Citation
The 2, 4-ethano- and 2,4-etheno-3-alkoxy-tris-homocyclopropenyl cations, orientation and stereochemistry of nucleophile capture

Material Information

Title:
The 2, 4-ethano- and 2,4-etheno-3-alkoxy-tris-homocyclopropenyl cations, orientation and stereochemistry of nucleophile capture
Creator:
Nielsen, Warren Charles, 1945- ( Dissertant )
Battiste, Merle A. ( Thesis advisor )
Tarrant, Paul ( Reviewer )
Dolbier, William R. ( Reviewer )
Stoufer, Robert C. ( Reviewer )
Allen, Charles M. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1974
Language:
English
Physical Description:
x, 131 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Absorption spectra ( jstor )
Aluminum ( jstor )
Chlorides ( jstor )
Ethers ( jstor )
Hydrides ( jstor )
Infrared spectrum ( jstor )
Mass spectroscopy ( jstor )
Protons ( jstor )
Solvents ( jstor )
Spectral bands ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Solvolysis ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Solvolyses of endo, anti- tri cycl o [3. 2 . 1 . 2 ' " ].octanyl derivatives, i.e. , hydrogen, vinyl, and p-anisyl substitution at the bridge C 8 position, have been shown to afford rearranged products {endo C 2 or C^ attack) almost exclusively (>99%). Concentration of positive charge at the C 2 and C^ positions of the respective intermediate tri s-homocycl opropenyl cations might be an explanation; an alternative view stresses the importance of strain relief in the transition state for solvent capture of the non-classical cations. In contrast to the above results, acidic hydrolysis of 8 ,8-dimethoxy-endo-tri cycl o [3. 2 . 1 . 2 ' k ] octane' (I) yields the expected ketone II with no evidence for the formation of a rearranged hydroxyether. This suggests that the charge stabilizing ability of the methoxyl group at the bridge carbon may well be overwhelming the skeletal bias of the transition state for solvent capture. Saturated ketal I was treated with dichloroal umi num hydride to yield the sz/n-methyl ether exclusive of its anti epimer. Under the same reaction conditions, the unsaturated ketal , 8,8-dimethoxy-endo-tricyclo [3.2.1.0 2,1+ ]oct-6-ene (III), gave the unsaturated syn- methyl ether IV as the major product along with l-methoxy-endo-6-chl orotri cycl o [3. 3 . . 2 ' 8 ] - oct-3-ene (V) as a minor product. Traces of endo- 6- me thoxyand endo-6-chlorotricyclo[3. 3. . 2 ' 8 ] oct-3-enes,(VI) and (VII), were also detected. The trace components are attributed to the intermediacy of unsaturated anti-methyl ether VIII as evidenced by its authentic synthesis and subsequent exposure to the reactions conditions. The predominance of bridge C 8 anti hydride attack and formation of endo unsaturated rearranged chloride V argue strongly for the intermediacy of a delocalized system with charge concentrated at C a due to methoxyl stabilization. The transient formation of the unsaturated anti methylether VIII could result from hydride attack upon a bis-homocyclopropenyl cation, whose mode of formation has several potential routes. Use of an 8:1 molar ratio of the aluminum chloride/ lithium aluminum hydride reagent with the respective saturated and unsaturated methyl ketals I and III produced good yields of the saturated and unsaturated rearranged methoxy chlorides IX and V, both of which were solvolyzed at 100° in aqueous ethanol. For the unsaturated chloride V, the major product, cycloheptatriene , and a mixed ethoxy-methoxy ketal fraction of the original tri cycl o [3. 2 . 1 . 2 » k ] octenyl structure were isolated. Cy cl ohep ta tr i ene is the expected product from the well-known decarbonyl ati on of the unsaturated ketone. According to PMR analysis the mixed ketal fraction consisted of 89% arcti-ethoxy-syrc-methoxy unsaturated ketal and 11% antimethoxy- st/n-ethoxy ketal. Solvolysis of the saturated chloride IX in aqueous ethanol produced the saturated ketone II as the major product with a mixed ketal fraction consisting of 96% anti-ethoxy-syn-methoxy ketal and 4% anti-methoxy-syn ethoxy ketal . The results obtained strongly indicate that the same cations are being generated from the respective reactions of the saturated and unsaturated ketals I and III with dichloroaluminum hydride and aqueous acid, as well as in the solvolyses of the saturated and unsaturated rearranged methoxy chlorides IX and V. The electronic structure of these ions has now been radically altered with respect to the parent ions in that positive charge del ocal i zati on is not as extensive, giving a more localized or concentrated charge at the methoxyl bridge carbon. The cyclopropyl del ocal i zati on may even have been weakened to the point where nucleophiles are able to penetrate and attack the bridge from the syn as well as anti face of the bridge. Thus an electronic effect, and not steric bias, is the overwhelming factor in determining the orientation and stereochemistry of solvent or nucleophilic attack on the 2,4-ethano- or 2 ,4-etheno-3-al koxy- tri s-homocycl opropeni urn cations
Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 125-130.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Warren Charles Nielsen. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000869114 ( AlephBibNum )
AEG6136 ( NOTIS )
014237878 ( OCLC )

Downloads

This item has the following downloads:


Full Text





THE 2,4-ETHANO- and 2,4-ETHENO-3-ALKOXYTRIS-HOMOCYCLOPROPENYL CATIONS, ORIENTATION AND
STEREOCHEMISTRY OF NUCLEOPHILE CAPTURE










By



WARREN CHARLES NIELSEN


A DISSERTATION PRESENTED TO THE GRADUATE
COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY










UNIVERSITY OF FLORIDA


1974



























DEDICATION To My Parents












ACKNOWLEDGEMENTS


The writer wishes to express his most sincere gratitude to Professor Merle A. Batiste for his excellent guidance, interest, and enthusiasm in the development of this research project. It has been a pleasure to work with a research

director who has the ability to maintain a high level of professionalism while extending a very real, personal friendship. Appreciation is expressed to Dr. Roy W. King for his many hours of advice and assistance. The author also feels a need to express his admiration and fondness for Gainesville the people, the land, and the free spirit. The writer's wife Judi deserves such an expression of gratitude that one feels incapable of the words. Her patience during hours of superb typing, and her overall giving nature defy description.








TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS.

LIST OF TABLES. vi

LIST OF FIGURES. vii

ABSTRACT viii

CHAPTER

I Introduction. 1

II Results and Discussion. 32

Synthesis of Precursors. 32
Statement of Problem . . . . 33
Reaction Analysis of 8,8-Dimethoxy-endo-tricyclof3.2.1.D2'4]octane, (69), with Dichioroaluminum Hydride Reagent . 36
Synthesis of anr-3-Methoxy-eno-tricyclo[3.2.1.02'4]octane, (90), and Subsequent
Reaction with Dichloroaluminum Hydride
Reagent . . . 37
Reaction Analysis of 3,8-Dimethoxy-endo-tricyclo[3.2. 1.02,4]oct-6-ene, (78), with
Dichloroaluminum Hydride Reagent. . 40
Synthesis of anti-8-Methoxy-endo-tricyclo[3.2.1.02,' oct-6-ene, (101), and Subsequent Reaction with Dichloroaluminum
Hydride Reagent. . 48
Synthesis of 1-Methoxy-endo-6-chlorotricyclo[3.3.D.02'8]oct-3-ene, (98), and 1-Methoxyendo-4-chlorotricyclo[3.3.0.02 ,']octane, (108), with an 8:1 Molar Ratio Aluminum
Chloride/Lithium Aluminum Hydride Reagent. 51
Solvolyses of 1-Methoxy- and l-Ethoxyendo-6-chlorotricyclo[3.3.0.02'8]oct-3enes, (98) and (114). . . 53
Solvolyses of 1-Methoxy- and 1-Ethoxy-endo4-chlorotricyclo[3.3.0.02 '6octanes, (108)
and (117) . . . 57
Reaction of 1-Methoxy-e:do-4-chlorotricyclo(3.3.0.02'8]octanes, (008), with Silver
Perchlorate . . . . . . 61
Synthesis of endo-6-Chlorotricyclo[3.3.0.D2"]oct-3-ene, (97). . 63







Chapter Page

III PMR Studies. 67

Syn and Anti Chemical Shifts of Alkoxyl
Groups at C8 in the Tricyclo[3.2.1.02']
octanyl-octenyl Systems. 67
Computer Andlyzed PMR Spectra. 70

IV Experimental 86

BIBLIOGRAPHY. 125

BIOGRAPHICAL SKETCH. 131







LIST OF TABLES


Table Page

I Hydride Reduction Product Distribution for
Ketone -(70). 61

II Syn and Anti Ce Alkoxyl Chemical Shifts Tricyclo(3.2. 1.02'4]octenyl Systems. 68

III Syn and Anti C8 Alkoxyl Chemical Shifts in
Tricyclo[3.2. 1.02,4] octanyl Systems . 69







LIST OF FIGURES

PMR Spectra

Figure Page

1 endo-4-Methoxytricyclo[3.3.0.Q2'8]octane, (92) 73 2 endo-4-Chlorotricyclo[3.3.0.02'8]octane, (93) 74

3 endo-6-Chlorotricyclo[3.3.0.02 ,8]oct-3-ene,
(97) 75

4 1-Methoxy-endo-6-chlorotricyclo[3.3.O.02 ' oct-3-ene, (98) 76

5 1-Methoxy-endo-4-chlorotricyclo[3.3.0.02 ' ]octane (108). 77

6 anti-8- Ethoxy-syn-8-methoxy-eno-tricyclo[ 3. 2 . 1. 02'"]oct-6-ene, (112) 78

7 anti-8-Me thoxy-syn-8-ethoxy-endo-tri cyclo[3.2.1.02', ] oct-6-ene, (113). 79

8 anti-8-Ethoxy-syn-8-methoxy-eu -tricyclo(3.2.1.02'4]octane, (115) 80

9 anti-8-Methoxy-syn-8-ethoxy-ctbo-tricyclo[3.2.I.02'4]octane, (116). 81

10 1,5,6,7-Tetrachloro-8,8-dimethoxy-endc-tricyclo(3.2.1.0"''Ioct-6-ene, (88), at 250 Hz. 82 11 1,5,6,7,8,8-Hexachloro-iufo-tricyclo[3.2. 1.02,1]oct-6-ene, (125), at 250 Hz. 83

12 1,5,6,7-Tetrachl oro-S-di chl oromethyl ene-en
tricyclo[3.2. 1.02'4]oct-6-ene, (126), at
250 Hz. 84

13 1,5,6,7-Tetrachloro-e'.Jo-tricyclo[3.2. 1.0211]oct-6-ene, (127), at 250 Hz . 85





Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy





THE 2,4-ETHANO- and 2,4-ETHENO-3-ALKOXYTRIS-HOMOCYCLOPROPENYL CATIONS, ORIENTATION AND
STEREOCHEMISTRY OF NUCLEOPHILE CAPTURE

By

Warren Charles Nielsen

December, 1974

Chairman: Dr. Merle Battiste Major Department: Chemistry

Solvolyses of endo,anti-tricyclo[3.2.1.042.octanyl

derivatives, i.e., hydrogen, vinyl, and p-anisyl substitution at the bridge Ca position, have been shown to afford rearranged products (endo C2 or C4 attack) almost exclusively (>99%). Concentration of positive charge at the C2 and C4 positions of the respective intermediate tris-homocyclopropenyl cations might be an explanation; an alternative view stresses the importance of strain relief in the transition state for solvent capture of the non-classical cations.

In contrast to the above results, acidic hydrolysis

of 8,8-dimethoxy-endo-tricyclo[3.2. 1.02'4]octane (I) yields the expected ketone II with no evidence for the formation of a rearranged hydroxyether. This suggests that the charge stabilizing ability of the methoxyl group at the bridge carbon may well be overwhelming the skeletal bias of the transition state for solvent capture.


viii





Saturated ketal I was treated with dichloroaluminum hydride to yield the syn-methyl ether exclusive of its anti epimer. Under the same reaction conditions, the unsaturated ketal, 8,8-dimethoxy-endo-tricyclo[3.2.1.02'4]oct-6-ene (III), gave the unsaturated syn-methyl ether IV as the major product along with 1-methoxy-endo-6-chlorotricyclo[3.3.0.02'8]oct-3-ene (V) as a minor product. Traces of endo-6-methoxyand endo-6-chlorotricyclo[3.3.0.02']oct-3-enes,(VI) and (VII), were also detected. The trace components are attributed to the intermediacy of unsaturated anti-methyl ether VIII as evidenced by its authentic synthesis and subsequent exposure to the reactions conditions. The predominance of bridge C8 anti hydride attack and formation of endo unsaturated rearranged chloride V argue strongly for the intermediacy of a delocalized system with charge concentrated at C8 due to methoxyl stabilization. The transient formation of the unsaturated anti methylether VIII could result from hydride attack upon a bis-homocyclopropenyl cation, whose mode of formation has several potential routes.

Use of an 8:1 molar ratio of the aluminum chloride/

lithium aluminum hydride reagent with the respective saturated and unsaturated methyl ketals I and III produced good yields of the saturated and unsaturated rearranged methoxy chlorides IX and V, both of which were solvolyzed at 1000 in aqueous ethanol For the unsaturated chloride V, the major product, cycloheptatriene, and a mixed ethoxy-methoxy ketal fraction of the original tricyclo[3.2.1.02,"Joctenyl structure were





isoldted. Cycloheptatriene Lhe expected product from the well-known decarbonylation of the unsaturated ketone. According to PMR analysis the mixed ketal fraction consisted of 89% anti-ethoxy-syn-methoxy unsaturated ketal and 11% antimethoxy-syn-ethoxy ketal Solvolysis of the saturated chloride IX in aqueous ethanol produced the saturated ketone II as the major product with a mixed ketal fraction consisting of 96% anti-ethoxy-syn-methoxy ketal and 4% anti-methoxy-syn ethoxy ketal

The results obtained strongly indicate that the same

cations are being generated from the respective reactions of the saturated and unsaturated ketals I and III with dichloroaluminum hydride and aqueous acid, as well as in the solvolyses of the saturated and unsaturated rearranged methoxy chlorides IX and V. The electronic structure of these ions has now been radically altered with respect to the parent ions in that positive charge delocalization is not as extensive, giving a more localized or concentrated charge at the methoxyl bridge carbon. The cyclopropyl delocalization may even have been weakened to the point where nucleophiles are able to penetrate and attack the bridge from the syn as well as anti face of the bridge. Thus an electronic effect, and not steric bias, is the overwhelming factor in determining the orientation and stereochemistry of solvent or nucleophilic attack on the 2,4-ethano- or 2,4-etheno-3-alkoxy-tris-homocyclopropenium cations.








CHAPTER I

Introduction


The neighboring group reactivity of the cyclopropyl

moiety has received intense investigation and documentation in the chemical literature. One of the more dynamic directions of these studies has involved the assistance to ionization rendered by the cyclopropyl carbon-carbon sigma (edge) bond at a remote site relative to the cyclopropyl group.

H +H
O T s k A - 0Ac

H (6) (3)
OTS_


k s
HOAc


-OTs


0OOTs
(2)


k -OTs
H
HOAc


(OOAc
(3)


OAc OAc



OAc (5)


OAc /CXH


0+0







In the vanguard of the early cyclopropyl studies was

work carried out by Winstein and his reports2 of the results of solvolytic studies involving cis- and trans-3-bicyclo[3.1.O]hexyl toluenesulfonates, (1) and (2).

The cis-3-tosylate (1), exhibited essentially complete retention of configuration under acetolysis with the nearly exclusive formation (99%) of cis-3-acetate (3). Trans-3acetate (4) amounted to less than 1% of the reaction product. Trans-3-tosylate (2), in contrast, produced a complex mixture of compounds, 54% of which consisted of cis-3-acetate (3), an inverted, solveot-assisted ionization product resulting from the shielding by the departing anion. The remaining components, i.e. olefins and acetates, apparently were derived from the competitive formation of cation (5), which would have orginated via hydride transfer. The facts that cis-3-tosylate (1) solvolyzed ca. thirty-five times faster than the trans-3-tosylate (2), and only the cis-3-tosylate (1) exhibited a special salt effect, prompted Winstein to state that the eis-3-tosylate (1) was ionizing to form the homoaromatica-c 3-bicyclo[3.1.O]hexyl cation (6), a non-classical tris-homocyclopropenyl3 cation (two sigma electrons delocalized over three equi-distant centers).

Both the cis- and trans-3-bicyclo[3.1.Ohexyl tosylates,

(1) and (2), were deuterated at the C3 position2a'd'e and solvolyzed under the original conditions. It was found that in the case of cis-3-tosylate (1), the deuterium label had completely scrambled in the C1, C3, C5 positions of the cis3-acetate product. There was essentially no scrambling







observed with the correspondingly deuterated trans-3-tosylate under the same conditions. This was interpreted2a,d,e as powerful evidence for the intermediacy of the tris-homocyclopropenyl cation (6)

Corey studied4 the deamination of cis- and trans-3bicyclo[3.1.0]hexylamine and found a complex product mixture for both amines consisting of epimeric 2- and 3-alcohols. There was only partial scrambling of a deuterium label for the cis-3-amine and essentially no scrambling for the trans3-amine deamination. Corey felt that both amines produced classical ions which did not leak over to the homoaromatic species, possibly because deamination produces a vibrationally excited cation. A number of other workers have also expressed5 the viewpoint that the diazonium ion is a poor model for solvolytic work.

In an accompanying paper,' Corey acetolyzed 1,5-diphenyl-cis-3-bicyclo[3.1.0]hexyl toluenesulfonate (7), expecting a rate increase from phenyl stabilization if the trishomocyclopropenyl cation (8) was indeed generated. In fact, he observed a small rate retardation; the products detected,



j 3 TH -Ac
(8) (7)

however, were totally rearranged. To explain his results and Winstein's, Corey suggested equilibrating classical ions

(9), the equilibration being the reason for deuterium scrambling and a "weak interaction involving the vacant






+



(9g)

orbital at C3 and the loose electrons of the three-membered ring"' responsible for the stereospecificity of cis-3bicyclo[3.1.O]hexyl tosylate (1).

Winstein has countered with the point that ionization with assistance from the cyclopropyl ring in the cim-3tosylate (1) should occur only in the chair conformation of the bicyclohexyl ring, and there is evidence2e,7,1 that bicyclohexyl derivatives have a marked preference for the boat conformation. As a consequence, there would be, at any one time, a proportionately small number of molecules in the system capable of employing the anchimeric assistance of the cyclopropane ring. The titremetric rate constant for the cis-3-tosylate (1) should be well below any valid estimate for the degree of anchimeric assistance. Winstein felt2e the diphenyl derivative (7) would be shifted even more into the boat conformation and therefore would be expected to ionize in essentially a classical manner. Extending this hypothesis further, generation of a positive charge at the C3 position of the bicyclo[3.1.O]hexyl system and its derivatives by processes which do not depend upon anchimeric assistance would, by very large odds, take place in the boat conformation. One would anticipate the classical ion formation to be a higher energy process than the non-classical solvolytic process, and this, coupled with the high probability for boat form







involvement, would result in a very reactive ion that may have little opportunity to leak into the tris-homocyclopropenium manifold.

Gassman attempted' the electroytic oxidative decarboxylation of cis- and trans-bicyclo(3.1.0]hexane-3-carboxylic acids (10) and determined the predominant products to be the cis- and trans-2-bicyclo[3.1.0]hexanols, (I) and (12) H0.~.H HOH

6t$O2H < + ~J+
(10) (1') (12)


0 OH H


Deuterium labeling at C6 produced no scrambling. Both Gassman and Winstein agreed that the reaction on the electrode surface was too complex to allow a valid mechanistic comparison with the solvolysis of the 3-bicyclo[3.1.0]hexyl tosylates

(1) and (2).

Freeman7 attempted to generate a carbonium ion at the C3 position of the bicyclo[3.1.0]hexyl system via the acid catalyzed addition of methanol to 2-bicyclot3.1.0]hexene

(13), anticipating the generation of Winstein's cation (6).



(13) QH
C3OCH3



9Q) +CH ,








All the products identified, however, were considered to be derivatives of the bicyclohex-2-yl cation (5), which is a cyclopropylcarbinyl cation that Winstein considers2e to be more stable than the tris-homocyclopropenyl cation (6).

Much of the solvolytic work following Winstein's original work2 generally added to the weight of evidence for the existance of the tris-homocyclopropenyl cation as a viable intermediate. Norin subjected'� the optically active thujyl tosylates, (14) and (15), to acetolysis and found they yielded their respective racemic acetates, paralleling Winstein's deuterium labeling studies in the parent bicyclohexyl series. As expected, the rates of (14) and (15) were greater than their trans analogs. Solvolyses of the mono- and dimethyl substituted bicyclohexyl tosylates, (16) and (17), have been recognized2eaa as examples for generation of tris-homocyclopropenyl species.
H H

.L1jJ Ts CH OTs CH OTs
CH3 HCM3 "H
H CH3
(14) (16) (17)




OTs
(18) (19)
Of the numerous options for the tricyclodecyl tosylate

(18) during solvolysis, HUckel LCAO-MO calculations predicted" the development of the tris-homocyclopropenium cation (19) as the intermediate. Again, by means of kinetics,






deuterium labeling, product identification and stereochemistry, the prediction was verified.''

Further evidence for the homoaromatic nature of the 3-bicyclo[3.1.0]hexyl cation (6) has been formulated by determining'2 the activation volumes from the solvolyses of the cis- and trans-3-bicyclo(3.1.0hexyl tosylates (1) and

(2). The trans-3-tosylate (2) exhibited an activation volume (-17.4 cc/mole) in line with cyclopentyl- and cyclohexyl tosylate. The cis-3-tosylate (1), however, had a volume of activation of -13.9 cc/mole. The authors viewed the difference as indicative of a diffuse charge, e a non-classical ion,

in the cis-3-tosylate (1) solvolysis transition state, which supports the concept of the intermediacy of the tris-homocyclopropenyl cation (6).


O H

(20) (21)

Broser and Rahn reacted'1 the epimeric alcohols (20) with boron trifluoride in polar solvents to yield deeply colored solutions which were unaffected by oxygen. The introduction of tropilidene gave a 51% yield of tropylium hydroxyfluoroborate. On the basis of direct observation via NMR, IR, and visible spectra, the authors favor the formation of the non-classical ion (21).

Sauers, in an earlier study, 14 reported the NMR spectrum of lactone (22) in concentrated sulfuric acid. It was Sauers' view that the tris-homocyclopropenyl cationic derivative (23) was a stable entity in this medium.











(22) H2SO, 1
03H
.H CO,H


The weight of kinetic evidence for the existence of the tris-homocyclopropenyl cation at this point was certainly not overwhelming, the maximum reported'0 rate enhancement being ca. 922. The experimental results fell short of the theoretical predictions and calculations previously published. As a result of early LACO calculations, Winstein stated~b "that the 3-bicyclo 3.1.0]hexyl cation prefers the non-classical trishomocyclopropenyl structure to a classical one. Extended HUckel calculations by Hoffmann predicted1s that Corey's "almost classical" non-classical ion (9) would be less stable than the classical ion on a planar five-membered ring. Remarkably, calculations did predict a deep minimum in energy of ca. 1 eV for the symmetrical tris-homocyclopropenyl cation (6).
+





Hoffman also reported'5 calculations on the hypothetical cation (24), which indicated an unsymmetrical double minimum resembling the 7-norbornadienyl cation calculations, except the energy well was deeper as bending of the bridge occurred

-toward the cyclopropyl group as compared to bending toward the double bond.

The consensus was that the full potential of the homoconjugative ability of the cyclopropyl group had yet to be






realized because of the geometric shortcomings of the various systems studied. The added "p" character of the cyclopropyl carbon-carbon sigma bond and the general "bent bond" nature of the orbitals"6 appeared to require a rather exacting alignment for efficient overlap with the developing "p" orbital of the ionization center during the initial stages of cation formation.

X X H











endo, anti- endo, synE exo, synH exo, anticis-chair trans-chair cis-boat trans-boat

(25) (26) (27) (28)


It occurred to a number of researchers that if the C2 and C4 positions of the chair and boat conformations of both the cis- and trans-3-bicyclo(3.1.0]hexanols were connected by an ethano bridge, the four possible combinations could effectively be frozen out, allowing an efficient-probe of the reactivity of the cyclopropyl group while holding the stereochemistry under virtually complete scrutiny

An added advantage to this series of tricyclo[3.2.1.02']octan-8-ols, (25)-OH through (28)-OH, was the fact that a direct comparison could be made to the reactivity of the cyclopropyl group versus the double bond.








Earlier, Winstein et al. solvolyzed both the syn and anti-7-norbornenyl tosylates, (30)-OTs and (31)-OTs, and


(29) 6(30) L~~Q1 (31) ?

relative
rate 110 1011

compared their rates to that of the saturated norbornyl tosylate (29)-OTs. The large rate enhancement for the antitosylate (3 )-OTs has been explained17,18 by Winstein in terms of a non-classical bis-homocyclopropenium intermediate

(32), although others have strongly endorsed'' the concept of

+


(32) (33)

rapidly equilibrating ions (33).






(28) (29)

Pincock first reported" the synthesis of the p-bromobenzenesulfonate derviative of exo-anti-tricyclo[3.2.1.02'4] octan-8-ol, (28)-OBs, and found that it underwent acetolysis at a rate 2.7 times slower than the corresponding 7-norbornyl brosylate (29)-OBs. It was felt that the slight rate retardation was due to the steric interference by the exo-methylene group with the solvation occurring at the backside of the leaving brosylate group at C8. A possible negative inductive








effect of the cyclopropyl group was also considered. The primary observation, that there was no unusual effect of the cyclopropyl group in the Pxo psoition, was attributed to the fact that the cyclopropyl sigma orbitals are directed down and away from the reaction site at the Cs position. Pincock anticipated that this would not be the case for the endo isomer (25).


X H



(29) (31) (28) (25)

k rel 1 10 ' .4 10x4-15


HH



(27)
x
15 104


The three laboratories of Pincock,21a Battiste,21a and Tanida2lbc issued the simultaneous report of the synthesis of the endo-anti- and endo-syn-8-tricyclo[3.2.1.02'4]octanols

(25)-OH and (26)-OH. Solvolysis of (25)-OPNB produced a rate enhancement factor of ca. I0"' relative to the exo-anti system (28) and the norbornyl derivative (29), a "new record for participation''22b With 70% aqueous dioxane as the solvent, only two major products were detected,22 the rearranged endo-4tricyclo[3.3.0.O2,4] octanol (34)-OH and its p-nitrobenzoate








ester (34)-OPNB. Tanida also reported2lc the formation of a trace (0.1%) amount of the retained alcohol (25)-DH. The overwhelming formation of rearranged products implies that

BNPO H H

+


(25)-OPNB (35) (34)-OH OH (25)-OH

most of the positive charge resides on the C2, C4 positions of the intermediate, analogous to the charge distribution in the ion generated from the anti-norbornenyl derivative

(31).
A very important point is that ion pair return or solvent capture occurs stereospecifically at C2 and C4 from the endodirection only, in sharp contrast to the observed21csa reduction of ketone (36) with lithium aluminum hydride which gave exclusive exo attack to yield endo-alcohol (34)-OH. Equilibration experiments have reportedly produced a mixture


(36) ( h(34)-OH 9>(37)

endo C0 OH H


of 38% end alcohol (34)-OH and 62% exo-alcohol (37). Thus, it would appear that exo-alcohol (37) is preferred both kinetically and thermodynamically. It is worth noting that despite an environment that is 67 mole percent water, the intermediate cation is stable and long-lived enough to permit the internal return of the p-nitrobenzoate anion, i.e.,










time enough for the anion to approach the cation from another direction. These facts are uniquely explained by the invocation of the non-classical 2,4-ethano-tris-homocyclopropenyl cation (35) as the intermediate.

Tanida prepared lc an optically active sample of the rearranged alcohol (34)-OH, and acetolyzed its tosylate
---- -- racemization



(34)-OTs OTs (35)

(34)-OTs. The rate of racemization was 3.2 times faster than the rate of acid formation, strongly indicating the occurance of internal return. Ultimately complete racemization occurred, and the existance of any potential hydride shift was eliminated via deuterium labeling. A comparison

(1) (34)-OTs



Ts~
O~s T s r

relative OTs
rate 1 13 10 540



(37)OTs





H OTs OTs









of the solvolysis rate of (34)-OTs with a number of appropriate compounds" c demonstrated that the rearranged derivative is still very reactive. Tanida was compelled to state that the tris-homocyclopropenyl cation (35) is the simplest and most economical intermediate to invoke.

The rate of acetolysis of endco-syn brosylate (26)-OBs is a little over ten times that of 7-norbornyl brosylate and gives a complex product mixture.21'22





(26) (38)





(27) 3

In contrast, exo-syn brosylate (27)-OBs produces a relative rate of lO The methylene group from the exc cyclopropyl ring could sterically aid the departure of the brosylate anion.22,24 Another possible factor could be the concerted shift with ionization of the C1 to C7 bond to the Cs position to produce a stabilized cyclopropylcarbinyl cation (39). There would not be initial stabilization in the cation (38) produced by the analogous process for (26)-OBs since the orbitals would not be aligned in parallel 22



a : (28)








The sluggishness of exo-anti (28) can be attributed to its lack of options, i.e., it cannot undergo the C,, C7 to C8 bond rearrangement to the cyclopropylcarbinyl cation, and, of course, the rigid geometry prevents the proper orientation of the cyclopropyl orbitals for participation.

Coates synthesized25 pentacyclo[4.3.0.02, .03', .05,7]nonan-9-ol, (40)-OH, and found a rate enhancement of 101o-Io12 relative to 7-norbornyl derivatives. Overall strain relief






(40)-OPNB (41) (40)-UH

cannot be a driving force since a homocyclopropylcarbinyl rearrangement produces a structure identical with the original The only hydrolysis product, in the presence of 2,6-lutidine, was determined to be the parent alcohol (40)-OH. These facts, coupled with the results of deuterium labeling, were interpreted in terms of the threefold symmetric tris-homocyclopropenyl cation intermediate (41). Apparently any secondary rearrangement is unable to compete with the attack of water. Coates suggested that the greater reactivity (ca. 80) of the tricyclic p-nitrobenzoate (25)-OPNB relative to the pentacyclic p-nitrobenzoate (40)-OBNB could be attributed to some strain relief in the solvolytic transition state and/or a somewhat less favorable orientation of the anti-cyclopropyl group in (40)-OPNB due to the bond connecting the two threemembered rings. Nevertheless, Coates argued that the








overwhelming bulk of the driving force for both (25)-OPNB and (40)-OPNB arose from the anchimeric assistance of the anti-cyclopropyl ring to form the symmetrical tris-homocyclopropenyl cations (35) and (41).

Ellen and Klumpp2E acetolyzed the interesting compound,

exo-tetracyclo.4.4.0.02' .03, ] dec-7-en-10-yl tosylate (42)-OTs,

and found exclusive attack at C6 yielding only exo-tetracyclo[4.3. 1.01,1.07, ]dec-4-en-2-yl acetate (43)-OAc. Conversion

H OTs
2 ' 7 HOAc
2 7


(42) (44) (43)






(45) (46)

of (43)-OAc into its tosylate (43)-OTs, followed by acetolysis, regenerated (43)-0Ac as the sole product. While the authors felt that the tris-homocyclopropenyl cation (44) was involved as the intermediate, they saw a high contribution from (45) in which the positive charge is concentrated at C3, resulting in less strain than cations having the charge bulk on C2 or CIO. The possibility of homoallylic stabilization, (46), was also recognized.


2 $

(47) (48)








Battiste and Winstein27 acetolyzed the tris-methanonaphthalene brosylate (47)-OBS, and found evidence for considerable (ca. I0'- 108) cyclopropyl participation, proposing the initial formation of the non-classical intermediate

(48). Product analysis revealed the absence of (47)-0Ac or any rearranged or unrearranged brosylate derivatives. The authors felt the cyclopropyl group, and therefore its sigma orbitals, are directed toward the cavity between the two bridges, decreasing the initial degree of orbital overlap with the developing positive center and resulting in some moderation of the relative rate.

NsO Ogs



(49) (50)
As a further example of the rather precise orientation requirements of the cyclopropyl ring needed for assistance to occur, the solvolysis28 of 2-(trans-3-bicyclo(3.1.O]) ethyl p-nitrobenzenesulfonate (49) gave no kinetic or product evidence whatsoever for participation. Tanida's study" of a-(tricyclo[3.2.1.024] oct-syn-8-yl) ethyl p-bromobenzenesulfonate (50) gave questionable evidence for cyclopropyl participation. The small rate factor of three for (50) relative to its anti-analog, was suggested to be due, at least in part, to repulsive hydrogen interactions in the transition state.

The solvolyses of derivatives of exo- and endo-antitricyclo[3.1.1.02'4lheptan-6-ol, (51) and (52), were studied3"









to provide additional insight into the geometrical requirements for participation of the cyclopropyl group. Interestingly, it was determined that (51)-OPNB was only fifty-six times faster than (52)-OPNB, although (5l)-OPNB yielded the two k rel 6 +
56k



11
7
3H H OH X
(51)
x OH


1 (52 )


rearranged products expected from a tris-homocyclopropenyl cation intermediate while (52)-OPNB produced an olefinic mixture. It was concluded3 ob that both compounds were solvolyzing with considerable (though different) neighboring group assistance. The fact that (5l)-OPNB was solvolyzing at a rate ca. ten times slower than endo-anti-8-tricyclo[3.2.1.02,loctanyl p-nitrobenzoate (25)-OPNB was attributed through X-ray studies to significant geometrical distortion caused by hydrogen interaction at C3 and C7.


(53)









Coates subjected31 the exo-tetracyclo[3.3.0.03,6.0, ] oct-4-yl tosylate (53) to acetolysis and detected the two expected acetate products. Kinetic studies demonstrated a rate acceleration from anchimeric assistance of ca. 109 Coates concluded that the methylene bridge at C7 directs the cyclopropyl orbitals away from the site of ionization.

Lustgarten reported2 the acetolysis of the endo- rearranged tosylate (54)-OTS, giving the endo-rearranged acetate

(54)-OAc as the sole product. It was his view that (54)-OTs


~OTs


OTs OAc
(54)-OTs (55) (54)-OAc
ionized with assistance from the C1-C2 bond to give trishomocyclopropenyl cation (55), the same intermediate ion derived from 51-OTs. A deuterium label at C4 gave an equal distribution of deuterium on C2 and C4 after solvolysis, however only half of the original label was accounted for at these two positions, compelling Lustgarten to conclude that the full nature of the intermediate cation had yet to be determined.

In separate papers, Masamune33 and Hart34 detailed direct spectral evidence for the exista.nce of the symmetric pyramidal delocalized cations, (56) and (57) respectively. Both cations accounted for the deuterium label discrepancy previously mentioned, and the higher symmetry was predicted by Stohrer.









H H


(57)

(56)

and ioffmann35to be relatively stable.







(59) (60)

krel 622 1 26.2




(61) 1 H(62)

x

Sargent probed36 the ability of a cyclopropyl moiety to function as a remote, nucleophilic neighboring group by interacting with a carbon-carbon double bond which is itself providing a source of electronic stabilization for a developing cation. Solvolysis of the tricyclic dinitrobenzoate

(58)-ODNB demonstrated a significant rate acceleration (krel= 622) relative to the allylic dinitrobenzoate (59)-ODNB, which lacks an internal remote nucleophile. The ester (58)-ODNB was also faster than the double bond analog (60)-ODNB by a factor of 23.6.

Product studies from the hydrolysis of (58)-ODNB indicated only two rearranged products, endo-alcohol (61)-OH and endodinitrobenzoate (61)-ODNB. The striking simplicity and








and stereospecificity of the products strongly suggest the intermediacy of an unusually stable cation. Both solvent and dinitrobenzoate anion attack exclusively from the more hindered direction at a position four bonds (ca. 4A) away from the initial site of ionization. The longevity of this cation is further demonstrated by the fact that the weakly nucleophilic dinitrobenzoate anion is able to compete with water in the product forming sequence. As a result of these observations, Sargent favored36 the intermediacy of the nonclassical cation (62).

From the solvolyses of syn-7-p-methoxyphenyl-anti-7norbornenyl p-nitrobenzoate and its saturated analog, Gassman37 determined that the p-anisyl group was capable of exerting a leveling effect of ca. 3 x 101 0 with regard to neighboring group participation. A p-anisyl group was substituted at the OCH3 OCHa
OCH 3






(63) (64)-OH OH (64)-OPNB OPNB

syn-Co position in the endo-anti-8-tricyclo[3.2.1.02'4loctanol system giving (63)-OH. Treatment of (63)-OH with acid yielded the rearranged alcohol (64)-OH as did the hydrolysis of the p-nitrobenzoate derivative (63)-OPNB. The hydrolysis of

(63)-OPNB also yielded the internal return product, the rearranged p-nitrobenzoate (64)-OPNB. Gassman had predicted








a rate acceleration by the cyclopropyl group of 3 x 10' over and above the leveling of the p-anisyl moiety, and observed a value of 3.8 x 10 It was emphatically noted that even though a cation at Ce would be tertiary and stabilized by the p-anisyl group, the bulk of the charge resides on C2 as determined by the formation of rearranged products, i.e., the cyclopropyl ring controlled product formation.
As has been seen up to this point, the generation of

the 2,4-ethano-tris-homocyclopropenyl cation (35) has resulted overwhelmingly in product formation at the endow C2 and C4 positions, suggesting, perhaps, that the bulk of positive charge in the intermediate resides at these positions as opposed to the bridge Ce position. Tanida calculated2tc the ground state energy difference between the two alcoholic



7


(35) 3 (32)

products of cation (35) based on a distribution of 99.9% (C2, C4 attack) for rearranged alcohol (34)-OH and 0.1% (Ca attack) for retained alcohol (25)-OH. The resulting value of 12.1 kcal for AF0 indicates a considerable amount of strain relief in the transformation to the rearranged alcohol (34)-OH

Pincock2" pointed out the analogous charge distributions for both the tris-homo- (35) and bis-homo- (32) cations, and referenced Winstein's description of the 7-norbornenyl cation

(32). The bridge carbon atom of cation (22) has "considerable








tendency to rehybridize from sp2 toward sp3 Such rehybridization increases the C7 Coulomb integral as well as C7-C2 and C7-C3 orbital overlap. This leads to net stabilization of the bridged ion, and these very features of rehybridization at C7 tend to diminish the charge on this atom.'23b
.CH3








As noted earlier, even substitution at C9 with a vinyl group (formally causing the generation of an allylic cation upon ionization) and an anisyl group resulted in rearranged product upon solvolysis. Substitution of a cationic stabilizing group at the C8 anti position along with a syn leaving group apparently allows leakage to the non-classical intermediate followed by nucleophilic capture to yield endo rearranged products. Thus, when Gassman treated the synOCH3



CH30 H H +/H20(64)-OH


(65) H
OH
alcohol (65) with acid,3" only rearranged alcohol (64)-OH was isolated. Baird and Reese36 reacted the anti-methyl-syn-bromo tricycle (66) with Ag+ and recovered only the rearranged









CH 3
OH3 Br OH
Ag+ H 2 0z
H,0

(66) (6) HH (68)


alcohol (67). Alcohol (67) was also the hydrolysis product of the syn-mesylate (68)

That there is little positive charge residing on the

bridge C8 position of the tris-homocyclopropenyl cation (35) is a conclusion which might be supported by the fact that substitution of various electron donating groups at Ce had no effect on the products, i.e., only endo rearranged species were observed. This concept is feasible, but it seems more likely that there is a steric bias in the transition state of ncleophilic capture by cation (35).

Tanida's ground state energy calculations, which favored

the rearranged alcohol (34)-OH over the retained alcohol (25)-OH by 12.1 kcal, should reflect the strain energy difference between the two possible transition states for solvent capture.
H




SOH

This being the case, the transition state energy barrier difference, paralleling the product alcohols, would be greater than zero and less than 12 kcal. The intensity of positive charge at the bridge C8 of cation (35) could very well have








been increased by the substitution of the electron donating groups previously discussed, but to a degree insufficient to swamp out the steric bias of the transition state for capture of weak nucleophiles.

Considering this point, one could envision stronger

electron donating groups at the bridge C6 position tipping the balance of solvent capture in favor of charge over steric bias. As can be seen below, the methoxyl group appears to be just such an entity and it is consequently used in this investigation as a further probe into the balance between charge versus steric strain relief in the tris-homocyclopropenyl cation manifold.

Part of the synthetic scheme employed to prepare the alcohols (25)-OH and (26)-OH involves the acidic hydrolysis of the saturated ketal, 8,8-dimethoxy-endo-tricyclo(3.2.1.02']octane (69). The only product reported21a,22,9 by three different laboratories was ketone (70), erdo-tricyclo(3.2.1.02,]

0
H +/H2Q


(69) (70)

octan-8-one. It is generally assumed,40 mechanistically,


R-- -ROH R-O-C+ - R-O=C
-1':1 ' 71 1

that the ease with which ketals suffer acidic hydrolysis is the result of resonance stabilization in the alkoxy carbonium ion intermediate (71).









The formation of only ketone (70) from ketal (69), or more specifically the absence of rearranged products, leads one to question the existence of participation involving the cyclopropyl group. There appear to be two extremes as to the nature of the carbonium ion intermediate generated from

(69): 1) essentially a classical oxo-carbonium ion (72), in which methoxy resonance stabilization of the positive charge at Ca swamps out any energy need for cyclopropyl interaction 2) a non-classical methoxy-trishomocyclopropenyl cation (73)in which most of the positive charge resides on the bridge carbon Cs, despite considerable cyclopropyl involvement in charge stabilization.


OCH3


(72)


CH,

(73


CH 6 H3




(74) rel
rate 1

CH3 CH3




(76) rel
rate 18


CH3O OCH3




(75)
2.3

CH30 OCH3




(77)

143


.OCH ,


(78) 320









C H CHO 0CH,3



(79) (80)

During the course of the research efforts reported in

this text, the kinetics of the acidic hydrolyses of the above ketal series were revealed by Lamaty et al. The rates of formation of the expected ketone products were monitered by means of the carbonyl ultraviolet absorption. The contrast between previous solvolytic studies of the respective alcoholic derivatives (rate acceleration up to 10i1) and the more subtle trends (rate acceleration ca. 102) observed for the hydrolysis of their ketal precursors is obvious. Apparently, the methoxyl group is capable of an even greater leveling effect than the p-anisyl group. Lamaty explained his results on the basis of the likely sites of protonation. It has been shown42 that the alcohols syn with respect to the double bond or the cyclopropyl group exhibit considerable hydrogen bonding, the strongest interaction occurring in the olefin case. As an example of Lamaty's argument, the norbornenyl ketal (75) hydrolyzes only 2.3 times faster than the norbornyl ketal

(74): Protonation of the anti-methoxyl group of (75) is required for departure with assistance, but due to the propensity for hydrogen bonding, protonation is more likely to occur at the syn-methcxyl moiety to yield the relatively stable hydrogen-bonded intermediate (79). Since hydrogen bonding is weaker for cyclopropane; thereby reducing the selectivity,








anti-methoxyl protonation competes more favorably in the case of saturated ketal (69). Participation of the cyclopropyl ring then becomes a more significant factor as is reflected in the relative rate of 120. The unsaturated endo-cyclopropyl ketal (78), if the above mentioned protonation factors were of little consequence, should have hydrolyzed at a rate intermediate to that of ketals (75) and (69), i.e., between the relative rates of 2.3 and 120. The observed relative rate was 320, which Lamaty felt4l was indicative of the fact that protonation of either methoxyl could result in effective participation from either the double bond or the cyclopropyl group. The rate factor of 18 for ketal (76) was attributed to protonation on the least sterically hindered methoxyl group (anti) and the subsequent tilting away of the syn methoxyl group from the exo-cyclopropyl methylene in the transition state. A relative rate factor of 143 was observed for the exo, endo-dicyclopropyl ketal (77) Noting that k77/k7, B while k69/k74 120, Lamaty argued that this was the best example of selective protonation occurring on the least sterically hindered methoxyl coupled withthe syn hydrogen bonding factor Kessler has, however, pointed out43 that according to his product studies, ketal (77) yielded, along with ketone, rearranged product upon acidic hydrolysis. This would affect Lamaty's results since he was following the rate of formation of ketone. Kessler estimated that Lamaty's rate was low by a factor of three.43a Later, Kessler repeated43b Lamaty's hydrolysis of ketals (69) and (77), but monitered








the product formations via PMR. Using Lamaty's relative value of 120 for ketal (69), Kessler reported a rate ratio of 120:5 for (69) and (77) respectively (120:143, Lamaty). Kessler politely decided not to speculate on the discrepancy between the two research groups. He does speculate that the exomethylene group hinders protonation on the anti-methoxyl, and resulting syn-methoxyl protonation does not induce participation by the endo-cyclopropyl moiety.

Lamaty has discussed the nature of the ketal hydrolysis intermediate. He found the cyclopentyl ketal (20) to be

5.1 x 103 times more reactive than the 7-norbornyl ketal (74), under the conditions of acid hydrolysis. The 7-norbornanone, however, is 1.7 x 104 more reactive toward nucleophilic addition of borohydride than cyclopentanone. Since he had already reported that the transition state for borohydride addition presents an entirely sp3 profile," Lamaty felt the reaction profile of the acidic hydrolysis of the 7-norbornyl ketal and the related bridge ketals passes from an initial sp3 state to a transition state very sp2 in character, i.e. a carbonium-oxonium ion. There was even speculation that the participation of a double bond or cyclopropyl group would in fact involve interaction with the ir" orbital of the carbonium-oxonium ion.
OCH3 0
CH3O OCH3



(77) 81) x(82)








When Kessler subjected" ketal (77) to acidic solvolysis in wdter or methanol, he found complete conversion to the respective rearranged products, (81)-OH and (81)-OCH3. Changing the solvent system to dioxane/water or acetone/water produced, in addition of (81)-OH, the ketone (82).

Acidic solvolysis of the methyl derivative (83) led entirely to rearranged products, (81), the methyl group OCH3
CHO 0 -OCH3
(83) E (84)
I~J~~Ii -~- IJCH3\

x

apparently adding some stabilization of positive charge at C2

H OCH3 OCH3 OCH3
AoB --- .-A BA BA




C C C C
(35) (73) (85) (86)

The formed products provide a means for determining the mode of nucleophilic attack on the intermediate carbonium ions, (35), (73), (85), (86), and a summary of these correlations has been made by Kessler. Path A for cation (35)

is not observed because of the interaction of the cyclopropyl sigma orbitals with Ca, while B (retention) becomes an extremely minor process in comparison to path C (rearrangement) For

cation (73), the methoxyl stabilization of the carbonium ion center allows a "normal" path of hydrolysis to ketone, probably






31


via path 8 because of cyclopropyl interaction. In contrast to cation (73), attack by path B for cation (85) is hindered by the methylene group of the exo cyclopropyl ring and consequently path C predominates over ketone formation. Kessler did state his belief that path B still dominates over A in the formation of ketone from (85). Cation (86) reacts onfly with rearrangement due to the stabilizing effect of the methyl group, i.e., via path C.








CHAPTER II

Results and Discussion


The synthetic scheme employed for the preparation of the precursors and compounds used in this study has b~en reported by several workers.z1a,22,39

C1 C1 CH30 OCH3 CH3O OCH3
C_ CI KOH Cl Cl A

/I KOH 1C14 (88)
C I C 1I
(87) INa, THF
I t-BuOH

0
CH.3O0 COCHC

L1111 H2 /(78)
H,0 Pd/C

(70) (69)


Hexachlorocyclopentadiene reacted with methanolic46

potassium hydroxide to give 1,2,3,4-tetrachloro-5,5-dimethoxycyclopenta-1,3-diene, (87). The Diels-Alder addition of cyclopropene to (87) yielded the 1,5,6,7-tetrachloro-8,8dimethoxy-endo-tricyclo[3.2.1.02,']oct-6-ene (88).2,a,22,39 Dechlorination of (88) was accomplished via Gassman's procedure47 giving the 8,8-dimethoxy-endo-tricyclo(3.2.1.02,41oct-6-ene (78).21a,22,39 Hydrogenation of the unsaturated ketal (78) with palladium/charcoal catalyst produced








8,8-dimethoxy-endo-tricyclo[3.2.1.02',]octane (69).21a,22,39 Hydrolysis of (69) with wet acetic acid'6c generated endotricyclo[3.2.1.02'4]octan-8-one (70) According to an analogous scheme, the diethoxy ketals were also prepared.

The acidic hydrolysis of the endo-saturated ketal (69), as described in the introduction, has been reported2'a,2239 as leading exclusively to endo-tricyclo[3.2.1.02,4]oct-8-one

(70) Obviously a cationic intermediate is involved, and


CH'O OCH3 H +/H20


(69) (70)


one is able to speculate upon at least two extremes as to the nature of this cation: 1) essentially a classical, oxo-carbonium ion (72) involving no delocalization of the cyclopropane ring, with the methoxyl stabilized charge at C8 attacked by nucleophiles only at either face of the bridge or 2) a delocalized non-classical methoxy-tris-homocyclopropenyj cation (73) in which there is both methoxyl and cyclopropyl stabilization resulting in charge concentration at C8 and nucleophilic attack at the bridge C8 from the anti face stereospecifically

CH30+ CHOCH


(72)


(73)








As an initial probe into the nature of the intermediate, both the dimethoxy-saturated and unsaturated ketals, (69) and (78) respectively, were dissolved in d4-methanol in the presence of catalytic amounts of either silver(I) perchlorate or trifluoroacetic acid. The disappearance of the syn- and anti-methoxyl signals in the PMR of both ketals was monitored. The syn and anti designations are assigned with respect to the endo cyclopropyl group, and in both ketals, the syn methoxyl singlet is assumed to be downfield from the anti methoxyl singlet, the verification of which will be reported in Chapter III.


(69) H;





C H D







CH30 OCH3

(78)



C H OCD 3


CD CH3







CD3O OCD3






CD30 OCH3






CD OO0CD3









The anti-unsaturated methoxyl signal diminished in

strength at a rate greater than the syn-unsaturated and antisaturated methoxyl signals, which disappeared at ca. the same rate. The syn-saturated methoxyl singlet was the most sluggish, but it too was eventually washed out by the d"-methanol

While any interpretation of these observations is

subject to potential ambiguities, the fact that the antimethoxyl groups for both ketals had a greater propensity for d4-methanol substitution than their syn counterparts would enhance speculation of at least some involvement of the cyclopropyl group with charge development at Cs.

At this point it was considered desirable to generate the intermediate cation by other means, be it (72) or (73), and study the orientation and stereochemistry of nucleophilic capture. Generation of oxo-carbonium ions by the reaction of ketals with dichloroaluminum hydride has been documented in the literature"8 The dichloroaluminum hydride, which exists in ether as an etherate, complexes with an oxygen of the ketal, which in turn fragments to give an oxo-carbonium ion. Subsequent attack by hydride or other nucleophiles gives Cl 2H~l
C OR + AlHCI2 C +1-R


C-+-OR + ROAlHC12 A1HC12 / OR

an ether product.kec,d Eliel suggested using a 4:1 molar ratio of aluminum chloride to lithium aluminum hydride as








the most efficient means of generating dichloroaluminum hydride, and the ratio is employed in this work.

The reaction of 8,8-dimethoxy-endo-tricyclo[3.2.1.02']octane (69) with a 4:1 molar ratio of aluminum chloride/lithium aluminum hydride for 1.75 hours produced syn-8-methoxy-endotricyclo[3.2.1.02'"]octane (89) in 82.8% yield. The colorless


CH H4:1AICI/LiAH

ether


liquid was identified by its PMR spectrum 6[3.82(1,m), 3.34 (3,s), 2.18 (2,m), and 1.67 to 0.67 (8, complex)], mass spectrum (m/e 138), infra-red spectrum, and elemental analysis. An alternative synthesis for (89) was accomplished by the reaction of diazomethane with an authentic sample9 of endosyn-tricyclo[3.2.1.02 4]octan-8-ol, (26)-OH, confirming the assigned structure of (89) via spectral comparison.

H OH

CH2N2
a i BF3/ether
(26)-OH (89)

Glpc/mass spectral analysis of the crude dichloroaluminum hydride product mixture resulted is the detection of four minor reaction proucts totaling 5.3% relative to the major product (89). The highest recorded m/e values were 136, 128 and 179 respectively for the third, fourth, and fifth eluted components.









The fact that none of these trace products were attribut able to the intermediate formation of any anti-8-methoxyendo-tricyclo[3.2.1.02',]octane (90) was demonstrated by the synthesis of (90) and its subsequent exposure to the reaction conditions. Reduction of ketone (70) with lithium metal in

0
HO HCHO H
L Li CH2N2

NH3(1) BF3

(70) (25)-OH (90)


liquid ammonia" gave product which consisted of 94.5% antialcohol (25)-OH. Treatment of (25)-OH with distilled diazomethane etherate and boron trifluoride catalyst produced anti-8-methoxy-endo-tricyclo[3.2.1.02,4])octane (90) which was characterized by its PMR spectrum 6[3.62 (1,m), 3.23 (3,s), 2.21 (2,m), and 1.80 to 0.28 (8, complex)], absolute mass measurement of [CG9H10]t, infra-red spectrum, and elemental analysis.

anti-Ether (90) was stirred for 1.75 hours with 4:1

aluminum chloride/lithium aluminum hydride reagent and three products were isolated via preparative glpc. Tricyclo[3.3.0.02,8]octane (91) was collected as a colorless liquid in 28.6% yield. The PMR and infra-red spectra were in accord with an earlier report,46c and the absolute mass measurement of [CaH12]t along with the elemental analysis confirmed the structure.











4:1 AICl,/LiAIH4 + 1

ether
(90) (91) OCH, Cl
(92) (93)
The second product isolated was endo-4-methoxytricyclo[3.3.0.02'8]octane (92) in 33.7% yield. The PMR spectrum 6[3.75 (l,ddd), 3.19 (3,s), and 2.70 to 0.87 (10, complex),], mass spectrum (138 m/e) and absolute mass measurement of [CqH140]t infra-red spectrum, and elemental analysis all confirmed the assigned structure. The endo configuration



HA (92) and (93)

H HAAXJA'X 9.0 Hz; Jmx=6.0 Hz

was verified by analysis of PMR couplings for Hx which agree with reported data21c,22,36,37,43b for endo analogs.

The third product (7 7%) was identified as endo-4chlorotricyclo[3.3.0.02',]octane (93) from its PMR spectrum 6[4.18 (1,ddd), and 2.86 to 103 (10, complex)], mass spectrum (142 m/e), and the absolute measured mass for (CaHieCljt The endo configuration was also verified by the coupling pattern for H '

The detection of only syn-ether (89) from saturated

ketal (69) and dichloroaluminum hydride with no evidence for the formaton of the epimeric anti-ether (90) adds considerable weight to the existence of the non-classical intermediate








(73). Following coordination of dichloroaluminum hydride H Cl2
Al H OCH,
CH3O OCH" H OCH3
-Cl2AI(OCH3)H


(69)73 (89)







(94)


with the anti-methoxyl group of (69), it could be anticipated that this anti complex would depart with assistance from the cyclopropyl group. The observation that the anti-ether (90) is labile under the reaction conditions while the syn-ether

(89) is stable lends credence to this hypothesis. There is, however, no compelling evidence to suggest which methoxyl group in 69 is intially lost. In any event ionization to the intermediate cation (represented as non-classical (73) for the sake of argument) and subsequent attack by one or each of three formal nucleophiles, i.e., methoxide, chloride, and hydride, would explain the observed product(s). Attack at C8 is apparently favored by a concentration of positive charge at this site in either the classical (72) or non-classical

(73) models. Methoxide attack would regenerate starting ketal

(69), while chloride attack would afford the highly reactive a-chloro-ether (94) which should quickly regenerate the cationic








intermediate under the reaction conditions. Hydride attack at C. accounts for the observed product, syn-ether (89), whose formation exclusive of C2 attack and any anti-ether (90) demonstrates stereospecific hydride capture from the anti face of C8. This is of course, indicative of interaction and delocalization of charge at CB with the cyclopropyl ring and the methoxyl grout, but with the methoxyl stabilization concentrating the charge at C8 to a point at which the steric bias for the transition state of endo rearranged product formation is overcome.

The formation of only endo-rearranged products from the reaction of dichloroaluminum hydride with ant- -ether (90) is in line with earlier results for nucleophillic capture of the solvolytically generated parent tris-homocyclopropenyl cation (35) In the case of (90), all three potential nucleophiles are observed in the product analysis, a relAtively rare result'ae for dichloroaluminum hydride reactions.

CH3O H



(90)

At this point it became desirable to investigate the reaction of dichloroaluminum hydride with 8,8-dimethoxyendo-tricyclo[3.2.1.02,4]oct-6-ene (78), particularly with respect to any potential anchimeric competition between the cyclopropyl group and the carbon-carbon double bond. Reaction of unsaturated ketal (78) with a 4:1 molar ratio of aluminum








chloride/lithium aluminum hydride gave four products which were isolated via preparative glpc (yields reported are relative glpc peak areas).


H -r 0H3


(96)


(95)
CH3O CH3
/ H4:1 AlCl3/LiAlH,
ether


qHm -H' A Hm Hx A'
C1
The major product (76.5%) isolated was syn-8-methoxyendo-tricyclo[3.2.1.02.4]oct-6-ene (95) which was identified from its PMR spectrum 6[0.53-0.94 (2, complex), 1.26-1.57 C2, complex), 2.66-2.88 (2,m), 3.25 (3,s), 3.50 (1,m) and

5.61 (2,t)], the mass spectrum (136 m/e), infra-red spectrum, and elemental analysis. Hydrogenation of (95) with 10% Pd/charcoal catalyst produced the previously characterized



nH OH H OCH3
10% Pd/C Z

(95) (89)

saturated syn-ether (89) to further verify the structure.

endo-6-Methoxytricyclo[3.3.0.02' 8]oct-3-ene (96) was isolated in 3.5% relative yield and was characterized by its PMR 6[5.65 (2,m), 3.95 (1,ddd), 3.24 (3,s), 2.79 (1,m), and 2.48 to 0.58 (5, complex)] and mass (136 m/e) spectra. The endo configuration was verified by the PMR couplings2 Ic,22,36,37,43b for Hx: JAX=9.25; JA'X=7.75; and JMX=5.0.








endo-6-Chlorotricyclo [3.3.0.02,f]oct-3-ene (97), collected in 1.0% yield, received the assigned structure on the basis of the literature authenticated5 PMR spectrum 5[5.70 (2,m), 4.28 (1, oct), 3.24 (1,m), 2.57 to 1.34 (5, complex)], mass spectrum (140 m/e), and spectral comparison with a sample derived for an alternative synthetic route to be reported later in the text. The endo configuration was confirmed21c,22,36,37,43b by the PMR couplings of HX: JAX 10.0; JA'X 8.0; and JMX 5.0 Hz.

The last component was identified as 1-methoxy-endo-6chlorotricyclo[3.3.0.02,']oct-3-ene (98) and was collected in 18.9% yield. The elemental analysis, infra-red spectrum, mass spectrum (170 m/e), and PMR spectrum 6[5.73 (2,t), 4.37 (1,m), 3.39 (1,m), 3.29 (3,s), and 2.60 to 1.27 (4, complex)] agreed with the structural assignemnt. The PMR coupling for Hx confirmed21a,22,36,37,4b the endo configuration: JAX 10.0; JA'X 7.75; and JMX 5.25 Hz. Further structural proof is given later.

At first glance, the product bulk appears explanable

on the basis of the intermediacy of the 8-methoxy-tris-homocyclopropenyl cation (99), however the detection of the minor products (96) and (97) requires the invocation of at least two or more intermediates. Cl OCH3

CHH0 'A 0CH
CH OCC (9


> (99))








The formation of unsaturated rearranged chloride (98) would appear to be the result of chloride attack upon the non-classical cationic intermediate (99), which would be generated by the complexation of dichloroaluminum hydride with the anti-methoxyl group of ketal (78) with subsequent departure of the anti complex via the anchimeric assistance of the cyclopropyl group. The chloride anion would then have path B (C8, anti-face) or path C (C2, endo) as possible routes to cation collapse. Path C, of course, would lead directly to product (98) but it would be anticipated that C8 attack (path B) would predominate in line with the observed-course of reduction (hydride attack). If formed, the resulting unstable bridge c-chloro-ether (100) could then suffer rearrangement to the thermodynamically more stable, rearranged chloride (98) via a tight ion pair intermediate.2 b,22 If the intermediate cation was essentially classical, i.e. (71), one would expect a mixture of both the syn (path A) and anti (path B) a-chloro-ether at C8, both of which should be very reactive. As already mentioned, generation of a classical cation via syn ionization followed by leakage into the nonclassical manifold has been domonstrated to occur when cationic stabilizing groups are located on C.37,38 OCH 3
CH,O H B .-A H OCH,
H: H
Path A Path B
(101) C (95)
(99)








Hydride attack upon (99) at the Ca position accounts for the major product, syn-unsaturated ether (95). The non-classical description of (99) would require the stereospecific approach from the anti face (path B). Approach from the syn face (path A) by hydride wo ld generate the anti-8-methoxy-endo-tricyclo[3.2.1.02'1]oct-6-ene, (101), whose transient existance, as will be seen, is highly probable. There are, however, potential alternative routes to the formation of (101) which will shortly be discussed. Hydride attack via path C was not observed which is entirely consistent with hydrolytic results for ketal (78), and makes the observations of rearranged chloride (98) all the more interesting.

Methoxide quenching of cation (99) at Ce would regenerate (path B likely) the starting material, unsaturated dimethyl ketal (78) Under the experimental conditions, this process would not be detected.








(102) (96) OCH3 (97) Cl

The characterization of the two non-methoxylated endo Yearlanged minor products (96) and (97) argues strongly for the intermediacy of the unsubstituted 2,4-etheno-trishomocyclopropenyl cation (102) which would suffer nucleophilic attack via path C to give the observed products (96)








and (97). There appear to be a number of viable mechanistic routes to (102).


H O3H+
? ?


(95 (103) (102)


One mechanistic pathway questions the stabilityof the

major product, syn-unsaturated ether (95), under the reaction conditions. Coordination of dichloroaluminum hydride as a Lewis acid with the syn-methoxyl group of (95) could result in cleavage of the ether with anchimeric assistance by the internal carbon-carbon double bond generating the unsubstituted bis-homocyclopropenyl cation, (103). The previously mentioned theoretical calculations of Hoffmann'5 predict a lower ground state energy for the tris-homocyclopropenyl cation (102) relative to the bis- homocyclopropenyl cation (103), suggesting the possibility of interconversion or equilibration, favoring (102), between the two ions, with subsequent nucleophile capture by cation (102) to give the observed products.

It has been reported,51 however, that the hydrolysis

of the p-nitrobenzoate ester of the syn unsaturated alcohol, (104)-OPNB, yields only the retained syn-alcohol (104), presumably via the bis-homocyclopropenyl intermediate (103). Apparently (103) and (102) do not interconvert under the reactions conditions because bridge flipping is not competitive with solvent capture."









OH

H OPNB . (104)




(104)-OPNB (103) (102)


The syn unsaturated ether (95) itself was subjected to the original reaction conditions and was exposed to the dichloroaluminum hydride reagent for a three hour period. Capillary glpc analysis revealed that 93% of the syn ether

(95) had not reacted, with the detection of only 6% of an unknown hydrocarbon and 1% of an unknown whose retention time was too great to be either (96) or (97). As a consequence of the above facts, the generation of the tris-homocyclopropenyl cation (102) from syn-ether (95) via the bis-homocyclopropenyl cation (103) is effectively eliminated as a mechanistic pathway.

One alternative approach to the formation of cation (102) could involve the intermediacy of 8-methoxy-bis-homocyclopropenyl cation (105). The generation of cation (105) might be accomplished by equilibration between (105) and the 8-methoxytris-homocyclopropenyl cation (99), since the energy barrier to bridge flipping should be lowered52 relative to cations (102) and (103) as a result of the methoxyl stabilization of positive charge at Ce in (99). Of at least equal probability is the dichloroaluminum hydride promoted cleavage of the sunmethoxyl group in unsaturated ketal (78), with concomitant







anchimerical assistance by the double bond to give cation (105) directly.

The viability of the bridged ion (105) as an intermediate is somewhat speculative considering reports53 that substitution of a p-anisyl group at the bridge (C7) carbon of norbornene cancels out (levels) stabilization provided by the double bond umn'thecation generated at the bridge. Direct spectral observation52 of the 7-methoxynorbornenyl and norbornadienyl cation, however, has provided some evidence that delocalization involving one or both double bonds, respectively, exists for these ions.

An 8-methoxy-bis-homocyclopropenyl cation (105) would be expected to suffer hydride attack from the nnti face (with respect to the double bond) at Ce to generate the anti8-methoxy-endo-tricyclo[3.2.1.02',]oct-6-ene (101).

Ether (101) could also have been generated by hydride attack on the tris-homo cation (99) from the syn face of C8, allowing for weakened interaction of the cyclopropyl group and incorporation of considerable classical character. By analogy, it has been shown that from strictly steric point of view, anti attack is already favored over syn attack without any potential complications of delocalization. Borohydride reduction39 of the unsaturated ketone (106) produced a 3: predominance of syn-alcohol (104) over anti-alcohol (10_7

Whatever the origin of (101), coordination of its anticipatorily labile anti-methoxyl group with dichloroaluminum hydride, followed by the cyclopropyl assisted ionization of the complex, would generate the unsubstituted non-classical








cation (102), with subsequent product formation.
OCH3



CH3 OCH CH,



(7)O H (101)E12


(99)
In order to test the later part of this hypothesis, it was deemed advantageous to synthesize the unsaturated antiether (101), subject this ether to the original reaction conditions, and analize the resultant product mixture for the presence of (96) and (97)

The synthesis of endo, anti-tricyclo[3.2.1.0'0]oct-6en-8-ol, (107) was accomplished via the method of Clark, Frayne, and Johnson.39 The unsaturated ketal (78) was sub-


BH.


(78)


(106)


jected to acidic hydrolysis at -50 solution of the thermally unstable tricyclo[3.2.1.02',]oct-6-en-8-one never isolated but was reduced in hydride to yield ca. 75:25 mixture and anti-alcohol (107). Isolation anti-alcohol (107), whose spectral


H H H O

+

(107) f(10O4)


to yield a concentrated (decarbonylation) endo(106). Ketone (106) was solution by sodium boroof the syn-alcohol (104) via preparative glpc of data vere in agreement








with the literature,39 was followed by methylation by diazomethane with boron trifluoride catalyst to yield the desired


HO H CH3O H

gCH2N2
BF3

.(107) (101)
anti-8-methoxy-endo-tricyclo[3.2. 1.02,4]oct-6-ene (101). Structural identification of (101) was established by its PMR spectrum 6[5.67 (2,m), 3.62 (1,m), 3.21 (3,s), 2.93 (2,m),

1.14 (2,m), .33 (2,m)], mass spectrum, and the measured mass for [C9H120].

Exposure of the anti-unsaturated ether (101) to the

original reaction conditions was followed by capillary glpc analysis of the resultant product mixture. Both the endorearranged methoxy- and chloro-octenes, (96) and (97), were established via authentic compound comparison to be present


CH;O H

g~I~ 4:1 AIC ./LiAIH
~HH

(101) (96) OCH3 (97) Cl

in a relative ratio of ca. 3.2.:1 compared to the original ratio of ca. 3.5:1 obtained from the ketal reduction The two products accounted for the bulk (61.1%) of the reaction mixture with no unreacted ether (101) detected. The results above strongly endorse the viability of anti-unsaturated ether (101) as an initial reduction product in the reaction of dichloro-










aluminum hydride reagent with the unsaturated dimethyl ketal

(78). A summary mechanistic scheme for the reduction of unsaturated ketal (78) is presented below.


CI H3


CH3O 0 C0CH3



(78)

OCH,



(105)


*1
C(H3H (101)


CH 3
B -- K ,c - - A


// J,/
(99)




A n 1


(102)


OCH3




C1




slow

(203)







(97) C-


(96) )H


OCH3


(98)








If the mechanistic rational presented for the formation of the endo-unsaturated rearranged chloro-ether (98) is correct, one should expect that an increase in the molar ratio of aluminum chloride/lithium aluminum hydride should lead to increased yields of chloro-ether product, since the availability of the chloride anion is increased while the hydride molar equivalents are decreased. An 8:1 molar ratio was employed in reactions with both the saturated and unsaturated ketals (69) and (78). The small amount of hydride present effectively serves to destroy any moisture or proton acid build-up.
OCH3
C H;O--, OCHa 8) C
C C 8:1 AIC13/LiAIH4

H
(78) (98) CI

Use of the 8:1 reagent with the unsaturated ketal (78)

gave a 71.8% yield of 1-methoxy-endo-6-chlorotricyclo[3.3.0.02'loct-3-ene (98) as a dark yellow oil which, despite the color, was glpc pure. The color was removed upon vacuum distillation.

CH3
8:1 AICI3/LiAlH4 HA (108)

HA
4HH A
(69) M C1


Treatment of the saturated ketal (69) with the 8:1

AlC13/LiAIH4 reagent gave an 81.1% yield of 88.5% glpc pure 1-metho







(7.7 1.7, and 2.1%) of slightly shorter retention time were not identified. The structure of the methoxy saturated chloride (108) was verified by its PMR spectrum 6[4.40 (1,ddd),

3.35 (3,s), and 3.02 to 1.13 (9, complex)], mass spectrum (172 m/e), measured mass for [C9H130C1]t and infra-red spectrum. The endo configuration was confirmed by the coupling patternzlc,22,36,37,k3b for Hx: JAX 9.75; JA'X 8.75; and JMX 6.5.


OCH3 OCH3


(98) /_ r(108)
HH

INa, THF Na, THF
t-BuOH -BuOH


OCH3 OCH3


(109)< H-N=N-H (110)


H H


As a further verification of the structure of the unsaturated and unsaturated chlorides (108) and (98), both compounds were dechlorinated using Gassman's procedure. Treatment of unsaturated chloride (98) with sodium metal in tetrahydrofuran/ tert-butanol gave a 44.2% yield of a pale yellow oil which was identified as 1-methoxytricyclo[3.3.0.02,8]oct-3-ene (109). Identification was accomplished from the PMR spectrum 6[5.53 (2,m), 3.33 (3,s), 3.13 (1,m), and 2.40 to 1.22 (6, complex)],







mass spectrum (136 m/e), measured mass for [CH1201t infrared spectrum, and elemental analysis.

In a similar manner, the sodium metal dechlorination of saturated chloride (108) produced a 49.6% yield of a colorless liquid identified as 1-methoxytricyclo[3.3.0.02'8]octane (110). The structural assignment was based on the PMR spectrum 6[3.37 (3,s), 2.74 (1,m), and 2.30 to 1.08 (10, complex)] mass-spectrum (138 m/e), infra-red spectrum, and elemental analysis.

The skeletal relationship between the saturated and unsaturated dechlorinated tricycles (110) and (109) was demonstrated by the diimide reduction of the olefinic bond of (109) to yield the saturated compound (110) as confirmed by spectral comparison. The use of the relatively mild diimide reduction procedure was necessitated by the fact that'hydrogenation of (109) over 10% Pd/C absorbed ca. twice the theoretical amount of hydrogen, leading one to assume that the methoxylated cyclopropyl ring was also being reduced.

Since the saturated and unsaturated rearranged chlorides (108) and (98) had become readily available from the synthetic point of view, and since their ionization should also lead directly into their respective tris-homocyclopropenyl cation manifolds, these systems provide excellent precursors to the identical intermediates involved both in the acidic hydrolyses of the saturated and unsaturated ketals (69) and (78), and the reaction of these ketals with dichloroaluminum hydride. This includes, of course, a rather confident assumption of a delocalized structure for these intermediates.









(1 1_ (112) (113)
OCH,
CH3CH2 CH CH30 OCH2CH3
60% aq. EtOH 0 J

H 104�,K2C03
(9 ) Cl 89 .2:10 .8


The first examined was the unsaturated rearranged methoxy chloride (98) which was solvolyzed in 60% aqueous ethanol in the presence of potassium carbonate at 1040 for ten hours. Capillary glpc analysis displayed only two components whose respective yields were determined by internal standards. Both fractions were isolated via preparative glpc. The first eluted component (53.6%) was determined to be cycloheptatriene (111) by spectral comparison with an authentic sample.

The second fraction consisted of mixed alkoxy ketals (21.5%) and was determined by PMR to be made up primarily (89.2%) of anti-8-ethoxy-syn-8-methoxy-e2o-tri cy lo[3.2.1.02 4]oct-6-ene (112). The PMR spectrum 6[5.72 (2,t), 3.38 (2,q), 3.27 (3,s), 2.86 (2,m), ca. 1.24 (2,m), 1.13 (3,t), and .47 (2,m)], mass spectrum (180 m/e), and absolute mass measurement for [C1iH1602l all agreed with the designated structure. The assignment of syn for the methoxyl group and anti for the ethoxy group is based on the fact that the cyclopropyl group's field effect causes a greater downfield shift than the double bond, a trend that will be summarized later with regard to all the ketals and methyl ethers encountered in this text. A singlet appearing at 63.13 in the PMR of the mixed ketal fraction was tentatively attributed the anti-methoxyl hydrogens








of anti-8-methoxy-syn-8-ethoxyendo-tricyclo[3.2.1.02"4]oct-6-ene

(113), and, through PMR integration, was determined to be 10.8% of the mixed ketal product. It is most noteworthy that the syn and anti methoxyl signals of both mixed ketals are identical to the chemical shift values of the respective syn and anti methoxyl signals for the dimethoxy unsaturated

ketal (78)

The need to firm up the PMR assignments for the two

mixed ketals (112) and (113) was immediately recognized, the complete characterization of ketal (113) being of prime importance.

The ethoxy analog of the methoxy rearranged unsaturated

chloride (89), i.e., 1-ethoxy-endo-6-chlorotricyclo[3.3.0.02,l]oct-3-ene (114) was synthesized by the identical procedure for (98), and subsequently subjected to solvolysis in 70% aqueous methanol at 1040 for 24 hours.

OCHCH, (112) 1113)
CH3CH0 OCH3 CHCH, 70% aq. CHJOH N+-H

1040, KC03
C 13.1:86.9

The mixed ketal fraction isolated was found by PMR to consist for the most part (86.9%) of anti-methoxy-synethoxy unsaturated ketal (113). In addition to the PMR spectrum 6[5.70 (2,t), 3.53 (3,q), 3.13 (3,5), 2.87 (2,m), ca. 1.26 (2,m), 1.22 (3,t) (3,g), and 0.48 (2,m)], the mass spectrum (180 m/e), and the absolute measured mass for








[C11H1602] were in agreement with the structural assignment. A singlet at 63.27 was attributed to the syn methoxyl PMR absorption of anti-ethoxy-syn-methoxy unsaturated ketal (112), and (112) was calculated through integration to comprise 13.1% of the mixed ketal fraction.

The solvolysis products are uniquely explained by cyclopropyl anchimeric assistance in the ionization of the rearranged unsaturated chloride (98) to the non-classical tris-homocyclopropenyl cation (99) If cation (99) should then be
OCH, OCH3 OCH,





Cl (99) (105)



0
CH3CH 0C H3 CH30 OCH2CH3
O -co /

CC 0

(111(106) (112) (1)


rapidly interconverted with the bis-homocyclopropenyl cation (105), the major ketal product, anti-ethoxy-syn-methoxy unsaturated ketal (112), would be a result of ethanol attack from the anti face of C8 in (99), (path B), and formation of the minor ketal product, anti-methoxy-syn-ethoxy unsaturated ketal (113) could be attributed to ethanol approach at C8 of bis-homo cation (105) via path B, i.e., from the anti face








with respect to the delocalized double bond. Of at least equal probability would be attack of cation (99) by ethanol along path A to give (113), assuming a weakly delocalized system or a symmetrically bridged intermediate ion. Leakage of (99) to a classical oxocarbonium ion, followed by solvent attack from both syn and anti faces, cannot be ruled out, but appears to be less consistent with all the results.

Attack by water at C8 of the intermediate(s) ion would generate the major product, unsaturated ketone (106). Under the experimental conditions, ketone (106) is known2]b,,54 to decarbonylate giving cycloheptatriene (111).

It is quite significant to note that, despite the indication that cation (99) was generated from the rearranged precursor (98), no rearranged products (path C) were observed. This point can only add emphasis to the concentration of positive charge at C8 in the delocalized cation (99).

The behavior of the saturated rearranged chloride (108)

under solvolytic conditions was subsequently studied. Chloride (108) was heated for 19 hours at 1000 in 60% aqueous ethanol

OCH3 (115) (116)
CH3CH2 CH3CH3 CH2CH30
60% aq. EtOH (70)
I 00�, K2CO 3 H
(108) Cl 96.0:4.0


in the presence of potassium carbonate. Capillary glpc analysis with an internal standard revealed two fractions. The largest fraction was determined to be endo-tricyclo-








[3.2.1.02'']octan-8-one (70), in a calculated yield of 68.1%. Ketone (70) was identified by spectral and glpc comparison with a known sample.

The ketal fraction (19.2%) was determined by PMR to consist mostly (96.0%) of anti-8-ethoxy-syn-8-methoxy-endotricyclo[3.2.1.02'4]octane (115). The PMR spectrum 63.47 (2,q), 3.33 (3,s), 2.13 (2,m), 1.78 to 0.52 (8, complex), 1.18 (3,t), mass spectrum (182 m/e), and absolute measured mass for [C1IH1802 ]. all supported the designated structure The antimethoxyl group of anti-8-methoxy-syn-8-ethoxy-endo-tricyclo(3.2.1.02,"'octane (116) was tentatively assigned to a PMR singlet appearing at 63.22, and comparative integration of the methoxyl signals of (115) and (116) revealed the minor component (116) to be 4.0% of the total ketal fraction.

As was the case with the unsaturated mixed ketals, there was a demonstrated need for the complete characterization of anti-methoxy-syn-ethoxy ketal (116) to add assurance to the above assignments. The ethoxy analog of the methoxy rearranged saturated chloride (108), namely, 1-ethoxy-endo-4-chlorotricyclo[3.3.0.02'8]octane (117), was synthesized via the identical method for (108), followed by solvolysis in 70% aqueous methanol for 27 hours at 1000

CH3CH20 0
70 CH3CH20 OCH3 CH30 OCHzCHa
70% aq. CHaOH

H 100 , K2COs
C 1
3.7:96.3


(117)


(115) (11._ 6)


( 709o)









The product analysis of the two fractions isolated

revealed the largest fraction (71.3%) to be tricyclic ketone

(70). The great bulk (96.3%) of the ketal fraction was determined to be anti-8-methoxy-syn-8-ethoxy-endo-tricyclo[3.2.1.02'"]octane (116) as demonstrated by the PMR spectrum 6[3.59 (2,q), 3.22 (3,s), 2.12 (2,m), 1.84 to .51 (8, complex), and 1.24 (3,t)], mass spectrum (182 m/e), and absolute measured mass for [C11H180j. A singlet at 63.33 was assigned to the syn-methoxyl PMR signal of anti-ethoxy-syn-methoxy saturated ketal (115), which was determined via integration to be 3.7% of the ketal fraction.

The anchimerically assisted ionization of saturated

rearranged chloride (108) is supported by the fact that only CB nucleophilic attack (paths A or B) on cation (73) was observed from the product analysis. Going from the rearranged


CHO
OCH3 B /

-Cl H20


Cl C (73) (70)

T EtOH

CH3CHj CH3+ CH3> H2CH3



96:4
(115) ( 116)









structure of (108) to the unrearranged products to the exclusion of path C adds to the concept of positive charge concentration at Ca, and the highly stereoselective attack (96%) of solvent on cation (73) from the anti face (path B) of Ce underlines the delocalized nature of the intermediate. The detection of 4% anti-methoxy-syn-ethoxy saturated ketal (116) leads to the rather obvious question of syn attack (path A) on what is conceived of as being a delocalized cation (73). It could be anticipated that since the methoxyl group competes so favorably with the cyclopropyl ring with regard to charge stabilization, any potential rehybridization of C8 from sp2 to sp3 would effectively be eliminated, resulting in a protruding, electron-deficient "p" orbital, occasionally accomplishing nucleophilic capture from the syn face of the bridge, despite residual cyclopropyl delocalization.

At the same time, the product distribution hardly argues for the intermediacy of the classical cation (72) since there is simply no explanation at this point for the overwhelming attack of nucleophiles via path B for (72). As
QCH3
B + A
(72)



can be seen from Table I, the reduction of tricyclic ketone

(70) with numerous reagents shows no overwhelming steric preference for hydride attack via paths A or B, and certainly not to the point of 96% anti-attack.








Table I. Hydride Reduction Product Distribution for Ketone (70)
0

%H 0



LiAIH, 5 67 33

NaBH446c 58 42

LiAl(Ot-Bu)H 4C 58 42

Ali-PrO) /i-PrOH46C 20 80

Li(secBu)3BHa 90.5 9.5

PMHS/DBATOa'b 63 37



aThis work

bPolymethylhydrogen siloxane/tetrabutyldiacetoxytin oxide dimer



Yet another means of generating the non-classical methoxy tris-homocyclopenyl cation (73) was found in the reaction of saturated rearranged chloride (108) with silver perchlorate in methanol/acetone. The tricyclic ketone (70)
OCH3 0


Ag+
CHOH

(108) Cl (118)

was isolated (30.7%) and identified via spectral comparison with an authentic sample. The second component (34.7%) was identified as methyl 4-cycloheptene-l-carboxylate (118) from its PMR spectrum 5[5.75 (2,t), 366 (3,s), and 2.82 to 1.21








(9, complex)], mass spectrum (154 m/e), and infra-red spectrum, all of which agreed with the published spectral data. Ester (118) was also synthesized from 4-cycloheptene-1-carboxylic acid (119) and diazomethane, followed by spectral comparison and authentication.
C02H C02CHs

CH2N2


(119) (118)


OCH~ 3iQ OCH


Ag+ + -A Cl H20 Ag+


Cl (73) (120)

(108) 0
11 CO2CH3

(70) [ ) (118)



Both ketone (70) and the cycloheptene ester (118) can be viewed as products derived from the hemi-ketal (120) Attack by water upon the charge rich C8 position of the trishomocyclopropenyl cation (73) (anti approach) would yield the hemi-ketal (120). Hemi-ketal (120) then has the option of losing methanol to form ketone (70), or, with the assistance of either Ag+ or proton cleavage of the cyclopropyl ring, rearranged7 in the manner shown to give the cycloheptene ester (118).







Mention was made earlier of an alternative synthesis of endo-6-chlorotricyclo[3.3.0.02"'8]oct-3-ene (97). Breslow has reported"8 the generation of 5-halo substituted cyclopentadienes, and in particular, 5-chlorocyclopentadiene (121), by the reaction of cyclopentadienylthallium with n-chlorosuccinimide. Cyclopropene was then bubbled through a solution of (121), and glpc analysis revealed the presence of three minor products and one major product (total yield ca. 25%). The major product (83.7% of product mixture), rearranged chloride (97), was identified by its spectral data and by literatureso comparisons already discussed.



NCS

(121)



(97) + (123)


CI (122) C 1

Two of three minor products were identified. Syn-8chloro-endo-tricyclo[3.2.1.02 ']oct-6-ene (22), in 8.0% relative yield, was identified from it's PMR spectrum 6[5.81 (2,t), 4.04 (1,m), 2.82 (2,m), 1.62 (2,m), 0.78 (2,m)], and glpc/mass spectrum (140 m/e). 1-Chloro-endo tricyclo[3.2.1.02'4]oct-6-ene (123), in 6.1% relative yield, was also identified from the PMR spectrum 6[5.80 (2,d), 2.77 (1,m), 2.38 (1,q), ca. 2.04 (1,m), 1.63 (2,m), ca. 0.68 (2, complex)], and glpc/mass spectrum (140 m/e).








Cyclopentadienes substituted in the C. position are knowns" to tautomerize readily to the more stable vinyl substituted cyclopentadienes; however, Breslow reportedsa that his method involving cyclopentadienylthallium reduces this problem. In fact, only one minor product, i.e. (123), could be attributed to tautomerization. Syn-chloride (122) appears to result simply from the Diels-Alder addition of cyclopropene to (121) on the chlorine side of the cyclopentadienyl plane.


Cl





(121) (124) (102) (97) Cl

The rearranged chloride (97) appears to be the result of cyclopropene addition to (121) giving the transient antichloro-endo-tricyclo[3.2.1.02,]oct-6-ene (124) which, through cyclopropyl assistance and possibly thallium (I) cationic coordination catalysis, would rearrange (probably through a tight ion pair) via the tris-homocyclopropenyl cation (102) to give the major product (97).

In summary, the results presented strongly indicate that the ionic intermediates generated by the reaction of dichloroaluminum hydride with, and the acidic hydrolysis of, the saturated and unsaturated ketals (69) and (78), are the same as those generated by the solvolysis of the saturated and unsaturated rearranged chloro ethers (108) and (98), i.e.,






the 3-methoxy-2,4-ethano- and 3-methoxy_2,4- etheno-trishomocyclopropenyl cations (73) and (99) respectively. Product studies have also raised the possibility that the unsaturated tris-homocyclopropenyl cation (99) is in equilibrium via a bridge flipping process with the bis-homocyclopropenyl (105).

The electron structure of these ions has been radically altered with respect to the parent ions in that positive charge delocalization is not as extensive with most of the charge localized or concentrated at the methoxyl bridge carbon. The cyclopropyl delocalization may even have been weakened to the point where nucleophiles are able to penetrate and attack the bridge from the syn face of the bridge. This electronic effect, not steric bias, is thus the overwhelming factor in determining the orientation and stereochemistry of solvent or nucleophilic attack.


C 3 CH3
CH H (69)



Al Cl 2H

OCH3 CH 3
CH3O OCH 3k
H+/H20 aq. ROH/1000
or Ag+/CH30H H

(69) (73) (108) Cl













CH30- OCH,
(78) Al ClH


CHl . OCH OCH3
~~~H+/HaO "


(99)
(78)
aq. ROH 100� OCH




cl


OCH3




(105)






(98)










CHAPTER III

PMR Studies
I and Anti Chemical Shifts of Alkbxyl' Groups at Ce in

Tricyclo[3.2.1.0''joctanyl-octenyl Systems

It is advantageous at this point to summerize the PMR chemical shifts of methoxy and ethoxy groups substituted at the C, position of the tricyclo[3.2.1.02'4]octane and tricyclo[3.2.1.02'4]oct-6-ene systems. Much of the work already presented depends upon the correct syn or anti assignment at Ca of these alkoxy groups.

Authors seem to have been hesitant to declare in the

literature their assignments of the methoxyl chemical shifts of the saturated and unsaturated ketals, (69) and (78). Pincock has stated22 that the difference in chemical shift caused by the field effect of the endo cyclopropyl ring in saturated ketal (69) upon the syn methoxyl relative to the anti methoxyl group is 0.11 ppm (observed here as 0.10 ppm). He compared22 this value with the shift induced by the double bond upon the syn methoxyl in norbornene dimethyl ketal (75), i.e. 0.07 ppm, relative to the anti-methoxyl group, and concludes that, within the specified geometry of these systems, the cyclopropyl group relative to the double bond, has the greater effective ability to deshield. This fact is demonstrated in Table II if one examines the methoxyl shifts for the syn and anti unsaturated ethers (95) and (101). The








Table II. Syn and Anti Cs AlkoxylChemical Shifts in
Tricyclo[3.2.1.02'4]octenyl Systems


CH OCH3

(78) /



H OCH3

(95)




C H 0 -1 H (101---) /


CHCH20 OCH3






CH3,O CHCH (113)


6an ti



3.13


6syn



3.27


Asyn-anti, ppm


3.27


3.21


3. 38
(-CH2)





3.13


3.27






3.53 (-CH2)


.14 (-OCH,) .15 (-OCHCH,)


syn-methoxyl of (95) is deshielded by the cyclopropyl group by .06 ppm more than the double bond deshielding of the antimethoxyl of (101). The observed syn and anti-methoxyl signals for the mixed ketals (112) and (113) are identical to the syn









and anti-methoxyl signals of the parent compound, unsaturated dimethoxy ketal (78), confirming the epimeric assignments of (112) and (113). An inspection of Table III listing the chemical shifts of the saturated analogs also confirms the epimeric assignments for the saturated ketals, (69), (115) and (116).


Table III. Syn and Anti Cs Alkoxyl Chemical Shifts in
Tricyclo[3.2.1.02,41 octanyl Systems
anti 6syn Asyn-anti, ppm
CH 30 CH 3

(69) 3.21 3.31 .10



H OH3

(89) 3.33


.06
CH

(90) 3.27




CH3CH 0 OCH

(115) 3.47 3.33


(11 (-CH3)
CH3 HCH3
.12 (-OCH2CH3)
(116) 3.22 3.59
(-CH2) (-CH2)








Computer Analyzed PMR Spectra

In an effort to add to the relatively scarce PMR data

concerning the cyclopropyl group in the tricyclo[3.2.1.02',4octyl systems, the four polychlorinated endo cyclopropene Diels-Alder adducts, (88), (125), (126), and (127) were prepared. Thoroughly degassed carbon tetrachloride solutions of these compounds were sealed under vacuum in NMR tubes. The individual spectra were expanded to 50 Hz and the absorptions calibrated via the TMS side-band technique. The observed absorption data was applied to a modified LAOCOON 1116' computer program, and coupling signs were based on convention.


CH30 OCH3

(88) C1 Hi


H3 Ha
The preparation of 1,5,6,7-tetrachloro-8-dimethoxyendo-tricyclo[3.2.1.02'4]oct-6-ene (88) has already been described in this text. The proton chemical shifts and standard deviations were reported for a 1M solution (CC14) in Hz as follows: Hi, -105.317�0.015; H2, -53.263�0.017; H3,

-24.715�0.016. Coupling constants were: Ji,2=7.073�O.020; Jl,,=3.469�0.018; J2,3= -7.434�0.023.



(125) C 4H
Ha4 Ha

The addition of cyclopropene to hexachlorocyclopentadiene gave 1,5,6,7,8,8-hexachloro-endo-tricyclo[3.2. 1.02'] oct-6-ene










(125) in 92.6% yield. The PMR and infra-red spectra were in agreement1 with the assigned structure, as was the elemental analysis. The computer analyzed proton chemical shifts for a IM solution (CCI4) were reported in Hz as: H,, -120.777�0.012 Hz; H2, -77.408�0.013; H3, -46.877�0.012. The coupling constants were revealed as: J1,2=7.175�0.014; J, =3.493�0.014; J2,3= -7.707�0.017




(126)H
C1l4

H3 H2
1,5,6,7-Tetrachloro-8-dichloromethylene-endo-tricyclo[3.2.1.02,4]oct-6-ene (126) was prepared in 80.4% yield from the addition of cyclopropene to hexachlorofulvene.2 The PMR and infra-red spectra agreed with the assigned structure, as did the mass spectrum (322 m/e) and elemental analysis. The computer analyzed PMR spectrum gave the chemical shifts of a 0.51 M solution (CCI,) in Hz as: H,, -121.390�0.009; H2, -67.772�0.010; H3, -39.458�0.012. The coupling constants were determined to be: J1,2=7.025�.011; J1,3=3.090�.011; J2,3= -7.804�.014.



(127) C I H3

Hs H4







The addition of cyclopropene to 1,2,3,4-tetrachlorocyclopentadiene gave 1,5,6,7-tetrachloro-endo-tricyclo[3.2.1.02']oct-6-ene (127) in 77.7% yield. The PMR and infra-red spectra, along with the elemental analysis, confirmed the structure.61 The computer analyzed PMR spectrum of a 1M solution (CCI4) gave the proton chemical shifts in Hz as follows: HI, -185.844�0.020; H2, -149.271�0.017; H3,

-119.841�0.018; H4, -56.641�-0.018; HS, -34.716�0.018. The

coupling constants were reported as: J1,2= -6.707�0.026; J3,9= -0.246�0.026; J,,=0.298�0.028; J213= -0.290�0.020; J2,,= 2.528�0.025; J3,4= 7.218�0.021; J3,5= 3.170�0.020; J,5= -7.503�0.024. The noteworthy (2.5 Hz) extended w'1. 3 coupling of H2 and H4 protons, and the small (0.30 Hz) coupling of H, and H, has potential use in confirming syn and anti assignments involving mono substitution at the bridge C, position.


















J .!H 3 I
I hjPy-


Figure 1. endo-4-Methoxytricyclo[3.3.0.02,8]octane, (92).


OCH3
















Cl


Figure 2. endo-4-Chlorotricyclo[3.3.0.02',]octane, (93).

















- h









iJil




J1 7

5-0 .0 50 PPM ( ) 4.0 2.0 Lo


Figure 3. endo-6-Chlorotricyclo[3.3.O.O2f8]oct-;3-ene, (97).


V 0-.',













I I-


50 . .~T) .0T


OCH, cl


I __ /J


Figure 4. 1Methoxy-endo6Chlorotricyclo[3.3.0.O2,8]oct-3-ene, (98).


7.0


9.0


I , . . I , i ,










" I . . . . . . . . I


OCH3




4H cl


JjUL


JL IL


~L] .~1~JZJ ZZZZL I .j~


- I . . . . -- - - ,J1 I . . . . . l. - . . - , . . .
0.0 7.0 o.0 3.0 PPM 1 6 .0 .0

Figure 5. 1-Methoxy-endo-4-chlorotricyclo[3.3.0.02,]octane, (108).


2.0


I 0.0


I ? I







do I I . . I IO P A fI i
do ' ' '3 .

CHCH20 001,











; IL

F., PM , (112)
_ I . .0t
Fiue6 at--tox-y--etoyedotiyl[321.2'ic--n , (2.


































Figure 7. anti-8-Methoxy-syn-8-ethoxy-endo-tricyclo[3.2.1.O2,4]oct-6-ene, (113).







.0 J. 4 3.0 rrA( I '0 0 .









Fi ur S . , -8- E h -y- . . , -ti [ .2.1 , ] o ctane , (115).'-'"
.)aCH 3C O OCH 3 6H'C 04ac
















3. 2.0 j 1.0 "0 ''
Figure 8. cnti-8-Ethoxy-sn-8-mthoxy-eno-tricyco[3.02,']octane, (11_5).








o .-,


CH3O OCH2CH3


Figure 9. anti-8-Methoxy-syn-8-ethoxy-endo-tricyclo[3.2.1.O2',4]octane, (116).










- rI 6



'"'i CH30 OCH,



I I H

'H H










II

,,1


10


S I


I - I


0.0


7.0


5.0 P'M (6J 4.0


2.U


Figure 10. 1,5,6,7-Tetrachloro-8,8-dimethoxy-endo-tricyclo[3.2.1.02,4toct-6-ene, (88), at 250 Hz.


I )"


,,v 1 . . . . ",' - - � . � . . ", '.


. . . . . . . . m . . . . . . . �


I


.v









272 I .Y - " � . . . .l .,: .-" ,


C14
H


II .~ I


Figure 11.


/ .0 6 .0 5 .0 V I M ' , 6( ) 4 .0 3 .0 2 .0 ).0

1,5,6,7,8,8-Hexachloro-endo-tricyclo[3.2.1.02,4]oct-6-ene, (125), at 250 Hz.











U - - PPM In II-o


I I1

C1




H
H H "















- -


U.0


Figure 12.


-" ,-,,.,! 'I


01' d opt






F I ii II



I ~



Ii I i i~* ~i I


4k' MUM. Li 1L4 K






- I I I


SI . W ~ -- Y -~ ~ - - I


/.0 6.0 5.0 PPM (j) 4.0 3.0 2.0 1.U


1,5.6,7-Tetrachloro-8-dichloromethylene-endo-tricyclo[3.2.1.0,4]oct-6-ene, (126),
at 250 Hz.


4
I0


I0










,o 1 * io , ' l . 0, . ' .






C141


H H




I I"
I______H___ I F














0.0 7.0 6.0 .0 P'M ) 4.0 2.0 2.0 1.0 0


Figure 13. 1,5,6,7-Tetrachloro-endo-tricyclo[3,2.1.02,4loct-6-ene, (127), at 250 Hz.


co
(n











CHAPTER IV

Experimental


General

Melting points were determined on a Thomas-Hoover unimelt capillary melting point apparatus and are uncorrected. Elemental analyses were performed by Atlantic Microlab, Inc. Atlanta, Georgia.

Vapor Phase Chromatography

The analytical gas/liquid phase chromatography was performed with an Aerograph Hy-Fi 600-D instrument with a flame ionization detector, using column A, which consisted of 15% FFAP on Chrom W, AW, DMCS, 1/8" x 5', with a helium flow rate of 300 cc/min. Column B consisted of a 100' x 1/100" DEGS capillary column fitted on a Varian Aerograph Series 1400 flame ionization instrument. For preparative work, a Varian Aerograph model 90-D with a thermal conductivity detector was employed using one of two columns: column C which consisted of 15% FFAP on Chrom W, AW, DMCS, 4.5' x 1/4", helium flow ca. 140 cc/min. column D which was packed with 20% DEGS on 45/60 Chrom W, 10' x 1/4", helium flow ca. 100 cc/min.

Spectra

Infrared spectra were determined on either a Beckman IR-1O or a Perkin-Elmer 137 Sodium Chloride Spectrometer. Proton








magnetic resonance spectra were recorded on Varian A-60A instrument. Low resolution mass spectra were provided by a Perkin-Hitachi RMU-6E instrument. High resolution spectra were determined by a AEI-MS-30 instrument which was also used in conjuction with a Pye Unicam Series 104 Chromatograph utilizing a 5' x 1/4" SE-30 column.


Generation of Cyclopropene

A slightly modified procedure of the original method reported by Closs and Krantz64 was employed. A typical run is as follows

To a magnetically stirred suspension of 50 grams of

fresh sodium amide in 70 ml of dry mineral oil at 85-900 were added dropwise (1 drop/4 sec.) approximately 75 ml of 3-chloropropene. A slow stream of dry nitrogen gas transported the evolved cyclopropene through a Friedrich condenser (water cooled), followed by a 4N sulfuric acid trap (200 ml) and then onto the reaction vessel via a scintered glass outlet. Preparation of 1,2,3,4-Tetrachloro-5,5-dimethoxycyclopenta1,3-diene (87)

To a mechanically stirred solution of 205 grams of hexachlorocyclopentadiene (0.751 mole, Aldrich) i'n 300 ml absolute methanol was added a solution of 112 grams of potassium hydroxide (2.0 moles) in 300 ml absolute methanol The addition was carried out at a rate commensurate with maintaining gentle reflux, and was followed with continued stirring at room temperature for ca. ten hours. The reaction mixture was filtered to remove precipitated potassium chloride, and







solvent was removed under water aspirator pressure after drying with anhydrous magnesium sulfate. Vacuum distillation (b.p. 63.065.50/0.05 mm) yielded 115 grams (0.436 mole, 58.0%) of pale yellow oil The spectral data were in agreement with those previously reported."6

Preparation of 1,2,3,4-Tetrachloro-5,5-diethoxycyclopenta1,3-diene

The synthetic procedure followed is identical to the preparation of the dimethoxy analog (87) already described. Using 205 grams (0.751 mole) of hexachlorocyclopentadiene, 112 grams (2.0 moles) of potassium hydroxide, and 600 ml total absolute ethanol, 107.4 grams (0.389 mole, 51.8%) of the diethoxy ketal was isolated after vacuum distillation (b.p. 1031050/2.6 mm). Spectral data were in accord with literature values.,

Preparation of 1,5,6,7-Tetrachloro-8,8-diemthoxy-endo-Tricyclo [3.2.1.04'f]oct-6-ene (88)

Cyclopropene was passed through a rapidly stirred solution of 40.0 grams (0.152 mole) of 1,2,3,4-tetrachloro-5,5dimethoxycyclopenta-1,2-diene (87) in ca. 250 ml of petroleum ether (20-400) at ambient temperature. The reaction was followed by PMR monitoring of the respective methoxyl methyl signals and was found to be complete after ten hours. The solvent was removed by passing a stream of nitrogen over the solution, yielding 34 grams (0.112 mole, 73.7%) of white crystals (m.p. 63-660, lit. m.p. 68-700). Spectral data were in agreement with literature values,21a,22,39 and a computer analyzed PMR spectrum is presented in Chapter III.








Preparation of 1,5,6,7-Tetrachloro-8,8-diethoxy-endotricyclo[3.2.1.02, ]oct-6-ene

The same procedure applied above for the preparation of

the dimethoxy adduct (88) was employed for the diethoxy analog. From 105 grams (0.380 mole) of 1,2,3,4-tetrachloro-5,5-diethoxycyclopenta-1,3-diene, 98.6 grams (0.297 mole, 78.1%) of the cyclopropene adduct was isolated after vacuum distillation as a yellow liquid (b.p. 136-145o/1.25 mm). The analytical sample was isolated from preparative glpc column C, at 1301 The PMR spectrum (CDC13) consisted of two overlapping two proton quartets at 63.95 and 3.83, a two proton quartet at

1.78, two overlapping three proton triplets at 1.26 and 1.14, a one proton sextet at 0.90, and a one proton quintet at 0.40. The infra-red (neat) gave absorption bands at 2990 (m), 2890

(w), 1597 (m), 1275 (m), 1160 (s), and 1025 (m) cm"1

Anal. Calcd. for C12H1,Cl,: C, 43.40; H, 4.25; Cl, 42.71.

Found: C, 43.43; H, 4.28; Cl, 42.65. Preparation of 8,8-Dimethoxy-endo-tricyclo[3.2. 1.02'4]oct6-ene (78)

The dechlorination of ketal (88) was accomplished via

Gassman's procedure. Into a 11 flask fitted with a mechanical stirrer, nitrogen inlet, and water-cooled Friedrich condenser, were added 25.0 grams (0.0822 mole) tetrachloroketal (88), 125 ml tert-butanol, 325 ml tetrahydrofuran, and chopped sodium (39.0 grams, 1.70 g-atoms) The solution was stirred and heated to gentle reflux for ten hours, cooled, and filtered through wire gauze to remove unreacted sodium. The filtered solution was added to 50D ml of water, followed by the addition of







250 ml of brine and 250 ml of ether. The organic layer was separated and the aqueous layer was extracted (6 x 100 ml) with ether. The combined ether fractions were dried over anhydrous magnesium sulfate, and the bulk of the solvent was distilled utilizing a steam bath and a 2' Vigreaux column. Traces of solvent were removed on a rotary evacuator. Vacuum distillation (30-350/4mm) yielded 7.62 grams (0.0458 mole, 56.0%) of light yellow product, whose spectral analysis agreed with that previously reported.2la2239 Preparation of 8,8-Diethoxy-endo-tricyclo[3.2. 1.02'"]oct-6-ene

The dehalogenation procedure employed for the dimethoxy analog (88) was followed using, in this case, 98 grams (0.295 mole) of 1,5,6,7-tetrachloro-8,8-diethoxy-ento-tricycloE3.2.1.02'"]oct-6-ene, 450 ml of tert-butanol, and 140 grams (3.59 g-atoms) of chopped sodium metal Due to the generation of a considerable amount of tar, work-up was more difficult and required the use of hexane to effect separation. Vacuum distillation (b.p. 65-70/1.9 mm) gave 18.0 grams (0.0927 mole, 31.4%) of a pale yellow oil The analytical sample was isolated on preparative glpc column D at 1450 (7 min.). The PMR spectrum (CDCI3) exhibited a two proton triplet at 65.67, two overlapping two proton quartets at 3.53 and 3.38, a two proton multiplet at 2.84, two overlapping three proton triplets at 1.20 and 1.09 which obscure a two proton multiplet at ca. 1.26, and two overlapping one proton multiplets at ca. 0.49. The calculated mass for [C12H80jO is 194.1306, while accurate mass measurement (70.eV) gave 194.1303, for an error of -1.39




Full Text

PAGE 1

THE 2,4-ETHAN0~ and 2 , 4ETHENO-3ALKOXYTRIS-HOMOCYCLOPROPENYL CATIONS, ORIENTATION AND STEREOCHEMISTRY OF NUCLEOPHILE CAPTURE By WARREN CHARLES NIELSEN A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1974

PAGE 2

DEDICATION To My Parents

PAGE 3

ACKNOWLEDGEMENTS The writer wishes to express his most sincere gratitude to Professor Merle A. Battiste for his excellent guidance, interest, and enthusiasm in the development of this research project. It has been a pleasure to work with a research director who has the ability to maintain a high level of professionalism while extending a \>ery real, personal friendship. Appreciation is expressed to Dr. Roy W. King for his many hours of advice and assistance. The author also feels a need to express his admiration and fondness for Gainesville, the people, the land, and the free spirit. The writer's wife Judi deserves such an expression of gratitude that one feels incapable of the words. Her patience during hours of superb typing, and her overall giving nature defy description. 1 1 1

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS iii LIST OF TABLE'S vi LIST OF FIGURES ; vii ABSTRACT viii CHAPTER I Introduction 1 II Results and Discussion 32 Synthesis of Precursors 32 Statement of Problem 33 Reaction Analysis of 8,8-Dimethoxy-endo-tri cyclo[3. 2. i.O 2 '^octane, (69), with Dichloro aluminum Hydride Reagent 36 Synthesis of ant-£-8-Methoxy-en
PAGE 5

Chapter Page III PMR Studies 67 Syn and Anti Chemical Shifts of Alkoxyl Groups at C 8 in the Tri cycl o [3 . 2 . 1 . 2 ' "*] octanyl -octenyl Systems 67 Computer Analyzed PMR Spectra ,\ 70 IV Experimental 86 BIBLIOGRAPHY 12 5 BIOGRAPHICAL SKETCH , 131

PAGE 6

LIST OF TABLES Table I II III Page Hydride Reduction Product Distribution for Ketone .(70) 61 Syn and Anti C 8 Al koxyl Chemical Shifts in Tricyc1o[3.2.1.0 2,,f ]octenyl Systems , 68 Syn and Anti C 8 Alkoxyl Chemical Shifts in Tricyclo[3.2.1.0 2 » ,f ]octanyl Systems 69 VI

PAGE 7

LIST OF FIGURES PMR Spectra Figure Page 1 endo-4-Methoxytricyclo [3. 3.0.0 2 ' 8 ]octane, (92) . 73 2 endo-4-Chlorotricyclo [3.3.0. 2 ' 8 ]octane, (9_3).. 74 3 endo-6-Chlorotricyclo [3.3.0.0 2 ' 8 ]oct-3-ene, (iZ.) 75 4 l-Methoxy-en5,6,7,8,8-Hexachloro-enao-tricyclo[3.2.1.0 2 '' 4 ]oct-6-ene, ( j_2_5_) , at 250 Hz 83 12 l,5,6,7-Tetrachloro-8-dichl oromethyl ene-endctricyc1o[3.2. 1 . 2 ' " ] oct-6-ene , (126), at 250 Hz 777 84 13 1, 5,6,7Tetrachloro-e -i.o-tr icy clo [3.2.1. 2 '"]oct-6-ene, (12_7_) , at 250 Hz 85 VI i

PAGE 8

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE 2,4-ETHANOand 2 , 4-ETHEN0-3-ALK0XYTRIS-HOMOCYCLOPROPENYL CATIONS, ORIENTATION AND STEREOCHEMISTRY OF NUCLEOPHILE CAPTURE By Warren Charles Nielsen December, 1974 Chairman: Dr. Merle Battiste Major Department: Chemistry Solvolyses of endo, antitri cycl o [3. 2 . 1 . 2 ' " ].octanyl derivatives, i.e. , hydrogen, vinyl, and p-anisyl substitution at the bridge C 8 position, have been shown to afford rearranged products {endo C 2 or C^ attack) almost exclusively (>99%). Concentration of positive charge at the C 2 and C^ positions of the respective intermediate tri s-homocycl opropenyl cations might be an explanation; an alternative view stresses the importance of strain relief in the transition state for solvent capture of the non-classical cations. In contrast to the above results, acidic hydrolysis of 8 ,8-dimethoxy-endo-tri cycl o [3. 2 . 1 . 2 ' k ] octane' (I) yields the expected ketone II with no evidence for the formation of a rearranged hydroxyether. This suggests that the charge stabilizing ability of the methoxyl group at the bridge carbon may well be overwhelming the skeletal bias of the transition state for solvent capture. VI 1 1

PAGE 9

Saturated ketal I was treated with dichloroal umi num hydride to yield the sz/n-methyl ether exclusive of its anti epimer. Under the same reaction conditions, the unsaturated ketal , 8,8-dimethoxy-endo-tricyclo [3.2.1.0 2,1+ ]oct-6-ene (III), gave the unsaturated synmethyl ether IV as the major product along with l-methoxy-endo-6-chl orotri cycl o [3. 3 . . 2 ' 8 ] oct-3-ene (V) as a minor product. Traces of endo6me thoxyand endo-6-chlorotricyclo[3. 3. . 2 ' 8 ] oct-3-enes,(VI) and (VII), were also detected. The trace components are attributed to the intermediacy of unsaturated anti-methyl ether VIII as evidenced by its authentic synthesis and subsequent exposure to the reactions conditions. The predominance of bridge C 8 anti hydride attack and formation of endo unsaturated rearranged chloride V argue strongly for the intermediacy of a delocalized system with charge concentrated at C a due to methoxyl stabilization. The transient formation of the unsaturated anti methylether VIII could result from hydride attack upon a bis-homocyclopropenyl cation, whose mode of formation has several potential routes. Use of an 8:1 molar ratio of the aluminum chloride/ lithium aluminum hydride reagent with the respective saturated and unsaturated methyl ketals I and III produced good yields of the saturated and unsaturated rearranged methoxy chlorides IX and V, both of which were solvolyzed at 100° in aqueous ethanol. For the unsaturated chloride V, the major product, cycloheptatriene , and a mixed ethoxy-methoxy ketal fraction of the original tri cycl o [3. 2 . 1 . 2 » k ] octenyl structure were i x

PAGE 10

isolated. Cy cl ohep ta tr i ene is the expected product from the well-known decarbonyl ati on of the unsaturated ketone. According to PMR analysis the mixed ketal fraction consisted of 89% arcti-ethoxy-syrc-methoxy unsaturated ketal and 11% antimethoxy-st/n-ethoxy ketal. Solvolysis of the saturated chloride IX in aqueous ethanol produced the saturated ketone II as the major product with a mixed ketal fraction consisting of 96% anti-ethoxy-syn-methoxy ketal and 4% anti-methoxy-syn ethoxy ketal . The results obtained strongly indicate that the same cations are being generated from the respective reactions of the saturated and unsaturated ketals I and III with dichloroaluminum hydride and aqueous acid, as well as in the solvolyses of the saturated and unsaturated rearranged methoxy chlorides IX and V. The electronic structure of these ions has now been radically altered with respect to the parent ions in that positive charge del ocal i zati on is not as extensive, giving a more localized or concentrated charge at the methoxyl bridge carbon. The cyclopropyl del ocal i zati on may even have been weakened to the point where nucleophiles are able to penetrate and attack the bridge from the syn as well as anti face of the bridge. Thus an electronic effect, and not steric bias, is the overwhelming factor in determining the orientation and stereochemistry of solvent or nucleophilic attack on the 2,4-ethanoor 2 ,4-etheno-3-al koxytri s-homocycl opropeni urn cati ons .

PAGE 11

CHAPTER I Introduction The neighboring group reactivity of the cyclopropyl moiety has received intense investigation and documentation in the chemical literature. 1 One of the more dynamic directions of these studies has involved the assistance to ionization rendered by the cyclopropyl carbon-carbon sigma (edge) bond at a remote site relative to the cyclopropyl group. >H (i) (6) (3) OTs 1 k s HOAc OTs OOAc H A) OAc or -*$ (2) HOA OAc OAc OTs HOAc l" & OAc (3) 0'&O-0~ P-Q-Q oac (5 ;

PAGE 12

In the vanguard of the early cyclopropyl studies was work carried out by Winstein and his reports 2 of the results of solvolytic studies involving oisand trans-3-b] cycl o[3.1.0]hexyl tol uenesul fonates , [1) and (2_) . The
PAGE 13

observed with the correspondingly deuterated trans-3tosyl ate under the same conditions. This was i nterpreted 23 » d > e as powerful evidence for the i ntermedi acy of the tri s-homocycl opropenyl cation (6_) . Corey studied 1 * the deami nati on of aisand tvans-2bi cycl o [3. 1 . 0] hexyl ami ne and found a complex product mixture for both amines consisting of epimeric 2and 3-alcohols. There was only partial scrambling of a deuterium label for the c-ts-3-amine and essentially no scrambling for the trans3-amine deamination. Corey felt that both amines produced classical ions which did not leak over to the homoaromati c species, possibly because deamination produces a vi brati onal ly excited cation. A number of other workers have also expressed 5 the viewpoint that the diazonium ion is a poor model for solvolytic work. In an accompanying paper, 6 Corey acetolyzed 1,5-diphenyl -cis-3-bicyclo [3. 1. Ojhexyl to! uenesul fonate (_7 ) , expecting a rate increase from phenyl stabilization if the trishomocyclopropenyl cation (8_) was indeed generated. In fact, he observed a small rate retardation; the products detected, (8) (7) • * * * * however, were totally rearranged. To explain his results and Winstein's, Corey suggested equilibrating classical ions (_9) , the equilibration being the reason for deuterium scrambling and a "weak interaction involving the vacant

PAGE 14

1.9) orbital at C 3 and the loose electrons of the three-membered ring" 6 responsible for the stereospeci f i ci ty of ais-3bicyclo [3. l.OJhexyl tosylate (l). Winstein has countered with the point that ionization with assistance from the cyclopropyl ring in the ais-2tosylate (_1) should occur only in the chair conformation of the bicycl ohexyl ring, and there is evidence 2e » 7 > 8 that bicyclohexyl derivatives have a marked preference for the boat conformation. As a consequence, there would' be, at any one time, a proportionately small number of molecules in the system capable of employing the anchimeric assistance of the cyclopropane ring. The titremetric rate constant for the cis-3-tosyl ate [I) should be well below any valid estimate for the degree of anchimeric assistance. Winstein felt 2e the diphenyl derivative {]_) would be shifted even more into the boat conformation and therefore would be expected to ionize in essentially a classical manner. Extending this hypothesis further, generation of a positive charge at the C 3 position of the bi cycl o [3. 1 . 0] hexyl system and its derivatives by processes which do not depend upon anchimeric assistance would by very large odds, take place in the boat conformation. One would anticipate the classical ion formation to be a higher energy process than the non-classical solvolytic process, and this, coupled with the high probability for boat form

PAGE 15

involvement, would result in a very reactive ion that may have little opportunity to leak into the tri s-homocyclopropeni urn mani fol d . Gassman attempted 9 the electroytic oxidative decarboxylation of aisand trans~bi cycl o [3 . 1 . 0] hexane-3-carboxyl i c acids (10) and determined the predominant products to be the aisand trans-2-bi cycl o [3. 1 . 0] hexanol s , (JJJ and (12). W^NpfCOzH (10) HCk^H H OH * + 11 (12) O O « * <> Deuterium labeling at Ce produced no scrambling. Both Gassman and Winstein agreed that the reaction on the electrode surface was too complex to allow a valid mechanistic comparison with the solvolysis of the 3-bi cycl o [3 . 1 . 0] hexy 1 tosylates (I) and (2). Freeman 7 attempted to generate a carbonium ion at the C 3 position of the bi cycl o [3. 1 . 0] hexyl system via the acid catalyzed addition of methanol to 2-bi cycl o [3 . 1 . 0] hexene (21), anticipating the generation of Winstein's cation (6). CHaOH t 0CH 0CH 0CH

PAGE 16

All the products identified, however, were considered to be derivatives of the bi cycl ohex-2-yl cation (5), which is a cyclopropylcarbinyl cation that Winstein considers 26 to be more stable than the tri s-homocycl opropeny 1 cation (6). Much of the solvolytic work following Winstein's original work 2 generally added to the weight of evidence for the existence of the tri s-homocycl opropeny! cation as a viable intermediate. Norin subjected 1 ° the optically active thujyl tosylates, (U) and ( j^) , to acetolysis and found they yielded their respective racemic acetates, paralleling Winstein's deuterium labeling studies in the parent bicyclohexyl series. As expected, the rates of (L4) and (_15J were greater than their trans analogs. Solvolyses of the monoand dimethyl substituted bicyclohexyl tosylates, (16) and ( 17) , have been recognized 2e ' 3a as examples for generation of tri s-homocyclopropenyl species. OTs CH 3 H 14 OTs CH 3 (15) OTs (18) CH OTs CH; CH : (16) (19) OTs (17) Of the numerous options for the tri cyclodecyl tosylate (1_8) during solvolysis, Huckel LCAO-MO calculations predicted 11 the development of the tri s-homocycl opropeni urn cation (_1_9) as the intermediate. Again, by means of kinetics,

PAGE 17

deuterium labeling, product identification and stereochemistry, the prediction was verified. 11 Further evidence for the homoaromatic nature of the 3-bicyclo [3. 1. 0]hexyl cation (6_) has been formulated by determining 12 the activation volumes from the solvolyses of the cisand trans-3-bi cycl o [3 . 1 . 0] hexyl tosylates (1_) and ( 2J . The trans-3-tosyl ate (2) exhibited an activation volume (-17,4 cc/mole) in line with cyclopentyl and cyclohexyl tosylate. The ais-3tosyl ate ( I) , however, had a volume of activation of -13.9 cc/mole. The authors viewed the difference as indicative of a diffuse charge, e.g. a non-classical ion, in the ais-3tosyl ate (_1) solvolysis transition state, which supports the concept of the intermediacy of the tris-homocycl opropenyl cation (6). so.; (20) 21) Broser and Rahn reacted 13 the epimeric alcohols (20) with boron trifluoride in polar solvents to yield deeply colored solutions which were unaffected by oxygen. The introduction of tropilidene gave a 51% yield of tropylium hydroxyf 1 uoroborate . On the basis of direct observation via NMR, IR, and visible spectra, the authors favor the formation of the non-classical ion (21) . Sauers, in an earlier study, 11 * reported the NMR spectrum of lactone (22) in concentrated sulfuric acid. It was Sauers 1 View that the tris-homocycl opropenyl cationic derivative (23) was a stable entity in this medium.

PAGE 18

(22) H 2 SCU 'fM (.23) H 'C0 3 H The weight of kinetic evidence for the existence of the tri s-homocycl opropenyl cation at this point was certainly not overwhelming, the maximum reported 10 rate enhancement being ca . 922. The experimental results fell short of the theoretical predictions and calculations previously published. As a result of early LACO calculations, Winstein stated 3 " "that the 3-bi cycl [3. 1 . 0] hexyl cation prefers the non-classical trishomocycl opropenyl structure to a classical one." Extended Hue ke 1 calculations by Hoffmann predicted 15 that Corey's "almost classical" non-classical ion (9j would be less stable than the classical ion on a planar f i ve-membered ring. Remarkably, calculations did predict a deep minimum in energy of c_a_. 1 eV for the symmetrical tri s-homocycl opropenyl cation (6_) . (24) Hoffman also reported 15 calculations on the hypothetical cation (2_4) , which indicated an unsymmetri cal double minimum resembling the 7-norbornadienyl cation calculations, except the energy well was deeper as bending of the bridge occurred toward the cyclopropyl group as compared to bending toward the double bond. The consensus was that the full potential of the homoconjugative ability of the cyclopropyl group had yet to be

PAGE 19

9 realized because of the geometric shortcomings of the various systems studied. The added M p" character of the cyclopropyl carbon-carbon sigma bond and the general "bent bond" nature of the orbitals 16 appeared to require a rather exacting alignment for efficient overlap with the developing "p" orbital of the ionization center during the initial stages of cation formation . (25) (26) 27) endo anti= endo , syn = exo 3 syn = exo 3 anti= ets-chair trans-chair cts-boat trans-boat (28) It occurred to a number of researchers that if the C2 and C i*. positions of the chair and boat conformations of both the aisand trans3-bicycl [3. 1 . 0] hexanol s were connected by an ethano bridge, the four possible combinations could effectively be frozen out, allowing an efficient -probe of the reactivity of the cyclopropyl group while holding the stereochemistry under virtually complete scrutiny. An added advantage to this series of tri cycl [3 . 2 . 1 . 2 ' h ] octan-8-ols, (25J-0H through (28)-0H, was the fact that a direct comparison could be made to the reactivity of the cyclopropyl group versus the double bond.

PAGE 20

10 Earlier, Winstein et al . x 7 solvolyzed both the synand anii-7-norbornenyl tosylates, (3_0)-0Ts and (3_l)-0Ts, and H^X H^X X-^-H (29) r^i^l (30) r'l ^7 (31) // rel ati ve rate 10' 10 compared their rates to that of the saturated norbornyl tosylate (29_)-0Ts. The large rate enhancement for the antitosylate (3JJ-0Ts has been explained 17 ' 18 by Winstein in terms of a non-classical bi s-homocycl opropeni urn intermediate (22 ), al though others have strongly endorsed 19 the concept of cm cb^c& 32 (33) rapidly equilibrating ions (33 ) . (28) (29) Pincock first reported 2 ° the synthesis of the p-bromobenzenesul fonate derviative of exo-antitri cycl o [3. 2 . 1 . 2 ' 4 ] octan-3-ol , (28)-OBs , and found that it underwent acetolysis at a rate 2.7 times slower than the corresponding 7-norbornyl brosylate (2_9)-0Bs. It was felt that the slight rate retardation was due to the steric interference by the ex
PAGE 21

11 effect of the cyclopropyl group was also considered. The primary observation, that there was no unusual effect of the cyclopropyl group in the exo psoition, was attributed to the fact that the cyclopropyl sigma orbitals are directed down and away from the reaction site at the C 8 position. Pincock anticipated that this would not be the case for the endo i somer (25 ) . X^H rel (29) 1 (31) 10 11 (28) .4 (27) 10* The three laboratories of Pincock, 213 Battiste, 219 and Tanida 21 > c issued the simultaneous report of the synthesis of the endo-antiand endo-syn-8tri cycl o [3 . 2 . 1 . 2 ' 4 ] octanol s (25)-0H and (26)-0H. Solvolysis of (25_)-0PNB produced a rate enhancement factor of ca_. 1 1 * relative to the exo-anti system (28) and the norbornyl derivative (2_9) , a "new record for participation" 225 . With 70% aqueous dioxane as the solvent only two major products were detected, 22 the rearranged endo-4tricyclo[3.3.0.0 2 >\]octanol (34)-0H and its p-ni trobenzoate

PAGE 22

12 ester (34)-0PNB. Tanida also reported 210 the formation of a trace (0.1%) amount of the retained alcohol (25)-0H. The overwhelming formation of rearranged products implies that BNP0_ H >25)-0PNB (35) (34)'OH OH (25) -OH most of the positive charge resides on the C2 , U positions of the intermediate, analogous to the charge distribution in the ion generated from the anti-norbornenyl derivative (31). 23 A very important point is that ion pair return or solvent capture occurs stereospeci f i cally at C 2 and C* from the endodirection only, in sharp contrast to the observed 210 ' 22 reduction of ketone (36) with lithium aluminum hydride which gave exclusive exo attack to yield endo-al cohol (3_4)-0H. Equilibration experiments have reportedly produced a mixture (36) (34)-0H endo^O exo (37) of 38% en* alcohol (34)-0H and 62% exo-alcohol (37). Thus, it would appear that exo-alcohol (37) is preferred both kinetically and thermodynami cal ly . It is worth noting that despite an environment that is 67 mole percent water, the intermediate cation is stable and long-lived enough to permit the internal return of the p-ni trobenzoate anion, i.e.

PAGE 23

13 time enough for the anion to approach the cation from another direction, 22 These facts are uniquely explained by the invocation of the non-classical 2 ,4-ethanotri s-homocyclopropenyl cation (35) as the intermediate. Tanida prepared 210 an optically active sample of the rearranged alcohol (3£)-0H, and acetolyzed its tosylate racemization (34)-OTs OTs (35) (3_4)-OTs. The rate of racemization was 3.2 times faster than the rate of acid formation, strongly indicating the occurance of internal return. Ultimately complete racemization occurred, and the existance of any potential hydride shift was eliminated via deuterium labeling. A comparison (1) (34)-OTs Q Q Pi OTs OTs H OTs relative rate l OTs 13 10 540 (37)0T: OTs 40 OTs OTs

PAGE 24

14 of the solvolysis rate of (34)-OTs with a number of appropri ate compounds 210 demonstrated that the rearranged derivative is still very reactive. Tanida was compelled to state that the tris-homocyclopropenyl cation (35) is the simplest and most economical intermediate to invoke. The rate of acetolysis of endo-syn brosylate (2_6)-OBs is a little over ten times that of 7-norbornyl brosylate and gives a complex product mixture. 21 ' 22 (26) H_ X (38) (27) If? In contrast, exo-syn brosylate (2_7)-OBs produces a relative rate of 10\ The methylene group from the exc cyclopropyl ring could sterically aid the departure of the brosylate anion. 22 , 2 " Another possible factor could be the concerted shift with ionization of the d to C 7 bond to the C 8 position to produce a stabilized cycl opropyl carbi nyl cation ( 3_9 ) . There would not be initial stabilization in the cation (38) produced by the analogous process for (26)-OBs since the orbitals would not be aligned in parallel. 22 (28)

PAGE 25

15 The sluggishness of exo-anti (28) can be attributed to its lack of options, i.e. , it cannot undergo the C ls C 7 to C 8 bond rearrangement to the cycl opropyl carbi nyl cation, and, of course, the rigid geometry prevents the proper orientation of the cyclopropyl orbitals for participation. 22 Coates synthesized 25 pentacycl o [4. 3 . . 2 » "* . 3 » 8 . 5 » 7 ] nonan-9-ol, (4_0)-0H, and found a rate enhancement of 10 I0 -10 12 relative to 7-norbornyl derivatives. Overall strain relief H^ _ X H. (40)-0PNB (41) (40)-UH cannot be a driving force since a homocycl opropyl carbi nyl rearrangement produces a structure identical with the original The only hydrolysis product, in the presence of 2 ,6-1 uti di ne , was determined to be the parent alcohol (40J-0H. These facts, coupled with the results of deuterium labeling, were interpreted in terms of the threefold symmetric tri s-homocycl opropenyl cation intermediate (£1) . Apparently any secondary rearrangement is unable to compete with the attack of water. 25 Coates suggested that the greater reactivity (ca. 80) of the tricyclic p-ni trobenzoate (2^5)-0PNB relative to the pentacyclic p-nitrobenzoate (40_)-0BNB could be attributed to some strain relief in the solvolytic transition state and/or a somewhat less favorable orientation of the anti-cyc} opropyl group in (40)-0PNB due to the bond connecting the two threemembered rings. Nevertheless, Coates argued that the

PAGE 26

16 overwhelming bulk of the driving force for both (25J-0PNB and (40J-0PNB arose from the anchimeric assistance of the anfci-cyclopropyl ring to form the symmetrical tri s-homocycl opropenyl cations (35_) and (41 ) . Ellen and Klumpp 26 acetolyzed the interesting compound, exo-tetracyclo[4.4.0.0 2 '^.O 3 ' 9 ] dec-7-en10-yl tosylate (42_)-OTs, and found exclusive attack at C 8 yielding only exotetracycl o[4.3.1.0 3 ' 6 .0 7 ' 9 ]dec-4-en-2-yl acetate (43)-0Ac. Conversion (45) (46) of (4_3)-0Ac into its tosylate (4_3)-OTs, followed by acetolysis, regenerated (43)-0Ac as the sole product. While the authors felt that the tri s-homocycl opropenyl cation (44) was involved as the intermediate, they saw a high contribution from (45) in which the positive charge is concentrated at C 3 , resulting in less strain than cations having the charge bulk on C2 or Ci 0, The possibility of homoallylic stabilization, (46) , was also recognized. -X (47) (48)

PAGE 27

17 Battiste and Winstsin 27 acetolyzed the tris-methanonaphtha.lene brosylate (47_)-0BS, and found evidence for considerable (c_a. 10 5 10 8 ) cyclopropyl participation, propos ing the initial formation of the non-classical intermediate (48). Product analysis revealed the absence of (4J)-0Ac or any rearranged or unrearranged brosylate derivatives. The authors felt the cyclopropyl group, and therefore its sigma orbitals, are directed toward the cavity between the two bridges, decreasing the initial degree of orbital overlap with the developing positive center and resulting in some moderation of the relative rate. NsOOBs (49) (50) As a further example of the rather precise orientation requirements of the cyclopropyl ring needed for assistance to occur, the solvolysis 28 of 2( trans-3-bi cycl o [3. 1 . ] ) ethyl p-nitrobenzenesul fonate (4_9) gave no kinetic or product evidence whatsoever for participation. Tanida's study 29 of B-(tricyclo[3.2. 1.0 2 ' "] oct-syn-8-yl ) ethyl p-bromobenzenesulfonate (5_0) gave questionable evidence for cyclopropyl participation. The small rate factor of three for (50) relative to its anti-analog, was suggested to be due, at least in part, to repulsive hydrogen interactions in the transition state. The solvolyses of derivatives of exoand endo-antitricyclo[3.1. 1. 2 ' 4 ] heptan-6-ol , (51) and (52), were studied 30

PAGE 28

18 to provide additional insight into the geometrical requirements for participation of the cyclopropyl group. Interestingly, it was determined that (5JJ-0PNB was only fifty-six times faster than (52J-0PNB, although (51J-0PNB yielded the two 'rel 56 rel 1 OH 1 \ H OH rearranged products expected from a tri s-homocycl opropenyl cation intermediate while (52J-0PNB produced an olefinic mixture. It was concluded 3 ob that both compounds were solvolyzing with considerable (though different) neighboring group assistance. The fact that (5JJ-0PNB was solvolyzing at a rate c_a. ten times slower than endo-anti-8tri cycl o [3. 2 . 1 . 2 » *] octanyl p-ni trobenzoate (2_5)-0PNB was attributed through X-ray studies to significant geometrical distortion caused by hydrogen interaction at C 3 and C 7 . H-^-OAc H _0Ac (53)

PAGE 29

19 Coates subjected 31 the exotetracycl o [3 . 3. 0. 3 > 6 . 2 ; » 8 ] oct-4-yl tosylate (5_3) to acetolysis and detected the two expected acetate products. Kinetic studies demonstrated a rate acceleration from anchimeric assistance of ca_. 10 9 . Coates concluded that the methylene bridge at C7 directs the cyclopropyl orbitals away from the site of ionization. Lustgarten reported 32 the acetolysis of the endorearranged tosylate (54J-0TS, giving the endorearranged acetate (54_)-0Ac as the sole product. ' It was his view that (54)-OTs OTs OTs /-,x OAc (54)-OTs (55) (54)-0Ac ionized with assistance from the C1-C2 bond to give trishomocyclopropenyl cation (55_) , the same intermediate ion derived from 51-OTs. A deuterium label at Ci gave an equal distribution of deuterium on C2 and C* after solvolysis, however only half of the original label was accounted for at these two positions, compelling Lustgarten to conclude that the full nature of the intermediate cation had yet to be determi ned . In separate papers, Masamune 33 and Hart 3l+ detailed direct spectral evidence for the exista.nce of the symmetric pyramidal delocalized cations, (56) and (57.) respectively. Both cations accounted for the deuterium label discrepancy previously mentioned, and the higher symmetry was predicted by Stohrer

PAGE 30

20 .C+ P\ (57 (56) and Hoffmann 3S to be relatively stable. /^ .'M rel (62) Sargent probed 36 the ability of a cyclopropyl moiety to function as a remote, nucleophilic neighboring group by interacting with a carbon-carbon double bond which is itself providing a source of electronic stabilization for a developing cation. Solvolysis of the tricyclic di ni trobenzoate (5_8)-0DNB demonstrated a significant rate acceleration ( k , = 622) relative to the allylic di ni trobenzoate (59J-0DNB, which lacks an internal remote nucleophile. The ester ( 58)-0DNB was also faster than the double bond analog (6TJ)-0DNB by a factor of 23.6. Product studies from the hydrolysis of (5_8)-0DNB indicated only two rearranged products, endo-e.} cohol (6JJ-0H and endodini trobenzoate (61)-0DNB. The striking simplicity and

PAGE 31

21 and stereospecificity of the products strongly suggest the intermediacy of an unusually stable cation. Both solvent and dinitrobenzoate anion attack exclusively from the more hindered direction at a position four bonds (ca. 4A) away from the initial site of ionization. The longevity of this cation is further demonstrated by the fact that the weakly nucleophilic dinitrobenzoate anion is able to compete with water in the product forming sequence. As a result of these observations, Sargent favored 36 the intermediacy of the nonclassical cation (62) . From the solvolyses of syn7-p-methoxyphenyl -anti-1norbornenyl p-nitrobenzoate and its saturated analog, Gassman 37 determined that the p-anisyl group was capable of exerting a leveling effect of ca. 3 x 10 1 ° with regard to neighboring group participation. A p-anisyl group was substituted at the OCH OCH ( — } (64) "OH OH (64)-0PNB OPNB sz/n-C 8 position in the endo-anti-8tri cycl o [3. 2 . 1 . 2 » "loctanol system giving (63)-0H. Treatment of (63J-0H with acid yielded the rearranged alcohol (64J-0H as did the hydrolysis of the p-nitrobenzoate derivative (62)-0PNB. The hydrolysis of (63J-0PNB also yielded the internal return product, the rearranged p-nitrobenzoate (64J-0PNB. Gassman had predicted

PAGE 32

22 a rate acceleration by the cyclopropyl group of 3 x 10 3 over and above the leveling of the p-anisyl moiety., and observed a value of 3.8 x 10 3 . It was emphatically noted that even though a cation at C 8 would be tertiary and stabilized by the p-anisyl group, the bulk of the charge resides on C 2 as determined by the formation of rearranged products, i.e. , the cyclopropyl ring controlled product formation. As has been seen up to this point, the generation of the 2,4-ethano-tris-homocyclopropenyl cation (35) has resulted overwhelmingly in product formation at the endo C 2 and C* positions, suggesting, perhaps, that the bulk of positive charge in the intermediate resides at these positions as opposed to the bridge Cs position. 22 Tanida calculated 210 the ground state energy difference between the two alcoholic (35) 3 (32) products of cation (35.) based on a distribution of 99.9% (Cz, U attack) for rearranged alcohol (3_4)-0H and 0.1% (Cs attack) for retained alcohol (25J-0H. The resulting value of 12.1 kcal for AF° indicates a considerable amount of strain relief in the transformation to the rearranged alcohol (34)-0H Pincock 22 pointed out the analogous charge distributions for both the tris-homo(35.) and bis-homo(3_2) cations, and referenced Winstein's description of the 7-norbornenyl cation (32.). The bridge carbon atom of cation (22) has "considerable

PAGE 33

23 tendency to rehybridize from sp 2 toward sp 3 . Such rehybridization increases the C 7 Coulomb integral as well as C 7 -C 2 and C 7 -C 3 orbital overlap. This leads to net stabilization of the bridged ion, and these \jery features of rehybri di zation at C 7 tend to diminish the charge on this atom." 23b QCH 3 As noted earlier, even substitution at C 8 with a vinyl group (formally causing the generation of an allylic cation upon ionization) and an anisyl group resulted in rearranged product upon solvolysis. Substitution of a cationic stabilizing group at the C 8 anti position along with a syn leaving group apparently allows leakage to the non-classical intermediate followed by nucleophilic capture to yield endo rearranged products. Thus, when Gassman treated the synOCH CH3O H + /H 2 (65) (64)-0H alcohol (65_) with acid, 37 only rearranged alcohol (64)-0H was isolated. Baird and Reese 38 reacted the anti-methyl -syn-bromo tricycle (6_6) with Ag + and recovered only the rearranged

PAGE 34

24 Ag + H 2 (67) OH H 2 alcohol (67). Alcohol (67) was also the hydrolysis product of the sz/n-mesyl ate (68) . That there is little positive charge residing on the bridge C 8 position of the tri s-homocycl opropenyl cation (35) is a conclusion which might be supported by the fact that substitution of various electron donating groups at C 8 had no effect on the products, i.e. , only endo rearranged species were observed. This concept is feasible, but it seems more likely that there is a steric bias in the transition state of nucleophil ic capture by cation (35) . Tanida's ground state energy calculations, which favored the rearranged alcohol (34J-0H over the retained alcohol (25J-0H by 12.1 kcal , should reflect the strain energy difference between the two possible transition states for solvent capture. This being the case, the transition state energy barrier difference, paralleling the product alcohols, would be greater than zero and less than 12 kcal. The intensity of positive charge at the bridge C 3 of cation (35) could very well have

PAGE 35

• 25 been increased by the substitution of the electron donating groups previously discussed, but to a degree insufficient to swamp out the steric bias of the transition state for capture of weak nucl eophi 1 es . Considering this point, one could envision stronger electron donating groups at the bridge C 8 position tipping the balance of solvent capture in favor of charge over steric bias. As can be seen below, the methoxyl group appears to be just such an entity and it is consequently used in this investigation as a further probe into the balance between charge versus steric strain relief in the tri s-homocycl opropenyl cation manifold. Part of the synthetic scheme employed to prepare the alcohols (25J-0H and (26_)-0H involves the acidic hydrolysis of the saturated ketal , 8 ,8-dimethoxy-endotri cycl o [3. 2 . 1 . 2 » "*] octane ( 69 ) . The only product reported 2 ! a 2 2 » 3 9 by three different laboratories was ketone (70), endotri cycl o [3. 2 . 1 . 2 » *] H + /H2p (70) octan-8-one. It is generally assumed, 1 * mechanistically, R-0-C-0-R 1 H ROH *R-0-C+ < — > 71 + ' R-0=C that the ease with which ketals suffer acidic hydrolysis is the result of resonance stabilization in the alkoxy carbonium ion intermediate (71).

PAGE 36

26 The formation of only ketone (70_) from ketal (5_9) , or more specifically the absence of rearranged products, leads one' to question the existence of participation involving the cyclopropyl group. There appear to be two extremes as to the nature of the carbonium ion intermediate generated from ( 6_9 ) : 1) essentially a classical oxo-carboni urn ion (72) , in which methoxy resonance stabilization of the positive charge at C 3 swamps out any energy need for cyclopropyl interaction 2) a non-classical methoxy-tri shomocycl opropenyl cation (7_3) i n whi ch most of the positive charge resides on the bridge carbon Ca, despite considerable cyclopropyl involvement in charge stabilization. (72) (73) CH,0^_0CH rel rate (74) CH,0 OCH (75) 2.3 CH 3 0. OCH CH3O---OCH3 CH 3 (K-0CH3 rel rate (76) 320

PAGE 37

27 CH 3 0CH 3 (79) 80 During the course of the research efforts reported in this text, the kinetics of the acidic hydrolyses of the above ketal series were revealed by Lamaty et al . " 1 The rates of formation of the expected ketone products were monitered by means of the carbonyl ultraviolet absorption. The contrast between previous solvolytic studies of the respective alcoholic derivatives (rate acceleration up to 10 11 *) and the more subtle trends (rate acceleration ca_. 10 2 ) observed for the hydrolysis of their ketal precursors is obvious. Apparently, the methoxyl group is capable of an even greater leveling effect than the p-anisyl group. Lamaty explained his results on the basis of the likely sites of protonation. It has been shown 1 * 2 that the alcohols syn with respect to the double bond or the cyclopropyl group exhibit considerable hydrogen bonding, the strongest interaction occurring in the olefin case. As an example of Lamaty's argument, the norbornenyl ketal (75) hydrolyzes only 2.3 times faster than the norbornyl ketal (_7_4)" Protonation of the anti-metboxyl group of (_75) is required for departure with assistance, but due to the propensity for hydrogen bonding, protonation is more likely to occur at the syn-methcxyl moiety to yield the relatively stable hydrogen-bonded intermediate (7_9) . Since hydrogen bonding is weaker for cyclopropane, thereby reducing the selectivity,

PAGE 38

28 anti-methoxyl protonation competes more favorably in the case of saturated ketal (69). , Participation of the cyclopropyl ring then becomes a more significant factor as is reflected in the relative rate of 120. The unsaturated endo-cyclopropyl ketal ( 7_8 ) , if the above mentioned protonation factors were of little consequence, should have hydrolyzed at a rate intermediate to that of ketals (75) and (69), i.e. , between the relative rates of 2.3 and 120. The observed relative rate was 320, which Lamaty felt" 1 was indicative of the fact that protonation of either methoxyl could result in effective participation from either the double bond or the • cyclopropyl group. The rate factor of 18 for ketal (76) was attributed to protonation on the least sterically hindered methoxyl group [anti) and the subsequent tilting away of the synmethoxyl group from the exo-cycl opropyl methylene in the transition state. A relative rate factor of 143 was observed for tne exo , en Lamaty's hydrolysis of ketals (69) and (77), but monitered

PAGE 39

29 the product formations via PMR. Using Lamaty's relative value of 120 for ketal (69), Kessler reported a rate ratio of 120:5 for (69) and (77) respectively (120:143, Lamaty). Kessler politely decided not to speculate on the discrepancy between the two research groups. He does speculate that the exomethylene group hinders protonation on the an *£me thoxyl , and resulting syn-methoxyl protonation does not induce participation by the endo-cycl opropyl moiety. Lamaty has discussed the nature of the ketal hydrolysis intermediate. He found the cyclopentyl ketal (30) to be 5.1 x 10 3 times more reactive than the 7-norbornyl ketal (74), under the conditions of acid hydrolysis. The 7-norbornanone , however, is 1.7 x 10 11 more reactive toward nucleophilic addition of borohydride than cycl opentanone . k 4 Since he had already reported that the transition state for borohydride addition presents an entirely sp 3 profile, 45 Lamaty felt the reaction profile of the acidic hydrolysis of the 7-norbornyl ketal and the related bridge ketals passes from an initial sp 3 state to a transition state very sp 2 in character, i.e ., a carboni um-oxonium ion. There was even speculation that the participation of a double bond or cyclopropyl group would in fact involve interaction with the rr* orbital of the carboni um-oxoni urn ion. 0CH 3 CHaO-^-OCHa (77)

PAGE 40

30 When Kessler subjected 1 * 3 ketal (77) to acidic solvolysis in water or methanol, he found complete conversion to the respective rearranged products, (81)-0H and (8_1)-0CH 3 . Changing the solvent system to di oxane/water or acetone/water produced, in addition of (81)-0H, the ketone (82) . Acidic solvolysis of the methyl derivative (83_) led entirely to rearranged products, (8JJ , the methyl group CHaO-v-OCHa OCH (83) (84) apparently adding some stabilization of positive charge at C 2 . 0CH 3 (86) The formed products provide a means for determining the mode of nucleophilic attack on the intermediate carbonium ions, ( 3_5 ) , (7_3 ) , (8_5) , (8_6 ) , and a summary of these correlations has been made by Kessler. 1 * 3 Path A for cation (35) is not observed because of the interaction of the cyclopropyl sigma orbitals with C 8 , while B (retention) becomes an extremely minor process in comparison to path C (rearrangement). For cation ( 7_3 ) , the methoxyl stabilization of the carbonium ion center allows a "normal" path of hydrolysis to ketone, probably

PAGE 41

31 via path B because of cyclopropyl interaction. In contrast to cation (73), attack by path B for cation (8_5) is hindered by the methylene group of the exo cyclopropyl ring and consequently path C predominates over ketone formation. Kessler did state his belief that path B still dominates over A in the formation of ketone from (85). Cation (86) reacts only with rearrangement due to the stabilizing effect of the methyl group, i.e. , via path C.

PAGE 42

. CHAPTER II Results and Discussion The synthetic scheme employed for the preparation of the precursors and compounds used in this study has bjen reported by seyeral workers. 213 ' 22 ' 39 CH 3 0. ~ o ^ v o 1 1 3 KOH CI CI CH3O OCH3 OCH Cl^Cl CH.3 0H CI CH..30 H 2 Pd/C CH3O (70) (69) Hexachl orocycl opentadiene reacted with methanol i c 1 * 6 potassium hydroxide to give 1 ,2 , 3 ,4tetrachl oro-5 , 5-dimethoxycyclopenta-l,3-diene, (87). The Diels-Alder addition of cyclopropene to (87) yielded the 1 ,5 ,6 ,7tetrachl oro-8,8dimethoxy-endo-tricyclo [3. 2 . 1 . 2 » h ] oct-6-ene (88) .213,22,39 Dechlorination of (8_8) was accomplished via Gassman's procedure 1 * 7 giving the 8,8-dimethoxy-endo-tri cycl [3 . 2. 1 . 2 » *] oct-6-ene (7_8) . 2 * a > 22 » 39 Hydrogenati on of .he unsaturated ketal (7_8) with palladium/charcoal catalyst produced 32

PAGE 43

33 8,8-dimethoxy-endc-tricyclo[3.2. 1. O 2 »*] octane (6_9). 2ia » 22 > 39 Hydrolysis of (6_9) with wet acetic acid 1 * 60 generated endotricyclo[3.2. 1. 2 ' 4 ] octan-8-one (70). According to an analogous scheme, the diethoxy ketals were also prepared. The acidic hydrolysis of the endo-saturated ketal (69) , as described in the introduction, has been reported 213 ' 22 ' 39 as leading exclusively to endotri cycl o [3 . 2 . 1 . 2 » "*] oct-8-one ( 7_0 ) . Obviously a cationic intermediate is involved, and CH 3 ^ OCH3 H + /H 2 (69) (70) one is able to speculate upon at least two extremes as to the nature of this cation: 1) essentially a classical, oxo-carbonium ion (7_2) involving no del ocal i zati on of the cyclopropane ring, with the methoxyl stabilized charge at Cs attacked by nucleophiles only at either face of the bridge or 2) a delocalized non-classical methoxytri s-homocycl opropenyj cation (7_3) in which there is both methoxyl and cyclopropyl stabilization resulting in charge concentration at Cs and nucleophilic attack at the bridge Ca from the anti face stereospeci f i cal ly . CH 3 0+ CH 3 < — (72 (73)

PAGE 44

34 As an initial probe into the nature of the intermediate, both the dimethoxy-saturated and unsaturated ketals, (69) and (_78) respectively, were dissolved in ck-methanol in the presence of catalytic amounts of either silver(I) perchlorate or trifluoroacetic acid. The disappearance of the synand anti-methoxyl signals in the PMR of both ketals was monitored. The syn and anti designations are assigned with respect to the endo cyclopropyl group, and in both ketals, the syn methoxyl singlet is assumed to be downfield from the anti methoxyl singlet, the verification of which will be reported in Chapter III. (69) CD 3 0__ 0CH 3 CD3O ,0CD 3 CH 3 CL^_0CH

PAGE 45

35 The anti-unsaturated methoxyl signal diminished in strength at a rate greater than the sz/n-unsaturated and antisaturated methoxyl signals, which disappeared at ca_. the same rate. The sz/n-saturated methoxyl singlet was the most sluggish, but it too was eventually washed out by the di-methanol . While any interpretation of these observations is subject to potential ambiguities, the fact that the antimethoxyl groups for both ketals had a greater propensity for dit-methanol substitution than their syn counterparts would enhance speculation of at least some involvement of the cyclopropyl group with charge development at Ca. At this point it was considered desirable to generate the intermediate cation by other means, be it (72_) or (73), and study the orientation and stereochemistry of nucleophilic capture. Generation of oxo-carboni urn ions by the reaction of ketals with di chl oroal umi num hydride has been documented in the literature 48 The di chl oroal umi num hydride, which exists in ether as an etherate, complexes with an oxygen of the ketal, which in turn fragments to give an oxo-carboni urn ion. Subsequent attack by hydride or other nucleophiles gives \ M / ^ + A1HC1 2 ^= OR C1 2 HA1 v?R / N 0R + A1HC1 2 v H C— OR + R0A1 HC1 2 *C / N 0R an ether product . k 8C > d Eliel suggested 11 8a using a 4:1 molar ratio of aluminum chloride to lithium aluminum hydride as

PAGE 46

36 the most efficient means of generating di chl oroal umi num hydride, and the ratio is employed in this work. The reaction of 8 ,8-dimethoxy-endotri cycl o [3 . 2 . 1 . 2 ' "*] octane (6_9) with a 4:1 molar ratio of aluminum chloride/lithium aluminum hydride for 1.75 hours produced syn-8-methoxy-endotricyclo [3.2. 1 . 2 ' 4 ] octane (89) in 82.8% yield. The colorless (69 4:1 A 1 CI 3 / L i A 1 H , ether (89 liquid was identified by its PMR spectrum 6[3.82(l,m), 3.34 (3,s), 2.18 (2,m), and 1.67 to 0.67 (8, complex)], mass spectrum (m./e 138), infra-red spectrum, and elemental analysis An alternative synthesis for (89) was accomplished by the reaction of diazomethane with an authentic sample" 9 of endosz/n-tricyclo[3.2. 1.0 2 > 4 ] octan-8-ol , (26J-0H, confirming the assigned structure of (8_9) via spectral comparison. CH 2 N. OCH BF 3 /ether (Z6J-0H (89 Glpc/mass spectral analysis of the crude di chl oroal umi num hydride product mixture resulted is the detection of four minor reaction proucts totaling 5.3% relative to the major product (89_) . The highest recorded m/e values were 136, 128 and 179 respectively for the third, fourth, and fifth eluted components .

PAGE 47

37 The fact that none of these trace products were attributable to the intermediate formation of any anti-8-methoxyendo-tricyclo[3. 2. 1.0 2 » ^octane (90) was demonstrated by the synthesis of (9_0) and its subsequent exposure to the reaction conditions. Reduction of ketone (70) with lithium metal in HO H CH 3 Li NH 3 (1) CH 2 N (70) (25J-QH (90J liquid ammonia 39 gave product which consisted' of 94.5% antialcohol (2_5)-0H. Treatment of (2_5)-0H with distilled diazomethane etherate and boron trifluoride catalyst produced anti-8-methoxy-endo-tricyclo [3. 2. 1. 2 ' "*] octane (90_) which was characterized by its PMR spectrum 6 [3.62 (l,m), 3.23 (3,s), 2.21 (2,m), and 1.80 to 0.28 (8, complex)], absolute mass measurement of [CsHinO]*, infra-red spectrum, and elemental analysis. anti-Ether (9_0) was stirred for 1.75 hours with 4:1 aluminum chloride/lithium aluminum hydride reagent and three products were isolated via preparative glpc. Tricyclo{3.3.0.0 2 8 ]octane (9_1) was collected as a colorless liquid in 28.6% yield. The PMR and infra-red spectra were in accord with an earlier report, tf6C and the absolute mass measurement of [CeHi2]« along with the elemental analysis confirmed the structure.

PAGE 48

38 (90) 4:1 AlCWLiAIH,, ether ._ , „,,, H 0CH 3 (£1) (93) The second product isolated was endo-4-methoxy tri cycl o(91) [3.3.0.0 2 ' 8 ]octane (92) in 33.7% yield. The PMR spectrum 5[3.75 (l,ddd), 3.19 (3,s), and 2.70 to 0.87 (10, complex),], mass spectrum (138 m/e) and absolute mass measurement of [CsHinO]', infra-red spectrum, and elemental analysis all confirmed the assigned structure. The endo configuration A (92) and (93): J AX =J A"X = 9 -° Hz > J mx =6 -° Hz was verified by analysis of PMR couplings for H x which agree with reported data 2 1C » 22 » 3 6 » 3 7 » h 3b for endo analogs. The third product (7.7%) was identified as endo-4chlorotricyclo[3.3.0.0 2 ' 8 ]octane (93) from its PMR spectrum 6 [4. 18 (l,ddd), and 2.86 to 103 (10, complex)], mass spectrum (142 m/e), and the absolute measured mass for [CsHisCl]*. The endo configuration was also verified by the coupling pattern for H . The detection of only syn-ether (8_9) from saturated ketal (6_9) and di chl oroal umi num hydride with no evidence for the formaton of the epimeric anti-ether (90) adds considerable weight to the existence of the non-classical intermediate

PAGE 49

39 (73_) . Following coordination of di chl oroal umi num hydride H Cl 2 \r s Al CH 3 0t^0CH 3 -C1 2 A1 (0CH 3 )H OC'H H OCH (94) with the anti-methoxyl group of (69_) , it could be anticipated that this anti complex would depart with assistance from the cyclopropyl group. The observation that the anti-ether (90) is labile under the reaction conditions while the synether (89) is stable lends credence to this hypothesis. There is, however, no compelling evidence to suggest which methoxyl group in 6_9 is intially lost. In any event ionization to the intermediate cation (represented as non-classical (73_) for the sake of argument) and subsequent attack by one or each of three formal nucl eophi 1 es , i.e. , methoxide, chloride, and hydride, would explain the observed product(s). Attack at C 8 is apparently favored by a concentration of positive charge at this site in either the classical (7_2_) or non-classical (73) models. Methoxide attack would regenerate starting ketal (69), while chloride attack would afford the highly reactive a-chloro-ether (94_) which should quickly regenerate the cationic

PAGE 50

40 intermediate under the reaction conditions. Hydride attack at C 8 accounts for the observed product, syn-ether (89), whose formation exclusive of C 2 attack and any anti-ether (90) demonstrates stereospeci f i c hydride capture from the anti face of Ce. This is of course, indicative of interaction and del oca! izati on of charge at Cs with the cyclopropyl ring and the methoxyl group, but with the methoxyl stabilization; concentrating the charge at Cs to a point at which the steric bias for the transition state of endo rearranged product formation is overcome. The formation of only endo-rearranged products from the reaction of di chl oroal umi num hydride with anti -ether (90) is in line with earlier results for nucl eophi 1 1 i c capture of the sol volyti cal ly generated parent tri s-homocycl opropenyl cation (35) . In the case of (9_0 ) , all three potential nucleophiles are observed in the product analysis, a relatively rare result 48e for di chl oroal umi num hydride reactions. CH,0. H A1C1 2 H (90) (35) At this point it became desirable to investigate the reaction of di chl oroal umi num hydride with 8 , 8-dimethoxyendctri cycl o [3.2. 1 . 2 » 4 ] oct-6-ene {]_§), particularly with respect to any potential anchimeric competition between the cyclopropyl group and the carbon-carbon double bond. Reaction of unsaturated ketal (78) with a 4:1 molar ratio of aluminum

PAGE 51

41 chloride/lithium aluminum hydride gave four products which were isolated via preparative glpc (yields reported are relative glpc peak areas). 4:1 A1C1 3 /LiA1H^ ether H A (96) OCH ., , H H A' m OCH* (98) fl The major product (76.5%) isolated was sz/n-8-methoxyendo-tricyclo [3.2. 1 . 2 ' " ] oct-6-ene (95.) which was identified from its PMR spectrum 6[0 . 5 3 . 9 4 (2, complex), 1.26-1.57 (2, complex), 2.66-2.88 (2,m), 3.25 (3,s), 3.50 (l,m) and 5.61 C2,t)], the mass spectrum (136 m/e), infra-red spectrum : and elemental analysis. Hydrogenati on of (95) with 10% Pd/charcoal catalyst produced the previously characterized H OCH 10% Pd/C saturated sz/n-ether (89) to further verify the structure. endc-6-Methoxytricyclo [3. 3.0.0 2 ' 8 ]oct-3-ene (96.) was isolated in 3.5% relative yield and was characterized by its PMR 6 [5.65 (2,m), 3.95 (l.ddd), 3.24 (3,s), 2.79 (l,m), and 2.48 to 0.58 (5, complex)] and mass (136 m/e) spectra. The endo configuration was verified by the PMR couplings 2 ^ , 22 , ^s , 3 7 t H 3b for Hx . j Ax = 9 .25; J A . X = 7.75; and J MX =5.0.

PAGE 52

42 endo-6-Chlorotricyclo [3 . 3. . O 2 • 8 ]oct-3-ene (97) , collected in 1.0% yield, received the assigned structure on the base's of the literature authenticated 50 PMR spectrum 6[5.70 (2,^1), 4.28 (1, oct), 3.24 (l,m), 2.57 to 1.34 (5, complex)], mass spectrum (140 m/e), and spectral comparison with a sample derived for an alternative synthetic route to be reported later in the text. The endo configuration was confirmed 210 ' 22 ' 36 ' 37 ' 4 ^ by the PMR couplings of H x : J AX = 10.0; J A i X = 8.0; and J MX 5.0 Hz. The last component was identified as l-methoxy-endo-6chlorotricyclo[3.3.0.0 2 ' 8 ]oct-3-ene (98) and was collected in 18.9% yield. The elemental analysis, infra-red spectrum, mass spectrum (170 m/e), and PMR spectrum 6 [5 . 73 (2,t), 4.37 (l,m), 3.39 (l,m), 3.29 (3,s), and 2.60 to 1.27 (4, complex)] agreed with the structural assignemnt. The PMR coupling for H x confirmed 213 ' 22 ' 36 ' 37 « it3b the endo configuration: J AX = 10.0; J A ' X = 7.75; and J MX = 5.25 Hz. Further structural proof is given later. At first glance, the product bulk appears explanable on the basis of the intermediacy of the 8-methoxytri s-homocyclopropenyl cation (9_9) , however the detection of the minor products (9_6) and (97_) requires the invocation of at least two or more intermediates.. CI— *— 0CH 3 (99) (98)

PAGE 53

43 The formation of unsaturated rearranged chloride (98) would appear to be the result of chloride attack upon the non-classical cationic intermediate (99) , which would be generated by the complexation of dichl oroal umi num hydride with the anti-methoxyl group of ketal (78_) with subsequent departure of the anti complex via the anchimeric assistance of the cyclopropyl group. The chloride anion would then have path B (C 8 , anti-face) or path C (C 2 , endo) as possible routes to cation collapse. Path C, of course, would lead directly to product (98) but it would be anticipated that C 8 attack (path B) would predominate in line with the observed course of reduction (hydride attack). If formed, the resulting unstable bridge a-chl oro-ether (J_0_0) could then suffer rearrangement to the thermodynami cal ly more stable, rearranged chloride (9_8) via a tight ion pair i ntermedi ate . 2 lb > 2 2 If the intermediate cation was essentially classical, i.e. (71), one would expect a mixture of both the syn (path A) and anti (path B) a-chloro-ether at Cs, both of which should be very reactive. As already mentioned, generation of a classical cation via syn ionization followed by leakage into the nonclassical manifold has been demonstrated to occur when cationic stabilizing groups are located on C„. 37 ' 38 H: Path A OCH H Path B (95)

PAGE 54

44 Hydride attack upon (99.) at the Ca position accounts for the major product, syn-unsaturated ether (9_5). The non-classical description of (9_9) would require the stereospecific approach from the anti face (path B). Approach from the syn face (path A) by hydride would generate the anti-8-methoxy-enio-tricyclo[3.2. 1.0 2,1 *]oct-6-ene, (101), whose transient existance, as will be seen, is highly probable. There are, however, potential alternative routes to the formation of (KH) which will shortly be discussed. Hydride attack via path C was not observed which is entirely consistent with hydrolytic results for ketal (78), and makes the observations of rearranged chloride (98) all the more in teres ti ng . Methoxide quenching of cation (99_) at Ca would regenerate (path B likely) the starting material, unsaturated dimethyl ketal (7_8). Under the experimental conditions, this process would not be detected. (102) (96) OCH (97) CI The characterization of the two non-methoxy 1 ated endo "rearranged minor products (96.) and (97) argues strongly for the intermediacy of the unsubsti tuted 2 ,4-ethenotri shomocyclopropenyl cation (102.) which would suffer nucleophilic attack via path C to give the observed products (96)

PAGE 55

.45 and (97_) . There appear to be a number of viable mechanistic routes to (102) . (95; (103) (102) One mechanistic pathway questions the stabilityof the major product, sz/n-unsaturated ether (95J , under the reaction conditions. Coordination of di chl oroal umi num hydride as a Lewis acid with the syn-methoxy 1 group of (9_5) could result in cleavage of the ether with anchimeric assistance by the internal carbon-carbon double bond generating the unsubs ti tuted bis-homocyclopropenyl cation, (103). The previously mentioned theoretical calculations of Hoffmann 15 predict a lower ground state energy for the tri s-homocycl opropeny 1 cation (102) relative to theL.\ bishomocyclopropenyl cation (103), suggesting the possibility of i nterconversi on or equilibration, favoring (102), between the two ions, with subsequent nucleophile capture by cation ( 102 ) to give the observed products. It has been reported, 51 however, that the hydrolysis of the p-ni trobenzoate ester of the syn unsaturated alcohol, (104)-0PNB, yields only the retained sj/nalcohol (104) , presumably via the bis-homocyclopropenyl intermediate (103). Apparently ( 103 ) and (102) do not interconvert under the reactions conditions because bridge flipping is not competitive with solvent capture, 51

PAGE 56

46 (104)-0PNB (103) ^ The syn unsaturated ether (95) itself was subjected to the original reaction conditions and was exposed to the dichl oroal umi num hydride reagent for a three hour period. Capillary glpc analysis revealed that 93% of the syn ether (95) had not reacted, with the detection of only 6% of an unknown hydrocarbon and 1% of an unknown whose retention time was too great to be either (96_) or (9_7_) . As a consequence of the above facts, the generation of the tri s-homocycl opropenyl cation ( 102 ) from syn-ether (95_) via the bishomocycl opropenyl cation ( 103 ) is effectively eliminated as a mechanistic pathway. One alternative approach to the formation of cation (102) could involve the intermediacy of 8-methoxy-bi s-homocycl opropenyl cation ( 1£5 ) . The generation of cation (10J_) might be accomplished by equilibration between ( 105 ) and the 8-methoxytris-homocyclopropenyl cation (99_) , since the energy barrier to bridge flipping should be lowered 52 relative to cations ( 102 ) and ( 103 ) as a result of the methoxyl stabilization of positive charge at Cs in ( 99_) . Of at least equal probability is the di chl oroal umi num hydride promoted cleavage of the synmethoxyl group in unsaturated ketal (78), with concomitant

PAGE 57

47 anchimerical assistance by the double bond to give cation ( 105 ) directly. The viability of the bridged ion (105) as an intermediate is somewhat speculative considering reports 53 that substitution of a p-anisyl group at the bridge (C 7 ) carbon of norbornene cancels out (levels) stabilization provided by the double bond upon 'the cation generated at the bridge. Direct spectral observation 52 of the 7-methoxynorbornenyl and norbornadi enyl cation, however, has provided some evidence that derealization involving one or both double bonds, respectively, exists for these ions. An 8-methoxy-bis-homocyclopropenyl cation ( 105 ) would be expected to suffer hydride attack from the nnti face (with respect to the double bond) at C 8 to generate the anti8-methoxy-endo-tricyclo[3.2. 1. 2 » * ]oct-6-ene (101). Ether ( 101 ) could also have been generated by hydride attack on the tris-homo cation (99) from the syn face of C 8 , allowing for weakened interaction of the cyclopropyl group and incorporation of considerable classical character. By analogy, it has been shown that from strictly steric point of view, anti attack is already favored over syn attack without any potential complications of del ocal i zati on . Borohydride reduction 39 of the unsaturated ketone (106.) produced a 3:1 predominance of sz/n-alcohol (104) over anti-alcohol (107 J ! •Whatever the origin of (KH) , coordination of its anticipatorily labile anti-methoxyl group with di chl oroal umi num hydride, followed by the cyclopropyl assisted ionization of the complex, would generate the unsubs ti tuted non-classical

PAGE 58

48 cation ( 102 ) , with subsequent product formation. 0CH 3 :H (99) In order to test the later part of this hypothesis, it was deemed advantageous to synthesize the unsaturated antiether ( HH ) , subject this ether to the original reaction conditions, and analize the resultant product mixture for the presence of (96_) and (97) . The synthesis of endo, antitri cycl o[3 . 2 . 1 . 2 ».*]oct-6en-8-ol, ( 107 ) was accomplished via the method of Clark, Frayne, and Johnson. 39 The unsaturated ketal (78) was subH^^OH jected to acidic hydrolysis at -5° to yield a concentrated solution of the thermally unstable (decarbonyl ati on ) endotricyclo[3.2.1.0 2 »-]oct-6-en-8-one (106). Ketone (106) was never isolated but was reduced in solution by sodium borohydride to yield ca_^ 75:25 mixture of the syn-al cohol ( 104 ) and anti-alcohol ( 107 ) . Isolation via preparative glpc of anti-alcohol { 107 ) , whose spectral dataware in agreement

PAGE 59

49 with the literature, 39 was followed by methylation by diazomethane with boron trifluoride catalyst to yield the desired CH 2 N; anti-8-raethoxy-endo-tn"cycl 13.2, 1 . O 2 ' "] oct-6-ene (101). Structural identification of (101) was established by its PMR spectrum 6 [5.67 (2,m), 3.62 (l,m), 3.21 (3,s), 2.93 (2,m), 1.14 (2,m), .33 (2,m)], mass spectrum, and the measured mass for [CsHi20]tExposure of the anti-unsaturated ether (10J_) to the original reaction conditions was followed by capillary glpc analysis of the resultant product mixture. Both the endorearranged methoxyand chl oro-octenes , (96) and (97_) , were established via authentic compound comparison to be present CH 3 0^H 4:1 A I C 1 , / L i A 1 H , (101) -^ iU \ +• // (96) 0CH 3 (97) CI in a relative rati of Cja. 3.2. : 1 compared to the original ratio of c_a. 3.5:1 obtained from the ketal reduction. The two products accounted for the bulk (61.1%) of the reaction mixture with no unreacted ether (roi) detected. The results above strongly endorse the viability of anti-unsaturated ether (101) as an initial reduction product in the reaction of dichloro-

PAGE 60

50 aluminum hydride reagent with the unsaturated dimethyl ketal (78). A summary mechanistic scheme for the reduction of unsaturated ketal (78) is presented below. (102) [ ^ CI (96)

PAGE 61

51 If the mechanistic rational presented for the formation of the endo-unsaturated rearranged chloro-ether (98) is correct, one should expect that an increase in the molar ratio of aluminum chloride/lithium aluminum hydride should lead to increased yields of chloro-ether product, since the availability of the chloride anion is increased while the hydride molar equivalents are decreased. An 8:1 molar ratio was employed in reactions with both the saturated and unsaturated ketals (69_) and (78_) . The small amount of hydride present effectively serves to destroy any moisture or proton acid build-up. OCH; OCH 8:1 AlCWLiAlFU (98) ci Use of the 8:1 reagent with the unsaturated ketal (_78) gave a 71.8% yield of l-methoxy-endc-6-chl orotri cyclo[3 . 3 . 0. 2 ' 8 ]oct-3-ene (98) as a dark yellow oil which, despite the color, was glpc pure. The color was removed upon vacuum distillation. 8:1 A1C1 3/LIAIH1 (69) OCH (108) Treatment of the saturated ketal (6_9) with the 8:1 A1C1 3/LiAHU reagent gave an 81.1% yield of 88.5% glpc pure 1-methoxy endo-tricyclo[3.3.0.0 2 ' 8 ]octane ( 1£8) . Three minor components

PAGE 62

52 (7.7. 1.7, and 2.1%) of slightly shorter retention time were not identified. The structure of the methoxy saturated chloride (108) was verified by its PMR spectrum 6[4 . 40 (l,ddd), 3.35 (3,s), and 3.02 to 1.13 (9, complex)], mass spectrum (172 m/e), measured mass for [C 9 H 13 0Cl]t, and infra-red spectrum. The endo configuration was confirmed by the coupling patted ic , 22 , 3 6 , 3 7 ,. 3b for Hx . j A)( = g>75 . Ja()( = 8 _ 75 . and J MX = 6 5 OCH (98; fi (109) Na, THF t-BuOH OCH H-N=N-H OCH Na, THF t-BuOH (108) (110) As a further verification of the structure of the unsaturated and unsaturated chlorides (108) and (98), both compounds were dechl ori nated using Gassman's procedure. 47 Treatment of unsaturated chloride (98) with sodium metal in tetrahydrof uran/ terr-butanol gave a 44.2% yield of a pale yellow oil which was •identified as 1 -methoxy tri cycl o [3. 3 . 0. 2 » 8 loct-3-ene (109) . Identification was accomplished from the PMR spectrum 6 [5 . 53 (2,m), 3.33 (3,s), 3.13 (l,m), and 2.40 to 1.22 (6, complex)],

PAGE 63

53 mass spectrum (136 m/e), measured mass for [C 9 H 12 0]t, infrared spectrum, and elemental analysis. In a similar manner, the sodium metal dechlorination of saturated chloride (108) produced a 49.6% yield of a colorless liquid identified as 1-methoxy tri cycl o [3 . 3 . . 2 » 8 ] octane ( 110 ) The structural assignment was based on the PMR spectrum
PAGE 64

54 OCH (111) (111) (113) CH 3 CH 2 (L^0CH3 CHsO^^OCHzCH (98) cl 60% aq. EtOH : >• ^H 104°,K 2 C0 3 19.2:10. The first examined was the unsaturated rearranged methoxy chloride (98) which was solvolyzed in 60% aqueous ethanol in the presence of potassium carbonate at 104° for ten hours. Capillary glpc analysis displayed only two components whose respective yields were determined by internal standards. Both fractions were isolated via preparative glpc. The first eluted component (53.6%) was determined to be cycl oheptatriene ( 111 ) by spectral comparison with an authentic sample. The second fraction consisted of mixed alkoxy ketals (21.5%) and was determined by PMR to be made up primarily (89.2%) of anti-S-ethoxy-syn-8-methoxy-endo-tri cycl o [3.2. 1.0 2j 4 ] oct-6-ene (112), The PMR spectrum 6 [5 . 72 (2,t), 3.38 (2,q), 3.27 (3,s), 2.86 (2,m), ca. 1.24 (2,m), 1.13 (3,t), and .47 (2,m)], mass spectrum (180 m/e), and absolute mass measurement for [Ci iHi 6 2 ] • all agreed with the designated structure. The assignment of syn for the methoxy! group and anti for the ethoxy group is based on the fact that the cyclopropyl group's field effect causes a greater downfield shift than the double bond, a trend that will be summerized later with regard to all the ketals and methyl ethers encountered in this text. A singlet appearing at 63.13 in the PMR of the mixed ketal fraction was tentatively attributed the anti -methoxy! hydrogens

PAGE 65

55 of : an ti -8-methoxysyn8ethoxy-endotricyclo [3 . 2 . 1 . 2 ' 4 ]oct6-ene ( 113 ) , and, through PMR integration, was determined to be 10.8% of the mixed ketal product. It is most noteworthy that the syn and anti methoxyl signals of both mixed ketals are identical to the chemical shift values of the respective syn and anti methoxyl signals for the dimethoxy unsaturated ketal (78). The need to firm up the PMR assignments for the two mixed ketals ( 112 ) and ( 113 ) was immediately recognized, the complete characterization of ketal ( 113 ) being of prime importance . The ethoxy analog of the methoxy rearranged unsaturated chloride (89_) , i.e. , l-ethoxy-endo-6-chl orotri cycl o[3 . 3 . . 2 • 8 ] oct-3-ene ( 114 ) was synthesized by the identical procedure for (98j , and subsequently subjected to solvolysis in 70% aqueous methanol at 104° for 24 hours. 0CH 2 CH ( 112 ) ( 113 ) CH 3 CH 2 0_0CH 3 CH 3 O_0CH 2 CH 3 70% aq. CH 3 0H fT ' H 104°, K 2 C0 3 13.1:86.9 The mixed ketal fraction isolated was found by PMR to consist for the most part (86.9%) of anti-methoxy-synethoxy unsaturated ketal ( 1J_3) . In addition to the PMR spectrum 6[5. 70 (2,t), 3.53 (3,q), 3.13 (3,5), 2.87 (2,m), ca. 1.26 (2,m), 1.22 (3,t) (3,g), and 0.48 (2,m)], the mass spectrum (180 m/e), and the absolute measured mass for

PAGE 66

56 C c n H i6°2]' were in agreement with the structural assignment. A singlet at 63.27 was attributed to the syn methoxyl PMR absorption of arcti-ethoxy-swn-methoxy unsaturated ketal ( 112 ) , and ( 112 ) was calculated through integration to comprise 13.1% of the mixed ketal fraction. The solvolysis products are uniquely explained by cyclopropyl anchimeric assistance in the ionization of the rearranged unsaturated chloride (98) to the non-classical tris-homocyclopropenyl cation (99). If cation (9_9) should then be OCH :98) // OCH 0CH 3 (99) I (105) CHsCHzO^-OCHaCHsO, 0CH 2 CH ( 112 ) (113) rapidly i nterconverted with the bi s-homocycl opropenyl cation ( 105 ) , the major ketal product, anti-ethoxy-sz/n-methoxy unsaturated ketal ( 112 ) , would be a result of ethanol attack from the anti face of C 8 in (9_9 ) , (path B), and formation of the minor ketal product, antt-methoxy-sz/n-ethoxy unsaturated ketal ( 113 ) could be attributed to ethanol approach at Cs of bis-homo cation ( 105 ) via path B, i.e., from the anti face

PAGE 67

57 with respect to the delocalized double bond. Of at least equal probability would be attack of cation (9_9) by ethanol along path A to give ( 113 ) , assuming a weakly delocalized system or a symmetrically bridged intermediate ion. Leakage of (99) to a classical oxocarbonium ion, followed by solvent attack from both syn and anti faces, cannot be ruled out, but appears to be less consistent with all the results. Attack by water at C 8 of the i ntermedi ate (s ) ion would generate the major product, unsaturated ketone ( 106 ) . Under the experimental conditions, ketone ( 106 ) is known 2 lb » 3 ^ > 5 k to decarbonyl ate giving cycl oheptatri ene ( 111 ) . It is quite significant to note that, despite the indication that cation (99) was generated from the rearranged precursor (9_8) , no rearranged products (path C) were observed. This point can only add emphasis to the concentration of positive charge at C 8 in the delocalized cation (99). The behavior of the saturated rearranged chloride ( 108 ) under solvolytic conditions was subsequently studied. Chloride ( 108 ) was heated for 19 hours at 100° in 60% aqueous ethanol 0CH: (115) ( 116 ) CH 3 CH 2 0-^0CH 3 CH30»_-0CH 2 CH3 ?, 60% aq. EtOH 100°, K 2 C0 3 (108) CI 96.0:4.0 in the presence of potassium carbonate. Capillary glpc analysis with an internal standard revealed two fractions The largest fraction was determined to be endo-tri cyclo-

PAGE 68

58 [3.2. 1.0 2,l, ]octan-8-one (70_) , in a calculated yield of 68.1%. Ketone (7_0) was identified by spectral and glpc comparison with a known sample. The ketal fraction (19.2%) was determined by PMR to consist mostly (96.0%) of ant£-8-ethoxy-st/n-8-methoxy-en
PAGE 69

59 The product analysis of the two fractions isolated revealed the largest fraction (71.3%) to be tricyclic ketone (70). The great bulk (96.3%) of the ketal fraction was determined to be anti-8-methoxy-sun-8-ethoxy-endotri cycl o[3.2.1.0 2 ' "]octane (H_6) as demonstrated by the PMR spectrum 5[3.59 (2,q), 3.22 (3,s), 2.12 (2,m), 1.84 to .51 (8, complex), and 1.24 (3,t)], mass spectrum (182 m/e), and absolute measured mass for [CnHieOjt. A singlet at 63.33 was assigned to the sz/n-methoxyl PMR signal of anti-ethoxy-syn-methoxy saturated ketal ( 115 ) , which was determined via integration to be 3.7% of the ketal fraction. The anchimerically assisted ionization of saturated rearranged chloride (108) is supported by the fact that only C 8 nucleophilic attack (paths A or B) on cation (73_) was observed from the product analysis. Going from the rearranged OCHs -CI EtOH CH 3 CH 2 CL_0CH 3 , CH 3 J H2O 0CH 2 CH (115) (116) {70)

PAGE 70

60 structure of ( 108 ) to the unrearranged products to the exclusion of path C adds to the concept of positive charge concentration at C 8 , and the highly stereoselective attack (96%) of solvent on cation (7J_) from the anti face (path B) of C 8 underlines the delocalized nature of the intermediate. The detection of 4% anH-methoxy-sz/n-ethoxy saturated ketal ( 116 ) leads to the rather obvious question of syn attack (path A) on what is conceived of as being a delocalized cation ( 7_3) • It could be anticipated that since the methoxyl group competes so favorably with the cyclopropyl ring with regard to charge stabilization, any potential rehybri di zati on of C 8 from sp 2 to sp 3 would effectively be eliminated, resulting in a protruding, electron-deficient "p" orbital, occasionally accomplishing nucleophilic capture from the syn face of the bridge, despite residual cyclopropyl del oca! i zati on . At the same time, the product distribution hardly argues for the intermediacy of the classical cation (72_) since there is simply no explanation at this point for the overwhelming attack of nucleophiles via path B for (72). As (72) can be seen from Table I, the reduction of tricyclic ketone (70) with numerous reagents shows no overwhelming steric preference for hydride attack via paths A or B, and certainly not to the point of 96% anti-attack.

PAGE 71

61 Table I. Hydride Reduction Product Distribution for Ketone (70) "~ H-rr-OH H0>r-H (70) LiAlH 4 55 NaBH^ 6 c L1Al(0t-Bu)H* 6 c Al(i-PrO) /i-Pr0H 46C Li (secBu) 3 BH a PMHS/DBATO a ' b This work Polymethylhydrogen si 1 oxane/tetrabutyl diacetoxy ti n oxide di mer Yet another means of generating the non-classical methoxy tris-homocycl openyl cation (7_3) was found in the reaction of saturated rearranged chloride ( 108 ) with silver perchlorate in methanol /acetone . The tricyclic ketone (70) OCHa n C0 2 CH 3 Ag + H CH 3 0H ( 108 ) CI — ( 118 ) was isolated (30.7%) and identified via spectral comparison with an authentic sample. The second component (34.7%) was identified as methyl 4-cycl oheptene1-carboxyl ate ( 118 ) from its PMR spectrum 6 [5.75 (2,t), 366 (3,s), and 2.82 to 1.21

PAGE 72

62 (9, complex)], mass spectrum (154 m/e), and infra-red spectrum, all of which agreed with the published spectral data. 56 Ester ( 118 ) was also synthesized from 4-cycl oheptene-1-carboxyl i c acid ( 119 ) and di azomethane , followed by spectral comparison and authentication. CO2H CH2N2 C0,CH (119) 118) OCH -AqCl (108) (118) Both ketone (7_0) and the cycloheptene ester ( 118 ) can be viewed as products derived from the hemi-ketal ( 120 ) . Attack by water upon the charge rich C 8 position of the trishomocyclopropenyl cation (_7_3) {anti approach) would yield the hemi-ketal (120). Hemi-ketal ( 120 ) then has the option of losing methanol to form ketone ( 7_0 ) , or, with the assistance of either Ag + or proton cleavage of the cyclopropyl ring, rearrange 57 in the manner shown to give the cycloheptene ester (118).

PAGE 73

63 Mention was made earlier of an alternative synthesis of endo-6-chlorotricyclo[3.3.0.0 2 ' 8 ]oct-3-ene (97). Breslow has reported 58 the generation of 5-halo substituted cyclopentadienes, and in particular, 5-chlorocyclopentadiene(121), by the reaction of cycl opentadi enyl thai 1 i urn with n-chlorosuccinimide. Cyclopropene was then bubbled through a solution of ( 121 ), and glpc analysis revealed the presence of three minor products and one major product (total yield c_a. 25%). The major product (83.7% of product mixture), rearranged chloride (£7), was identified by its spectral data and by 1 i terature 5 ° comparisons already discussed. o NCS H CI (121) A (97)* ^ l ( 123 ) CI ( 122 ) Two of three minor products were identified. Syn-8chloro-endo-tricyclo[3.2. 1.0 2 *]oct-6-ene ( 22), in 8.0% relative yield, was identified from it's PMR spectrum 6[5,81 (2,t), 4.04 (l,m), 2.82. (2, m), 1.62 (2,m), 0.78 (2,m)], and glpc/mass spectrum (140 m/e). 1-Chl oro-endo tricyclo[3.2.1.0 2,lt ]oct-6-ene (123), in 6.1% relative yield, was also identified from the PMR spectrum 6 [5 . 80 (2,d), 2.77 (l,m), 2.38 (l,q), ca. 2.04 (l,m), 1.63 (2,m), c_a. 0.68 (2, complex)], and glpc/mass spectrum .( 140 m/e).

PAGE 74

64 Cyclopentadienes substituted in the C s position are known 59 to tautomeri ze readily to the more stable vinyl substituted cyclopentadienes; however, Bres 1 ow reported 583 that his method involving cycl opentadi enyl thai 1 i urn reduces this problem. In fact, only one minor product, i.e. (123), could be attributed to tautomeri zati on . Syn-chl ori de ( 122 ) appears to result simply from the Diels-Alder addition of cyclopropene to (12_1) on the chlorine side of the cyclopentadienyl plane. CI H 6 A -*• //. (121) (124) (102) (97) CI The rearranged chloride (97) appears to be the result of cyclopropene addition to (,121.) giving the transient antichloro-endo-tricyclo[3.2. 1.0 2 ' * ] oct-6-ene ( 124 ) which, through cyclopropyl assistance and possibly thallium (I) cationic coordination catalysis, would rearrange (probably through a tight ion pair) via the tri s-homocycl opropenyl cation ( 102 ) to give the major product (97) . In summary, the results presented strongly indicate that the ionic intermediates generated by the reaction of dichloroaluminum hydride with, and the acidic hydrolysis of, the saturated and unsaturated ketals (69) and (78), are the same as those generated by the solvolysis of the saturated and unsaturated rearranged chloro ethers (108) and (98), i.e. ,

PAGE 75

65 the 3-methoxy-2,4-ethanoand 3-methoxy _2 ,4etheno-trishomocyclopropenyl cations (7_3) and (99_) respectively. Product studies have also raised the possibility that the unsaturated tri s-homocyclopropenyl cation (9_9) is in equilibrium via a bridge flipping process with the bi s-homocycl opropenyl ( 105 ) . The electron structure of these ions has been radically altered with respect to the parent ions in that positive charge del oca! i zati on is not as extensive with most of the charge localized or concentrated at the methoxyl bridge carbon. The cyclopropyl del ocal i zati on may even have been weakened to the point where nucleophiles are able to penetrate and attack the bridge from the syn face of the bridge. This electronic effect, not steric bias, is thus the overwhelming factor in determining the orientation and stereochemistry of solvent or nucleophilic attack. CH,0^OCH (69) (69) aq. ROH/100' or Ag + /CH 3 0H 73) (108)

PAGE 76

66 CH 3 Q>^0CH OCH + OCH H + /H 2 // (78) (99) aq. ROH 100° ' (105) OCH

PAGE 77

CHAPTER III PMR Studies Sun and Anti Chemical Shifts of Alkbxyl"! Groups at C 8 in the Tricyclo|_3.2. l.Q z ' H ]octany 1 -octeny 1 Syste•s" It is advantageous at this point to summerize the PMR chemical shifts of methoxy and ethoxy groups substituted at the C 8 position of the tri cycl o[3. 2. 1 . 2 "joctane and tricyclo[3.2. 1. 2 » l *]oct-6-ene systems. Much of the work already presented depends upon the correct syn or anti assignment at Ce of these alkoxy groups. Authors seem to have been hesitant to declare in the literature their assignments of the methoxy! chemical shifts of the saturated and unsaturated ketals, (69) and (78). Pincock has stated 22 that the difference in chemical shift caused by the field effect of the endo cyclopropyl ring in saturated ketal (69) upon the syn methoxyl relative to the anti methoxyl group is 0.11 ppm (observed here as 0.10 ppm). He compared 22 this value with the shift induced by the double bond upon the syn methoxyl in norbornene dimethyl ketal (7_5) , j e 0-07 ppm, relative to the anti-methoxyl group, and concludes that, within the specified geometry of these systems, the cyclopropyl group relative to the double bond,, has the greater effective ability to deshield. This fact is demonstrated in Table II if one examines the methoxyl shifts for the syn and anti unsaturated ethers (9_5) and ( 101 ) . The 67

PAGE 78

68 Table II. Syn and Anti Cs AT koxyl Chemi cal Shifts in Tricyc1o[3.2. 1.0 2 ' "]octeny1 Systems Santi 3.13 6 syn 3.27 hsyn-anti , ppm .14 (95) fj 3.21 3.27 06 CH,0^0CH,CH 113) 3.38 (-CH 2 ) 3. 13 3.27 3.53 (-CH 2 ) > 14 (-OCH3) 15 (-0CH 2 CH 3 ) sz/n-methoxyl of (95_) is deshi el ded by the cyclopropyl group by .06 ppm more than the double bond deshielding of the antimethoxyl of ( 101 ) . The observed syn and anti-methoxyl signals for the mixed ketals {U2_) and (l_13j are identical to the syn

PAGE 79

69 and anti-methoxyl signals of the parent compound, unsaturated dimethoxy ketal (78), confirming the epimeric assignments of (Hi.) and ( 113 ) . An inspection of Table III listing the chemical shifts of the saturated analogs also confirms the epimeric assignments for the saturated ketals, (69) , ( 115 ) and (116). Table III. Syn and Anti C 8 Alkoxyl Chemical Shifts in Tricyclo [3.2, 1.0 2j "joctanyl Systems CH 3 (69) OCH 6anti 3.21 Ssyn 3.31 tssyn-anti , ppm 10 H ^_0CH 3 (89) (90) 3.27 3.33 > 06 CH 3 CH 2 0^ r _0CH 3 (115) CH 3 0^_0CH 2 CH 3 (116) 3.47 (-CHJ 3.22 (-CH 2 ) 3.33 > 3.59 (-CH 2 ) 11 (-OCH3) 12 (-0CH 2 CH 3 )

PAGE 80

70 Computer Analyzed PMR Spectra In an effort to add to the relatively scarce PMR data concerning the cyclopropyl group in the tri cycl o [3 . 2 . 1 . 2 ' * ] octyl systems, the four polychl ori nated endo cyclopropene Diels-Alder adducts, (88), (125), (126), and (122) wer e prepared. Thoroughly degassed carbon tetrachloride solutions of these compounds were sealed under vacuum in NMR tubes. The individual spectra were expanded to 50 Hz and the absorptions calibrated via the TMS side-band technique. The observed absorption data was applied to a modified LA0C00N III 6 ° computer program, and coupling signs were based on convention. (88) H 3 H 2 The preparation of 1 ,5 ,6 ,7-tetrachl oro-8-dimethoxyendo-tricyclo[3.2.1.0 2 '^]oct-6-ene (88) has already been described in this text. The proton chemical shifts and standard deviations were reported for a 1M solution (CC1<*) in Hz as follows: Hi, 105. 317±0. 015 ; Ha, -53. 263±0. 017 ; Hs, -24.715±0.016. Coupling constants were: J i , 2 = 7 . 073±0. 020; J 1)3 =3.469±0.018; J 2>3 = -7 . 434±0 . 023 . CI -~-Cl !!«> c\ H 3 H 2 The addition of cyclopropene to hexachl orocycl opentadiene gave 1 ,5 ,6 ,7 ,8, 8hexachl ovoendotri cycl o [3.2.1.0 2 ' 1| ]oct-6-ene

PAGE 81

71 (125_) in 92.6% yield. The PMR and infra-red spectra were in agreement 61 with the assigned structure, as was the elemental analysis. The computer analyzed proton chemical shifts for a 1M solution (CC1 J were reported in Hz as: H x , 120 . 777±0. 012 Hz; H 2 , -77.408±0.013; H 3 , -46 . 877±0. 012 . The coupling constants were revealed as: J 1 , 2 = 7 . 175 + 0. 014 ; J 1 , 3 = 3. 493±0. 014 ; J 2 »3= -7. 707±0. 017. (126) CI H 3 H : 1,5,6, 7-Tetrachloro-8-di chl oromethyl ene-endo-tri cycl o[3.2. 1.0 25i *]oct-6-ene ( 126 ) was prepared in 80.4% yield from the addition of cyclopropene to hexachl orof ul vene . 6 2 The PMR and infra-red spectra agreed with the assigned structure, as did the mass spectrum (322 m/e) and elemental analysis. The computer analyzed PMR spectrum gave the chemical shifts of a 0.51 M solution (CC1 J in Hz as: H : , 121 . 390±0 . 009 ; H 2 , -67. 772 + 0.010; H 3 , -39 . 458±0 . 012 . The coupling constants were determined to be: J x , 2 = 7. 025 + . 011 ; J x , 3 = 3 . 090±. 011 ; J 2 »3= -7.804±.014. (127) H CI "X 1 H 5 H^

PAGE 82

72 The addition of cyclopropene to 1 ,2 ,3 ,4tetrachl orocyclopentadiene 1 * 9 gave 1 , 5 ,6 ,7tetrachl oro-endotri cycl o[3.2.1.0 2 '"]oct-6-ene ( 127 ) in 77.7% yield. The PMR and infFa-red spectra, along with the elemental analysis, confirmed the structure. 61 The computer analyzed PMR spectrum of a 1M solution (CCU) gave the proton chemical shifts in Hz as follows: Hi, -185.844±0.020; H 2 , -149.271+0.017; H 3 , -119.841±0.018; H^, -56.641 + 0.018; H 5 , 34 . 716±0 . 018 . The coupling constants were reported as: J x , 2 = -6.707+0.026; J lJ3 = -0.246±0.026; Jj , s =0. 298±0. 028 ; J 2 , 3 = -0.290+0.020; J 2 ,„ = 2.528±0.025; J 3 ,,= 7.218±0.021; J 3 , 5 = 3.170±0.020; J k , 5 = -7.503±0.024. The noteworthy (2.5 Hz) extended "w" 6 3 coupling of H 2 and H^ protons, and the small (0.30 Hz) coupling of ti x and H„ has potential use in confirming syn and anti assignments involving mono substitution at the bridge C 8 position.

PAGE 83

73 o o

PAGE 84

74

PAGE 85

75

PAGE 86

76

PAGE 87

77

PAGE 88

78 1 *r J

PAGE 89

79 ^§ X o

PAGE 90

80 o

PAGE 91

8) -5 :M ' L£_8_S J „ is o o >> o I o 1

PAGE 92

82

PAGE 93

83

PAGE 94

84 n

PAGE 95

85

PAGE 96

CHAPTER IV Experimental General Melting points were determined on a Thomas-Hoover unimelt capillary melting point apparatus and are uncorrected. Elemental analyses were performed by Atlantic Microlab, Inc., Atlanta, Georgia. Vapor Phase Chromatography The analytical gas/liquid phase chromatography was performed with an Aerograph Hy-Fi 600-D instrument with a flame ionization detector, using column A, which consisted of 15% FFAP on Chrom W, AW, DMCS, 1/8" x 5', with a helium flow rate of 300 cc/min. Column B consisted of a 100' x 1/100" DEGS capillary column fitted on a Varian Aerograph Series 1400 flame ionization instrument. For preparative work, a Varian Aerograph model 90-D with a thermal conductivity detector was employed using one of two columns: column C which consisted of 15% FFAP on Chrom W, AW, DMCS, 4.5' x 1/4", helium flow c_a. 140 cc/min.; column D which was packed with 20% DEGS on 45/60 Chrom W, 10' x 1/4", helium flow c_a. 100 cc/min. Spectra Infrared spectra were determined on either a Beckman IR-10 or a Perkin-Elmer 137 Sodium Chloride Spectrometer. Proton 86

PAGE 97

87 magnetic resonance spectra were recorded on Varian A-60A instrument. Low resolution mass spectra were provided by a Perkin-Hi tachi RMU-6E instrument. High resolution spectra were determined by a AEI-MS-30 instrument which was also used in conjuction with a Pye Unicam Series 104 Chromatograph utilizing a 5' x 1/4" SE-30 column. Generation of Cyclopropene A slightly modified procedure of the original method reported by Closs and Krantz 64 was employed. A typical run is as follows: To a magnetically stirred suspension of 50 grams of fresh sodium amide in 70 ml of dry mineral oil at 85-90° were added dropwise (1 drop/4 sec.) approximately 75 ml of 3-chloropropene. A slow stream of dry nitrogen gas transported the evolved cyclopropene through a Friedrich condenser (water cooled), followed by a 4N sulfuric acid trap (200 ml) and then onto the reaction vessel via a scintered glass outlet. Preparation of 1 , 2 ,3 ,4-Tetrachl oro-5 , 5-dimethoxycycl openta 1 , 3-di ene (87) To a mechanically stirred solution of 205 grams of hexachlorocyclopentadiene (0.751 mole, Aldrich) in 300 ml absolute methanol was added a solution of 112 grams of potassium hydroxide (2.0 moles) in 300 ml absolute methanol. The addition was carried out at a rate commensurate with maintaining gentle reflux, and was followed with continued stirring at room temperature for ca. ten hours. The reaction mixture was filtered to remove precipitated potassium chloride, and

PAGE 98

solvent was removed under water aspirator pressure after drying with anhydrous magnesium sulfate. Vacuum distillation (b.p. 63.065.5°/0.05 mm) yielded 115 grams (0.436 mole, 58.0%) of pale ' yellow oil. The spectral data were in agreement with those previously reported. 1 * 6 Preparation of 1 ,2 , 3 ,4-Tetrachl oro-5 ,5-di ethoxycycl op enta1 ,3-di ene ; The synthetic procedure followed is identical to the preparation of the dimethoxy analog (8_7) already described. Using 205 grams (0.751 mole) of hexachl orocycl opentadi ene , 112 grams (2.0 moles) of potassium hydroxide, and 600 ml total absolute ethanol , 107.4 grams (0.389 mole, 51.8%) of the diethoxy ketal was isolated after vacuum distillation (b.p. 103105°/2.6 mm). Spectral data were in accord with literature val ues . 6 5 Preparation of 1 ,5 , 6 , 7-Tetra chl oro-8,8-di emtboxv-endn-Jri rvrl n [3.2.1.0^]oct-6-ene~TW Cyclopropene was passed through a rapidly stirred solution of 40.0 grams (0.152 mole) of 1 ,2 ,3 ,4tetrachl oro-5 , 5dimethoxycyclopenta-1 ,2-diene (87) in c_a. 250 ml of petroleum ether (20-40°) at ambient temperature . The reaction was followed by PMR monitoring of the respective methoxyl methyl signals and was found to be complete after ten hours. The solvent was removed by passing a stream of nitrogen over the solution, yielding 34 grams (0.112 mole, 73.7%) of white crystals (m.p. 63-66°, lit. m.p. 68-70°). Spectral data were in agreement with literature values, 213 ' 22 ' 39 and a computer analyzed PMR spectrum is presented in Chapter III.

PAGE 99

89 Preparation of 1 , 5 ,6 ,7-Tetrachl oro-8 ,8-di ethoxv-endotricyclo[3.2. 1.0^ -]oct-6-en"e The same procedure applied above for the preparation of the dimethoxy adduct (88) was employed for the diethoxy analog. From 105 grams (0.380 mole) of 1 ,2 , 3 ,4tetrachl oro-5 ,5-di ethoxycyclopenta-l,3-diene, 98.6 grams (0.297 mole, 78.1%) of the cyclopropene adduct was isolated after vacuum distillation as a yellow liquid (b.p. 136145°/ 1 . 25 mm). The analytical sample was isolated from preparative glpc column C, at 130°. The PMR spectrum (CDC1 3 ) consisted of two overlapping two proton quartets at 53.95 and 3.83, a two proton quartet at 1.78, two overlapping three proton triplets at 1.26 and 1.14, a one proton sextet at 0.90, and a one proton quintet at 0.40. The infra-red (neat) gave absorption bands at 2990 (m), 289C (w), 1597 (m), 1275 (m) , 1160 (s), and 1025 (m).cm -1 . Anal. Calcd. for Cj 2 H llf Cl ,, : C, 43.40; H, 4.25; CI, 42.71. Found: C, 43.43; H, 4.28; CI, 42.65. Preparation of 8 ,8-Dimethoxy-endo-tri c ycl o [3.2.1.0 2,1 *]oct6-ene ( 7~8~) ~ ' ~~ The dechlorination of ketal (88) was accomplished via Gassman's procedure. 1 * 7 Into a 11 flask fitted with a mechanical stirrer, nitrogen inlet, and water-cooled Friedrich condenser, were added 25.0 grams (0.0822 mole) tetrachl oroketal (88), 125 ml tert-butanol , 325 ml tetrahydrof uran , and chopped sodium (39.0 grams, 1.70 g-atoms). The solution was stirred and heated to gentle reflux for ten hours, cooled, and filtered through wire gauze to remove unreacted sodium. The filtered solution was added to 500 ml of water, followed by the addition of

PAGE 100

90 250 ml of brine and 250 ml of ether. The organic layer was separated and the aqueous layer was extracted (6 x 100 ml) with ether. The combined ether fractions were dried over anhydrous magnesium sulfate, and the bulk of the solvent was distilled utilizing a steam bath and a 2 1 Vigreaux column. Traces of solvent were removed on a rotary evacuator. Vacuum distillation (30-35°/4mm) yielded 7.62 grams (0.0458 mole, 56.0%) of light yellow product, whose spectral analysis agreed with that previously reported. 219 ' 22 ' 39 Preparation of 8, 8Pi ethoxy-en
PAGE 101

91 ppm. The infra-red (neat) gave absorption bands at 3060 (m), 2970 (s), 2925 (m) , 2870 (m), 1565 (w), 1260 (s), and 1100 (s) cm" 1 Anal . Calcd for C 12 H 18 2 : C, 74.19; H, 9.34. Found: C, 74.26; H, 9.36. Preparation of 8^8-Djm e t ho xy -e ndotri cyclo [ 3.2 1.0 2 "Moctane (69j ~" " — A stirred suspension consisting of 3.0 grams (17.8 mmole) of unsaturated dimethoxy ketal (78), 0.268 gram of 10% Pd/C catalyst, and 150 ml of absolute metahnol was subjected to a partial positive pressure of hydrogen gas. After about 3.5 hours, 460 ml of hydrogen were consumed (440 ml theoretical), and the catalyst was remove'd by filtration. The solvent was removed at room temperature under water aspirator pressure. Vacuum distillation (b.p. 62-69°/12 mm) produced 1.76 grams (10.5 mmole, 58.8%) of saturated ketal (6_9) as a colorless oil whose spectral data agreed with earlier reports. 213 ' 22 ' 39 Preparation of 8,8-Diethoxy-ewd.o -tricyclo [ 3. 2 . 1 . 2j * ] octane The method of catalytic hydrogenati on of the dimethoxy ketal (69_) was employed to prepare the diethoxy ketal. The reaction mixture consisted of 1.079 grams (5.55 mmole) of unsaturated diethoxy ketal, 0.100 gram of 10% Pd/C, and 25 mi of absolute ethanol . Hydrogen absorption amounted to 140 ml as compared to 125 ml calculated for the theoretical. The saturated product was isolated as a colorless oil via preparative glpc column D at 130° (10 min), giving 0.549 gram (2.80 mmole, 50.5%) of product. The PMR spectrum (CDC1 3 ) consisted of two overlapping two proton quartets at 63.61 and 3.49, a

PAGE 102

92 two proton multiplet at 2.13, and a fourteen proton complex from 1.82 to 0.52, which includes two overlapping three proton triplets at 1.23 and 1.17. The mass spectrum (70 eV) gave peaks at m/e (rel. intensity) 129 (100), 101 (47.0), and 73 (68.0). The calculated mass for [C12H20O2]* is 196.1462, while accurate mass measurement gave 196.1460, for an error of 1.53 ppm. Anal . Calcd. for C 12 H 20 2 : C, 73.43; H, 10.27. Found: C, 73.50; H, 10.30. Preparation of endo-Tri c.ycl o [3 . 2 . 1 . 2 » h ] octan-8-one (70) A solution of 42.60 grams (0.253 mole) of saturated dimethoxy ketal (69), 4.7 ml (.261 moles) water, and 110 ml glacial acetic acid was heated at 69° with stirring for 24 hours. The reaction mixture was cooled with an ice-water bath, and a chilled solution of 72.5 grams sodium hydroxide in 440 ml water was slowly added to affect neutralization of the acetic acid. Extraction with ether (5 x 130 ml), drying with anhydrous magnesium sulfate, removal of solvent, and vacuum distillation (b.p. 65-70°/0.2 mm ) yielded 29.4 grams (0.241 mole, 92.2%) of ketone (70) as a slushy semi-solid with spectral properties in agreement with previously reported data. 213 ' 22 ' 39 Preparation of syn-8-Methoxy-endotri cycl o [3 . 2 . 1 . 2 "joct ane H9 from 8,8-Dimetnoxy-ewdo-tricyc I o [3 . 2 . 1.0 Z '^octane (69) Reaction procedure followed was that reported by Eliel. 1 * 83 Into a dry 300 ml flask was placed 6.66 grams (0.05 mole) of anhydrous aluminum chloride, and, with drying tube in place,

PAGE 103

93 the flask was cooled with an external ice-water bath. Chilled anhydrous diethyl ether (50 ml) was cautiously added with stirring, and was followed thirty minutes later by the addition of 0.450 gram (0.0125 mole) of lithium aluminum hydride dissolved in 20 ml ether. After another thirty minutes, 4.21 grams (0.025 mole) of saturated methoxy ketal (69), dissolved in 50 ml of ether, was added dropwise to the suspension. Upon completion of the last addition, the icewater bath was removed and the reaction mixture was stirred for 1.75 hours at room temperature. The reaction was again cooled, and subsequently quenched with 50 ml of 10% sulfuric acid. After filtration, the ether and aqueous layer were separated, and the aqueous layer was extracted with ether (3 x 50 ml). The combined ether fractions were dried over anhydrous potassium carbonate and magnesium sulfate, followed by solvent removal on rotary evacuator, to yield 3.02 grams of 94.7% glpc pure (0.0207 mole, 82.2%) szyn-methyl ether (89). The analytical sample was collected via preparative glpc on column D at 136° (6 min.). The PMR spectrum (CDC1 3 ) revealed a one proton multiplet at 63.82, a three proton singlet at 3.34 a broadened two proton multiplet at 2.18, and an eight proton complex from 1.67 to 0.68. The mass spectrum (70 eV) gave peaks at m/e (rel. intensity) 138.(9.0), 106 (20.5), 91 (23.0), 84 (77.5), and 71 (100). The infra-red (neat) exhibited absorption bands at 3072 (m), 3035 (s), 2960 (s), 2880 (s), 2825 (s), 1480 (s), 1370 (s), 1210 (s), 1135 (s), 1105 (s), and 1008 (s) cm -1 .

PAGE 104

94 Anal . Calcd. for C 9 H llt 0: C, 78.21; H, 10.21. Found: C, 78.16; H, 10.22. Glpc/mass spectral analysis of the above reaction mixture (5' SE-30, 100°) revealed, in order of product elution from the column, the following data: (a) 0.8%. (b) 94.7%; synether (89), data above. (c) 0.9%; m/e (rel. intensity) 136 (51.0), 121 (100), 91 (86.5), 79 (6.5), 78 (12.5), 77 (19.5). (d) 1.8%; 128 (5.0), 110 (22.2), 95 (27.8), 85 (33.7), 69 (43.0), 78 (58.5), 77 (47.0), 57 (100), 56 (25.0), and 54 (61.0). (e) 1.7%; 179 (43.0), 139 (82.0), 107 (61.0), and 79 (100.0). Reaction of 8 , 8-Dimethoxy -erdotri cyclo[3.2.1.0 2,lt ]oct-6-ene (78) wilh 4:1 Molar Ratio Aluminum Chloride/Lithium Aluminum Hydri de" The reaction procedure was developed by Eliel. 1 * 83 Into a dry 300 ml flask was charged 6.66 grams (0.05 mole) of anhydrous aluminum chloride. After cooling with an external ice-water bath, 50 ml of chilled, anhydrous diethyl ether was slowly added, followed in 20 minutes by 0.450 gram (0.0119 mole) of lithium aluminum hydride in 30 ml ether. After a further 20 minute stirring period, 4.32 grams (0.0260 mole) of unsaturated dimethyl ketal (_7_8) , dissolved in 50 ml of ether, was slowly added and the reaction mixture subsequently stirred at room temperature for 1.75 hours. The reaction mixture was again cooled and quenched with 50 ml of 10% sulfuric acid. The ether and aqueous layers were separated, the aqueous layer extracted with ether (3 x 50 ml), and the combined ether layers dried over anhydrous potassium carbonate

PAGE 105

95 and magnesium sulfate. After solvent removal, 3.96 grams of a pale-yellow oil was recovered and subjected to preparative glpc on column D at 124°. Yield precentages are based on relative peak areas of the four components detected. 1) syn-8-Methoxy-endo-tricyclo [3.2.1.0 2,lt ]oct-6-ene (95), 76.5%; 6.5 minutes. The PMR spectrum (CDC1 3 ) consisted of a two proton multiplet from 60.53-0.94, a two proton multiplet from 1.26-1.57, a two proton multiplet from 2.662.88, a three proton singlet at 3.27, a one proton multiplet at 3.50, and a two proton triplet at 5.61. The mass spectrum (70 eV) showed peaks at m/e (rel. intensity) 136 (10.3), 135 (17.9), 122 (38.0), 104 (100), 91 (100), 77 (92.3) and 44 (100). The infra-red (neat) showed absorption bands at 3060 (s), 2990 (s), 1570 (w), 1440 (m) , 1220 (s), and 1100 (s) cm" 1 Anaj_. Calcd. C 9 H 12 0: C, 79.37; H, 8.88. Found: C, 79.37; H, 8.92. 2) ercdo-6-Methoxytricyclo[3.3.0.0 2 ' 8 ] oct-3-ene (96), 3.50%, 80 minutes. The PMR spectrum (CDC1 3 ) showed a two proton multiplet centered at 65.65, a one proton multiplet at 3.95, a three proton singlet at 3.24, a one proton broad multiplet at 2.79, and a complex five proton multiplet from. 2.48 to 0.58. The mass spectrum (70 eV) showed peaks at m/e (rel. intensity) 136 (23.0), 121 (28.5), 105 (100), 91 (55.1), 78 (74.5), and 58 (71.2). 3) endo-6-Chlorotricyclo [3. 3.0.0 2 ' 8 ]oct-3-ene (97), 1.04%, 110.0 minutes. The PMR spectrum (CDC1 3 ), which agreed with an earlier literature report, 50 consisted of a two proton

PAGE 106

96 multiplet at 65.70, a one proton octet at 4.28, a one proton multiplet at 3.24, and a complex five proton. mul ti pi et from 2.57 to 1.34. The mass spectrum (70 eV) showed peaks at m/e (rel. intensity) 140 (6.5), 105 (100), 79 (27), 78 (35), and 77 (28). The spectral data were also identical to that of (97) prepared via an alterante route reported later. 4) l-Methoxy-en