against 0.0192 M perchloric acid in acetic acid using one drop of bromophenol blue indicator. First order rate data was determined by measuring consumption of tosylate by acetolysis relative to the experimental infinity point. Duplicate runs agreeing within 5% were made at all temperatures.

Into a 15 ml one-necked flask equipped with a drying tube mounted above a reflux condenser and magnetic stirrer was placed a solution of tosylhydrazine

(410 mg, 2.2 mmol) and ketone 49 (200 mg, 1.47 mmol) in 10 ml of absolute ethanol containing one crystal of £-toluenesulfonic acid. Heating at the reflux temperature was maintained for 3 hr prior to cooling a -20° for 12 hr. The crystalline solid was separated by filtration and recrystallized from ethanol to give 210 mg (47%) of white solid, mp

163-165°.

Anal. Calcd for Ci6H20N 2^2S: Cf ^3.13; H * 6*62; N, 9.20. Found:

C, 63.05; H, 6.65; N, 9.42.

s^n-3^^M^hojaocyc_loheptadienone^To^yl2iy(^razone^^(166_)^ A solution containing 200 mg (1.'47)' mmol) of ketone 46, tosylhydrazine (410 mg,

2.2 .mmol), and a single crystal of p-toluenesulfonic acid in 10 ml of absolute ethanol was heated at reflux for 3 hr. The resulting mixture was cooled at -20° for 12 hr and the precipitated solid was separated by filtration. Subsequent recrystallization from ethanol afforded 272 mg (61%) of white crystals, mp 145-147°. 168 Anal. Calcd for C. _H„„No0.,S: C, 63.13; H, 6.62; N, 9.20. Found: — — iu ^ ^ C, 62.90; H, 6.65; N, 9.16.

syn,anti-2,5-Bishomocycloheptadienol (167^^ Into a 250 ml three­ necked flask equipped with rubber septa, nitrogen inlet tube, and magnetic stirrer was placed 740 mg (6.16 mmol) of anti-1 ,5-bishomocycloheptatriene

(31) dissolved in 20 ml of dry tetrahydrofuran. After cooling this solution to 0°, 4.20 ml of a 0.95 M solution of diborane in tetrahydro­ furan was introduced dropwise via syringe. The reaction mixture was then allowed to warm slowly with stirring to room temperature during 3 hr, recooled to 0°, and treated sequentially with 15% sodium hydroxide solution (20 ml) and 30% hydrogen peroxide (20 ml). The resulting mixture was stirred overnight, the aqueous phase was saturated with potassium carbonate, the organic layer was separated, and the aqueous phase was extracted with ether (2 x 20 ml). The combined organic extracts were dried and evaporated and the residue was subjected to molecular distillation at 100° and 0.3 torr. There was obtained 740 mg (87%) of a colorless oil, VPC analysis (0.25 in. x 12 ft 5% DEGS on Chromosorb G,

155°) of which denoted the alcohol to be > 95% pure; NMR (CDC^)

5 4.50 (m, 1), 3.50-2.30 (series of m, 4), 3.00 (s, 1), 1.65 (m, 6), and 1.32 (m, 2); m/e 138.1047 (138.1045).

Anal. Calcd for CgH^O: C, 78.21; H, 10.21. Found: C, 78.02;

H, 10.12.

anti-2,5-Bishomocycloheptadienone (168a). Chromium trioxide (3.0 g,

30 mmol) was added carefully to 4.74 g (60 mmol) of pyridine dissolved in 30 169 of methylene chloride and the mixture was stirred under nitrogen at room temperature for 0.5 hr. A solution of 167 (414 mg, 3.0 mmol) in 5 ml of methylene chloride was introduced by syringe and after 30 min the organic phase was decanted and concentrated in vacuo. The inorganic solids were washed with ether (50 ml) which was added to the concentrate. A chalky brown precipitate was separated by filtration prior to washing with 10% hydrochloric acid (2 x 50 ml), 5% sodium hydroxide solution

(50 ml), and saturated sodium bicarbonate solution (50 ml). After drying , the filtrate was concentrated in vacuo to afford 310 mg (76%) of JL68a as a colorless oil; 1680 cm~^ NMR (CDCI3) 6 2.75 (m, 2) , 1.85 (m, 2) ,

1.53 (t, J = 6 Hz, 2), 1.20 (m, 2), 0.38 (m, 3), and 0.13 (m, 1); m/ie

136.0890 (136.0888).

anti-2,5-Bishomocycloheptadienone Tosylhydrazone^Q68b)j^ Reaction of 150 mg (1.1 mmol) of 168a with 205 mg (1.10 mmol) of tosylhydrazine in 5 ml of ethanol in the presdescribed manner (2 hr reflux period) and three-fold recrystallization of the product from ethanol gave 210 mg

(68%) of white crystals, mp 145-147°.

Anal. Calcd for C^gH2QN2®2^: 63.13; H, 6.62; N, 9.20. Found:

C, 62.89; H, 6.65; N, 8.98.

Lithium Aluminum^Hydride Induction of A solution of 166a

(100 mg, 0.74 mmol) in 3 ml of dry ether was added dropwise to a stirred slurry of lithium aluminum hydride (30 mg, 3.2 mol-eq) in 5 ml of ether.

After 1 hr at the reflux temperature, the mixture was cooled in ice while 0.05 ml of water, 0.05 ml of 15% sodium hydroxide solution, and 170

0.05 ml of water were introduced sequentially. The ether layer was decanted and the white precipitate was washed with 10 ml of ether. The

combined organic layers were dried and concentrated to yield 85 mg (84%) of 167 whose purity was > 95% (VPC analysis).

anti,anti-2 ,5-Bishomocycloheptadienol (169). As before, a solution of diborane in tetrahydrofuran (1.68) ml of 0.95 M) was added dropwise to a solution of 33^ (215 mg, 1.79 mmol) in 1 ml of the same solvent cooled to 0°. Workup and molecular distillation (100°, 0.3 torr) afforded 210 mg (89%) of 1£9, as a colorless oil (> 95% purity). An neat analytical sample was obtained by preparative VPC purification; v I U a X

3350 cm-1; 1H NMR (CDC13) 5 3.27 (m, 1), 2.73 (s, 1), 1.4-2.9 (m, 4),

0.70 (m, 6), and -0.13 (m, 2); m/e 138.1047 (138.1045).

Anal. Calcd for CgH^O: C, 78.21; H, 10.21. Found: 77.79; H,

10.32.

syn-2,5-Bishomocycloheptadienone (170a). The oxidation of 169.

(200 mg, 1.45 mmol) dissolved in 5 ml of methylene chloride with

chromium trioxide (1.45 g, 14 mmol) and pyridine (2.29 g, 29 mmol) in methylene chloride (20 ml) was performed as described previously to neat _i i give 140 mg (71%) of 170a as a colorless oil; v 1680 cm ; H NMR A ^ V W n>lcl2C

(CDClj) 6 3.80-2.80 (m, 2), 2.70-1.00 (m, 9), and -0.10 (m, 1); m/e^

136.0890 (136.0888).

syn-2,5-Bishomocycloheptadienone Tosylhydrazone (170b). Reaction

of 170a (136 mg, 1.0 mmol) in 5 ml of absolute ethanol with 200 mg (1.1 mmol) of tosylhydrazine as above afforded 188 mg (62%) of whit^ 1 crystals, mp 168-174°, after three recrystallizations from ethanol.

Anal. Calcd for C, 63.13; H, 6.62; N, 9.20. Found:

C, 62.95; H, 6.67; N, 9.10.

syn,syn-2,5-Bishomocycloheptadienol (171). Sodium borohydride

(38 mg, 1.0 mmol) was added to 31 mg (0.23 mmol) of 170a dissolved in

2 ml of methanol and the resulting solution was stirred under nitrogen for 2 hr. After dilution with 5 ml of water, the product was extracted with ether (3x5 ml). VPC analysis of the dried and concentrated organic phase (the 12 ft DEGS column, 155°) showed 171 and 169 to be present in a 4:1 ratio. Product separation gave 16 mg (53%) of 171 and

2 mg (7%) of 169. For 171: v"®** 3330 cm'1 ; NMR (CDCl ) 5 4.28 (m, /w-v- lucLX 3 2.80-1.55 (m, 3), 1.85 (s, 1)', 1.50-0.33 (m, 7), 0.16 (m, 1), and -0.16

(m, 1); m/e 138.1047 (138.1045).

Anal. Calcd for C^H^O: C, 78.21; H, 10.21. Found: C, 77.96;

H, 10.13.

Phot ode aminat ion_ o 165^ i n_ Methano1. A solution of 165 (170 mg,

0.56 mmol) and sodium hydroxide (160 mg, 4.0 mmol) in 25 ml of absolute methanol contained in a 50 ml Pyrex flask equipped with a magnetic stirring bar and reflux condenser was irradiated for 3.5 hr at ambient temperature with a 450W Hanovia lamp. Water (50 ml) was introduced and the resulting solution was extracted with pentane (3 x 20 ml). The combined organic phases were dried and carefully concentrated to leave

115 mg of a colorless oil. Preparative scale VPC separation (12 ft x

0.25 in. 10% XF-1150 on Chromosorb P, 120°) of this mixture afforded PLEASE NOTE: This page not included in material received from the Graduate School. Filmed as received. UNIVERSITY MICROFILMS 173 (m, 1), 2.27 (m, 2), 1.70 {m, 2), 0.67 (m, 6), and 0.28 (m, 2); m/e_

152.1204 (152.1201).

^ ^ t o ^ ^ m ji^i o n qf l65^ in Water . A 1 mmol sample of the

tosylhydrazone was added to 50 ml of 0.2 N potassium hydroxide solution

contained in a Pyrex flask and the resulting solution was irradiated for

3 hr as before. The products were extracted into ether (3 x 15 ml) and

the combined ether extracts dried and concentrated prior to VPC analysis

and product separation (6 ft x 0.25 in. 5% SE-30 on Chromosorb G, 115°).

Only two components were observed and these were identified as 40, (16%)

and 167 (84%). /

photodeamination of 166, in Methanol. A solution of 166 (220 mg,

0.75 mmol) and sodium hydroxide (160 mg, 4.0 mmol) in absolute methanol was irradiated for 3.5 hr as predescribed to give 90 mg of a colorless oil. For isolation purposes, the resulting mixture of products was

subjected to an initial separation on the 10% XF-1150 column at 120°.

Under these conditions, hydrocarbon 48, (12%) was readily separated from

four ether fractions. The two components of longest retention time were shown (further VPC work and NMR analysis) to be isomerically

pure and subsequently characterized as 175, (11%) and X 1 A * (19%)

respectively. Spectral examination of the remaining two fractions showed

them to consist of three isomers each. Their successful resolution was

achieved on a 12 ft x 0.25 in. 5% Bentone-5% SF-96/Chromosorb G column

at 120°. Fractionation of the most rapidly moving peak gave pure

samples of ^173, (3%), 178 (22%), and ^79 (2%). The percentage 174

composition was ascertained by peak area measurements and integration of methoxyl signals in expanded NMR spectra. Comparable handling of the

remaining fraction furnished 176^ (22%), 177 (6%), and a third component

(3%) which has eluded characterization.

Methylation of 132^ -OH (50 mg, 0.36 mmol) as before and VPC isolation

from the 10% XF-1150 column at 130° gave pure 174, as a colorless oil (41 neat “9' 75%); vmax 3030, 2995, 2920, 2855, 2815, 1460, 1360, 1085, and 1018 cm"1;--1. 1XH NMR (CDCl-j) 6 3.27 (s, 3), 3.07

and 0.00 (m, 2); m/e 152.1204 (152.1201).

syn^syn-3^^^ lohePt adi enyl^Methyl Ether _(3^75J_. Methylation of (50 mg, 0.36 mmol) in the predescribed manner and purification on

the 5% DEGS column at 155° gave 31 mg (56%) of JLV5 as a colorless oil; neat -1 max 3032, 2995, 2920, 2818, 1465, 1451, 1090, 1015, 832, and 810 cm ; NMR (CDC13) 6 3.52 (m, 1), 3.29 (s, 3) 2.29 (d of t, J = 15 and 5 Hz,

2), 1.52 (d of t, J = 15 and 7.5 Hz, 2), 0.8-0.3 (m, 4) and 0.14 (m, 2);

m/e 152.1204 (152.1201).

syn^syn-2^J>-gi^^ tadienyl^ Methyl^ Ether (JL76)_. From 16 mg

(0.12 mmol) of 171 and 24 mg (0.50 mmol) of sodium hydride-mineral oil

dispersion (50%) with an excess of methyl iodide, there was isolated

(the 10% XF-1150 column, 120°) 5.3 mg (29%) of .176 as a colorless oil; neat _i V 3060, 2995, 2920, 2815, 1470, 1155, 1093, 1072, 1020, 915, and 730 cm max NMR (CDC13) 6 3.87 (m, 1), 3.43 (s, 3), 2.53 (m, 1), 2.05 (m, 1), 1.17 (m, 2), 175 1.00-0.33 (m, 6), 0.10 (m, 1), and -0.11 (m, 1); m/e 152.1204

(152.1201).

Methylation of 145, (60 mg, 0.45 mmol) in the predescribed manner furnished

33 mg (50%) of 177 as a colorless oil (using the 5% DEGS column at 155°); v £ ! f 3060, 2990, 2920, 2860, 1465, 1090, and 700 cm-1; 1H NMR (CDCI3)

6 5.67 (m, 2), 3.55 (s, 3), 3.28 (m, 1), 2.28-2.00 (m, 5), 1.58 (m, 1),

0.77 (m, 1), and 0.06 (m, 1); m/e 152.1204 (152.1201).

Methylation of 141 (13.8 mg, 0.10 mmol) in the usual manner yielded (VPC on the 10% XF-1150 column, 120°) 5 mg (33%) of 179, as a colorless oil;

1H NMR (CDCI3) 5 5.83 (m, 2) 4.67 (m, 1), 3.67 (s, 3) 3.00-2.00 (m, 4),

1.33-0.42 (m, 3), and -0.10 (m, 1); m/e^ 152.1204 (152.1201).

Hiotodeamijiation^ofJJ6^JLn^Jtfater^ The general procedure (1 mmol of 166)was followed and the alcohol mixture (no evidence for hydrocarbon formation was seen) was oxidized directly with Collins reagent (prepared from 500 mg of chromium trioxide and 790 mg of pyridine in 15 ml of dichloromethane) for 1 hr at room temperature. The dichloromethane solution was decanted and concentrated. The inorganic solids were leached with ether (2 x 20 ml) and these were added to the organic residue prior to washing with 10% hydrochloric acid and saturated sodium bicarbonate solutions. Drying, evaporation, followed by VPC analysis and separation (12 ft x 0.25 in. 15% DEGS on Chromosorb G 165°) gave the following ketones: 46 (17%), ^ 68a (25%) , 170a, (27%) , JL8Q, (5%) , and

181 (26%).

l,5-Cyclooctadien-4-ol (ZL82). A stream of oxygen was bubbled through 40 g (0.37 mol) of 1 ,5-cyclooctadiene (115) warmed to 60° for

10 hr. Methanol (30 ml) was added, the solution was cooled to 0°, and sodium borohydride (1.20 g, 0.13 eq) slowly added. After 1 hr at room temperature, the reaction mixture was diluted with water (150 ml) and extracted with hexane-ether (4:1, 2 x 75 ml). The combined phases were washed with water, dried, and concentrated. Unreacted starting material was removed by distillation at 60-70° and 40 mm. The oxygenated products were distilled at 48-55° and 0.1 mm to give 2.51 g of colorless oil.

VPC analysis indicated the presence of four components which were separated on the 15% DEGS column at 165° and identified as cis-9-oxabicyclo[6.1.0]- non-4-ene (40%), 9-oxabicyclo[6 .1.0]non-4-en-2-ol (20%), 1,4-cycloocta- dien-6-ol (12%), and ^182, (28%). For 182: 3340, 3020, 2940, 2890,

2830, 1650, 1428, 1030, 805, and 720 cm”1; % NMR (CDC13)

4.88 (m, 1), 2.9-1.7 (m, 6), and 2.17 (s, 1). 7

cis-Bicyclo[6.1.0]non-4-en-anti-2-ol (183a). Zinc powder (325 mg,

5.0 mg-at) was added with stirring to 20 mg of silver acetate dissolved

in 10 ml of hot acetic acid. The acetic acid was decanted and the zinc- silver couple was washed with several portions of ether. The flask was now charged with 10 ml of anhydrous ether and diiodomethane (804 mg,

3.00 mmol) was introduced by syringe. After being heated to gentle reflux, the reaction mixture was treated with 124 mg (1.0 mmol) of 182,

in dropwise fashion (syringe) and refluxed for 12 hr. The ethereal 177 solution was decanted and the inorganic salts rinsed with ether. The combined organic phases were washed with cold 10% hydrochloric acid

(25 ml) and saturated sodium bicarbonate solution (25 ml), dried, and concentrated. VPC analysis showed two monocyclopropanated alcohols to be present in a 4:1 ratio. Preparative VPC collection of the major component (the 15% DEGS column at 165°) afforded 43 mg (31%) of 183a. as n&at. a colorless oil; vmax 3340, 3000, 2920, 1645, 1440, 1020, 780, and 710 cm"1; 1H NMR (CC14) 5 5.47 (m, 2), 3.40 (m, 1), 2.65 (s, 1), 2.57-1.48

(series of m, 6), 1.48-0.42 (m, 3), and 0.13 (m, 1); m/e 138.1047

(138.1045).

Anal. Calcd for CgH^O: C, 78.21; H, 10.21. Found: C, 78.61;

H, 9.87.

cis-Bicyclo[6.1.0]non-4-en-anti-2-yl Methyl Ether (183b). A 40 mg

(0.29 mmol) sample of JL83a^ was methylated in the predescribed manner and the resulting ether (183b) was obtained as a colorless oil (30 mg, 68%) after purification on a 12 ft x 0.25 in. 10% XF-1150 on Chromosorb P column at 125°; XH NMR (CDC13) 6 5.73 (m, 2), 3.42 (s, 3), 3.02 (m, 1),

2.43 (t, J = 6 Hz, 2), 2.15 (m, 4), 0.95 (m, 3), and 0.17 (m, 1); m/e

152.1204 (152.1201).

\

cis-Bicyclo[6.1.0]non—4-en—2—one (181). A 138 mg (1.0 mmol) sample

was oxi3ized with 1.00 g (10 mmol) of chromium trioxide and 1.58 g (20 mmol) of pyridine as described above. Standard workup and isolation (15% DEGS column, 165°) afforded 53 mg (39%) of 181; vneat max 3010, 2925, 1685, 1365, 1275, 1110, ]035, and 710 cm-1 ; NMR (CC14) 5 5.50 (m, 2), 3.40-2.60 (m, 2), and 2.6-0.7 (m, 8); m/e 136.0890

(136.0888).

cis-Bicyclo[6 .1.0]non-4-en-syn-2-yl Methyl Ether (178). Sodium borohydride (100 mg) was added to a solution of J-81^ (53 mg, 0.39 mmol) in 5 ml of methanol and the mixture was stirred at room temperature for

2 hr before dilution with water (20 ml) and extraction with dichloro- methane (3x5 ml). The usual workup afforded an oil, -*-H NMR analysis of which showed the presence of two alcohols in a approximate ratio of

2:1. The major portion of this material (33 mg) was methylated in the conventional manner and the resulting ethers were separated by preparative

VPC methods (10 ft x 0.25 in. 15% XF-1150 on Chromosorb P, 125°). The minor component (38%) was identical to 183b and the major component (62%) identified as 178 was identical to one of the major deamination products

Of 166: v j g * 3000, 2940, 2880, 2820, 1484, 1355, and 1090 cm"1; XH NMR

(CDC13) 6 5.58 (m, 2), 3.78 (m, 1), 3.40 (s, 3), 2.50-1.80 (m, 5),

1.80-1.33 (m, 1), 1.12-0.50 (m, 3), and 0.30 (m, 1); m/e 152.1204

(152.1201).

Anal. Calcd for C1QH160: C, 78.89; H, 10.59. Found: C, 78.97; H,

10,54.

non-4-ane m-Chloroperbenzoic acid (624 mg, 4.00 mmol) in 20 ml of dichloromethane was added dropwise to a stirred slurry of 117 (431 mg, 3.53 mmol) and sodium carbonate

(1.06 g, 10 mmol) in 20 ml of dichloromethane at 0°. After 5 hr at room temperature, 10% sodium bisulfite solution (2 ml) was added and the reaction mixture poured into water (40 ml). The organic phase was 179 separated, washed with saturated sodium bicarbonate solution (3 x 30 ml), dried, and concentrated. The resulting two epoxides (ratio 2:3) were separated by preparative VPC (10 ft x 0.25 in. 15% QF-1 on Chromosorb G,

150°). neat The less dominant isomer is considered to be 184 : vm_LX; 3060, 2990,

2910, 2860, 1470, 1450, 1430, 1010, 911, 902, and 765 cm"1 ; 1H NMR (CC14)

5 2.87 (m, 2), 1.92 (m, 6), 1.5-0.2 (m, 5), and -0.28 (m, 1); m/e^

138.1047 (138.1045).

Anal. Calcd for CgH^^O: C, 78.21; H, 10.21. Found: C, 77.99;

H, 10.20. For 185: vneat 3030, 2980, 2920, 2860, 1469, 1443, 1005, 983, max 870, and 770 cm"1; 1H NMR (CC1 ) 5 2.72 (m, 2), 2.07 (m, 4), 1.5-0.2 4 (m, 7), and -0.17 (m, 1); m/e 138.1047 (138.1045).

Anal. Calcd for CgH^O: C, 78.21; H, 10.21. Found: C, 77.98; H,

10.18.

cis-Bicyclo[6.1.0]non-3-en-anti-5-ol (186a). A. Lithium

Dimethviamide Promoted Ring Opening of 184. A solution of epoxide 184_

(90 mg, 0.65 mmol) in 2 ml of ether was added dropwise at room temperature to lithium diethylamide (prepared from 730 mg (10 mmol) of diethylamine dissolved in ether (2 ml) to which was added 2.15 ml (5 mmol) of 2.5

M n-butyllithium at 0°). After 72 hr, the reaction mixture was poured

into 25 ml of cold 10% hydrochloric acid, the organic phase was separated, and the aqueous layer extracted with ether. The combined extracts were washed with 10% hydrochloric acid and saturated sodium bicarbonate solutions, dried, and concentrated to leave an oily alcohol spectro­ scopically identical to the major photooxygenation product (186a)

synthesized in part B. ^_°n- To a solution of Rose Bengal (loJ'^mg)

in methanol {20 ml) was added 488 mg (4.0 mmol) of 2iZ in 130 ml of

dichloromethane. A slow stream of oxygen was bubbled through this

solution for 5 hr with concomitant irradiation from a 600W DYV tungstem

lamp. Sodium borohydride (380 mg, 10 mmol) was added with stirring and

after 1 hr the reaction mixture was washed with water (3 x 100 ml),

allowed to stand over magnesium sulfate and charcoal, filtered, and evaporated. Preparative VPC isolation (the 15% QF-1 column at 130°) was employed to separate the two resulting alcohols (ratio 43:57).

The minor component was isolated as a colorless crystalline solid, mp 65-67.5°, and identified as 287a; 5.83-5.35 (m, 2), 5.30 (m, 1) ,

2.18 (s, 1), 2.50-0.40 (series of m, 9), and 0.02 (m, 1); m/e 138.1047

(138.1045).

Alcohol 286a was obtained as a colorless oil; 3340, 3060,

2990, 2925, 1655, 1453, 1427, 1040, 1021, 1010, and 841 cm”1; 2h NMR

(CDCI3) 6 5.42 (m, 2), 5.18 (m, 1), 2.60-2.10 (m, 1), 2.22 (s, 1), 2.3-1.1

(series of m, 5), 0.83 (m, 3), and -0.10 (m, 1); m/e 138.1047 (138.1045).

Anal. Calcd for CgH^O: C, 78.21; H, 10.21. Found: C, 78.09;

H, 10.26.

cis-Bicyclo[6.1.0]non-3-en-anti-5-yl Methyl Ether (186b).

Methylation of jL86a (42 mg, 0.30 mmol) as before provided 38 mg (83%) of

286)^ as a colorless oil after preparative VPC purification on a 3 ft x

0.25 in. 5% QF-l/Chromosorb G column at 85°; vneat 306O, 2990, 2925,

2815, 1655, 1452, 1120, 1105, 1083, 1013, 840, and 704 cm-1; XH NMR

(CDCI3) <5 6.0-5.4 (m, 2), 3.77 (m, 1), 3.28 (s, 3), 2.70-0.50 (series of m, 9), and 0.00 (m, 1); m/e 152.1204 (152.1201). 181

cis-Bicyclo [6 .1.01 non-3-en-5-one (188) . A mixture of 186a^ and ^ 86b as obtained from the photooxygenation (330 mg, 2.4 mmol) was oxidized in the customary fashion with 1.00 g (10 mmol) of chromium trioxide and 1.58 g (20 mmol) of pyridine in 30 ml of ether. Preparative VPC isolation

(the 15% DEGS column at 165°) afforded 225 mg (70%) of 188; v° 1680 3 max cm-1; XH NMR (cci4) 5 5.58 (m, 2), 2.8-2.0 (m, 3), 2.0-1.0 (m, 3), 0.8-0.2

(m, 3), and -0.1 (m, 1); m/e 136.0890 (136.0888).

6-Chloro-cis_-l^ Thionyl chloride (5 ml) was added to alcohol 141^ (305 mg, 2.22 mmol) under a nitrogen atmosphere and the resulting solution was warmed at 40° for 3 hr. The reaction mixture was concentrated in vacuo to yield 325 mg (94%) of the chloride as a pale yellow oil which because it darkened rapidly upon exposure to air was reduced without further purification.

cis-Bicyclr>J6^ J ^ (19£j . Sodium sand was prepared from molten sodium (920 mg, 40 mg-at) in hot xylene (25 ml) tinder a nitrogen atmosphere. The sand was washed with anhydrous tetrahydrofuran and added to a solution of 3-89 (325 mg, 2.08 mmol) and tert-butyl alcohol

(300 mg, 4 mmol) in 15 ml of tetrahydrofuran under nitrogen. After an initial exothermic reaction, the mixture was stirred overnight at room temperature. The solution was decanted from the unreacted sodium which was washed with ether (4 x 10 ml). The combined organic layers were concentrated in vacuo and the residual oil was taken up in hexane (50 ml) .

The hexane solution was washed with water (3 x 30 ml), dried, and concentrated. Molecular distillation afforded 216 mg (85%) of 190 as a 182 colorless oil, bp 100° (bath temp) at 40 mm; 3060, 2990, 2915, 2850,

1650, 1460, 1010, 840, 780, and 700 cm-1; 1H NMR (CC14) <5 6.0-5.2 (m, 2),

3 .8 -0 .4 (series of m, 11), and 0.1 to -0.2 (m, 1); m/e, 122.1097 (122.1095)

Anal. Calcd for CgH^: C, 88.45; H, 11.55. Found: C, 88.52; H,

1 1 .4 7 .

Photooxygenation^ of 3^90,. To a solution of Rose Bengal (100 mg) in methanol (20 ml) was added a solution of 190, (122 mg, 1.0 mmol) in dichloromethane (130 ml). A slow stream of oxygen was bubbled through this solution with concomitant irradiation for 13 hr as described for jL17. VPC analysis (the 15% DEGS column, 165°) showed greater than 90% conversion to a mixture of alcohols with one isomer predominating (> 90%).

Chromatography on Florisil (elution first with hexane, then with ether) afforded 63 mg (46%) of a colorless oil; vneat 3550, 3000, 2930, 1455, max 1042, and 845 cm-1; 1H NMR (CCl4>

2.43-0,33 (.series of m, 9), and -0.13 (m, 1); m/e 138.1047 (138.1044).

cis-Bicyclo[6.1,0]non-4-en-3-one (1^2). A 50 mg (0.36 mmol) sample of was oxidized with 200 mg (2.0 mmol) of chromium trioxide and 316 mg (4.0 mmol) of pyridine as described earlier. Standard workup and

VPC purification (the 15% DEGS column, 165°) gave 35 mg (70%) of the pure ketone 19.2; vneat 3060, 3000, 2920, 1655, 1460, 1320, 1170, and 1025 cm-1, max NMR (CCl^) 6 6.2-5.4 (m, 2), 3.0-0,5 (series of m, 9), and -0.15 (m, 1) m/e 136.0890 (136.0888), 183 Photodeamination of 168b in Methanol. The general procedure described earlier was followed (304 mg, 1.00 mmol of^168b). VPC analysis and separation (the 5% Bentone/5% SE-96 column, 120°) gave a single hydrocarbon identified as 34. (11%) and three methyl ethers which were subsequently characterized as 173. (24%) , 179. (11%) and 194b (54%) .

Because 179^ and 194b proved difficult to separate under these conditions, they were collected together and the percentage composition determined by integration of NMR spectra.

PhoJiodeamination of^ 168b _in^Water.^ A 1.0 mmol sample of 168b was irradiated according to the general procedure. VPC analysis and separation (6 ft x 0.25 in. 5% SE-30 on Chromosorb G, 120°) revealed the formation of three alcohols but no hydrocarbon. By suitable comparisons, the three purified products were identified as 167. (20%),

/VW141 (10%), and 194a (70%). For 194a: Vneat 3370, 3060, 2995, 2805, 1458, 1050, 1050, 980, max 938, and 846 cm~^; NMR (CDCl3) 6 4.35 (m, 1), 2.70-1.60 (series of m, 3), 1.92 (m, 1), 1.60-0.20 (series of m, 8), and 0.10 (m, 1); m/e

138.1047 (138.1044).

Anal. Calcd for CgH140: C, 78.21; H, 10.21. Found: C, 78.09;

H, 10.19.

Collins^ Oxidation of 194,a._ A 28 mg (0.20 mmol) sample of 194a was oxidized with chromium trioxide (200 mg, 2.0 mmol) and pyridine

C316 mg, 4.0 mmol] in 10 ml of dichloromethane as described earlier

C30 min). Standard workup and preparative VPC purification (the 5%

SE-30 column, 130°) yielded 20 mg (74%) of ketone identical in all 184 respects to 168a

anti,syn-2,5-Bishomocycloheptadienvl Methyl Ether (194£) • Methyl-

ation of 10 mg of JJ^4g^ and isolation of the resulting ether by preparative

VPC (the 5% SE—30 column, 120°) furnished 6 mg (54%) of 194b whose spectral

properties were identical to those of the sample isolated from the photo­ neat deamination of 168b, vmax 3060, 2990, 2905, 2850, 2810, 1460, 1095, 1085,

1035, and 1018 cm-1; ^ NMR (CDC13) 6 3.85 (m, 1), 3.28 (s, 3), 2.68-2.05

(m, 3), 1.88 (m, 1), 1.33-0.4 (series of m, 7), and 0.03 (m, 1); m/£

152.1204 (152.1201).

Photodeamination of 170b, in Methanol and Water. The generalized

procedure was used in both experiments. VPC analysis (the 10% XF-1150

column, 120°) of the methanolic reaction mixture revealed the formation

of > 99% of a single methyl ether and < 1% of a single hydrocarbon whose

retention time proved identical to that of 33 under these conditions. /v v .

The isolated ether had spectral properties identical to those of 176.

From the aqueous photodeamination, a single product was detected.

VPC isolation (5 ft x 0.25 in. 5% QF-1 on Chromosorb G, 145°) afforded

crystalline ^171/ mp 72.74°, whose spectra were superimposable upon those o f the authentic sample.

Methanolysis of 131. A 200 mg (0.69 mmol) sample of 131 and 121 mg (1.06 mmol) of 2,4,6-collidine were dissolved in 10 ml of dry methanol

(distilled from magnesium methoxide) and sealed under reduced pressure

in a thick walled glass ampoule. The ampoule was immersed in a 145° 185 bath for 2.0 hr. The contents of the ampoule were diluted with 20 ml of cold water. The products were extracted with hexane (3 x 10 ml) and the combined extracts were washed with 10% hydrochloric acid (2 x 20 ml), saturated sodium bicarbonate (20 ml), dried, and concentrated in vacuo.

VPC analysis and separation (10 ft x 0.25 in. 15% XF-1150, 135°) est­ ablished the formation of ^8, (4.4%), YT1_ (85.3%), ,17£. (8.0%), ,179, (1.7%),

.6%). The spectra of the major products were identical to those of authentic samples. The minor products ( 1 ^ and ,175) were identified by their retention time behavior.

Methanolysis of A 200 rag (0.69 mmol) sample of ^130, and 121 mg Cl.00 mmol) of 2,4,6-collidine were dissolved in 10 ml of dry methanol and sealed under reduced pressure in a thick walled glass ampoule. The ampoule was immersed in a 145° oil bath for 2.0 hr. Workup in the pre­ described manner revealed the formation of 51^ (1%) , 172, (19%) , JJX (69%) r and X J S b (H%) » All but ,51, were characterized by direct spectral com­ parisons . 186

2-Chloro-syn-l, 5-bishomocycloheptatriene J^40) ^ i d ^ r5-Dichloro- syn-1,5-bishomocycloheptatriene (241). A solution of £ 3 , (300 mg, 2.5 mmol) and tert-butyl hypochlorite (400 mg, 36 mmol) in 3 ml of Freon 11 contained in a 10 mm Pyrex nmr tube was placed in a chloroform-liquid nitrogen slush bath (-63°) contained in a Pyrex Dewar flask. The reaction mixture was irradiated for 45 minutes with a Sylvania sunlamp, diluted with 15 ml of ether, washed with saturated sodium bicarbonate solution, and dried. Concentration and preparative VPC purification showed two products (ratio 85:15) to dominate the mixture (85%). The more rapidly eluted major component (176 mg, 46%) was a colorless oil identified as monochloride ,240; 3070, 3000, 2920, 2860, 1650, 1450,

1130, 825, and 725 cm-1; ^ NMR (CC14) 6 5.62 (s, 2), 2.53 (d of t J =

13 and 4 Hz, 1), 2.08-1.03 (br m, 4), 0.80 (d of t, J = 7 and 4 Hz, 1),

0.50 (t, J = 5.5 Hz, 1), and 0.00 (m, 2); 13C NMR (CDC13) 131.4 (d) ,

128.9 (d), 41.2 (s), 34.3 (k), 28.7 (d), 25.1 (t), 15.0 (d), 14.2 (t), and 12.8 ppm (d); m/e_ calcd 154.0549, found 154.00554.

Anal. Calcd for CgH^Cl: C, 69.90; H, 7.17. Found: C, 69.74;

H, 7.22.

The second component proved to be dichloride 241: NMR (CCl^)

6 5.82 (s, 2), 2.60 (d of t, J = 14 and 5.5.Hz, 1), 1.92 (m, 2), 1.30

(d of d, J = 9 and 5 Hz, 2), 0.58 (t, J = 5 Hz, 2), and 0.07 (d of t,

J a 14 and 11 Hz, 1); l3C NMR (CDCl.^) 130.1 (d) , 39.4 (s) , 34.9 (t) ,

28.6 (d) , and 25.1 ppm (t) ; m/ej 188.

Chlori^j:io i ^ ^ ^ 24p^ A solution of (20 mg, 0.13 mmol) and tert-butyl hypochlorite (17 mg, 0.16 mmol) in 0.4 ml of Freon 11 was 187

irradiated as described above for 45 min. The reaction mixture was

diluted with ether (10 ml), washed with saturated sodium bicarbonate

solution, dried, and concentrated to give 17 mg (71%) of 241, identical

in all respects to the above sample.

2-Chloro-anti-l ,5-bishomocycloheptatriene (£42^. A solution of £4,

(240 mg, 2.00 mmol) and tert-butyl hypochlorite (220 mg, 2.04 mmol) in I 2 ml of Freon 11 was allowed to react as described previously. Molecular

distillation (65°, 0.2 torr) afforded 119 mg (62%) of a colorless oil,

VPC analysis of which showed it to be composed chiefly (86%) of one

component. Preparative VPC purification under carefully controlled

conditions (6 ft x 0.25 in. 5% SE-30 on Chromosorb G, 120°) gave pure

,242,- 3020, 2920, 1645, 1450, 1123, 1037, 810, and 785 cm-1;

NMR (CC14) 5 5.70 (m, 2), 2.50 (br m, 1), 1.40 (br m, 1), 1.50-1.10 13 (m, 5), 1.05 (m, 1), and 0.70 (m, 1); C NMR (CDC13) 130.6 (d), 126.5 (d),

(d), 41.8 (s), 27.2 (t), 25.3 (d), 24.1 (t), 16.9 (d), 14.5 (d), and

12.8 ppm (t); m/

Anal. Calcd for C^H^Cl: C, 69.90; H, 7.17. Found; C, 69.74;

H, 7.33.

4-Chloromethylbicyclo[5JU01octa-2,4-diene (258) A. ^ JThermal

Rearrangement of 242,. a small sample of chloride 242, was sealed neat

in a small glass ampoule which was immersed in an oil bath preheated to

150°. After 4 hr, the ampoule was opened and the residual liquid was

examined by NMR. The spectrum was identical to that of 2S8_ as

prepared on a preparative scale below.

B. Florisil-Promoted Rearrangement of 242,. Chloride £42 (50 mg, 188 0.32 mmol) was adsorbed onto 2g of activated Florisil using hexane as solvent. After 15 min, the product was eluted with hexane to give 50 mg

(100%) of j£58; NMR (CCl4) 6 6.22 (d of d, J = 11 and 5 Hz, 1),

6.00-5.40 (m, 2), 3.87 (br s, 2), 2.43-1.97 (m, 2), 1.80-0.70 (m, 3), cvclohexane and 0.43 (d of t, J= 9 and 3 Hz, 1); X 232 niu (e 3100); m/e max calcd 154.0549, found 154.0552.

Anal. Calcd for CgH^Cl: C, 69.90; H, 7.17. Found: C, 69.48;

H, 7.11.

Methanolysis Experiments. A. Silver (I)-Assisted Ionization. To a 0.1M solution of 240 or 242 in anhydrous methanol (purified by distillation from magnesium methoxide) was added 1 .1-1.2 equiv of silver trifluoroacetate (Aldrich). The resulting mixture was allowed to stir at room temperature for 1 to 4 hr. After filtration and dilution with water, the products were extracted into petroleum ether and the combined organic layers were washed with brine, dried, and concentrated. Product analysis, accomplished by VPC methods on a 10 ft x 0.25 in. 15% XF-1150 column (Chromosorb P, 120°), revealed conversion to a mixture of ethers ,£60, £61, and 262. The individual components were purified by preparative scale separation. For 24Q,,

73.5, 7, and 19.5%; for ^4g, 67.5, 15, and 17.5%).

For 260: vneat 3000, 2920, 2880, 2830, 2810, 1610, 1440, and max 1090 cm"1 ; xJ^£lohexane 251 nm (c 2900); ^ NMR (CDCI3) 6 6.23 (d of d,

J = 11.5 and 4.5 Hz, 1), 5.75 (m, 1), 5.60 (d, J = 11.5 Hz, 1), 3.82

(br s, 2), 3.29 (s, 3), 2.60-2.20 (m, 2), 1.63 (m, 1), 1.40-0.80 (m, 2), and 0.80-0.30 (m, 1); m/e^ calcd 150.1045, found 150.1048. 189 Anal. Calcd for C-inH ,0: C, 79.95; H, 9.39. Found: C, 79.83; ------j - u 24

H, 9.56.

For 261: XH NMR (CDC13) 6 5.80 (m, 2), 5.03 (sf 2), 3.80 (d, J =

8 Hz, 1), 3.34 (s, 3), 2.61 (m, 1), 1.70-0.80 (m, 4), and 0.15 (m, 1); ra/e calcd 150.1045, found 150.1048.

For 262: NMR (CDCl-j) 5 5.83 (s, 1), 5.80 (br s, 1), 5.27 (m, 1) ,

4.98 (m, 1), 3.93 (m, 1), 3.43 (s, 3), 2.47 (m, 1), 1.70-0.70 (m, 4), and 0.20 (m, 1); m/e calcd 150.1045, found 150.1048.

Comparable treatment of 258 afforded 56% of 260, 24% of 261, and

20% of 262.

B^^_Uncatalyzed_J3olvol£sis^ A solution of 242 (25 mg) in 2 ml of purified methanol was sealed in a glass ampoule and heated in an oil bath at 100° for 4 hr. The contents of the ampoule were concentrated in vacuo and analyzed on the XF-1150 column: 94% of 260, 4% of 261, and

2% of 262.

Comparable reaction of 240 but at 150° gave an identical distri­ bution of the three ethers.

Photooxygenation of J>1. To a solution of 200 mg of Rose Bengal in 20 ml of methanol, 1.00 g (8.35 mmole) of ^51. in 180 ml of methylene chloride was added. The resulting solution was added to a photolysis vessel equipped with gas inlet frit. The vessel was immersed in an ice water bath, and a slow stream of oxygen was bubbled through the solution. Irradiation of the reaction mixture with a 600 watt DYV tungsten projector lamp at 80 volts in a quartz well began at the 190 same time. After 11 hr of irradiation at 0°, 0.76 g (80 meq) of sodium borohydride was added, and the resulting mixture was allowed to stir

for 2 hr at room temperature. The organic solution was washed with water

(3 x 100 ml), dried over magnesium sulfate and activated carbon, and

concentrated to yield 0.94 g (83%) of a pale yellow oil that was a

6:1 mixture of epimeric alcohols. Crystallization and recrystallization

from low boiling petroleum ether gave 0.26 g (23%) of a white crystall­

ine solid, mp 65-67°, identified as 248. NMR (CCl^) 6 5.70 (d x d,

1H, J = 10.5, 4 Hz), 5.38 (d, 1H, J = 10.5 Hz), 4.03 (m, 1H), 2.78 KBr (broadened s, 1H), 0.98 (m, 6H), 0.53 (m, 1H), 0.05 (m, 1H); v UlctX 3360,

3070, 3035, 3000, 1650, 1440, 1390, 1300, and 1040 cm"1 , m/e 136.0890.

anti-2 ,4-Bishomotropone (251). Manganese dioxide (1.74 g, 0.020

mole) was added to a solution of 0.30 g (2.2 mmole) of 248,in 40 ml of

cyclohexane. The resulting slurry was brought to reflux for 24 hr. The

reaction mixture was filtered through a pad of Celite and concentrated

to give 226 mg (76%) of a pale yellow oil. 1H NMR (CDCl^) <5 6.48 (m, 1H),

5.38 (broadened d, 1H, J = 12 Hz), 2.30-1.00 (m's, 7H), 0.75 (m, 1H)

3080, 3000, 1635, 1410, 1355, 1282, 1206, and 942 cm-1; *Et0H max max 263 nm (e 4500); m/e 134.0733.

Anal. Calcd for CgH^O: C, 80.56; H, 7.51. Found: C, 80.42;

H, 7.90,

syn,anti-2,4-Bishomocycloheptatrienol (249). To a -78° solution

fo 110 mg (0.82 mmole) of ,£51 in 5 ml of ether was added 0.80 ml of

a 25% solution of diisobutylaluminum hydride in heptane. The reaction 191 mixture was allowed to stir for 0.5 hr at -78° before wanning to room temperature. Reaction was stopped by the addition of 0.5 ml of IN sodium hydroxide. After the resulting mixture had stirred for 1 hr, the organic solution was decanted, dried over magnesium sulfate, and concentrated to yield 98 mg (88%) of a colorless oil. Crystallization and recrystallization from low boiling petroleum ether gave 66 mg (60%) of 249. as a white crystalline solid, mp 57-58°. NMR (CCl^) 6 5.57

TTRt* 2.68 (broadened s, 1), 1.50-0.20 (m's, 8); 3240, 3010, 2975, 1665, ulciX

1440, 1376, 1278, 1050, 1010 835, and 728 cm”1 ; m/e 136.0890.

Photoxygenation of ^8 . A 1.00 (8.3 mmole) sample of 4B was photo- oxygenated as described for 21 hr. The reaction mixture was treated with sodium borohydride (0.76 g, 80 meq) at 0° before workup as described. 1H NMR and mass spectral analyses were indicative of a mixture of allylic alcohols 252 and 253 and epoxy alcohols of the type

255. Careful HPLC separation on Florisil (elution with 65:35 hexane- ether) gave 0.35 g (44%) of a 3:1 mixture of /252. and 253^. The alcohol mixture was dissolved in 15 ml of dry pyridine and 0.74 g (4.0 mmole) of p-nitrobenzoyl chloride (recrystallized from carbon tetrachloride) was added. After stirring for 3 hr at room temperature, the reaction mixture was diluted with 50 ml of water. The product was extracted with ether (2 x 25 ml). The combined ether extracts were washed with cold

10% hydrochloric acid and saturated sodium bicarbonate solutions, dried over magnesium sulfate, and concentrated to give an off-white crystalline solid. Recrystallization three times from hexane gave pure 274^ (250 mg, 36%) as a white crystalline solid, mp 99-100°. NMR

(CC14) 6 8.30 (s, 4H), 6.02 (d, 1H, J = 10 Hz), 5.83 (m, 1), 5.62 (d x d

1H, J = 10, 6.5 Kz/, 1.70-1.15 (m, 4), 1.10-0.35 (m, 2H), 0.35-0.00 (m,

2H) ; 3110, 3070, 3000, 1713, 1605, 1515, 1350, 1280, 1110, 1100, and 710 cm"1; m/e 285.1003.

Anal. Calcd for C1cH,-NO.: C, 67.36; H, 5.30; N, 4.91. Found: Id 15 4

syn-2,4-Bishomotropone C256) . A 0.54 g (4.0 mmole) sample of the mixture of 252 and 253 was treated with manganese dioxide (3.5 g, 40 mmole) in 40 ml of cyclohexane as described. Product yield was 0.35 g

(65%) of 25&. ^ NMR (CDCI3) 5 6.33 (d x d, 1H, J = 12, 3 Hz), 5.61

(d x t, 1H, J = 12, 0.8 Hz), 2.08 (m, 2H), 1.53 (m, 2H), 1.02 (m, 2H),

0.70 (m, 1H) , 0.25 (m, 1H) ; max 3000, 2970, 2930, 2860, 1650, 1405, 1350, 1087, and 885 cm"1 ; AEt0H 225 nm (t 5200); m/e 134.0733. max

Anal. Calcd for Cq H^q O: C, 80.56; H, 7.51. Found: C, 80.72;

H, 7.54.

syn,syn-2,4-Bishomocycloheptatrienol (25^J. A 250 mg (1.86 mmole) sample of 253. was treated with 2.0 ml of 25% diisobutylaluminum hydride in heptane as described. Crystallization and recrystallization from petroleum ether yielded 137 (54%) of 253 as a white crystalline solid, mp 76.5-77.5°. 1H NMR (CCl^) 5 5.50 (broadened d, 1H, J = 11 Hz)

5.05 (broadened d, 1H, J = 11 Hz), 4.82 (m, 1H), 2.35 (broadened s, 1),

1.67-0.55 (m's, 6H) , -0.20-0.55 (m, 2H) ; v*®17 3400, 3075, 3015, 3005, max 2880, 1668, 1450, 1390, 1280, 1040, 1010, 905, 855, 835, and 728 cm"1 ; 193 m/e 136.0890.

^neral ^Proc^^e__fpr_.Saniple .Preparation^ in juperacidic ttedia.

A^^Proton^Sangles^ A thin-walled 5 nun NMR tube was immersed in an iso-penta.ie-liguid nitrogen slush bath (-140°) while being flushed with a stream of dry nitrogen. A 0.10 ml sample of spectrograde fluoro- sulfuric acid was introduced via pipette into the bottom of the tube where it quickly solidified. Sulfuryl chloride fluoride (0.30 ml) was next added. The organic compound was dissolved in 0.10 ml of methylene chloride-d^ and was introduced on top of the sulfuryl chloride fluoride.

A thin glass stirring rod was introduced. The NMR tube was allowed to warm until the methylene chloride layer dissolved with stirring. The tube was then stored at -140° until use.

B^^arbon^£5am|^ A 0.4 ml sample of fluorosulfuric acid was added to a nitrogen flushed 10 mm NMR tube immersed in a -140° slush bath., followed by the addition of 1/2 ml of sulfuryl chloride fluoride.

The organic sample was introduced on top of the sulfuryl chloride fluoride. A glass insert containing methylene chloride-d? was inserted and .served as both lock signal and reference as well as a stirring rod.

The NMR tube was warmed until a homogeneous solution was realized. The sample was stored at -140° until use.

jmti-anti-2 ^4-B ishomqcyc lohep tatr ieno 1^ g_-N it rob en zo ate ^ A

224 mg (1.65 mmole) sample of 248 was dissolved in 10 ml of pyridine and 400 mg (2.0 mmole) of g-nitrobenzoyl chloride was added. The resulting solution was allowed to stir at room temperature for 4 hr. The reaction mixture was diluted with 50 ml of water, and the product was extracted with ether (2 x 25 ml). The combined ether extracts were washed with cold 10% hydrochloric acid and saturated sodium bicarbonate solutions, dried over magnesium sulfate, and concentrated to yield an off-white crystalline solid. Recrystallization twice from hexane yielded 284 mg (60%) of .272 as a white, crystalline solid, m.p. 125.0-

125.5°. 1H NMR (CDCI3) 5 8.18 (s, 4H), 5.78-5.25 (m’s, 3H), 1.67-0.42

(m's, 7H), 0.30 (m, 1); m/e 285.1003

Anal. Calcd for C-^H^NO^: c » 67.36; H, 5.30; N, 4.91. Found:

C, 67.39; H, 5.43; H, 4.89.

A

200 mg (1.47 mmole) sample of.249 was treated with 400 mg (2.2 mmole) of £^nitrobenzoyl chloride as described. Workup and recrystallization twice from hexane yielded 280 mg (67%) of 273 as a white crystalline solid, mp 91.5-92.0°. ^ NMR (CDC13) 6 8.27 (s, 4H), 6.08 (m, 1H),

5.92 (m, 1H), 5.38 (m, 1H), 1.70-0.50 (m's, 8H); m/e 285.1005.

Anal. Calcd for C. ..H.. CN0.: C, 67.36; H, 5.30; N, 4.91. Found: 1 0 xd 4 C, 67.26; H, 5.41; N, 4.75.

A

272 mg (2.00 mmole) sample of 2.5.3, was treated with 740 mg (4.0 mmole) of

£-nitrobenzoyl chloride as described. Workup and recrystallization twice from hexane yielded 346 mg (61%) of 275 as a white, crystalline solid, mp 108-109°. ^ NMR (CDCl3) 6 8.30 (s, 4H), 6.20 (m, 1H), 5.75

(broadened d, 1H, J = 11 Hz), 5.20 (broadened d, 1H, = 11 Hz), 195 KB y* 1.65-1.05 (m, 4H) , 1.05-0.0 (m's 4H) ; 1720, 1610, 1530, and 1275 can-1; m/e 285.1005.

Anal. Calcd for ClcHlcN0 : C, 67.36; H, 5.30; N, 4.91. Found: 16 15 4 C, 67.49; H, 5.30; N, 4.92.

Kinetic Studies. A. Preparation of Reagents. Acetone was prepared by distillation from potassium permanganate. Doubly distilled water was employed. Tetrahydrofuran was prepared by distillation from sodium.

Methanol was prepared by distillation from magnesium methoxide. The aqueous acetone and 75:25 methanol-tetrahydrofuran solutions were pre­ pared on a volume to volume basis. Standard 0.0200 N sodium hydroxide wag prepared by the Ohio State University Reagents Laboratory.

D^^rmination^of^Data^ Solutions of ^-nitrobenzoates

273, 274, and 275 in 80% aqueous acetone were prepared by weighing the appropriate ester into a 10.0 ml volumetric flask and filling to the mark with 80% aqueous acetone. The concentration of £-nitrobenzoate varied from 0.0100M to 0.0200M over all runs. The resulting solution was divided into 8 glass ampoules which were sealed under partial

vacuum. All ampoules were simultaneously immersed into a constant

temperature bath. After 5 min, one ampoule was removed from the rate bath and placed in an ice-water mixture. A timer was started upon

removal of the first ampoule. The remaining ampoules were removed and

cooled at appropriate intervals covering two half lives. The final

ampoule was removed after at least 10 half-lives to give an infinity

point. The individual ampoules were allowed to warm to room temperature,

at which point a standard aliquot was removed, diluted with 1 ml of 196 acetone, and titrated against 0.0200 N sodium hydroxide using phenolph-

thalein indicator. First order rate data was determined by measuring

the amount of £-nitrobenzoic acid generated by solvolysis relative to

the experimental infinity point. Duplicate runs agreeing within 5%

were made at all temperatures.

GeneralJProcedure_ for^Preparative Seale^So1volyses.

A. 80% Aqueous Acetone. An 80 mg (0.28 mmole) sample of £-nitro-

benzoate was dissolved in 10 ml of aqueous acetone and 46 mg (0.40 mmole)

of tetramethyl urea. The resulting solution was sealed in a glass

ampoule and immersed in a bath for 10 half-lives. The reaction mixture

was concentrated, and the residue was taken up in 25 ml of ether. The

ether solution was washed with water, dried over magnesium sulfate and

concentrated.

B. 75 :25^ Methanol-Te: trahydrofbran. A 50 mg (0.18 mmole) sample of

j>-nitrobenzoate and 29 mg (0.25 mmole) of tetramethyl urea were

dissolved in 10 ml of 75:25 methanol-tetrahydrofuran, were sealed in a

glass ampoule, and were immersed in a constant temperature bath for an

estimated 10 half-lives. The reaction mixture was concentrated and the

residue taken up in 25 ml of ether. The ether solution was washed with

water, dried over magnesium sulfate, and concentrated.

C. ^ Thermodynamic_Pj~oducts^ A 20 mg sample of each of the alcohols

^248, J249, 252, and 253^ was dissolved in 5 ml of 70:20 tetradydrofuran-

10% hydrochloric acid. The resulting solution was brought to reflux

for 12 hr. The reaction mixture was diluted with 20 ml of water. The

aqueous solution was extracted with ether. The ether solution was washed with saturated sodium bicarbonate solution, dried over magnesium

sulfate, and concentrated. All four alcohols gave identical product

mixtures containing > 95% one alcohol identified as 268. NMR (CCl^)

6 6.0-5.0 (m, 6H), 4.18 (m, 1H), 3.0-2.0 (m, 4H), 1.77 (broadened s, 1H)

3350, 3020, 2920, 2860, 1630, 1455, 1025, 780, and 720 cm-1; BlcLX cyclohexane ^nax 223 nm (e 4400); m/e 136.0890.

Anal. Calcd for CQH,.0: C, 79.37; H, 8 .88. Found: C, 79.44; H, a 12 9.05.

Preparative Scale Solvolysis of ^272^ Solvolysis of 2,72 in aqueous

acetone gave an 88.12 mixture of^248 and ^249. Methanolysis yielded only

276.

Preparative Scale Solvolyses of 27J3-. Solvolysis of 273 in aqueous

acetone gave an 89:11 mixture of 24J3 and ^249^ Methanolysis yielded

only J276,.

anti,anti-2,4-Bishomocy c 1 oh ep t a t r i en °-*- Me thy1 _Eth e r^ JE.276)^. A 24 mg sample of 50% sodium hydride in mineral oil was washed with pentane

and dried under a stream of nitrogen. A 38.8 mg (0.286 mmole) sample

of 248 in 4 ml of tetrahydrofuran was added, and the resulting mixture was brought to gentle reflux for 2 hr. Methyl iodide (142 mg, 1.00 mmole) was added, and the resulting mixture was allowed to stir for

1 hr at room temperature. The reaction mixture was diluted with water

(20 ml) , and the product was extracted with hexane (30 ml) . The hexane

solution was washed with water, dried over magnesium sulfate, and 2 198 concentrated to give 35 mg (82%) of ^276^ as a colorless oil. H NMR

(CC14) 6 5.67 (d x d, 1H, J = 11.4 Hz), 5.40 (d, 1H, J = 11 Hz), 3.57 neat (m, 1), 3.30 (s, 3H), 1.60-0.40 (m's, 7H) , 0.13 (m, 1H) ; VTTiav 3075,

3000, 2930, 2820, 1655, 1405, 1105, 830, and 705 cm-1; m/e 150.1048.

Preparative Scale Solvolyses of 274. Solvolysis of 274 in aqueous acetone gave > 95% of 268. Methanolysis gave > 95% of 277.

Preparative Scale Solvolyses of ,275. Solvolyses of ,275^ gave product mixtures identical to those from 274.

2,4,7-Cyclononatrienol Methyl Ether (277). A 136 mg (1.00 mmole) sample of 268, was treated with 48 mg (2.00 mmole) of sodium hydride and 300 mg (2.1 mmole) of methyl iodide as described. Product yield was 120 mg (80%) of J277, as a colorless oil. NMR (CDCI3) 5 6.15-5.30

(m, 6H) , 3.72 (m, 1H) , 3.50 (s, 3H) , 3.20-2.20 (m, 4H) ; 3010, . cyclohexane 2920, 1630, 1450, 1090, 905, 780, and 620 era-1; Xmax 222 nm

(e 3700); m/e 150.1048.

Determination of Solvolytic Rate Constants. The raw data from kinetic experiments consisted of volumes of perchloric acid (Ve) or volumes of sodium hydroxide (Vt) used to titrate residual sodium acetate

(tosylate acetolysis) or p-nitrobenzoic acid (p-nitrobenzoate solvolysis) as well as the elapsed times at which those ampoules were quenched, including a reading at zero-time (Vc) and one at 10 half-lives (Vc).

The fraction of unreacted tosylate at any given time (Ft) is given by: 199

Ft *» Vt ~ Vco Vo - Vo, and the fraction of unreacted ]D-nitrobenzoate ester at any given time

(Ffc) is given by:

Ft = Vt ~ V« Voo - V0

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 intercept zero. The method of least squares was used to obtain the best slope (-K) of In Ft vs ;t. The error included

•with the rate constants is an average deviation for two runs. The half-life is determined from the relationship

tjy2 = (0.693/k).

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

k = RT AS+/Re-AH*/RT (1) Nh e where k = first order rate constant at temperature “ T 00 R = gas constant

N = Arogadro's constant

h = Plank’s constant

AS^ = entropy of activation

AH+ = enthalpy of activation

Eqn. (1) rearranges to

R[ln(k/T) - In (R/Nh) = AH*(Vt) + Ast (2) 200 which is a linear relationship between the left side of the equation and (V t ) with slope -AH^ and intercept A s V

A program (ACTIPAR-CAL (X)) written for the Wang 360 calculator by Dr. Ian R. Dunkin converts reaction temperatures in °C and rate constants into pairs of dummy variables Xqi and Y^ :

34, = R(k/T) - In (R/Nh)

Yt = 1000/T

The X^, 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.

syn,anti-2,4-Bishomocycloheptatrienol-l-d. jp-Nitrobenzoate (281).

Sodium borodeuteride (17 mg, 0.4 mmole) was added to a 0° solution of

251 in 1 ml of methanol-0-d. After 1 hr, the reaction mixture was diluted with water, and the product was extracted with ether. The combined ether extracts were washed with water, dried and concentrated.

The crude product mixture was treated with p-nitrobenzoyl chloride

(60 mg, 0.3 mmole) in pyridine as described. Recrystallization from hexane yielded a white crystalline solid, mp 90-92°. Methanolysis of this solid as described gave a mixture of methyl ethers displaying a

2.7:1 ratio of olefinic protons to protons alpha to the methoxyl. REFERENCES

1. (a) S. Winstein, Chem. Soc. Spec. Publ., 21, 5 (1967); (b) S. Winstein, Quart. Rev. Chem. Soc., 23, 141 (1969); (c) P. M. Warner, "Topics in Nonbenzoid Aromatic Chemistry;" Vol. II, in press; (d) L. A. Paquette, Angew. Chem., in press.

2. G. Maier, Angew. Chem. Intern. Ed. Engl., 6^, 402 (1967).

3. M. Traetteberg, J. Am. Chem. Soc., 86, 4265 (1964).

4. F. A. L. Anet, J. Am. Chem. Soc., 86, 458 (1964).

5. F. R. Jensen and R. A. Smith, J. Am. Chem. Soc., 86, 956 (1964).

6. R. Huisgen, Angew. Chem., 76, 928 (1964).

7. W. G. Woods, J. Org. Chem., 23, 110 (1958).

8. K. N. Klump and J. P. Chesick, J. Am. Chem. Soc., 85, 130 (1963).

9. E. P. Kohler, M. Tishler, H. Potter, and H. Thompson, J. Am. Chem. Soc., 6!L, 1057 (1939) .

10. K. Alder and G. Jacobs, Chem. Ber., 86, 1528 (1953).

11. R. C. Cookson, S. S. H. Gilani, and I. D. R. Stevens, Tetrahedron Lett., 615 (1962).

12. M. J. Goldstein and A. H. Gervitz, Tetrahedron Lett■, 4413 (1967).

13. M. J. Goldstein and A. H. Gervitz, Tetrahedron Lett., 4417 (1967).

14. E. Ciganek, J. Am. Chem. Soc., 87, 1149 (1965).

15. E. Ciganek, J. Am. Chem. Soc., 93, 2207 (1971).

16. C. J. Fritchie, Jr., Acta Cryst., 20, 27 (1966).

17. H. Gunther, Z. Naturforsch, 20b, 948 (1965).

18. H. J. Dauben, Jr., J. D. Wilson, and J. L. Laity, J. Am. Chem. Soc., 91, 1991 (1969).

19. L. A. Paquette, T. G. Wallis, T. Kempe, G. G. Christoph, J. P. Springer, and J. Clardy, J. Am. Chem. Soc., in press (1977).

201 202

20. N. Bodor, M. J. S. Dewar, and S. D. Worley, J. Am. Chem. Soc., 92, 19 (1970).

21. A. P. Ter Borg, H. Kloosterziel, and A. Van Meurs, Recueil Trav. Chim. Pays-Bas, 82, 717 (1963).

22. A. P. Ter Borg and H. Kloosterziel, Recueil Trav. Chim. Pays-Bas, 84, 245 (1965).

23. A. P. Ter Borg, H. Kloosterziel, and N. Van Meurs, Proc. Chem. Soc. (London), 359 (1962).

24. W. R. Roth, Angew. Chem. Intern. Ed. Engl., 2, 688 (1963).

25. A. P. Ter Borg and H. Kloosterziel, Recueil Trav. Chim. Pays-Bas, 84, 241 (1965).

26. (a) W. G. Dauben and R. L. Cargill, Tetrahedron, 12, 186 (1961); (b) M. V. Evans and R. C. Lord, J. Am. Chem. Soc., 83, 3409 (1961).

27. W. von E. Doering and W. R. Roth, Tetrahedron, 19, 715 (1963).

28. (a) L. Birladeanu, D. L. Harris, and S. Winstein, J. Am. Chem. Soc., 92, 6387 (1970); (b) H. Gunther and J. Ulmen, Chem. Ber., 108, 3132 (1975); (c) H. Gunther, J. B. Pawliczek, J. Ulmen, and W. Grimme, ibid., 108, 3141 (1975); (d) R. Bicker, H. Kessler, and W. Ott, ibid., 108, 3151 (1975).

29. R. B. Woodward and R. Hoffmann, Angew. Chem. Intern. Ed. Engl., 8, 781 (1969).

30. H. E. Zimmerman and G. L. Grunewald, J. Am. Chem. Soc. , 88, 183 (1966).

31. G. Schroeder, Chem. Ber. 97, 3140,3150 (1964); (b) G. Schroeder and J. F. M. Oth, Angew. Chem. Intern. Ed. Engl., 6, 414 (1967).

32. C. G. Overberger and A. E. Borchert, J. Am. Chem. Soc., 82, 1007, 4896 (1960).

33. (a) W. R. Roth and B. Peltzer, Ann., 685, 56 (1965); (b) L. A. Paquette and 0. Cox, J. Am. Chem. Soc., 89, 5633 (1967).

34. T. Sasaki, K. Kanematsu, and Y. Yukimoto, J. Org. Chem., 39, 455 (1974).

35. H. Prinzbach and C. Rucker, Angew. Chem. Intern. Ed. Engl., 15, 559 (1976). 203

36. S. Sawada and Y. Inoue, Bull. Chem. Soc. Japan, 42, 2669 (1969).

37. D. I. Schuster, J. M. Palmer, and S. C. Dickerman, J. Org. Chem., 31, 4281 (1966).

38. J. B. Lambert, A. P. Jovanovich, J. W. Hamersma, F. R. Koeng, and S. S. Oliver, J. Am. Chem. Soc., 95, 1570 (1973).

39. E. Le Goff, J. Org. Chem. , 29^, 2048 (1964) .

40. J. Furukawa, N. Kawabata, and J. Nishimura, Tetrahedron Lett., 3353 (1966); Tetrahedron, 24, 53 (1968).

41 (a) S. Winstein and J. Sonnenberg, J. Am. Chem. Soc., 83, 3235 (1961); (b) W. G. Dauben and G. H. Berezin, J. Am. Chem. Soc., 85, 468 (1963).

42. K. L. Williamson, D. R. Clutter, R. Emch, M. Alexander, A. E. Burroughs, C. Chua, and M. E. Bogel, J. Am. Chem. Soc., 96, 1471 (1974).

43. R. W. Thies, M. Gasic, D. Whalen, J. B. Grutzner, M. Sakai, B. Johnson, and S. Winstein, J. Am. Chem. Soc. , 94_, 2262 (1972) .

44. E. J. Corey and A. Venkateswarlu, J. Am. Chem. Soc., 94, 6190 (1972) .

45. (a) J. C. Collins, W. W. Hess, and F. J. Frank, Tetrahedron Lett., 3363 (1968); (b) R. E. Ireland, D. C. Muchmore, and U. Hengartner, J. Am. Chem. Soc., 94_, 5098 (1972); (c) M. Fetizon, M. Jurion, and N. T. Anh, Chem. Commun., 112 (1969).

46. (a) R. E. Ireland and G. Pfister, Tetrahedron Lett., 2145 (1969); (b) R. E. Ireland, D. C. Muchmore, and U. Hengartner, J. Am. Chem. Soc., 94, 5098 (1972) ; (c) M. Fetizon, M. Jurion, and N. J. Anh, Chem. Commun., 112 (1969).

47. E. M. Burgess, H. R. Penton, Jr., E. A. Taylor, J. Org. Chem., 38, 26 (1973).

48. M. S. Newman, A. Arkell, and T. Fukunaga, J. Am. Chem. Soc. , 82, 2498 (1960).

49. C. Walling and L. Bollyky, J. Org. Chem., 28, 256 (1963); C. H. Snyder and A. R. Soto, ibid. , 29_, 742 (1964) .

50. (a) K. Conrow, M. E. Howden, and D. Davis, J. Am. Chem. Soc., 85, 1929 (1963); (b) J. B. Lambert, L. J. Durham, P. Lepoutere, and J. D. Roberts, ibid., 87, 3896 (1965); (c) R. W. Murray and M. L. Kaplan, ibid., 88^ 3527 (1966); (d) H. Gunther, M. Gorlitz, and H. H. Hinrichs, Tetrahedron, 24, 5665 (1968); (e) W. E. Heyd and C. A. Cupas, J. Am. Chem. Soc., 93, 6086 (1971). 204

51. (a) T. Norin, Acta Chem. Scand., 17, 738 (1963); (b) R. R. Sauers and P. E. Sonnet, Chem. Ind. (London), 786 (1963); (c) S. Forsen and T. Norin, Tetrahedron Lett., 2845 (1964); (d) J. Tadanier and W. Cole, J. Org. Chem., 27, 4610 (1962); (e) K. Tori and K. Kitahonoki, J. Am. Chem. Soc. , 87., 386 (1965) ; (f) Y. E. Rhodes, P. E. Schueler, and V. G. DiFate, Tetrahedron Lett., 2073 (1970); (g) W. E. Volz and L. A. Paquette, J. Org. Chem., 41, 57 (1976).

52. (a) J. B. Grutzner, M. Jautelat, J. B. Dence, R. A. Smith, and J. D. Roberts, J. Am. Chem. Soc. , 92_, 7107 (1970) ; (b) K. Tori, M. Ueyama, T. Tsuji, H. Matsumura, and H. Tanida, Tetrahedron Lett., 327 (1974); (c) M. Christl, Chem. Ber., 108, 2781 (1975).

53. L. A. Paquette, M. J. Carmody, and J. M. Geckle, Tetrahedron Lett., 2753 (1976).

54. D. S. Glass, R. Boikess, and S. Winstein, Tetrahedron Lett., 999 (1966).

55. (a) D. S. Glass, J. Zirner, and S. Winstein, Proc. Chem. Soc., 276 (1963); (b) K. Heger and W. Grimme, Angew. Chem., 88, 60 (1976); Angew. Chem. Intern. Ed. Engl., 15, 53 (1976).

56. W. R. Roth, Ann., 671, 10 (1964).

57. P. G. Gassman and W. J. Greenlee, Org. Syn., 53, 38 (1973).

58. (a) E. Muller, H. Fricke, and W. Rundel, Z. Naturforsch, 15b, 753 (1960); (b) W. v. E. Doering and P. P. Gaspar, J. Am. Chem. Soc., 85, 3043 (1963).

59. J. A. Berson, R. A. Boettcher, and J. J. Vollmer, J. Am. Chem. Soc., 93, 1540 (1971).

60. D. L. Garin, J. Am. Chem. Soc., 92^ 5254 (1970).

61. K. G. Untch and D. J. Martin, J. Am. Chem. Soc., 87, 4501 (1965).

62. K. Untch and R. J. Kurland, J. Am. Chem. Soc., 85, 346 (1963).

63. R. H. Shapiro and M. Heath, J. Am. Chem. Soc., 89^ 5734 (1967).

64. (a) A. de Meijere, Tetrahedron Lett., 1845 (1974); (b) A. de Meijere and L. U. Meyer, ibid., 1849 (1974); (c) H. Hogeveen and H. C. Volger, J. Am. Chem. Soc., 89, 2486 (1967); (d) H. C. Volger, H. Hogeveen, and M. M. P. Gaasbeek, ibid., 91, 218 (1969); (e) L. Cassar, P. E. Eaton, and J. Halpem, ibid. , 92, 3515 (1970); (f) K. C. Bishop III, Chem. Rev., 76, 461 (1976). 205

65. (a) F. A. L. Anet, A. J. R. Bourn, and Y. S. Lin, J. Am. Chem. Soc., £6,3576 (1964); (b) L. A. Paquette, Tetrahedron, 31, 2855 2855 (1975).

66. E. Vogel, H. Konigshofen, K. Mullen, and J. F. M. Oth, Angew. Chem. Intern. Ed. Engl., 1£, 281 (1974).

67. L. A. Paquette, J. M. Gardlik, and J. M. Photis, J. Am. Chem. Soc., 98, 7096 (1976).

68. (a) J. S. McKennis, L. Brener, J. S. Ward, and R. Pettit, J. Am. Chem. Soc., 93, 4957 (1971); (b) L. A. Paquette, R. P. Davis, and D. R. James, Tetrahedron Lett., 1615 (1974); (c) L. A. Paquette, D. R. James and G. Klein, J. Org. Chem., in press; (d) G. Klein and L. A. Paquette, J. Org. Chem., in press.

69. R. S. Boikess and S. Winstein, J. Am. Chem. Soc., 85, 343 (1963).

70. H. Prinzbach, V. Wessely, and H. Fritz, Tetrahedron Lett., 2765 (1976).

71. P. Radlick and S. Winstein, J. Am. Chem. Soc., 85, 344 (1963).

72. K. G. Untch, J. Am. Chem. Soc., 85, 345 (1963).

73. G. I. Fray, J. Chem. Soc., 4284 (1963).

74. P. G. Gassman, J. Seter, and F. J. Williams, J. Am. Chem. Soc., 93, 1673 (1971).

75. J. W. H. Watthey and S. Winstein, J. Am. Chem. Soc., 85, 3715 (1963).

76. W. R. Roth, W. B. Bang, P. Goebel, R. L. Sass, R. B. Turner, and A. P. Yu, J. Am. Chem. Soc., 86, 3178 (1964).

77. P. Bischof, R. Gleiter, and E. Heilbronner, Helv. Chim. Acta, 53, 1425 (1970).

78. (a) R. H. DeWolfe and W. G. Young in "The Chemistry of Alkenes", Vol. I, S. Patai, Ed., Wiley-Interscience, New York, N. Y., 1964, pp 681-710; (b) N. C. Deno in "Carbonium Ions," Vol. II, G. A. Olah and P. v. R. Schleyer, Ed., Wiley-Interscience, New York, N. Y., 1970, Chapter 18; (c) H. G. Richey in "The Chemistry of Alkenes," Vol. II, J. Zabicky, Ed., Wiley-Interscience, New York, N. Y., L(&), PP 56-67.

79. (a) P. R. Story and B. C. Clark in "Carbonium Ions," Vol. Ill, G. A. Olah and P. v. R. Schleyer, Ed., Wiley-Interscience, New York, N. Y., 1972, Chapter 23; (b) H. G. Richey in ref. 78c, pp 77-101. 206

80. (a) H. G. Richey in ref. 79a, Chapter 25; (b) K. B. Wiberg, B. A. Hess, and A. J. Ashe, ibid., Chapter 26; (c) M. Hanack and H.-J. Schneider, Angew. Chem. Intern. Ed. Engl. , 6_, 666 (1967) ; (d) M. Hanack and H.-J. Schneider, Fortschr. Chem. Forsch., 8, 554 (1967).

81. J. Haywood-Farmer, Chem. Rev., 74, 315 (1974)

82. (a) C. U. Pittman, Jr., and G. A. Olah, J. Am. Chem. Soc. , 87, 3000 (1965); (b) G. A. Olah, C. L. Jeuell, D. P. Kelley, and R. D. Porter, ibid., 94, 146 (1972) ; (c) G. A. Olah and G. Liang, ibid., 95, 3792 (1973); (d) D. P. Kelley and H. C. Brown, ibid., 97, 3897 (1975); (e) W. J. Hehre and P. C. Hiberty, ibid., 96, 302 (1974); (f) D. F. Eaton and T. G. Traylor, ibid., 96, 1226 (1974).

83. (a) P. v. R. Schleyer and V. Buss, J. Am. Chem. Soc., 91, 5880 (1969); (b) J. C. Martin and B. R. Ree, ibid., 91, 5882 (1969); 92, 1660 (1970).

84. Consult, for example, (a) A. F. Diaz and S. Winstein, J. Am. Chem. Soc., 91, 4300 (1969); (b) H. C. Brown, C. J. Kim, C. J. Lancelot, and P. v. R. Schleyer, ibid., 92, 5244 (1970); (c) P. Ahlberg, D. L. Harris, M. Roberts, P. Warner, P. Seidl, M. Sakai, D. Cook, A. Diaz, J. P. Dirlam, H. Hamberger, and S. Winstein, ibid., 94, 7063 (1972).

85. (a) S. Winstein, E. C. Friedrich, P. Baker, and Y. Lin, Tetrahedron, Suppl. , 8^, 621 (1966) ; (b) S. Winstein, J. Sonnenberg, and L. deVries, J. Am. Chem. Soc., 81, 6523 (1959); (c) S. Winstein and J. Sonnenberg, ibid., 83, 3244 (1961).

86. (a) H. Tanida, T. Tsuji, and T. Irie, J. Am. Chem. Soc., 89, 1953 (1967); (b) M. A. Battiste, C. L. Deyrup, R. E. Pincock, and J. Haywood-Farmer, ibid., 89_, 1954 (1967); (c) J. Haywood-Farmer and R. E. Pincock, ibid., 91, 3020 (1968).

87. ■ (a) M. J. S. Dewar and J. M. Harris, J. Am. Chem. Soc., 90, 4468 (1968); 92, 6557 (1970); (b) Y. E. Rhodes and T. Takino, ibid., 90, 4469 (1968).

88. (a) A. Streitweiser, J. Org. Chem., 22, 861 (1957); (b) W. Kirmse, Angew. Chem. Intern. Ed. Engl. , 15^, 251 (1976) .

89 (a) J. A. Berson and P. Reynolds-Warnhoff, J. Am. Chem. Soc., 84, 682 (1962); 86, 595 (1964); (b) J. A. Berson and D. Willner, ibid., 84, 575 (1962); 86, 609 (1964).

90 C. J. Collins, Chem. Soc. Revs., 4_, 251 (1975).

91. L. A. Paquette, R. P. Henzel, and R. F. Eizember, J. Org. Chem., 38, 3257 (1973). 207 92. P. Radlick and S. Winstein, J. Am. Chem. Soc., 86, 1866 (1964).

93. S. Winstein, P. Bruck, P. Radlick, and R. Baker, J. Am. Chem. Soc., 86, 1867 (1964).

94. A. C. Cope, S. Moon, and C. H. Park, J. Am. Chem. Soc., 84, 4850 (1962).

95. (a) C. C. Lee and L. K. M. Lam, J. Am. Chem. Soc., 88, 2834 (1966); (b) C. J. Collins and M. H. Lietzke, ibid., 89, 6570 (1967); (c) C. J. Collins and C. E. Harding, Ann., 745, 124 (1971).

96. (a) H. A. Corver and R. F. Childs, J. Am Chem. Soc., 94, 6201 (1972); (b) L. A. Paquette, M. J. Broadhurst, P. Warner, G. A. Olah, and G. Liange, ibid., 95, 3386 (1973); (c) D. Whalen, M. Gasic, B. Johnson, H. Jones, and S. Winstein, ibid. , 89^, 6384 (1967).

97. (a) W. G. Dauben and F. G. Willey, J. Am. Chem. Soc., 84, 1497 (1962). (b) W. Kirmse and R. Siegfried, ibid., 90_, 6564 (1968); (c) W. Kirmse and G. Voigt, ibid., 96, 7598 (1974); (d) W. Kirmse and T. Olbricht, Chem. Ber., 108, 2616, 2629 (1974); (e) R. Siegfried, Tetrahedron Lett., 4669 (1975); (f) W. Kirmse and H. A. Rinkler, Ann., 707, 57 (1967).

98. W. J. Farissey, Jr., R. H. Perry, Jr., F. C. Stehling, and R. F. Chamberlain, Tetrahedron Lett., 3635 (1964).

99. C. D. Poulter, E. C. Friedrich, and S. Winstein, J. Am. Chem. Soc., 91, 6892 (1969); 92, 4274 (1970).

100. S. J. Cristol, G. C. Schloemer, D. R. James, and L. A. Paquette, J. Org. Chem., 37, 3852 (1972).

101. S. Winstein, J. Am. Chem. Soc., 81, 6524 (1959).

102. (a) S. Winstein, Chem. Soc. Spec. Publ., No. 21, 5 (1967); (b) Quart. Rev., Chem. Soc., 23, 141 (1969); (c) P. Warner, Topics in Nonbenzenoid Aromatic Chemistry, Vol. II, in press; (d) L. A. Paquette, Angew. Chem. Intern. Ed. Engl., in press.

103. (a) J. L. Rosenberg, J. E. Mahler, and R. Pettit, J. Am. Chem. Soc., 84, 2842 (1962); (b) C. E. Keller and R. Pettit, ibid., 88, 604 (1966); (c) J. D. Holmes and R. Pettit, ibid., 85, 2531 (1963).

104. (a) S. Winstein, C. G. Kreiter, and J. I. Brauman, J. Am. Chem. Soc., 88, 2047 (1966); (b) J. A. Berson and J. A. Jenkins, ibid., 94, 8907 (1972). 208

105. L. A. Paquetter, M. J. Broadhurst, P. Warner, G. A. Olah, and G. Liang, J. Am. Chem. Soc., 95, 3386 (1973) .

106. (a) H. D. Kaesz, S. Winstein, and C. G. Kreiter, J. Am. Chem. Soc., 88, 1319 (1966); (b) R. Auman and S. Winstein, Tetra­ hedron Lett., 903 (1970).

107. (a) S. Winstein, H. D. Kaesz, C. G. Kreiter, and E. C. Friedrich, J. Am. Chem. Soc., 87, 3267 (1965); (b) P. Warner, D. L. Harris, C. H. Bradley, and S. Winstein, Tetrahedron Lett., 4013 (1970).

108. 0. L. Chapman and R. A. Fugiel, J. Am. Chem. Soc., 91, 215 (1969).

109. M. Oda, Y. Kayama, H. Miyasaki, and Y. Kitahara, Angew. Chem. Intern. Ed. Engl., 1£, 418 (1975).

110. J. D. Holmes and R. Pettit, J. Am. Chem. Soc. , 65_, 2531 (1963) .

111. M. Brookhart, M. Ogliaruso, and S. Winstein, J. Am. Chem. Soc., 89, 1965 (1967).

112. (a) G. Boche, W. Hechtl, H. Huber, and R. Huisgen, J. Am. Chem. Soc., 89, 3344 (1967); (b) R. Huisgen, G. Boche, and H. Huber, ibid., 94, 6541 (1972).

113. L. A. Paquette, J. R. Malpass, and T. J. Barton, J. Am. Chem. Soc.,91, 4714 (1969).

114. (a) L. A. Paquette, U. Jacobsson, and S. V. Ley, J . Am. Chem. Soc., 98, 152 (1976); (b) J. Gasteiger, and R. Huisgen, ibid., 94, 6541 (1972).

115. W. Kitching, K. A. Henzel, and L. A. Paquette, J. Am. Chem. Soc., 97, 4643 (1975).

116. P. Warner, Ph.D. Dissertation, University of California at Los Angeles, Los Angeles, California, 1970.

117. P. Ahlberg, D. L. Harris, and S. Winstein, J. Am. Chem. Soc., 92, 2146 (1970).

118. D. Cook, A. Diaz, J. P. Dirlam, D. L. Harris, M. Sakai, S. Winstein, J. C. Barborak, and P. von R. Schleyer, Tetrahedron Lett., 1405 (1971).

119. P. Ahlberg, D. L. Harris, M. Roberts, P. Warner, P. Seidl, M. Sakai, D. Cook, A. Diaz, J. P. Dirlam, H. Hamberger, and S. Winstein, J. Am. Chem. Soc. , 94^, 7064 (1972) .

120. L. A. Paquette, M. Oku, W. B. Famham, G. A. Ohah, and G. Liang, J. Org. Chem., 40, 700 (1975). 209

121. A. S. Kende and T. L. Bogard, Tetrahedron Lett., 3382 (1967).

122. G. Schroder, U. Prange, N. S. Bowman, and J. F. M. Oth, Tetra­ hedron Lett., 3251 (1970).

123. M. Roberts, H. Hamberger, and S. Winstein, J. Am. Chem. Soc., 92, 6346 (1970).

124. G. Schroder, U. Prange, and J. F. M. Oth, Chem. Ber., 105, 1854 (1972).

125. G. Schroder, U. Prange, B. Putze, J. Thio, and J. F. M. Oth, Chem. Ber., 104, 3406 (1971).

126. L. A. Paquette and M. J. Broadhurst, J. Org. Chem., 38, 1886 (1973).

127. H. A. Corver and R. F. Childs, J. Am. Chem. Soc., 94, 6201 (1972).

128. P. Warner and S. Winstein, J. Am. Chem. Soc., 93, 1284 (1971).

129. L. A. Paquette, P. B. Lavrik, and R. H. Summerville, J. Org. Chem., 42, 2659 (1977).

130. P. Hildebrand, G. Schroder, and J. F. M. Oth, Tetrahedron Lett., 2001 (1976).

131. (a) R. W. Thies, M. Gasic, D. Whalen, J. B. Grutzner, M. Sakai, B. Johnson, and S. Winstein, J. Am. Chem. Soc., 94, 2262 (1972); (b) R. W. Thies, M. Sakai, D. Whalen, and S. Winstein, ibid., 94, 2270 (1972).

132. (a) G. C. Fettis and J. H. Knox in "Progress in Reaction Kinetics," Vol. 2, G. Porter, ed., Macmillan, New York, N. Y., 1964; (b) E. S. Huyser in "Advances in Free Radical Chemistry," Vol. 1, G. H. Williams, ed., Logos Press, London, England, 1965; (c) G. Lanchec, Chim. Ind. (Paris), 94, 46 (1965); (d) N. Coleboum and E. S. Stern, J. Chem. Soc., 3599 (1965); (e) B. Blouri, G. Lanchec, and P. Rumpf, Compt. Rend., 257, 3609 (1963); (f) G. A. Russell, J. Am. Chem. Soc., 80, 4997 (1958).

133. For an excellent review, consult M. L. Poutsma in "Methods Free Radical Chemistry," 1, 79 (1969).

134. Inter alia: (a) G. A. Russell and H. C. Brown, J. Am. Chem. Soc., 77, 4031 (1955); (b) G. A. Russell, Tetrahedron, 8_, 101 (1960); (c) J. M. Tedder, Quart. Rev. (London), 14, 336 (1960).

135. See for example K. B. Wiberg and E. L. Notell, Tetrahedron, 19, 2009 (1963). 210

136. C. Walling and P. S. Fredricks, J. Am. Chem. Soc., 84, 3326 (1962).

137. For other examples which reflect the reduced reactivity of cyclo­ propane rings, see D. E. Applequist and J. A. Landgrebe, J. A m . Chem. Soc., 86, 1543 (1964).

138. J. D. Roberts and R. H. Mazur, J. Am. Chem. Soc., 73, 2509 (1951)

139. K. B. Wiberg, G. M. Lampman, R. P. Ciula, D. S. Connor, P. Schertler, and J. Lavanish, Tetrahedron, 21, 2749 (1965).

140. R. T. LaLonde, J. Am. Chem. Soc., 87, 4217 (1965).

141. M. L. Poutsma, J. Am. Chem. Soc., 87, 4293 (1965).

142. C. Walling and W. Thaler, J. Am. Chem. Soc., 83, 3877 (1961).

143. A. de Mieiere, 0. Schallner, and C. Weitemeyer, Angew. Chem., 84, 63 (1972); Angew. Chem. Intern. Ed. Engl., 11, 56 (1972).

144. O. Schallner, Ph.D. Thesis, University of Gottingen (1974).

145. For a brief review, see J. W. Wilt in "Free Radicals, Volume I," J. K. Kochi, Ed., John Wiley and Sons, New York, 1973, p. 398.

146. J. A. Landgrebe and L. W. Becker, J. Am. Chem. Soc., 89, 2505 (1967); 90, 395 (1968).

147. For excellent reviews, consult (a) R. H. Denney and A. Nickon, Org. Reactions, 20, 133 (1973); (b) D. R. Kearns, Chem. Rev., 71, 395 (1971); (c) K. Gollnick, Adv. Photochem., 6, 1 (1968).

148. C. D. Poulter, E. C. Friedrich, and S. Winstein, J. Am. Chem. Soc., 91, 6892 (1969); 92, 4274 (1970).

149. (a) M. J. Jorgenson, Tetrahedron Lett., 559 (1962); (b) H. C. Brown and H. M. Hess, J. Org. Chem., 34, 2206 (1969); (c) W. L. Dilling and R. A. Plepys, ibid., 35, 2971 (1970); (d) K. E. Wilson, R. T. Seidner, and S. Masamune, Chem. Commun., 213 (1970) .

150. N Heap and G. H. Whitman, J. Chem. Soc. B . 164 (1966).

151. R. T. Taylor, Ph.D. Thesis, The Ohio State University, 1977. Hydrocarbon 163 was prepared by addition of methyllithium to cis-bicyclo[5.1.01oct-2-en-4-one [L. A. Paquette, G. V. Meehan, R. P. Henzel, and R. F. Eizember, J. Org. Chem., 38, 3250 (1973)] and subsequent dehydration with iodine in acetic acid. 211

152. P. v. R. Schleyer, W. F. Sliwinski, G. W. Van Dine, U. Scholl- kopf, J. Paust, and K. Fellenberger, J. Am. Chem. Soc., 94, 125 (1972) ; W. F. Sliwinski, T. M. Su, and P. v. R. Schleyer, ibid., 94, 133 (1972) and earlier references cited in these papers.

153. X. Creary, J. Am. Chem Soc., 98, 6608 (1976); D. B. Ledlie, T. Swan, J. Pile, and L. Powers, J. Org. Chem., 41, 419 (1976); D. B. Ledlie, W. Barber, and F. Suritzen, Tetrahedron Lett., 607 (1977), and references contained therein.

154. R. B. Woodward and R. Hoffman, J. Am. Chem. Soc., 87, 395 (1965); H. C. Longuet-Higgins and E. W. Abrahamson, ibid., 87, 2045 (1965).

155. B. A. Howell and J. G. Jewett, J. Am. Chem. Soc. , 93^ 798 (1971) .

156. L. Radom, J. A. Pople, and P. v. R. Schleyer, J. Am. Chem. Soc., 95, 8193 (1973).

157. W. Hanstein, H. J. Berwin, and T. G. Traylor, J. Am. Chem. Soc., 92, 829 (1970).

158. L. Radom, J. A. Pople, P. C. Hariharan, and P. v. R. Schleyer, J. Am. Chem. Soc. , 95_, 6531 (1973) and references therein; V. Buss, P. v. R. Schleyer, and L. C. Allen, Top. Stereochem. , 7_, 253 (1973); D. Aue, W. R. Davidson, and M. T. Bowers, J. Am. Chem. Soc. , 9£[, 6700 (1976) .

159. E. M. Kosower, "An Introduction to Physical Organic Chemistry," John Wiley and Sons, Inc., New York, 1968; p. 233; M. J. S. Dewar and S. Kirschner, J. Am. Chem. Soc., 93, 4290 (1971).

160. J. D. Roberts and D. Schuster, J. Org. Chem., 27, 51 (1962).

161. K. B. Wiberg and T. Nakahira, Tetrahedron Lett., 1769 (1974); 1773 (1974).