Heteroatom Effects on 2, 8-Annulated Semibullvalene Equilibria

INFORMATION TO USERS

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

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

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

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

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

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

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

University Microfilms International 300 North Zeeb Road Ann Arbor, Michigan 48106 USA St. John's Road, Tyler's Green High Wycombe, Bucks, England HP10 8HR BURSON, Richard Leslie, 1949- HETEROATOM EFFECTS ON 2,8-ANNULATED SEMIBULLVALENE EQUILIBRIA. SYNTHESIS AND STUDIES OF NEUTRAL, POTENTIALLY HOMOAROMATIC MOLECULES.

The Ohio State University, Ph.D., 1976 Chemistry, organic

Xerox University Microfilms,Ann Arbor, Michigan 48106 HETEROATOM EFFECTS ON 2 , 8-ANNULATED SEMIBULLVALENE EQUILIBRIA.

SYNTHESIS AND STUDIES OF NEUTRAL, POTENTIALLY HOMOAROMATIC MOLECULES.

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Richard Leslie Burson, B. S.

The Ohio State University

1976

Reading Committee: Approved by

Dr. Leo A. Paquette Dr. P h ilip D. Magnus Dir. Robert J. Ouellette

Adviser Department of Chemistry To Kathy and Heather ACKNOWLEDGMENTS

I would like to extend my sincere thanks to the numerous individuals who have assisted me in so many ways during my graduate career. In particular, I thank Professor Leo A. Paquette for his guidance. I am also deeply indebted to my colleagues in this problem area. Their close collaboration and stimulating influences made the completion of the work p o s s ib le .

I gratefully acknowledge financial support from The Ohio State

University in the form of a Proctor and Gamble Fellowship, from the

Chemistry Department in the form of a Teaching Associate position, and from Professor Paquette in the form of Research Associate positions.

Lastly, the most profound influence upon this work and my life has been my family. It is truly impossible to detail the contributions that ray wife has made in helping me to reach this point. Without her constant support none of th is would have been possible.

i i i VITA

J u l y 26, 1 9 k 9...... Born - W ellington, Kansas

1 9 7 1 ...... B. S., Summa Cum Laude, Washburn University of Topeka, Topeka, Kansas

1971-1972 ...... Graduate Teaching Associate, Department of Chemistry, The Ohio State University, Columbus, Ohio

1972-197 4 ...... Research Associate, Department of Chemistry, The Ohio State University, Columbus, Ohio

1974-1975 ...... Proctor and Gamble Company Fellow, The Ohio State University, Columbus, Ohio

1975-1976 ...... Research Associate, Department of Chemistry, The Ohio State University, Columbus, Ohio

PUBLICATION

" Probins of the Interrelationship between Heteroatomic Substitution and Equilibrium Imbalance in 2,8-Trimethylenesemibullvalene D erivatives,"!. A. Paquette, R. K. Russell, and R. L. Burson, J. Amer. Chem. Soc. , 2Z> 6124 (1975)*

FIELDS OF STUDY

Major Field: Organic Chemistry TABLE OF CONTENTS

Page

DEDICATION...... i i

ACKNOWLEDGEMENTS...... i i i

VITA ...... iv

LIST OF FIGURES ...... v i

LIST OF ILLUSTRATIONS...... v i i i

LIST OF TABLES...... ix

INTRODUCTION...... ■...... 1

RESULTS AND DISCUSSION

PART I. Synthesis of Annulated Sem ibullvalenes ...... 2k

PART II. Analysis of the Equilibrium Imbalances in Selected Annulated Semibullvalene ...... 69

PART III. Synthesis and Evaluation of Potentially Homoaromatic M olecules ...... 106

EXPERIMENTAL...... ikb

APPENDIX...... *...... 208

REFERENCES...... 218

v LIST OF FIGURES

Figure Page

1 The 60 MHz 1H NMR spectrum of sulfide 110 in CDCI3 at 500 Hz sweep w idth ...... 68

2 The 60 MHz 1H NMR spectrum of sulfoxide 112 in CDCI3 at 5°° Hz sweep w idth ...... 68

3 The 60 MHz 1H NMR spectrum of sulfoxide 115 in CDCI3 at 500 Hz sweep w idth ...... 68

!)• The 100 MHz 1H NMR spectrum of 2,8-trimethylenesemibullvalene 32. in CDC13 at 500 Hz sweep width . . . • 102

5 The 100 MHz 1H NMR spectrum of oxasemibullvalene 6l in CS2 at 500 Hz sweep w id th ...... 102

6 The 100 MHz 1H NMR spectrum of thiasemibullvalene 63, in CS2 at 500 Hz sweep w idth ...... 102

7 The 100 MHz 1H NMR spectrum of azasemibullvalene 62 in CDCI3 at 1000 Hz sweep w id th ...... 103

8 The 100 MHz 1H NMR spectrum of azasemibullvalene 95 in CDCI3 at 1000 Hz sweep w idth ...... 103

9 The 100 MHz 1H NMR spectrum of oxasemibullvalene in CS2 at 500 Hz sweep w id th ...... lOlt-

10 The 100 MHz 1H NMR spectrum of thiasemibullvalene §6 in CS2 at 500 Hz sweep w idth ...... lOlt-

11 The 100 MHz 1H NMR spectrum of 2,8-pentamethylene semibullvalene 5J+ in CDC13 at 500 Hz sweep width . . 105

12 The 100 MHz 1H NMR spectrum of sulfoxide §8 in CD2CI2 at 500 Hz sweep w id th ...... 105

13 The 100 MHz ^H NMR spectrum of sulfone £9 in CDC13 a t 500 Hz sweep w i d t h ...... 105

lit- The 100 MHz 1H NMR spectrum of sulfoxide 97 in CD2CI2 at 500 Hz sweep w id th ...... 105

v i Figure Page

15 The 100 MHz 1H NMR spectrum of pentaene in CDCI3 a t 500 Hz sweep w i d t h ...... lh j

16 The 100 MHz 1H NMR spectrum of pentaene 168 in CDCI3 a t 500 Hz sweep w i d t h ...... , ...... 1^3

17 The 100 MHz ^H NMR spectrum of pentaene 166 i n CDCI3 a t 500 Hz sweep w i d t h ...... 1^3

v i i LIST OF ILLUSTRATIONS

Scheme Page

I Synthetic Approach to 2,8-Trimethylene- semibullvalenes jjl, 6l, and 6 3 ...... 26

II Synthetic Approach to Intermediate 8l ...... 28

III Transformations of Intermediate 8l ...... 33

IV Synthetic Approach to 2,8-Trimethylene - semibullvalene 6 2 ...... 37

V . Synthetic Approach to 2,8-Pentamethylene- semi bull valenes %k and 9 j 5 ...... Ml-

VI Synthetic Approach to 2,8-Pentamethylene- semibullvalene s $6 and 9J9 ...... 51

VII Synthetic Approach to 2,8-Pentamethylene- semibullvalene s 9J. an<^ 9 8 ...... 53

VIII Chemical Epimerization of Sulfoxide 1 1 2 ...... 55

IX Stereoselectivity in Photooxygenation of 68a and 87a ...... 59

X Fhotooxygenation of Sulfides 110, llj, and 122 6l

XI Proposed Mechanism of Photooxygenation of Sulfides . . 6k

XII Synthetic Approach to Pentaene jjjj. and Attempted Transformations of Intermediate 1 0 8 ...... 109

XIII Attempted Synthetic Approaches to Dialdehyde l^jl • • • 115

XIV Synthetic Approach to Pentaene 166 and Attempted Synthetic Approaches to Pentaenes 1£1, L52 j 164, and 165, ...... 119

XV Attempted Synthetic Approach to Pentaene l6l ...... 121

XVI S y n th e tic Approach to Pentaene 1 6 8 ...... 12^

XVII Attempted Synthetic Approaches to Pentaene 176 130

v i i i

t LIST OF TABLES

Table Page

I Summary o f A c tiv a tio n P aram eters f o r S e le c te d Homotropilidenes ...... 10

II Assignment of 1H and l3C NMR Shifts of Semibullvalene ( S i ) ...... ^ III Equilibrium Distribution of Annelated Semibullvalenes 5i-5iL...... 18 IV Bridging Effects in Systems Capable of Valence Isomerization ...... TO

V 1H NMR Data (60 MHz, CDC13), Computed Equilibrium Constants ('Kgq), and Gibbs Free Energy Values (AG°) f o r J51, 6l , 62, and 63. (IfO0 ) ...... 78

VI Summary of CMR Data (22.6 MHz, CDC13, ambient temperature) for 6l, 62, and 63...... 79

VII Equilibrium Distributions for 6l, 62, and 63. Calculated by ^-H and 13C NMR ...... 79

VIII Selected Molecular Parameters ...... 82

IX Variable Temperature 1H NMR Data, Computed Equilibrium Constants (Keq) and Gibbs Free Energy Values (AG°) for the Fluxional System 9j+b 52 209

X Variable Temperature ^-H NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9j+b 25 s ...... 210

XI Variable Temperature *11 NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for th e F lu x io n a l System 25)3 ~ 2 5 a ...... 211

XII Variable Temperature XH NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 255 ~ 555...... 212

X III Variable Temperature XH NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for th e F lu x io n a l System 96b ~ 96a ...... 213

ix T able Page

XIV Variable Temperature "hi NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 96b ~ ^6a ...... 2lk

XV Variable Temperature 1H NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for- the Fluxional System gjb 9 Ja ...... 215

XVI Variable Temperature 1H NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 98b ^ 9 8 a ...... 216

XVII Variable Temperature 1H NMR Data, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9£b ^ 9 § a ...... 217

XVIII 1H NMR Data (100 MHz, CDC13 ), Confuted Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for £2, 6l, 62, and 6g. (^0° ) ...... 90

XIX Summary of Thermodynamic Data for Semibullvalenes 5j+> Sib 21, 98 > art 29...... 91

XX Summary of 13C NMR Data (22.6 Hz, CDC13, ambient temperature) for Semibullvalenes £4, 2it> 25.’ 2)2.’ 97, 98 and 9 2 ...... 92

XXI Comparison of Equilibrium Distributions Calculated by XH and •13 C NMR (CDC13 a t l K > ° ) ...... 93

XXII Adjusted 13C NMR Shifts of Carbons cv, 2, and 8 ...... 9k

XXIII Summary of 1H NMR Data—Chemical Shifts of 166 and 1 6 8 ...... 136

XXIV Summary of XH NMR Data—Coupling Constants for 166, and 168 ...... 137

XXV Summary of 13C NMR Data (22.6 MHz, CDC13, ambient temperature) for i)2)2’ an<^ ^-38

x INTRODUCTION

As the scientist seeks to increase his understanding of the

vagaries of the universe, he must classify the data which have "been

accumulated so that the scientific method may he applied. The chemist,

in particular, often seeks to categorize information: ions as clas

sical or non-classical, bonds as ionic or covalent, and reactions as homolytic or heterolytic. Nature, however, does not recognize these arbitrary dichotomies and w illingly provides an abundance of exceptions

to the rule. Ironically, it is often these anomalies which prove to be the most interesting and challenging to study.

An exception to the homolytic-heterolytic distinction would be a unimolecular reaction which is thermally induced but is essentially

insensitive to the influences of free radical initiators or inhibitors,

solvent, and acid or base catalysis. The reaction should also be

general over a wide variety of structural variations. The Cope rearrangement is an example of such a " no mechanism " reaction.

In the course of demonstrating the generality of their synthesis

of (dialkylvinyl)alkylcyanoacetic esters (.l),1 Cope and coworkers

NaOEt KX CH3CH2C(CH3 ) =C( CN)C02E t ------> — > CH2CH=C(CH3 )C(CN)C02E t R 1

attempted to prepare the ally l derivative (l E = -CH2CH=CH2). Surprisingly, the product was thermally labile, rearranging in k h o u rs at I5O-I600 or in 20 minutes at 260° to an isomer (2) with a boiling point elevated by 10°. The analogy to the previously known Claisen rearrangement, e. g., - 4,2 was clear to Cope.3

CH-CH^"" ^C(CN)2C0 Et CH-CH^^CtCNlCOjEt

h^ ; ch; 3 ^

analogous to

h 2c ^ c ^ c h 2 h K h ''C"'

It was soon evident that, although this reaction is facilitated by the presence of " negative groups " such as cyano and carboalkoxy on one or both of the central methylene groups,3^"c the rearrangement is a general reaction. For example, even the very simple 3-methylhexa-

1,5-diene (3.) undergoes rearrangement to 6, albeit incompletely at 3

CH. 3

5 6

Doering and coworkers elegantly provided mechanistic insight into the Cope rearrangement with a series of studies. Confirmation that the basic structural requirement for this " no mechanism " reaction is a

1,5-hexadiene moiety (ja) was made with the study of the 1,1-dideuterio labeled analogue (jb). The reaction, monitored by infrared spectroscopy, was found to proceed with a AH* of 33*5 ±0.5 kcal/mole and a As* of

-15.8 ± 1 eu, neglecting deuterium isotope effects. The large negative entropy value was in agreement with the highly ordered

7 7 ' a , R = H b , R = D transition state previously demonstrated. In an analysis of the rearrangement of meso- and raeemic-5,4-dimethylhexa-l,5-dienes (8),5

Doering and Roth established that the reaction occurs via a four center quasi-chair transition state which is some 5-7 kcal/mole more favorable than a six-center path. k

trans, trans CH-

+ Four-center

CH.

c i s , c is CH. S ix -c e n te r Racemic 8 CH.

CH.

c i s , t r a n s

The structural and theoretical subtleties offered by such a reaction are truly boundless and the chemical literature is replete 6 *7 with examples of what is now termed a [3j3j sigmatropic rearrangement.

Obvious subjects for investigation are those factors which govern the ease or difficulty with which such reactions occur. One driving force is electronic, as partially detailed by Cope's early work. A second and ofben overriding influence is ring strain.

When the 1,5-hexadiene unit is incorporated into a cis-1,2-divinyl- cycloalkane (9)>8 the major determinant governing the position of the equilibrium is the relief of Baeyer ring strain. At one extreme is cis-1,2-divinylcyclopentane (ll) which is stable at 250°.9 At even higher temperatures, a cis-trans equilibrium (11 t± 1^_) is established in preference to a Cope rearrangement. The lack of strain in 11, as well

13 11 1 2 as the presence of unfavorable transannular hydrogen interactions in 12, account for the observed reluctance of 11 to undergo the sigmatropic reaction. At the other extreme is cis-1,2-divinylcyclopropane (lA) which was originally thought to transform to 1, h-cycloheptadiene (l£) even at very low temperatures.50 Recently, two groups have isolated lU prepared by two independent routes.9^’c The relief of cyclopropyl ring strain comprises such a driving force that lJ+ readily rearranges to 1£ at 11° with a half-life of 25 min, even though attainment of the thermo dynamically favored four-center transition state is geometrically p re c lu d e d . 14

Another interesting subtlety offered by Cope systems is the amusing H potential for an appropriately structured molecule to react to give itself as the product! The simplest case where the reaction is

" degenerate " is the unlabeled parent, 1,5-hexadiene (j) . The d iv e r sity of possible degenerate Cope systems is lim itless and includes

(R'

16 17 hypostrophene (l6)10 and any arrangement corresponding to formulation

IX-11 Even more fascinating is the consequence of incorporating a rapid

Cope reaction into a degenerate system. Bicyclo[5.1.0]octa-l,5-diene

(18, homotropilidene) was the first such compound.5*1 Doering and Roth found that, although the infrared spectrum was constant (indicating the 7

presence of only one compound), the 1H NMR spectrum varied markedly

with temperature. At +20°, two very broad peaks were seen1, but at -50°

and +l80°, well defined spectra were obtained. The authors realized

th at NMR spectroscopy measures very small energy changes which, by the

uncertainty principle, require long times while IR transitions have high

energy and short time spans. At -50°, they reasoned, the rate of

interconversion is very slow and both forms are observed*, while at

+l80°, the sharp spectrum represents an average of signals, e. g. the mean value of Ha/H s ' and H7/h 7 '.

This system was clearly not the ultimate. The inherent flexibility

in the molecule allows some dissipation of energy via facile ring

18a 18b 8 flipping (l8a & l8b). The propensity for rearrangement is further diminished by the fact that reaction can occur best from transoid conformer lib which is thermodynamically disfavored by the steric repulsions of hydrogens Hs and H7. Clearly, the ability for rearrange ment could be significantly enhanced by tying together the ends of the ring with bridging group X, thereby removing the offending interactions and assuring a transoid form as shown below. X

19-24

Organic chemists quickly responded to the synthetic challenge as a whole fam ily of bridged homotropilidenes appeared within the decade.

These include bullvalene (19, X = -CH=CH- ) ,12’13 dihydrobullvalene

(20, X = -CHsCHs-),12®'’138' bullvalone (21, X = -CH2C(=0)-),14 barbaralane (22, X = >CH2),14,15 barbaralone (2£, X = >C=0),14 and methylenebarbaralane (2j+, X = >C=CH2).16 Heteroatoms also have been introduced into the system with the synthesis of azabullvalenes 2J1 1 7 X

25a. X=OCH3 26a, X=0 27. X = S b, X=SCH3 b, X=S 28. X =S02 29. X = NH 30. X=NCN and 2617a’c>d and 9-hetero barbaralanes 2J-J0.18 Perhaps the most

unique of these molecules is the fully degenerate bullvalene 19 which

can undergo 101/3 or 1,209,600 unique Cope rearrangements!

etc. etc.

An interesting trend appears within this family of compounds which

suggests that subtle factors can exert dramatic effects on the rate of

Cope rearrangements. This is mirrored in the activation parameters for

isomerization (Table I). Specifically, it appears that as the length

of the bridging chain decreases, the reaction becomes more facile. The

ultimate case should be when tropilidene carbons 7 and 8 (for numbering,

see 18) are directly attached as in the hydrocarbon semibullvalene (Jl)*

Semibullvalene and its derivatives are of substantial theoretical

interest and present interesting synthetic challenges. Calculations of

the barrier to Cope rearrangement by both Hoffmann and Dewar (extended

Huckel and MIIIDO/2, r e s p e c ti v e ly )24 had p re d ic te d the AG* to be betw een

2-3 and 3*6 kcal/mole. Both groups investigated the theoretical

possibility that semibullvalene (31a), a substituted semibullvalene (32), a Table I. Summary of Activation Parameters for Selected Homotropilidenes

E a, AH*, AS*, AG*, Temp, Compound R eferen ce k c a l/m o le k c a l/m o l e . u. k c a l/m o l °C

B u llv a le n e (1)9) 1 4 .5 ± 0 .7 13-9 ± 0 . 7 4.4 ± 2.3 12.6 ± 0.1 25 19

13.9 ± 0.1 13.3 ± 0.1 3 o4 ± 0 .4 20

Dihydrolhullvalene (20) 1 2 .6 9-5 -40 21

Homotropilidene (l8) 1 3 .0 ± 0 .1 1 2 .3 ± 0 .1 - 5 .9 ± 0 .6 1 4 .1 ± 0 .2 25 22

Methyleneharbaralane (24) 1 0 .5 1 0 .0 2-3 9-5 -67 l 6a

Barbaralane (22) 1 0 .4 9 -8 11.5 7-8 -77 13a , l 6a

Barbaralone (2%) 8 .1 • 1 0 .5 - 4 l 16a , 23 a - AG* is expressed for the temperature indicated. 11 or a diazasemibullvalene (%j) might have either a vanishingly small energy of activation or a negative activation energy and, therefore, be better represented as a mesovalent or neutral homoaromatic molecule.25

31a 32 33

Introduction of substituents on the semibullvalene framework would be expected to perturb the equilibrium between the tautomeric forms in significant ways. Similar studies have been carried out with other members of the homotropilidene family with varying degrees of success. X3SL 5 Investigations of substituent effects in the bullvalene series are handicapped by the plethora of possible isomers which render analysis difficult. The cases that have been successfully analyzed show that there is a strong preference for attachment of a ll substituents except fluorine to a vinyl site.

Twofold-degenerate systems are easier to analyze and exhibit much more dramatic effects. Sehleyer, for example, investigated the influence of methyl and deuterium substitution on barbaralone equilibria

(3iL“3X) • The isom er in d ic a te d below was found to dom inate th e

luncertairO 37 12 equilibrium to the extent shown. The preferential attachment of deuterium to sp3 rather than sp2 carbons is well established,27 but this investigation extends the ordering to sp2’" (eyclopropyl) with sp3 > sp2-25 > sp2. The inverse order is established for methyl which bonds preferentially sp2 > sp2,25 > sp3', such is in accord with thermo dynamic data.

Using extended Huckel calculations, Hoffmann predicted that, as in the norcaradiene-cycloheptatriene system,28529 it-donating substituents would weaken the eyclopropyl (2,8) bond and accordingly prefer attach ment at C5. On the other hand, it-acceptors, because they exert the opposite influence, would tend to bond at Ci 8). The reverse effects were predicted for substituents attached at the 2,8 and

donor ^ acceptor

acceptor^ donor

positions (5£). Further, if the proper choice of substituents was made, Hoffman reasoned, then the semibullvalene could be delocalized With this wealth of theoretical background at hand, a formidable synthetic challenge was evident. Initial preparations of semibullvalene

(;ll) involved interconversions of other (CH)S hydrocarbons. Zimmerman synthesized semibullvalene first in 1966 by acetone sensitized di-it-methane rearrangement of barrelene (k0)3° and later by sensitized

4 0 . A 41

42 O 31 irradiation of cyclooctatetraene (kl) at -U0 to -60 . Mild thermal isomerization of tricyclo[3*3*0.02’63octa-3,7_&iene (b2) also gives 31*32

The synthesis of £1 via an isomer rearrangement process is undesirable, however, because of the tedious synthesis of starting material and/or the frequently required laborious separation of the desired product from contaminants.

A recent synthesis by Malherbe gives a 20$ overall yield of semibullvalene from diazoketone In this method, diazoketone undergoes decomposition in acetic acid to yield acetate which con tains the semibullvalene carbon skeleton. Subsequent reduction with lithium aluminum hydride gave diol bji which was converted to the corresponding bismesylate j+6. Elimination with potassium t-butoxide lit-

provides semibullvalene (jl). Unfortunately, this method still lacks

HOAc, O LiAIH^ 40° ‘ AcO 43 4 4

1)n-BuLi t-BuO oso 2ch3 3 1 2)CH3SQ2CI DMF CH 0 SO 3 2 4 6

Clearly, the most flexible and best synthesis of is that re ported independently by Paquette34 and Askani.35 Key to this method is

the silver(i)-catalyzed rearrangement of diazabasketane j+8 to diaza-

hv Ag

47 48 49

1) hydrolysis. -N, 2)ox idation ' ■> 11

50 15 snoutane 4g_. Hydrolysis and oxidation of 4^ provides azo compound 50., which spontaneously loses nitrogen to afford £1 in good yield.

Derivatives of £1 could then he prepared by synthesizing appropriately substituted adducts of type 4j[.

Semibullvalene has indeed proven to be a remarkable molecule, undergoing Cope rearrangement with unparalleled ease. Zimmerman found that the "hi NMR consisted of three signals at 6 5-08, 4.17, and 2.97 in a ratio of 2:4:2 and was temperature independent from +117 to - 110° .

Although UV and 1H WMR arguments eliminated the delocalized structure

Jla from consideration, it was only recently that Anet and coworkers 36 0 were able to observe non-fluxional semibullvalene. At -1 6 7 , the Cope rearrangement is slow on the NMR time scale at 220 MHz such that a ll five resonances are seen anu assignments can be made for each position

(Table II). Line shape analysis and coalescence studies gave

Table II. Assignment of 1H and 13C NMR Shifts of Semibullvalene (2JL)a

Chemical Shift13 (6) 13C (ppm) Assignment

2 .7 9 4 2 .2 2 ,8

2 .8 3 4 8 .0 1

3 .1 6 55-1 5

5 .0 8 121.7 5 ,7

5.-59 131-8 4 ,6 aIn CF2CI2 at -167°• ^Referenced to TMS. 16

Ea = 5*1 ±0.2 kcal/mole, AG* (at -1^3°) = 5>5 ±0.1 kcal/mole,

AH* = ^ .8 ±0.2 kcal/mole, and AS* = ^.8 ± 0.2 eu as the activation parameters. These are lower than the values observed for other homotropilidenes.

Another theoretical prediction received experimental scrutiny with the synthesis of a variety of substituted semibullvalene s.37 All of the 1(5)-monosubstituted semibullvalenes (3 j8 ) show preferential

38a. R=CH, b, R = Ph c, R = CH20CH3 d, R=CH,0H e, R=CN

attachment at the cyclopropyl site, which if methyl is regarded as a rt-acceptor,38 agrees well with the predictions. In the 2(8)-monosubstituted semibullvalenes ( 3 9 .), attachment at the vinyl position is favored in a ll cases. This is in agreement with predictions except for

39c (R = F), which is also in conflict with the aliphatic > olefinic

39a. R-CH3 b, R=CH2OCH3 c, R=f IT 2 0 -p > cyclopropyl ordering seen in fluorobullvalene. The interpretation of this result is uncertain.

Substituent effects of a different type were assessed with a series of 2,8(U,6)-bridged semibullvalenes (jjl-jjji) • 39 A study of the

51, X=-CH2CH=CH CH2~ 52, X = -(CH2 )3- 53, X=-(CH2)4- 54, X=-(CH2)5-

temperature dependent 1H NMR of 2, 8 -pentamethylenesemibullvalene revealed coalescence at - 8 5 . 5° and, at lower temperatures, the presence 39C of the non-degenerate isomers a and b. The values for the chemical shift of the permanently vinylic protons ( 6y, form b H 4 and Hs) and the permanently cyclopropyl protons ( 6C, form a H* and Hs), which agree extremely well with the values determined later by Anet and coworkers 36 for unsubstituted semibullvalene, provide model values that allow determination of the position of the equilibrium for the 2 , 8 -b rid g e d semibullvalene s 5>l-5 ^«

The observed shift of the fluxional protons H 4 and Hs ( 6) can be approximated as a weighted average of the model values:

6 = Pb6c + (1-Pb^8c

< 18 where p-^ is the mole fraction of tautomer h. Thus,

6 -6 C pb = 6^ 6- '

The equilibrium distributions of semibullvalenes obtained by this method are listed in Table III.

Table III. Equilibrium Distribution of Annelated Semibullvalenes

Compound $ tautomer b

5 1 , x = -ch2ch=chch2- 7^

5 2 , X = -(ch2)3- 57

3 2 , X = - ( ch2 )4- 10

3 ib x = -( ch2 )5- h2 aBased upon 60 MHz NMR in CDC1 3 a t k0°.

Paquette and coworkers concluded that the positions of equilibria were the result of an interplay of the thermodynamic preference for attachment of an alkyl group or chain to a vinylic site (which should a priori favor tautomer a) and various strain forces and non-bonded interactions. As the length of the bridging " belt " is shortened, tautomer a should become increasingly disfavored. Further, tautomer b, with the cyclopropane ring in the " center " of the molecule, possesses a greater degree of conformational freedom. This factor becomes 19 increasingly important as form a becomes increasingly strained as in

5jL and £2, or as further molecular constraint is introduced with the addition of a double bond (see _|l). Annulated semibullvalene 5^. with

X = -(0112)4- is seen to possess a minimum of these influences and thus is found to exist largely as tautomer a. As the ring size increases

(5b ) , non-bonded interactions typical of medium-sized rings become

significant and form a_ gains lesser importance.

As Paquette and coworkers pointed out, when the bridging moiety is a 1,3-butadiene unit, several distinct structures are possible.39a

The molecule could simply be a semibullvalene {^Qa) or perhaps a semi bullvalene with negative energy of activation for the Cope rearrangement (.§£b). Equally possible is pentaene 5£c or, if overlap of the

5 5a 55 b

5 5c 5 5 d

orbitals at C4 and Cs is sufficient, the 10 jt neutral homoaromatic

molecule 553.* 20

The question of what " homoaromaticity " means or indeed what

" aromaticity " connotes is a knotty one. The term " aromatic "

originally was descriptive of the odor of a given group of molecules—

a meaning hardly appropriate for the likes of pyridine or benzene

itself! Later, the term implied that the compounds had unusual

stability—a quite relative term. Differences in reactivity help to

define "aromatic ", but still there are exceptions. These cyclic molecules are usually planar (although cyclophanes and bridged

annulenes violate this) and have tai + 2 « electrons, but this does not

sufficiently define the term.

The most important criteria for aromaticity are the physical properties of the molecule in question. The most definitive of these

is x-ray crystallography which readily differentiates between the

alternate single and double bond pattern of the polyolefin and the uniform intermediate length bonds of the aromatic molecule. Since x-ray

structure determination is not always feasible, other criteria, such as

the detection of a ring current by NMR methods or by diamagnetic

susceptibility studies and ultraviolet data, are frequently utilized.40

It is perhaps ironic that no single criterion can be judged necessary

and sufficient to determine whether or not a compound is aromatic.

The situation is only ascerbated with regard to the question of homoaromaticity. Homoaromaticity requires that the carbon skeleton of

an aromatic system (A) be interrupted (in one or more places) by the

imposition of a sigma bound bridging unit (HA). Although there are numerous well documented examples of ionic homoaromatic compounds,

e.g. the homotropylium ion (^6) which results from the protonation of cyclooctatetraene (4l), the question of homoaromaticity in neutral

\ = / 41 56

(closed shell) molecules remains very much in doubt.

In 1956, Doering proposed that the 1 and 6 positions of tropilidene

(5J.) were dose enough to interact and gain stabilization through

delocalization (jj7a ).42 Although it was subsequently shown that

cycloheptatriene is non-planar,43 much uncertainty remains as to the

4 4 exact electronic nature of tropilidene-norcaradiene systems. To

date, a universally accepted example of a neutral homoaromatic compound has yet to be found.

It was into this atmosphere that V7as thrust by Paquette and

coworkers. The XH NMR spectrum of quickly eliminated the semi-

bullvalene-like structures 5J5a, since the vinylic region was shifted

significantly downfield (6.0-6.8 6) and thus was clearly devoid of any

cyclopropyl contribution.39a Further, bridgehead protons Hg and are

also substantially shifted, 0.9 ppm upfield and 0.6 ppm downfield,

respectively, from their usual semibullvalene locations. This 22

5 5c

differential shifting could be attributed to the relative positions inside and beyond the rim of an aromatic ring shielding cone. The authors reasoned that the ultraviolet spectrum also was in agreement with a mesovalent structure jjjjb or _^3d. Subsequently, 13C NMR data was

5 5b 55 d presented as further proof of structure, i.e. 55 b , c, o r d_.39b

The suggestion that could be the first neutral homoaromatic molecule met with bitter opposition. Vogel, et^al. ^5a contended that the

1H NMR spectrum m s more like cycloheptatrienes anc* "than the generally accepted aromatic bridged annulenes 5£ and 60. This, in fact, skirts the issue.

57 _5B_ 59 fin It would be unlikely that a neutral homoaromatic molecule would experience the driving force for delocalization observed for aromatic molecules or for positively or negatively charged homoaromatic molecules. What is at issue, it seems, is whether or not there is electronic interaction between the k and 6 positions and how much stabilization does such an interaction provide. Vogel's argument hinges, as he admits, on " the still unsolved problem of the extent to which cycloheptatrienes possess homoaromatic character. "

The most clearcut evidence that is 10-m homoaromatic comes from the photoelectron spectrum,46 which is remarkably sim ilar to that of naphthalene in that four distinct bands are observed in the 7-H eV region. Extended Huckel and perturbation molecular orbital calculations on model, systems were used to interpret the results. Neither a bridged semibullvalene model nor a model constructed from olefin and cyclo propane fragments adequately explain the data. The best agreement is with a model in which all nonbridgehead atoms are confined to a common plane allowing maximum it overlap.

Thus, while the 1H NMR and photoelectron spectra argue for a delocalized structure, there is not an overwhelming accumulation of evidence which supports this conclusion. In a like manner, there is a question as to the value of some of the evidence, i.e. the 1H and

13C NMR spectra; but there is no definitive proof that is not homo aromatic. As discussed earlier, such is the state of the art. The answer to questions of homoaromatic (or aromatic for that matter) lies in the collection of enough information, x-ray crystal structure, other types of NMR studies, diamagnetic susceptibility studies, etc., so that im partial judgments can be made. RESULTS AND DISCUSSION

Part I. Synthesis of Annulated Semibullvalenes

The homotropilidene moiety has been the chemical proving ground for a number of theoretical concepts. For example, studies of sub stituted barbaralones26t^ and semibullvalenes37 have demonstrated the sensitivity of these systems to substituent influences and have provided invaluable information on the nature of chemical bonding.

Annulation of the semibullvalene framework at the 2(k) and 8(6) positions allowed examination of the interplay of ground state pertubational effects. In this light, a systematic investigation of the equilibrium imbalances of a series of 2,8-annulated semibullvalenes in which one of the methylene units has been replaced by a heteroatomic functionality was expected to yield interesting information not readily obtainable by other methods.

The first compounds to be prepared were the trimethylene-bridged semibullvalene s in which the central methylene group had been replaced

52, X=CH2 61, X=0 62, X=NCH2Ph 63, X=S

2^ by x47’48. These compounds were prepared by two routes, both of which have as their basis the method of semibullvalene framework construction developed by Paquette34 and Askani.35

The first synthetic approach47’48 paralleled the previous preparations of 2,8-annulated semibullvalenes .39 As depicted in

Scheme I, the appropriately functionalized propellane (64) was treated with a triazolinedione to give a Diels-Alder adduct (jo§)• Sensitized photoclosure of 6^ provided a diazabasketane 66, which isomerized to the corresponding diazasnoutane 6£ in the presence of silver(i) ion.

The d e s ire d s e m ib u llv a le n e s (jjjL, 6l , and 63.) were th en o b tain e d b y hydrolysis and oxidation of the diazasnoutane.

In principle, this approach is quite serviceablej however, in practice it is unnecessarily duplicative. The major problem is that the key propellane (64) is not readily available and a multi-step synthesis is required for each propellane desired (64a requires 13 steps

64b requires 15 steps; and 64c requires 10 steps). The four illustrated reactions must subsequently be performed for each derivative. If the desired functionality could be introduced at some point late in the sequence via a common intermediate, then substantial savings of time, effort, and money could be achieved. Diazasnoutene 68 was viewed as a

R to. o 3* X *< q! a 0) o’ *< 3 w .'

1® •0) Z-3J

lo ID- o> O) Ol > W -* -k (Q + XXX II II II to o o X W o o 0 ) >1 IP 1!F 03 X X X r r II ii II (0 o o X X a II X o o II X X X o u 3 r lo io-

0 03] ro o\ 27 likely common intermediate with the olefinic unit serving as a moiety which could he transformed into the desired new functionality. Efforts were then directed to the large scale synthesis of 68.

Many of the details of the synthesis of 68 had heen previously reported;39’49-51 but, as is so often the case, many improvements have heen made in the intervening time. As outlined in Scheme II, the monopotassium salt of acetylenedicarboxylic acid (69) was acidified and condensed with excess 1,3-butadiene, according to the procedure of

Scott,52 to give a mixture of monoadduct (£0 and hisadduct Jl. Bulb-to- hulh distillation of 70 and J1 served to remove higher molecular weight

0 7 0 hy-products. Extraction of an ethereal solution of 70 and 71 with a

10$ solution of sodium carbonate removed 70 and left the sterically hindered anhydride 71 (22$ yield). Reduction of J1 with lithium aluminum hydride gave hexalin diol 72a (91$) when an acidic workup was used.50’52 Treatment of diol J2a with excess methanesulfonyl chloride

in pyridine at temperatures below 0° afforded bismesylate 72b in essen tially quantitative yield.

When allowed to react with anhydrous sodium sulfide in hexamethylphosphoramide (HMPA), J2b was smoothly converted to sulfide £5/ The high cation-solvating capacity of HMPA increases the nucleophilicity

of the sulfide anion and allows S^2 displacement to proceed, despite 28

Scheme II

CO.H 1^Li AIH4 * 2)CH3SQpi pyr COK O 69 71 7 2 a . R=H b, R = S02CH3

1) NCS A t-BuOK HMPA 2) RC03H‘ THF

73 74

j) C5H5N*HBr3 2) Li Cl, Li2C03 HMPA

h v sens R*N

78 a. R=Ph 81a, R=Ph b, R^CH3 b, R=CH3 29 the bisneopentyl character of the reaction centers, in nearly quan titative yield. Some drawbacks to the use of HMPA are the possible health hazard (potentially tumor inducing),53 its cost, and the tedious reaction workup required. Recovery and recycling of the HMPA reduces the cost54 and the workup may be improved by continuous pentane extraction of the product from the crude reaction mixture (diluted with water) followed by the usual processing.

Sulfide 75. was chlorinated with commercial N-chlorosuccinimide and oxidized with a freshly prepared ethereal solution of monoperphthalic acid55 to give or-chlorosulfone 7j+ in quantitative yield. Ramberg-

Backlund rearrangement56 of 7^ with commercial potassium t-butoxide in tetrahydrofuran (THF) provided propellatriene 7)5 (80$).

Bromination of triene 7)5 with pyridinium hydrobromide perbromide was originally thought to give dibromide j6 cleanly.50 Bisdehydro-

76. 79 80

bromination of 76 with lithium chloride and lithium carbonate in HMPA57

smoothly gave 7J. (72$), but subsequent addition of N -alkyltriazoline-

dione to proceeded in only 75$ yield.50 This last result was hard to

justify, especially since the addition of triazolinedione could be performed like a titration, with the red color of the azo compound utilized as indicator. Careful chromatography provided an answer to the 30 problem when starting triene ms isolated in addition to adduct j8_-

To account for the quantitative recovery found in the bromination reaction, the formation of tetrabromide 72. must be postulated.58 The excess base in the elimination reaction would then also convert to pentaene 80 in part. The residue which remained after distillation of tetraene 7£ (contaminated with sons triene Tj) was then logically attributed to unreacted or partially reacted tetrabromide 79, and/or the decomposition products of pentaene 80. (A bistriazolinedione adduct from pentaene 80 would be extremely insoluble and was not observed.)

The use of molecular bromine at low temperature (-78°) did not improve the selectivity of bromination and l,5-diazabicyclo[5«^- 0]undec-5-ene was less effective as a base in the elimination reaction. However, continuous extraction of the HMPA reaction with pentane does improve the yield of tetraene from 66$ to 76.3

The addition of triazolinedione to tetraene was originally car ried out in acetone at -78°. The mixture was then evaporated and the residue was chromatographed. Considerable time and money can be saved by merely using a different solvent system. If a solution of triazolinedione in a minimum volume of ethyl acetate is added to a solution of tetraene 7£ in pentane at 0°, most of the product (78) is precipitated as a pure white solid. Filtration then serves to remove much of the product in a pure state. More pure product can be obtained by recrystallization of the mother liquor. As a last measure, chromatography provides the unreacted triene 72 and the rest of the adduct 78. In this manner a 97-5$ yield (based on recovered triene jj) of phenyl adduct 78a was realized. Initially N-phenyl adducts were used in the synthesis of diaza snoutane 68 and subsequent derivatives. Russell found that N-methyl- triazolinedione was, like the phenyl derivative, a powerful dieno- phile.47 However, the methyl adducts are generally much easier to work with since they are more soluble than their phenyl counterparts. More over, the by-product in the semibullvalene preparation is volatile meth- ylamine rather than aniline. For these reasons, a number of N-methyl derivatives of previously characterized compounds were prepared during the later stages of this research. Hie characterization of these compounds is based upon the close spectral analogies between the two series and accurate mass measurements.

Triene 78, the first of these cases, serves to illustrate the spectral sim ilarities. Phenyl isomer jSa exhibits a 1H NMR spectrum

(CDCI3) which consists of a singlet at 6 7. 43 (5, aryl), a triplet at

6.28 (J = 3*5 Hz, 2, vinyls adjacent to the hetero ring), a multiplet a t 5-78 (4 , v i n y l) , a t r i p l e t a t 4 .7 3 (J = 4 .0 Hz, 2 , >CHN<), and a multiplet at 2.20-2.70 (4, allyl) and an IR spectrum characterized by peaks at 1760, 1720, 1690, 1490, and 1400 cm-1. On the other hand, the methyl derivative j 8b, formed from tetraene 7X in 99* 7$ yield, has a

1H NMR which contains a quartet at 6 6.17 (J = 3-5 and 3*5 Hz, 2, vinyls adjacent to the hetero ring), a multiplet at 5*74-5-85 (4, vinyl), a triplet at 4.58 (J =3-5 Hz, 2, >CHN<), a singlet at 3-00

(3, methyl), and a multiplet at 2.22-2.40 (4, allyl). The IR spectrum of 78b has maxima at 1770, 1700, 1450, 1390, and 1190 cm-1. Further, a parent ion at m/e_ 269.11689 is observed (calcd 269.11639) in the high resolution mass spectrum. The reactivities of methyl and phenyl isomers are usually similar, but this is not the case in the photocyclization of to diazabasketene 8l. Wingard found that irradiation of J8a in acetone with a

200 watt Hanovia lamp through a Vycor filte r for 2 hr gave diazabasketane 8la in 44$ yield;50 however, repetition of this procedure provided only a 20-30$ yield. Optimization of conditions revealed that j8a (1.2 g) in acetone-benzene (1:1) could be cyclized more efficiently by irradiation through a 2 cm Corex filte r with a 450 watt Hanovia lamp for 75 min. Photoadduct 8lb was then isolated in an average yield of

78$. Surprisingly, these same conditions were not suitable for methyl isomer 78b. it was found59 that irradiation of 1.2 g of j 8b in acetonitrile-acetone (l:l) through a 2 cm Corex filte r for 45 min gave

8lb in essentially quantitative yield at ca. 50$ conversion. The spectral properties of 8lb compared well with known 8la.

As detailed in Scheme III, the next reaction in the sequence is the silver(i)-promoted rearrangement of diazabasketane 8la. Occasionally this reaction proceeded much as Wingard50 found to give diazasnoutane

68a; however, more often than not, only decomposition of both product and starting material was seen. It therefore appeared that 8l would have to be the common intermediate. Hopefully, the modified diaza basketane s would be more amenable to rearrangement.

The first structural modification attempted was the bromination- dehydrobromination of 8l to provide diene 82. Bromination of 8l in methylene chloride at -78° using a solution of molecular bromine in carbon tetrachloride gave a quantitative recovery of dibromide 82.

Without purification, 82 was stirred with excess DBU in dry THF 33

Scheme I II

81 a. R=Ph 68 a. R = Ph b, RrCH3 b, R=CH3

Br,

Ph Ph

DBU

82 83

1) hydrolysis 2)oxidation ?h overnight at room temperature to give the desired diene 83. contaminated with some starting dibromide. The diene was characterized by its XH

NMR in CDCI3 which consisted of a multiplet at 6 7-1-7*5 (5, aryl), an

AA "’BB/ pattern at 5-1-5-7 (^, diene), a m ultiplet at 4.60 (2, >CHN<) and a multiplet at 3•3-3•8 (^, methine).

Quite surprisingly, diene 83, was thermally labile, undergoing cyclobutane cleavage in refluxing 2-propanol to give 8^ quantitatively.

The 1H NMR spectrum now contained a m ultiplet at 6 7« 1-7*5 (9, aryl), a singlet at 5-7^ (2, vinyl), a doublet at 5• -1-9 (J = ^.0 Hz, 2,

>CHN<), and a d o u b le t a t 3*56 (J = k. 0 Hz, 2, methine). The structural assignment for Qh was further confirmed by the IR spectrum and combustion analysis. Removal of the urazole protecting group by the standard hydrolysis-oxidation sequence gave benzocyclooctatetraene (85) which wa,s identical to an authentic sample.60

The remarkable ease with which this cyclobutane cleavage occurs is presumably attributable to the relief of strain in the starting bishomocubyl derivative (83.) and to the gain in stability from the aromatization of the diene ring. A major kinetic study of this process was outside the scope of this project, but a brief study was made by

^■H NMR. In pyridine-dg-benzene-dg (l:l), the loss of diene 83 could be measured by integration of the signal (n) at 6 if.98 (>CHN<) vs a constant signal (c) at 3*°° (the methine multiplet present in both 8l and 82).

The reaction was run in duplicate at 55-7, 67*7, and 76.6° C. A least s q u a re s f i t o f I n N/C v s time provided the rate constants, k = 1.0, 3-2, and 5*8 x 10“s sec"1, respectively. A plot of Ink vs l/temperature revealed an = 19*5 ± 1.2 kcal/mol and other standard treatment of the 35 data gave AH* = 19-1 kcal/mol and AS = -19*2 eu. The ridiculously large AS* can be attributed to a factor overlooked in itially —that the aromatic solvents used may not solvate &3 and 8U to the same extent.

In fact, aromatic solvent-induced shifts are seen in the 1H HMR spectra of both 83. and 8h (see Experimental Section). An effect of this sort must be present since it would be impractical to suggest a transition state as highly ordered as the AS* value would indicate. Nevertheless, the magnitude of is seemingly reasonable. As such it is well below the 50-60 kcal/mol observed for many cyclobutane sS1 and is coup arable to the 3°. 5 kcal/mol seen in the thermal ring opening of syn-tricyclo- octadiene (86).62

Since investigation of the rearrangement was not a goal of this project, an approximate value is considered sufficient and further investigation is left for a later date.

The instability of 83. Quite naturally prompted a reexamination of the silver perchlorate induced rearrangement of 8la. (Treatment of

8la with silver nitrate in refluxing 2-propanol-water (l:l) or with silver triflate in refluxing benzene resulted in the decomposition of product and/or starting m aterial.) One complication with the silver perchlorate reaction was the apparent complexation between the olefin unit of 8la and silver(l). Evidence for this belief was gained when the grey precipitate from an incomplete reaction was processed separately. A good yield of pure 8la was thereby recovered. Since

acetone dissolved the precipitate readily, it was added as a co

solvent for the reaction. Another problem is the trace of perchloric

acid found in silver perchlorate.63 (A trace of 60$ perchloric acid

causes complete decomposition of 8la in a matter of minutes in refluxing

benzene. ) To alleviate this difficulty, the silver perchlorate was

dried under high vacuum prior to use. A marked improvement a s noted, but the reaction required 9 days in refluxing acetone-benzene (1:4) and was s till somewhat unreliable.

After Russell had found that a silver nitrate-dioxane-water system

at 130° in a sealed reactor effected rearrangements not previously

realized,47 a sim ilar mixture was used for the rearrangement of 8la.

An excess of silver nitrate in dioxane-water (4:l) at reflux rearranged

8la to diazasnoutane 68a in 95$ yield over a 48 hr period. This

procedure proved to be very reliable and the crude product requires only

minimal purification. Similarly, methyl isomer 8lb was isomerized to

snoutene 68b in 98-3$ yield under the same conditions over a 53 hr

period. Further support for structure 68b was gained by accurate mass

measurement, together with its IR spectrum (1755> 1685, 1455, and 1390)

cm-1) and 1H NMR spectrum (CDC13) which consists of trip lets at 6 5*48

( j = 1 .5 Hz, 2 , v in y l) and 4 .7 8 ( j = 2 .5 Hz, 2 , >CHN<), s in g l e ts a t

3.05 (3, methyl) and 2.46 (4, allyl), and a multiplet at 1.96 (4,

cyclopropyl) all of which compare well with the previously characterized

phenyl isomer (68a).

With a workable synthesis of 68 now in hand, work began on the

functionalization of 68 as outlined in Scheme IV. Diol 89a was 37

Scheme IV

l)Bra 1)0 S 'OR IDBU 2)NaBH4 OR

89a. R = H (68.) 87a. R = Ph b, R=COCH3 b , R -C H 3

1)Ts CI 2) NaH, THF 38 envisioned as the key intermediate for construction of the hetero trimethylene series because the hydroxyl functionalities could, in theory, be transformed into good leaving groups. Provided that the solvolytic reactivity of these derivatives was not excessively high, bivalent uniparticulate nucleophilic displacement could result in introduction of the desired heteroatomic functionality. Diol 89a, viewed as arising frcm diene 8^ via ozonolysis and reductive workup, was then the first target molecule.

First, diene was prepared. Bromination of 68a proceeded quantitatively with molecular bromine in methylene chloride at -78°.

Dehydrobromination of the crude dibromide with excess DBU in dry THF gave diene 8ja in 99$ yield. Diene 8ja was fully characterized by its combustion analysis, IR spectrum, and 1H NMR spectrum which consists of a multiplet at 6 7*37 (5> aryl), a distinctive AA'BB' pattern at

5.75-6.40 (U, diene), a doublet of doublets at 5*20 (j = 2. 5 and 3-0 Hz,

2 , >CHN<), a m u ltip le t a t 1 .9 2 -2 .6 0 (3 , c y c lo p ro p y l), and th e c h a r a c te r istic doublet, J = 4.0 Hz, at 0. 56 (l, cyclopropyl syn to the norcaradiene ring). The 13C NMR in CDC13 confirms the norcaradiene structure with cyclopropyl carbon signals at 4-7-84, 28.41, 25-74, and

25-35 ppm- These compare reasonably well with the values found for

6 4 •known diene 88, given the sizeable difference in substitution plan.

$— 33.1 ppm *r-'38.6 ppm

88 39

In a like manner, methyl snoutene 68b was brominated and bis- dehydrobrominated to produce diene 8jb in 8&fo yield. Hie 1H MMR and IR spectra of 8jb were analogous to those of 8ja and the accurate mass was correct within experimental error.

Ozonolysis of diene 87a followed by reductive workup with sodium borohydride afforded diol 89a in 81$ yield.65 The diol exhibits charac teristic spectra: IR (3^-20, 175°, 1685 cm-1) and 1H RMR ($, py-d5)

7 .1 - 7 .^ (m, 5, aryl), 5-55 (s, 2, -OH), 5-30 (m, 2, >CM<), 3-90 (.s, 4,

-CHgOH), and 1.80-2.10 (m, 4, cyclopropyl). The corresponding diacetate

89b, a crystalline solid, gave an acceptable combustion analysis and

1H HMR and IR spectra consistent with the assigned structure.

The synthetic u tility of diol 8ga was initially demonstrated by the preparation of ether 9£. Treatment of 8ga with one equivalent of tosyl chloride in refluxing pyridine66 gave 90 together with an Q0Q uncharacterized product. In a second and similar fashion, the preformed monotosylate 91 was stirred with excess sodium hydride in THF to give an 85.5$ yield of 90. The spectral (1H NMR and IR) sim ilarities

TsCI base 8 9 a ------> 90 pyr

91 of gO to its N-methyl counterpart 6jb and the accurate mass verify the structural assignment. Both reactions are seen to be intramolecular nucleophilic displacements of tosylate by either a hydroxide group or the corresponding alkoxide. It is perhaps interesting to note that J»Q attempted cyclization of the diol with a catalytic amount of tosyl acid

gave no characterizable products, presumably because the cationic process gives rise to deep seated structural rearrangements.

Attention was now turned to the synthesis of a suitable pyrrol

idine derivative. A method developed by Ottenbrite and Alston appeared particularly attractive.67 These workers synthesized a number

of pyrrolidines by treatment of 0,6-dibromides with an excess of an

alkylamine. The amine acts both as a nucleophile, effecting a double

displacement, and as a base, removing the hydrobromic acid by-product.

This appeared applicable if diol 8ga could be modified to contain two

good leaving groups. Dimesylate 92 was therefore chosen as the target

s u b s tr a te .

Attempts to convert 8ga to 92 with methanesulfonyl chloride in pyridine were met with low yields of complex mixtures. On the other hand, the " sulfene " method of mesylate preparation developed by

Crossland and Servis68 proved quite workable, providing dimesylate £2

in 87% yield. The white crystalline material gave a correct elemental

analysis and was further characterized by its entirely consistent

1H NMR spectrum : ( 6, CDC13 ) 7.2-J.5 (m, 5, a r y l ) , 5*15 (m, 2 , >CHN<),

k.JO ( s , b, -CHgO-), 3-00 (s, 6, methyl), and 1.8-2.3 (m, b, c y c lo -

propyl). The IR and mass spectra were also in keeping with the assigned

s t r u c t u r e .

As hoped, dimesylate j22 readily underwent double nucleophilic

displacement in the presence of benzylamine in acetonitrile at room

temperature. Pyrrolidine gja, obtained in 63$ yield, has a 1H NMR

spectrum which notably contains an AB quartet ( = 9*^ Hz and Avab = 37*4 Hz) at 6 2.59 arising from the pyrrolidine ring protons, a singlet at 3-53 due to the two benzyl protons, and a multiplet at 2.4-3 attributable to the eyclopropyl proton syn to the pyrrolidine ring which is deshielded by 0,48 ppm relative to the other three eyclopropyl hydrogens. Also present in the NMR spectrum are two singlets at 6 J.4-5 and 7.20 due to the aryl protons and a m ultiplet of area 2 at 5*02 arising from >CHN<. The XR and m ss spectra and combustion analysis served to further verify the structural assignments.

Since semibullvalenes are, as a rule, acid, air, and/or temperature sensitive,47 the hydrolysis and oxidation of protected derivatives must be performed using appropriate precautions. The reactions and workups were always done under argon under conditions where contact with air was minimized as much as possible. All solutions and solvents were purged with argon prior to use and washes during workup were performed with solutions which were neutral or slightly basic. The mild oxidizing agent manganese dioxide69 served to oxidize the incipient hydrazide to the unstable azo compound without harming the semibullvalene product. Lastly, once the semibullvalene had been liberated, a very minimum of heat was used for distillations, etc. These properties quickly lim it the available methods of purification (and as mentioned earlier make the methyl precursors all the more desirable) although, fortunately, the compounds usually sublimed quite nicely. Extraction of the crude product from the reaction mixture with pentane rather than ether or dichloromethane47’50 followed by thorough washes (water) of the combined organic layers also served to facilitate purification. kz

Given the above precautions, fQa was hydrolyzed and oxidized to give following sublimation (^-0-95°, 1.2 x 10“3 mm) semibullvalene 62 as a low melting (ca 10°), light yeliow solid in 5 yield. The a c c u ra te mass (calcd m/e 235-1361, found 235-1355) and spectral properties provided the justification for structural assignment. The ultraviolet spectrum exhibits relatively low intensity shoulders at 215 (e 67OO),

231 (5000), 2^1 (2800), and 252 nm (1600) and the IR spectrum contains a host of vinylic and aliphatic absorptions. The NMR spectra will be discussed in detail in Part II.

In principle, it should be possible to obtain thiolane 95,b from dimesylate 92 by cyclization with sulfide ion under suitable conditions; however, since the methyl congener (6ja) was already available via the propellane route,47 this was unnecessary.

In summary, the common intermediate approach had provided two members of the he ter 0-trimethylene series with a minimum number of transformations. The synthetic u tility of diazasnoutene J08 was there- ✓ fore well established. As w ill be evidenced, this was only the b eg in n in g .

With the heteroatom located p to the semibullvalene framework, it would not be surprising to observe perturbation of the semibullvalene tautomer equilibrium as a result of electronic effects as well as the

special steric requirements of the heteroatomic functionality.

Preparation of a series of pentamethylene-bridged semibullvalenes, in which the heteroatom is further insulated from the fluxional framework by an additional methylene group might serve to minimize electronic interactions so that the full thrust of the steric question might be addressed.

A likely progenitor for the hetero-pentamethylene bridged series of compounds would also be snoutene 68. Key to the synthetic strategy

(as outlined in Schemes V-VIIl) would be functionalization of diol 100a, to be obtained from 68 via an ozonolysis-reduction sequence like that used previously for diol 8ga. After good leaving groups had been introduced, a variety of nucleophilic displacement reactions could give rise to the desired semibullvalene precursors.

Cleavage of the olefin functionality in snoutene 68a with excess ozone, followed by reductive workup with sodium borohydride as described in the preparation of diol 89a, gave diol 100a in 97- 8$ yield. The yield of 100a was improved (ca 10$) over that seen for 8§a because there probably is less tendency for the ozonolysis intermediates to rearrange to cyclic products inert to reduction. (The tendency for

intermediates from the ozonolysis of diene 87 to cyclize is further

discussed in Part III. ) Diol 100a exhibits in its 1H NMR spectrum in Scheme, V Ph Nv/.O

OR OR

68a 102a. X-OTs b, R=COCH3 b, X=I Ph

A g20 2)MnO'

9 4 1 0 1

N v^O

CH3S02CI 0M PhCHNH. 100a ------> OMs------NCH2Ph EtoN CH3CN

1 0 8 109

l) OH* NCH2Ph 2) MnOs

95 CDC13 the usual snoutane features plus the diagnostic trip let at 6 3*72

(j = 6.5 Hz, it-, -CH2OH) and the signal at 3.00 (s, 2, -OH) which varies in chemical shift with concentration. Although the IR spectrum was also in agreement with the structural assignment, diol 100a was further characterized as the diacetate 100b, prepared using acetic anhydride and pyridine. The diacetate is a white crystalline solid with compatible IR and mass spectra. The 1H NMR spectrum in CDC13 contains a m ultiplet at 6 4.00-4.50 (4, -CH202CCH3) and a singlet at 2.03

(6, Cg3C02-) which serve to confirm the presence of the desired functionalization. The remainder of its NMR spectrum and the combustion analysis also are in agreement with the assigned structure. With diol

100a in hand, the first synthetic objective was oxepane 101. Ironically this was to be the most difficult member of the series to synthesize.

Various methods of nucleophilic cyclizations were attempted and, at best, low yields of oxepane 101 were realized. The approach of

Reynolds and Kenyon,ssa using one equivalent of methanesulfonyl chloride and excess 2,6-lutidine in refluxing THF gave only starting diol. The method of Wolff, Smith, and Agosta,66C which had proven applicable in the preparation of ether ^0, was only marginally successful. The first step, conversion of diol 100a to monotosylate 102a, proceeded in 88$ yield using tosyl chloride and excess triethylamine. Spectroscopic characterization, particularly the 1H NMR spectrum [(6, CDC13) 7.15-

7 .8 3 (m, 9, aryl), 5-10 (t, J = 2,5 Hz, 1, >CHN<), 4.88 (t, J = 2.5 Hz,

1, >CHN<), 4.18 (t, J = 6.0 Hz, 2, -CH20Ts), 3-73 (t, J = 6.5 Hz, 2,

-CHgOH), 2 .5 8 ( s , 1, -OH), 2 .3 8 ( s , 3 , C§3-A ry l), and 1 .7 5 -2 .2 0 (m, 8, eyclopropyl and cyclopropylcarbinyl) ], verified the structural assignment. However, treatment of tosylate 102a either with excess triethylamine in refluxing THF or with excess sodium hydride in refluxing THF gave only a low yield of the desired oxepane 107» The use of the corresponding mesitylsulfonate (102c), in hopes of minimizing any

102c

0-S cleavage side reactions, proved sim ilarly unsuccessful.

One last nucleophilic displacement method was attempted. A recently developed ether synthesis utilized N-methyl-U,N'-di-tert- butylcarhodiimidium tetrafluoroborate (10^.) and triethylamine to produce oxepane (10^) from 1,6-hexanediol (105_) in 22$ yield as shown

7 0 below. Unfortunately, application of this method to diol 100a gave

ii N 105 + less than a 10$ yield of ether 101. Formatidn of the seven membered ring by an intramolecular nucleophilic displacement route did not seem to be a viable route to ether 101. Attention was turned to the opposite approach, an intramolecular electrophic cyclization. Earlier work had shown that 6-iodo-3-hexen- l-ol (106) could he cyclized to 2,3,6,7-tetrahydrooxepin (107.) in 25$

71 yield using silver oxide in ether. With this in mind, tosylate 102a

HV _ / H Agp

106 107 was converted to iodohydrin 102b in 73$ yield via the agency of sodium iodide in refluxing acetone. Since 102b was viewed merely as an inter mediate, it was characterized only by spectral means including the definitive 1H MR spectrum in CDC13: o 7*^8 (m, 5, aryl), 5-15 (t,

J = 2 .5 Hz, 1, >CHN<), 5-02 ( t , J = 2 .5 Hz, 1, >CHN<), 3-78 ( t ,

J = 6.0 Hz, 2, -CH20H), 3-29 (t, J = 8.0 Hz, 2, -CH2l), 2.95 (br s, 1,

-OH), and 1.60-2.20 (m, 8, eyclopropyl and cyclopropylcarbinyl).

Iodohydrin 102b was stirred with freshly prepared silver oxide in refluxing THF to give oxepane 101 smoothly in 57$ yield. Ether 101 was fully characterized by combustion analysis, accurate mass, and IR and

XH NMR spectroscopy.

The 1H NMR spectrum in CDCI3 is particularly unique. At 100 MHz, the protons a to the ether functionality are widely separated and strongly coupled. One pair appears at 6 3*85 as a doublet of quartets

(j = 13*0, 5-0, and 3*0 Hz), while the other pair of a protons appears at 3. lit- as a doublet of quartets (j = 13»0, 8.0, and 2.0 Hz). If, as is often the case, the upfield pattern is assumed to be the signals due 48 to the axial protons,72 then the I3.O Hz splitting may be assigned as geminal coupling constant ( J a x - E Q .), the 8.0 Hz splitting as the trans coupling and the 2.0 Hz splitting as the gauche coupling value (Jax-EQ O (see 1 0 1 a below). The signals at 3.85 are the result of

Hgq with | JQeml = 0 Hz and iJQauche' = 5*° and 3*0 Hz. These values

agree well with those seen in analogous systems.72

Hydrolysis and oxidation of oxepane 101 gave the ether in the pentamethylene-bridged semibullvalene series in 60. 4$ yield after two sublimations. Semibullvalene 94, a light yellow solid, mp 42-53°> was characterized by its accurate mass, IR spectrum (3035j 2920 , 2860, and

1 1 1 0 cm-1), ultraviolet spectrum [212.5 (e 9-° x 103), 23O.O (4.7 x 103), and 250.0 nm (2.8 x 103)3, and NMR (1H and 13C) spectra. (The NMR spectra are discussed in Part II.)

Next, attention was turned to the functionalization of diol 100a so that double nucleophilic displacements could be used to prepare the precursors of the thia- and aza-derivatives. In this light, dimesylate

108 was viewed as a desirable goal.

20 As in the preparation of dimesylate §2, the " sulfene " method worked well to provide 108 in quantitative yield. Dimesylate 108 was not crystalline, but was characterized by its spectral properties, the b9 most notable of which is the 1H IMR spectrum in CDC13: 6 7*^3 (m, 5> a r y l ) , 5 -0 1 ( t , J = 2. 5 Hz, 2, >CHN<), b .3 7 (t, J = 6.5 Hz, It,

-CH203SCH3), 2.97 (s, 6, -S03CH3), and 1.75-2.10 (m, 8, eyclopropyl and cyclopropylcarbinyl).

With dimesylate 108 in hand, the hexahydroazepine derivative was available via the method used to prepare pyrrolidine ^gci. Stirring dimesylate 108 with excess benzylamine in acetonitrile for 4^ hr, afforded the desired amine in 6l$ yield. Hie IR and 1H NMR spectra were in agreement with the structural assignment, as were the accurate mass and the elemental analysis.

Upon application of the standard hydrolysis-oxidation sequence, hexahydroazepine 10*3 provided the aza derivative of the pentamethylene- bridged semibullvalene series as a low melting (ca 10-15°), light y ello w s o l i d in 53 1° yield after sublimation. The IR spectrum of ££ is characterized by an abundance of medium intensity peaks while the UV spectrum displays shoulders at 23O (e 7*5 x 103), 2^0 (5.1 x 103), and

255 nm (2.5 x 103). The accurate mass and the NMR (1H and 13C) spectra also confirm the structural assignment. As before, the NMR spectra will be discussed in Part II.

The final intermediate required for the series was thiepin 110.

From this compound could be obtained sulfur derivative ^6 as well as the sulfoxide and sulfone members of the series. (See Schemes VT and

V II.) Since the mesylate leaving groups in 108 are both primary, the use of a powerful cation solvent such as HMPA was deemed unnecessary; however, since the reaction involved the synthesis of a seven-membered ring, precautions were taken to prevent formation of dimers.73 With these principles in mind, a solution of sodium sulfide in

N,N-dimethylacetamide-ethanol and a solution of dimesylate 108 in

N,N-dimethyacetamide-ethanol were slowly added simultaneously to refluxing ethanol. By this method, yields of thiepin 110 in the range of 55-65% were consistently realized following chromatography. The spectral properties were in agreement with the structure assigned, as m s the combustion analysis. The 60 MHz 1H NMR of 110 in CDCI3 is reproduced in Figure I.

Thiasemibullvalene 96 was then available in 60$ yield (after sublimation) via the standard hydrolysis-oxidation of thiepin 110. A light yellow solid melting at room temperature, 96 exhibits the usual large number of medium intensity peaks in the IR and a typical UV spectrum [225 (s 1.7 x 103), 255 (1-5 x 103), and 2^0 nm (1.2 x lO3)^.

The accurate mass and NMR (1H and 13C) spectra further verify the structure assigned. The NMR spectra are discussed in Part II.

Oxidation of thiepin 110 with two equivalents of m-chloroperbenzoic acid provided the sulfone derivative 111 in 83% yield. Characterization by the IR spectrum and elemental analysis helped verify the structural assignment. Fourier transform NMR at 90 MHz m s necessary to obtain the 1H NMR spectrum as 111 was extremely insoluble in CDC13 and most other solvents. The spectrum is, however, in accord with the assigned s t r u c t u r e .

Hydrolysis and oxidation of sulfone 111 gave the corresponding

semibullvalene 92.} however, in this case, the product-ms only slightly soluble in pentane and had to be isolated by ether extraction. As a result, after removal of ether, a large amount of 2-propanol remained. 51

Scheme VI

9 6

A 1)0H“ 2)Mn03

SO.

108 1 1 0 1 1 1

'l) 0H“ 2) Mn02

2 x -

99 52

Fortunately, trituration with pentane gave pure as a ^an solid,

dec > 120°. Semibullvalene £9 exhibits a typical UV spectrum [225

(e 1.9 x 103), 235 (1-5 x 103), and 2^5 nm (1.3 x 103)] and an IR

spectrum replete with nedium intensity absorbtions. The accurate mass

served as further characterization, as did the NMR (1H and "^C) spectra which are discussed in Part II.

Sulfoxides 97 and 98 present an opportunity to monitor bridging and substituent effects at a still finer level. In these two isomers, the bridge character (length of bonds, bond angles, electronic demands

of the heteroatom, conformation of the ring, etc. ) is the same except

for the slight changes induced by the relative positioning of the lone pair of electrons and the oxygen atom. It would be of interest to

determine what, if any, perturbations in the semibullvalene equilibrium are seen in these isomers.

Thiepin 110 was stirred with one equivalent of m-chloroperbenzoic

acid to give a 'J&fo yield of product (see Scheme VII). The crude product

appeared to be homogeneous by thin layer chromatography and by 1H NMR

in CDCI3 at 60 MHz. At first glance, this might be in keeping with the known tendency of m-chloroperbenzoic acid to oxidize sulfides with preferential formation of the kinetic product.74 Sodium metaperiodate, known to preferentially give the thermodynamically favored sulfoxide, might provide at least an enrichment of the thermodynamic product.74

Oxidation of thiepin 110 with one equivalent of sodium metaperiodate in methanol gave, however, a crude product which was identical to that

isolated from peracid treatment. Incremental additions of tris

(6,6,7,7,8,8,8-heptafluoro-2,2-dimethyl-3,5-octanedionato)europium 55 Scheme VII

A "0 OH** 2) M n ,02

1 1 0 11 2

1 ) OH- 2) MnOa

98 [Eu (fod)3] to the CDC13 solution of the crude peracid product caused the

expected differential shifting of 1H NMR absorbtions such that two

X!M< signals, in an 80:20 ratio, were seen. Similar Eu(fod)3 studies

with the sodium metaperiodate product revealed the same 80:20 ratio of

isomers. Apparently the kinetic product is also the thermodynamically

favored sulfoxide.

By preparative thin layer chromatography (five elutions with 10$

acetone-chloroform) it was possible to separate the two isomers. The

major isomer, Rp = O.35, is a white solid, mp 292.5-293*5° (dec). The

major peaks in the IR spectrum (KBr) are at 1755, 1700, 1^95, 1^15,

1278, 1258, 1130, 1068, 1027, 1015 , 785 , 770 , 765 , 717, and 710 cm'1.

The 1H NMR in CDC13 consists of a singlet at 6 7*5° (5, aryl), a triplet

at ^.95 (J = 2.5 Hz, 2, >CHN<), a m ultiplet at 2.80-3.ll-0 (te), a multi plet at 2.30-2.60 (lffl), and a multiplet at 2.08 (4, eyclopropyl). (The

spectrum is reproduced in Figure 2 at the end of this section. ) Further,

an acceptable combustion analysis was obtained. Not surprisingly, the minor sulfoxide product, Rp = 0.2k, was found to interconvert thermally

to the major product. The interconversion is extremely facile and the o thermodynamic equilibrium is essentially established at 70*0 in 2 hr.

The pure minor product, however, has an IR spectrum with maxima at 1755,

1700, 1500, lh95, 1^10, 1125, 1020, and 765 cm'1. The ^ NMR spectrum

in CDCI3 (see Figure 3) differs markedly from that of the other isomer, particularly with the presence of m ultiplets at 5 1.20-2.00 (2H) and with distinctive splittings in the 2.20-3.60 region (6h). Clearly,

these comprise the methylene protons with the upfield pair arising from

shielding induced by the anisotropy of the sulfoxide functionality. 55

The rest of the spectrum is easily detailed: a singlet at 7*50 (5? a r y l ) , a t r i p l e t , J = 2 .5 Hz, a t Ik 95 (2 , >CHN<), and a m u ltip le t a t

2.05 (i)-, eyclopropyl).

Further proof of the interrelation of the two sulfoxide isomers

(despite the striking differences in their 1H NMR spectra) was gained by the chemical interconversion of the major sulfoxide to the minor

sulfoxide using a method developed by Johnson and McCants (Scheme

VTTT ) .75 The major sulfoxide was treated with Meerwein's reagent in methylene chloride to obtain the O-methylated salt (ll4). The salt Ilk

Schena VIII

11 3

has an IR spectrum with maxima at 1755? 1700, 1^907 1^20, 1^10, 1275?

1120, 1075? 1020, and 780 cm"1. Treatment of 3JA with a solution of

sodium hydroxide causes loss of methanol by nucleophilic displacement.

The result is, therefore, a net inversion of configuration and the

product is the minor sulfoxide.

In attempting to assign the sulfoxide configuration, some insight

may be obtained from briefly considering the trimsthylene sulfoxides 11£_

and ll6 .47 In this system, comparisons of the 1H NMR spectra of the 56

CH3 N sxO J.

H H H H 11 5 1 1 6 1 1 7 1 1 8

sulfoxides with those of sulfide ITT and sulfone 118 lead to straight

forward assignments of configuration. In sulfide 117, the eyclopropyl proton syn to the hetero ring appears as a doublet (j = ^ Hz) at 6 2.U6, well separated from the remaining eyclopropyl protons at 1.8^-2.17- On the other hand, in the spectrum of sulfone 118, which necessarily has

an oxygen syn to the cyclopropane ring, all four eyclopropyl protons

appear as a m ultiplet at 1.99-2.4^1-. The diagnostic eyclopropyl proton

again appears with the other eyclopropyl protons in a multiplet at

1.87-2.3^ in the spectrum of syn-sulfoxide 116. Lastly, the anti

sulfoxide deshields the syn eyclopropyl proton which is a doublet

(J = k. 5 Hz) at 2.8^4* distinct from the remaining eyclopropyl protons at

1.81-2.1)-. Similar arguments cannot be used for the seven-membered

ring sulfoxides, however, due to the increased complexity of the

spectra. Nevertheless, there are several interesting parallels between

the two series. First, oxidation of sulfide ITT with sodium meta

periodate or m-chloroperbenzoic acid gives the anti-sulfoxide (lL?) as

the major product. Secondly, 115 has the higher Rf in 10$ acetone- chloroform. Such circumstantial evidence would then lead to assign ment of 112 as the structure of the major product from oxidation of sulfide 110 and 11^ as the configuration of the minor product.

Johnson's study of sulfoxides 74 has revealed that 4—substituted\ thiane 1-oxides (IT?) with an axial oxygen are more stable than their

1 1 9 epimers. If this trend also holds true for thiepins, then 112 would be the axial isomer and 1T5 would be the equatorial sulfoxide.

Several criteria for assessing sulfoxide configurations have been advanced.76 Buck and coworkers77 established the fact that an axial sulfoxide group deshields a f3 axial proton in six-membered ring compounds. Others78 have extended th is to four- and five-membered rings. A second criterion is the greater nonequivalence of cv-methylene hydrogens when the sulfoxide is equatorial.77’79 Lambert and Keske also noted that the axial sulfoxide has a more negative (larger in absolute value) geminal coupling constant.79 Solvent induced shifts have been used to determine configuration with some success.80 Cooper and coworkers also claimed some success with nuclear Overhauser effects

(WOE) and hydrogen bonding studies.800 Also, several groups have used lanthanide shift studies to assign configurations.70*3’800’81 Even infrared spectroscopy'has been used to discern the sulfoxide orientation.82 More recently, two groups83 independently demonstrated that in the 13C NMR spectra of thiane oxides the carbon a to an axial

sulfoxide resonates at higher field than the a carbon of the equatorial iso m er.

Unfortunately, many of these techniques are not applicable in our system. The greatest handicap is that the 1H NMR spectra of

112 and UJ5. are too complicated to permit detailed analysis. It is even impossible to decide with any certainty which signals are the result of protons q> to the sulfoxide and which signals are due to protons P to the sulfoxide. As a result, all of the techniques which rely upon the 1H NMR spectra of the sulfoxides are useless in this

case. Similarly, IR spectroscopy is not definitive since the spectra

of 112 and ll^ are too similar to be of use.

Fortunately, the NMR technique provides some insight into the problem. As outlined in Part II, it can be shown that the sulfoxide

in semibullvalene §7 (the semibullvalene derived from sulfoxide 112)

is axial and the sulfoxide in semibullvalene 98 (the semibullvalene

derived from 113) is equatorial. Since the 1H NMR spectra of 9£ and

£8 indicate that the sulfoxide configuration has been maintained

during hydrolysis, it may be inferred that sulfoxide 112 is axial as

previously suspected and that sulfoxide 1P3_ is indeed equatorial.

•These rationales are clearly not definitive and further proof was

so u g h t.

Recent work84 has shown that singlet (1Ag) oxygen stereo-

specifieally reacts with snoutene 68a to give, after workup, only the

allylic alcohol syn to the cyclopropane ring (120a) (see Scheme DC.)

Intriguingly, photooxidation of diene 87a. gives rise to products 59

Scheme IX

68 a 1 2 0 a

N v^O

8 7a 1 2 1 a

derived from endoperoxide 121a. A unifying explanation is found using

frontier molecular orhital theory.

To a first approximation, transition state energies are a function

of orhital energy differences. On this basis, the significant inter

action during photooxygenation is between the substrate HOMO and the

singlet oxygen LUMO. In olefin 68a, the HOMO of the molecule is the

as (n-) band of the urazole ring at -7*95 eV. On the other hand, the

HOMO of the norcarene moiety is at -9«1 eV. The net result is that the

urazole ring quenches singlet oxygen when it approaches from the less 6o hindered anti side. Oxidation on the sterically encumbered side (syn to the cyclopropane) becomes the only available reaction pathway. In diene 8ja, conjugative interactions raise the a2 (n-) energy to -8.6 eV.

More important, however, is the norcaradiene HOMO at -J.8 eV. R eactio n with the diene functionality is now energetically more rewarding than urazole quenching and the oxidation takes place from the less crowded anti direction.

Since the ionization energies of sulfides are commonly in the range of 8.1-8.7 eV,85 similar quenching might be observed in the oxidation of thiepin 110 with singlet oxygen (see Scheme x). T his was in fact the case, as only sulfone 111 and minor or syn-sulfoxide (1T5) were formed in a k0:60 ratio. Control experiments demonstrated that sulfone 111 was formed directly from sulfide 110 and was not the result of further oxidation of either of the sulfoxide isomers. Further, the sulfoxides did not undergo isomerization, as each isomer was returned unchanged following exposure to light initiated singlet oxygen production. Lastly, it should be recalled that the syn-sulfoxide is the minor product of other oxidations and thermally epimerizes. That this is the only isomer seen must clearly be the result of the quenching of singlet oxygen from the anti face by the urazole moiety. In this regard, the minor or equatorial sulfoxide must be 112, the major or axial sulfoxide is required to be llg.

Since there is very little literature precedent for oxidation of sulfides by singlet oxygen,86 several other systems were investigated.87

Notably, photooxygenation of sulfide 11£ gave only the syn-sulfoxide

116 and sulfone 118 in approximately equal amounts by NMR analysis. Scheme X

Ph P h

11 0 11 3 111

’O,

1 1 7 1 1 6 1 1 8

CH CH. CH

1 2 2 1 2 3 1 2 4 1 2 5

CH CH

A Pt

8.1 b 1 2 7 62

Anti-sulfoxide 115, the major isomer produced in the oxidation of 117 hy either m-chloroperbenzoie acid or sodium metaperiodate, was clearly absent from the product mixture. Further evidence for reagent quenching from the anti face was obtained from the photoelectron spectrum of 117 which shows the urazole HOMO at -8.0 eV while the sulfide HOMO is at -8.5 eV.88 As in previous examples, reagent quenching can be calculated to be energetically more favorable than oxidation of sulfur from the anti side. It would be reasonable to assume that similar energy levels are to be found in the bishomologated sulfide 110.

The case for selective quenching was further strengthened by the results of the photooxygenation of sulfide 122. Here no quenching is seen and both sulfoxides 12j3_ and 12k and sulfone 125 are obtained.

These products comprise a 1:1:2 ratio which compares with the J>\2 r a t i o of sulfoxides observed from m-chloroperbenzoic acid oxidation of sulfide 122. This is expected because the urazole n electrons can no longer experience a stabilizing through space interaction with the cyclopropane Walsh orbitals. A value of -8.6 eV like that for the

1 2 7 as (n_) band of 126 should be expected. In this case, the sulfide HOMO is of similar energy and sulfoxidation should be at least competitive with quenching. To test this thesis, the photoelectron spectra of 122 and 127 (obtained by reduction of diazabasketene 8lb) were recorded. 63

Surprisingly, the spectra of 122 and 12£ were superimposable upon the

spectra of 117 and 128, respectively. Preliminary results indicate 88 that rearrangement had apparently occurred on the spectrometer probe.

(This is often seen with molecules which have a propensity for thermal

rearrangement.89)

With hopes of further generalizing.this phenomena, sulfides

129-r52_ were exposed to singlet oxygen. The synthetic u tility of a

.CH 'CH OH CH

1 2 9 1 3 0

PhOCHCNHJ^S

NO. N 0.

1 3 1 1 3 2

stereoselective sulfoxidation in these systems would be lim itless.

Unfortunately, none of the penicillin or cephalosporin sulfides

122-13£ reacted with singlet oxygen. Whether the lack of reactivity is

attributable to steric or electronic factors is unknown; whatever the

cause, no further investigations were made. The above results seem to be slightly at variance with the mechanism proposed by Foote and Peters86e (see Scheme Xl). These workers suggest that oxidation proceeds with the formation of an inter-

Scheme XI

l ° 2 + - R2S — > R 2S 0 0 or R2SOO or R2S C ? 2 R2S -» 0 1 33 134 1 3 5

mediate, which could be a persulfoxide (1^2,), a diradical (1^4), or a cyclic peroxide (TJjO- The intermediate then reacts with another molecule of substrate to give, via disproportionation, the sulfoxide product. Sulfones are seen to arise from peroxide 1^5.. The quenching work would indicate that the intermediate must form from syn attack. s The question would be in the manner in which the intermediate collapses.

It would appear that the mechanism does not allow for the selectivity that is observed. Since the intermediate is a high energy species, selective attack on the sulfide might be involved. Also, the possibility of an entirely different mode of collapse cannot be ruled out. Nevertheless, the selectivity observed in the above systems appears to be real and, therefore, serves to decide the sulfoxide orientations.

Once the question of sulfoxide configurations had been answered, only the preparation of semibullvalenes and 2$ remained to complete the series. Since experience with sulfone 22 suggested that the sulfoxides might he only sparingly soluble in pentane, following hydrolysis of 112 in the usual manner, ether was used as a cosolvent for the oxidation reaction and as a solvent for extractions. Recrystal lization from ether-petroleum ether (65-110°) of the crude residue remaining after workup gave only a 16$ yield of semibullvalene ££ as a tan solid, mp > 120° (dec sealed tube). The low yield may be due to decomposition which occurred during hydrolysis when the reaction became uncharacteristically dark. Nevertheless, the IR spectrum contained the usual large number of medium intensity peaks and the accurate mass further confirmed the structure. The ultraviolet spectrum, however, was not obtained due to the extreme lack of solubility of £X in iso octane. The NMR spectra (XH and ^C) completed the structure veri fication and are discussed in detail in Part II.

The low yield obtained for the hydrolysis of 112 prompted an attempt to use a milder method of hydrolysis for 11)5.. Since a method developed by Erker and Roth 90 using potassium t-butoxide in dimethylsulfoxide-water at room temperature had proven successful earlier (see

Part III), this method was tried on pure sulfoxide 11)5. Upon workup, only starting sulfoxide 113. (8.9)0 was recovered. However, when a mixture (ca l:l) of sulfoxides 112 and 1T5. was hydrolyzed with a larger than normal excess (20 fold vs 10 fold) of sodium hydroxide in refluxing 2-propanol (*4-0 min vs the usual 60 min) and subsequently oxidized with activated manganese dioxide, a 36/0 yield of semibull valene £8 was realized. Apparently sulfoxide 112 was selectively destroyed during the hydrolysis reaction or perhaps 98 was selectively isolated during the ether and dichloromethane workup; but, by NMR 66

(1H and ^C), the product was pure at least to the 90f> level following recrystallization from dichloromethane-petroleum ether (65-110°).

Semibullvalene §8, a tan solid, mp > 170° (dec, sealed tube), has the fam iliar multitude of medium intensity peaks in the IR spectrum.

Again, the product was insoluble in isooctane and the ultraviolet spectrum was not obtained. The accurate mass further verified the structure as did the 1H and 13C NMR (discussed in Part II). In particular, the upfield multiplet (6 1.50-2.20) in the NMR confirms the retention of the equatorial sulfoxide configuration. Europium shift studies on 97 and §8 proved sim ilarly uninterpretable because of the complexity of the 1H NMR spectra and because there still are multiple sites for lanthanide coordination.

In summary, the common intermediate diazasnoutene 68 has given rise to a series of trimethylene- and pentamethylene-bridged semibullvalenes with a minimum of synthetic manipulation. Further, common intermediate 68 w ill be shown in Part III to be a suitable progenitor for a number of potentially homoaromatic molecules. 67

Figure 1. The 60 MHz 1H NMR spectrum of sulfide 110 in CLCI3 at 500 Hz sweep width.

Figure 2. The 60 MHz 1H NMR spectrum of sulfoxide 112 in CDCI3 at 500 Hz sweep width.

Figure 3. The 60 MHz 1H NMR spectrum of sulfoxide 11^ in CDCI3 a t 500 Hz sweep width. F igure 1 Ph

F ig u re 2

8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0 8 F ig u re 3 RESULTS AND DISCUSSION

Bart II. Analysis of the Equilibrium Imbalances

in Selected Annulated Semibullvalenes

Studies of annulation effects have traditionally provided useful

information about systems capable of valence isomerization. Some

examples of such studies are summarized in Table IV. Vogel demon

strated that bridging a norcaradiene-cycloheptatriene system (136.) at the 1 and 6 positions with a trimethylene chain introduces sufficient

strain to confine the molecule to the norcaradiene form.91 On the

other hand, when the bridging " belt " is lengthened to four methylene

units, only the cycloheptatriene valence isomer is present. A similar

result is seen in the aza analogue of this system (l^X)*92

As concerns the cyclooctatetraene-bicyclo[U.2.0]octatriene valence

isomerization pair (ljS), the additional unsaturation in the central

bridge exerts a "pinching effect" .51 As a result, the propellane form

is favored until n = 5- The aza analogue (3^3) in this case is even more

strained and remains in the tricyclic form until n = 6.93 . 39 The study of bracketing effects in semibullvalenes £l-5j+ proved

even more rewarding for several reasons. First, the recognized sen

sitivity of the semibullvalene nucleus to substituent influences37 is

further reflected in the delicate responses to annulation. Also, the

doubly degenerate nature of the Cope rearrangement in semibullvalene

requires that a cis-divinylcyclopropane moiety be present in both 69 70

Table IV. Bridging Effects in Systems Capable of Valence Isomerization.

B rid g e S ize When System Tricyclic Bicyclic Ref.

n = 5 n = 4 90 (CH2)n 136

^cooch3 .C00CK; N ^I^TcH2)n n = 3 n = 4 91 1 3 7

n = 3 ,h , n = 5 51 (CHa )n 1 38

n = 3 ,^ ,5 n = 6 92

1 39 5 J., X=-CH2CH = CH C H 2-

52, X = -(CH2)3"

5 3 , X = - ( C H 2}4-

5 4 , X = - ( C H 2)s -

tautomers. As a result, there exists no heavily weighted ground state preference for one of the valence isomers as is the case in the examples in Table IV. Wide variations in strain are, therefore, no longer the dominant factor. The degenerate nature of the semibullvalene framework should also minimize the entropy contributions from such factors as differential solvation of the two tautomers. (This factor is assumed to be minimal although this does not imply that it may be entirely over looked. ) lastly, because of the symmetrical substitution of the semi bullvalene framework, analysis is facilitated and evaluation of sub stituent influences can be made without concern for unsymmetrical perturbational effects.

The results of the above study demonstrated that the semibullvalene nucleus is an excellent fine-tuning device for the probing of ground state perturbational effects. For this reason, a systematic investigation of substituent influences seemed warranted. First to be prepared was a series of derivatives of trimethylene-bridged semi bullvalene 52 in which the methylene group most remote from the .. ' 72 semibullvalene framework (i.e ., x) has been replaced by 0, NCHaPh, and s (61-63.).

5 2 . 6JT> X'- O 6 2. X=NCH2Ph 6 3 . X = S

Before the results of this study are interpretable, however, a basis for the analysis of the data from these systems must be established. Pentamethylene semibullvalene provided the solution to the problem.

A low temperature 1H NMR study of in CF2C12-CD2C12 revealed that the time averaged signal due to H4, Hs and H4 /, He' broadens at

4 v

a b 5 4

-5 1 ° , coalesces at -85.5°* and separates into two distinct peaks at

-120°.39c This was the first direct spectral observation of two nondegenerate semibullvalene isomers. At -120°, the permanently vinylic protons, H4/ and H6' , resonate at 6 5-59 (6V) and the permanently cyclopropyl pair, H4 and Hs, appears at 6 2-5^ (Sc)> These values are in good agreement with the chemical shifts of 6 5*59 and 2.79 observed some 36 time later by Anet and coworkers for semibullvalene (Jl). The values obtained for may then be used as models for the permanent resonances of other closely related annulated semibullvalenes.

The observed time averaged shift for fluxional protons H4 and Hs (6m)

can be expressed as a weighted average of the permanent values based

upon the population (expressed by the mole fraction p) of each tautomers

a and b. That is,

^m = Pa^c + Pb^v ^

or 6m = (1-p-b)6C + pb5v . (2)

Rearranging equation (2) gives the mole fraction of b (pb) directly

(equation 3) and the equilibrium constant (Kgq) for (b ~ a) indirectly

(equation Ip) Pb = (^m"^c)/(^v~^c) (5)

Keq = pa / p b = ( l “pt>)/p b (k)

Once Keq is known, the enthalpy (AH0) and entropy (AS0) are available

from a least squares fit of In Kgq vs l/T, where T is temperature in

°K, as shown by equation (5):

in Kgq = -(A H °/r ) ( 1 / t ) + AS°/R . (5) 7^

With these basic equations established, the equilibrium studies as based on the 41 NMR spectra may be evaluated.

The 1H NMR of trimethylene semibullvalene ^2_39j47j48 (see FigUre 1)-) in CDCI3 a t b0° consists of a multiplet at 6 k.99 (2, H3 and H7), a

5 i a b 5 2 multiplet at 4.12-4. 2.6 (2, H4 and Hs), a multiplet at 2.73-2.93 (2, Hi and H5), and a multiplet at 1.5)4-2.28 (6, a and p). As gauged from the nonfluxional values of 6V and 6C, an equilibrium distribution of 57$ of tautomer b and by}a of tautomer a is present. Variable temperature

studies in CS2 over a range from -A-5 to -91° revealed an equal isomer

distribution to exist at -29° and isomer a to be favored at yet lower temperatures. Further processing of the data disclosed a AH° of -315

cal/mol and a As° of -1.29 eu for the conversion of b into a.

Replacement of the central carbon atom of the trimethylene bracket with an oxygen atom produces a marked shift in the equilibrium.47’48

5 a 1 b 6 1 75

This is demonstrated in the room temperature spectrum in CS2 (Figure 5) where the chemical shift of the time-averaged protons H4 and Hs (6 5-58) indicates that tautomer 6lb comprises 99-7$ of the equilibrium mixture,

Due to the symmetry of this molecule and the wide chemical shift differences of each set of protons, the 1H MR spectrum of 6l is first order, thus allowing for ready analysis of coupling constants.

Specifically, the spectrum exhibits a doublet of doublets at 6 5*58

(J3,4 = Js ,t =5-2 Hz, J4,5 = J6,7 = 2.0 Hz, 2, H4 and Bfe), a doublet at 5*26 (J3j4 = Js,7 = 5-2 Hz, 2, H3 and H7), an AB quartet at b.20

(j^B = 9*5 Hz, = 3.95 Hz, -CHgO-), a doublet of triplets at

3*35 (J.1,5 = 7*1 Hz, J4j5 = J5j6 = 2.0 Hz, 1, H5), and a doublet at

2.62 (Jis5 =7*1 Hz, 1, Hi). Variable temperature 1H MR studies from

-100° to +35° (CS2) and +35° to +100° (C2C14) revealed a lack of temperature dependence in the isomer distribution.

Mich like ether 6l, the XH MR spectrum of thiasemibullvalene 6^ is largely first order (Figure 6) and consists of a doublet of doublets,

5 = J5 —2.0 Hz and = Jg}7 = ^• 0 Hz, at 6 5*3^' (2, H4 and H©), a doublet, 0^4 = J.6,7 55 Hz, at 5-20 (2, H^ and H7), an AB quartet,

J^B = 10-0 Hz and AvAB = 1 6 .3 Hz, at 3*3^ (^> -CHgS-), a multiplet at

3.26 (l, Hi), and a doublet, JijS = 7-0 Hz, at 2.93 (l> Hi).

Also like ether 61, thiasemibullvalene 63. was found to favor tautomer b.47’48 In this case, however, the dominance by structure b is somewhat less, with the " closed " form comprising nearly 90$ of the equilibrium population. This is seen to be an isomer ratio opposite to that experienced by tetramethylene derivative 53..39 Unlike 33.j ^He thia conpound exhibited weak temperature dependence, with isomer 63b 76

7 A

4 .s

5 1 a b 6 3 increasing in population as the temperature was lowered from +100°

(89.85k, C2CI4) to -120° (96.9/0, CS2). Further treatment of the data reveals that AH° ^ 0 and As° = -8.8 eu. The entropy value is probably substantially in error, however, due to the systematic errors intro duced in Keq when one isomer substantially dominates the equilibrium.

In aza derivative 62 the equilibrium-diagnostic signal due to protons H4 and Hs is a doublet of doublets (J3?4 = £6,7 = 5-0 Hz and

J4,5 = J3?6 = 2.0 Hz) centered at 6 5*36 (Figure j ) . On the basis of

s a 6 2 b this chemical shift, the approximate isomer distribution in 62 is seen to be intermediate between that of the oxa and thia compounds (Table V).

Application of the standard equations revealed that isomer 62b comprises 92*9$ of the equilibrium mixture at ^-0°. The resonances of the various sets of protons are again nicely resolved with the aryl protons appearing as a broad singlet at 6 7.15 and the permanently 77 vinylic (H3 and H7) protons appearing as a doublet, £3^4 = £3,7 = 5*0Hz, at 5-08. The benzylic protons are a singlet at 3*53 while the protons or to the semibullvalene framework are an AB quartet, J^g = 9.0 Hz and

Avab = ^9-6 Hz, centered at J.08. Remaining absorbtions are the doublet of triplets, £1,5 = 6.5 Hz and £4^5 = £5,6 = 2.0 Hz, centered at 3*12 which is due to H5 and the multiplet at 2.76 which is due to Hi« It is important to note that the 1H NMR spectrum of 62 in CS2 is shifted upfield an average of 0.2 ppm relative -to the spectrum in CDC13 (see

Experimental Section). Alterations in chemical shift due to changes in solvent systems is a general phenomenon.94 For this reason, comparisons can only be made for equilibrium values and other information derived from NT© data in similar solvent systems. The spectra displayed in the figures are typical spectra, but are primarily for illustrative purposes. Based upon the findings in the oxa and thia analogues, alterations in the distribution of the a and b forms with changing temperature were expected to be snail and prone to systematic error.

For these reasons, a temperature study was not conducted.

The results of the 1H NMR investigations in CDC13 at ^0° are summarized below in Table V.

13 In like manner, the C NMR spectra of the hetero trimethylene semibullavlenes 6l-6j5 reflect the equilibrium distribution of tautomers.

As noted earlier, non-fluxional (at -167°) semibullvalene (31) contains five resonances at 131.8 (C4 and C6), 121. 7 (C3 and C7), 53-1 (C5),

I4-8.0 (Ci), and h2.2 ppm (C2 and Ca). On the basis of these model values,, assignments can be made for 61-62.* The relevant data are summarized in

Table VI. In the case of 62, the single frequency off resonance 78

Table V. 1H MR Data (60 MHz, CDC13), Computed Equilibrium Constants (Kgc), and Gibbs Free Energy Values (AG°) for £2, 6l, 62, and 6^, (4 0 ° ).

B ra c k et C hem ical Mol f r a c tio n ^ Keq , AG°, Substituent S h i f t , a 6 isom er b s/k c a l/m o l

ch2 (52) 4.21 0.57 0.75 175

0 5 .5 8 0.997 3 x 1 0 -3 36OO

NCH2Fh (62) 5 .3 6 O.929 7 .6 x 10“2 1600

s (63.) 5.25 0.895 0 .1 2 1320 a b Protons H4 and He relative to internal TMS', The method employed is not considered to provide accuracy levels of better than 1 °fo a t b e s t.

decoupled (SFORD) spectrum was used to firmly establish the level of substitution of each carbon, i.e., quaternary carbons appear as singlets, methines as doublets, methylenes as triplets, and methyls as quartets (see Experimental Section).

Especially noteworthy are the observations that the central carbon atoms (Ca,a) have considerable cyclopropane character while the peri pheral carbons (C4j6) are highly olefinic in nature. Use of the cor responding non-fluxional semibullvalene values (131.8 and 42.2 ppm) as models for 6V and 6C> respectively, permits calculation of the mole fraction of tautomer Jo as before (Table VIl). The excellent agreement of the tautomer distributions calculated by 1H and 13C MR serves to underscore the earlier trend of 0 > NCH2Ph > S for tautomer b and substantiates an earlier use of 13C MR chemical shifts to determine 79

Table VI. Summary of Cmr Data (22.6 MHz, CDC13, ambient temperature) f o r 6l , b2 , and 63..

Chemical S hifts, ppm from TMS Semibullvalene Cx cs,e C3j7 C4j6 C5 aa

6 1 (X = 0) 53.6b 61.7 121.0 127.6 52.8b 6 3 .k

62 (X = NCH2Ph)d 53.8b 61.5 121.9 125.lj. 52.7b 53.913

63 (X = s) 53.lb 73.2 123.3° 119.3C 52. 9° 3 1 .7

delates to the carbon atoms of the bracket with the a position being most proximal to the semibullvalene nucleus. bThese values may be interchanged. cThese values may be interchanged. d59*^: benzyl} 12b.9, 128.1, 128.7, and 139*0: aryl carbons.

Table VTT. Equilibrium Distributions for 6l, 62_, and 63 Calculated by 's~'v ' and 13C NMR.

Semibullvalene Mol fraction Mol fraction b ( 13C NMR) b (^H NMR)

61 (X == 0) 0.9b5 0 .9 9 7

62 (X == NCH2Ph) 0.930 0.929

63 (X == s) 0.905 0.895

equilibrium positions. While carbons 2 and 8, i.e., those positioned

P to the heteroatom, exhibit more varied shifting, C4 and C6 (positioned

on the opposite side of the molecular frame) are reasonably well insul

ated from such factors. 80

Earlier studies37 on 2(h)-methyl'semibullvalene (Jga) substantiated the theoretical predictions24 that an electron-releasing group38 exerts an electrical effect within the molecule which alters the relative tautomer energies in such a way as to favor bonding to an olefinic carbon as in 39a (by ca 1 kcal/mol at 4o°). Direct extrapolation of

39a. R = CH3

this analysis to 2,8-dialkylated semibullvalenes would predict still greater partiality for substituent attachment to olefinic sites in the absence of steric considerations. The study of annulated semibullvalenes £1-^4 fully demonstrates, however, that the position of the equilibrium is dependent upon factors supplementary to the major forces of bonding preferences and strain. In this context, the trimethylene derivative £2. is exemplary. Here the thermodynamic preference for vinylic attachment of the alkyl chain is partially obscured by the strain introduced in tautomer a. Thus, tautomer a is favored by only

315 cal/mol. On the other hand, the greater degree of conformational freedom is seen in the isomer in which the cyclopropyl ring occupies a position central, to the system, that is the " closed form " tautomer b.

Accordingly, at low temperatures the enthalpy factors are more impor tant (favoring 52a ), whereas at higher temperatures entropy factors and 52b become dominant. If the bracketing belt were shortened, the amount of strain in form a would increase until it equals or surpasses the thermodynamic bonding preferences. At this point, the equilibrium m il be largely controlled by the entropy factors which favor tautomer b.

As seen in Table VIII, the bond lengths and atom size increase in the order 0 < NR < CH2 < S. This trend almost parallels the trend for the increase in the amount of tautomer a predicted above. The anomaly is sulfide derivative 63. In this case, there is another steric factor that must be considered, i. e.^ the bond angle. The average angle observed in sulfides is some 3° less than that observed for typical saturated amines, ethers, or hydrocarbons. This smaller bond angle in effect shortens the C-S bond distance somewhat. The degree of impor tance which this factor holds is uncertain however.

Superimposed on these factors are the electronic effects generated by the positioning of various heteroatoms p to the semibullvalene framework. A previously prepared semibullvalene, 39b_, provides a clue to the nature of these effects. While the equilibrium for methyl

39 b. R = CH2OCH3

isomer 39a contains 82$ of the vinylic isomer, the methoxymethyl substituent is bound to the olefinic site only 57$ of the time at the same temperature. Similarly, the equilibrium is shifted in favor of tautomer b in all of the heteroatom analogues and the effect becomes Table V III. Selected Molecular Parameters.

van der Waals Covalent single Average C-Xo E le c tr o - Bond X radius (A)a bond radius (A)a bond le n g th (A) negativitya a n g le

-CH2 - 2.00 0.77 1.5^b 2.5 1 0 9b

- 0- 1 A 0 0.7^ 1.^3b 3.5 n o b

-NR- 1.5 0.75 l.VTb 3 .0 1 0 8b

-S- 1.85 1.02 1.82b 2.5 105b

-so - 1 . 8oc,d 98c ’d’

-SOs- 1.76® 107e

aL. Pauling, "Nature of the Chemical Bond ", 3rd ed., Cornell University Press, Ithaca, N.Y. (i960); 13" Tables of Interatomic Distances and Configuration in Molecules and Ions” , Chem. Soc. Spec. Publ. , No. 11, (1958)j CR. Thomas, C. B. Shoemaker, and K. Eriks, Acta. Cryst. , 21, 12 (1966); dM, A. Viswamitra and K. K. Kannen, Nature, 299., 10l6 (1966); eR. K. Bullough and P. J. Wheatley, Acta. Cryst., 10, 233 (1957). more pronounced as the electronegativity of the heteroatom increases.

These effects would appear to be explicable in terms of adjacent bond weakening arguments. Clearly, the net electronic effect of the p heteroatom is the diminution of vinylic bonding preferences.

The combination of the above influences provides a unified theme as the structural enlargement of the bracket and the lessening of «• electronegativity demands by group X work in concert to increasingly favor tautomer a. The study of the temperature dependence of the equilibrium for sulfide reveals that the electronic and steric factors almost equally oppose the thermodynamic bonding preference as

AH° « 0. Here the entropy factors are dominant. The removal of eclipsing interactions of vicinal hydrogens should enhance the greater conformational freedom seen in tautomer b,

In summary, both enthalpy and entropy factors combine to ■ increasingly favor tautomer b in 2,8-trimethylene semibullvalenes in ac cord with the ordering CH2 < S < HR < 0. Thus, by means of spectro scopically detectable equilibrium imbalances, subtle libration effects not visible by other techniques have been detected. The trimethylene

series, then, proved to be interesting and heuristic, prompting the syn thesis of an analogous series of 2,8-pentamethylene semibullvalenes (9^_~

22.). The results of the NMR studies of these compounds are now detailed.

The parent 2,8-pentamethylene semibullvalene 53. has been previously reported.39>47 The ^-H NMR spectrum of 53_ in CDC13 (Figure 8) is characterized by a doublet, J3,4 = J6,7 =2.7 Hz, at 6 ^.98 arising from the permanently olefinic protons H3 and H7, an equilibrium- diagnostic multiplet at 3.75 Hue to H4 and Hs, another m ultiplet at 8k

7 OC P 6 ^ 8

5 1 a b 53

3.00 attributable to H5, and a doublet, £1,5 = 6.0 Hz, at 2.J8 from Hi*

The a protons appear as a complex m ultiplet at 2.38-2.65 while the p and v protons comprise the m ultiplet at 1.^0-2.10. It w ill be remem bered that it is this compound which in its non-fluxional state at -120° provides the model values which are used to determine the equilibrium distributions for other semibullvalene s. At +W3, 53, was found to be almost evenly distributed between form a (58$) and form b. The temperature study also provided (for b -* a) AH° = 1.13 kcal/mol and

AS° = 3.9 eu in CD2C12-CF2C12 and AH° = 700 ± 20 cal/mol and

AS° = 2.90 ± 0.09 eu in CS2.

The first member of the hetero series, ether 9^,, strikingly favors tautomer 9^a. The very distinctive 1H NMR spectrum of in CDC13 at

7 ■ 8 a P

5 a 9 4 b 85

+*t-0° (Figure 9) reveals the time-averaged H4 and Hs signal at 5 2.87 as a doublet of doublets further split by long-range coupling with J3j4 =

J6,y = 1.5 Hz and J4j5 = J5j6 =7-0 Hz. At lowest field (5.20) are the permanently vinylic protons H3 and H7 which appear as a doublet,

£3,4 = £s,7 = -*"5 a^ equatorial p protons are strongly split much as in precursor oxepane 101. In this case, the axial protons, a doublet of triplets at 3*195 reflect a geminal coupling of 11.6 Hz and an additional coupling of 6.5 Hz. The equatorial protons, also a doublet of triplets, occur at k.0 3 w ith | jQem] = 1 1 .6 Hz and a d d itio n a l splitting, J = 5-0 Hz. Unlike oxepane 101, it is not possible to discern the trans and gauche couplings. The spectrum is completed with a doublet,

£1,5 = 7.0 Hz, at 3.55 Hue to H^ a triplet, J1?5 = 7.0 Hz and

£4,5 “ £5,6 = 5.5 Hz, at 2.9*1- arising from H5; and a m ultiplet at 2.50 attributable to the sx protons. The position of the H4 and H6 signal suggests that isomer ^la is favored to the extent of 8*1-$.

Variable temperature 1H NMR studies carried out in CS2 (Table IX) and CD2C12-CF2C12.revealed increasing amounts of tautomer a as the temperature was lowered. In CS2, the isomer distribution ranged from

12$ b at +32.3° to 3-7$ b at -99*9°. Graphic analysis of these data gave (for b -» a) AH° = -1.0b ± .10 kcal/mol and aS° = 0.*t-3 ± .*t-0 eu.

In CD2C12-CF2C12, there was 18$ of isomer b at +32.3° while at -99*9° there was 10$ of form b. Further analysis of the data gave AH° = -593

± 6l cal/mol and AS° =1.05 ± 0.28 eu.

The aza derivative 95 is more equitably distributed. As seen in the

1H NMR spectrum of 95 in CDC13 at *J-0° (Figure 8) the H'4 and H6 protons appear at 8 3 * 75s the same position observed for the fluxional protons 86 in the parent hydrocarbon 5|+. The permanently vinylic protons H3 and

H7 appear as a well-defined doublet, £3,4 = £6,7 = 5.0 Hz at 5*00.

Spin decoupling at 3.75 and 5.00 confirmed the assignments. The

4 NCH2Ph

5 1 95 b

benzylic functionality was characterized by the singlet due to the aryl protons at 7.17 and the singlet at 3-52 from the benzylic protons. The remaining protons could not be specifically located and are contained in a multiplet at 2.10-3.20. The spectrum was temperature dependent as the studies in CS2 (Table XI) and CD2CI2-CF2CI2 (Table XII) exhibit.

In fact, in CS2, the equilibrium distribution ranges from 37$ b. at

+3^*2° to 52$ b at -85.6°. At -75-^° an equal distribution, of tautomers is seen. Further treatment of these data reveals AH° = 582 + 20 cal/mol and AS° = 2.91 + 0.09 eu for b -* a. The temperature range studied in

CD2CI2-CF2CI2 was narrower (-3^.3 to -85.6°). Temperatures above -3^.3° were not utilized due to the low boiling point of the freon. (Later studies showed that in a well-sealed NMR tube, measurements could be made even at room temperature.) Below -85.6° considerable resolution was lost and below -100° the solutions separated into layers. Neverthe less, the enthalpy and entropy values were calculated to be 427 ± 17 cal/mol and 1.37 - 0.18 eu, respectively. Surprisingly, thiasemibullvalene 96 favors tautomer a to a greater extent than aza derivative 95. This is in contrast to the ordering observed in the trimethylene series. In this case, the equilibrium-

5 1 a b 9 6

diagnostic protons H4 and H6 appear as a doublet of doublets, J3j4 =

Je, 7 = 3«° Hz and J4j5 = J5>6 = 5*9 Hz, at 6 3-57 (Figure 10). The

other signals are the doublet, J3,4 = £6,7 = 3.0 Hz, at 5.17 due to. H3

and IIV and the complex pattern at 2.10-3.55 generated by the remaining protons. Application of the standard equations reveals that, at ^0° in

CDCI3, isomer 95a comprises 62$ of the mixture. Temperature studies

in CS2 further show that the equilibrium distribution ranges from 65$

of tautomer 96a at +33*5 to an even distribution of isomers at -92.5°.

Treatment of these data (Table XIIl) gives AH° = 555 ± 13 cal/mol and

AS° = 3.09 ± 0.06 eu. On the other hand, the equilibrium distribution

in CD2C12-CF2C12 is only weakly temperature dependent (Table XIV) with

AH° = 187 ± 29 cal/mol and AS0 = l.Olj- + 0.1*1- eu.

Tautomer b was found to be the favored form of the sulfoxides and

sulfone. Of these, the syn sulfoxide (£8) was the most equitably •

distributed, favoring 98b -to the extent of 6 6 fo. As shown in F ig u re 12,

the 1H NMR in CD2C12 at +*10° exhibits several distinctive features.

The permanently vinylic protons H3 and H7 comprise the doublet, 88

J3,4 = £e,7 = 3*5 Hz, a-fc 5 5.20. Next, the trip let (£3,4 = £6,7 = 3*5

Hz) a t h.57 is due to the fluxional protons H4. and H6. Particularly significant is the multiplet at 1. 50-2.20 which is similar to the upfield multiplet seen in precursor 115. It is clear, based upon arguments similar to those used in the analysis of the configuration of

115, that the sulfoxide in 98 must also be equatorial. The remaining protons generate the complex multiplet at 2.U-^.0. Since sulfoxide 98 was•v'N* only slightly soluble in CD2C12, 90 MHz Fourier transform 1H NMR was utilized to study the temperature dependence of the equilibrium. As the temperature decreased, the population of 98b increased from 68$ to

+50.9° to 89$ at -73.0°, corresponding to a AH° of 1.10 - 0.12 kcal/mol and a AS° of 3-15 ± 0.51 eu (Table XVl).

The sulfone 99 is biased even further toward tautomer b as seen by its 1H NMR in CDC13 at +^0° (Figure 13). Spin decoupling verified the assignment of the doublet. (j3}4 = £6,7 = 3.8 Hz) at 6 5*19 to the

H3 and H7 protons, the triplet (£3,4- = £ 6 ,7 = 3 .8 Hz and £4^ = £5,6

= 3*5 Hz) at 5.22 to protons H4 and H6, and the m ultiplet at 3*30 as proton H5. The remaining protons were assigned as follows: 3.20-3.25

(m, t, -CH2S02-), 3.00 (m, 1, Hi), and 2.75-2.90 and 2.20-2.50 (m, -CH2CH2S02“). Calculation of the equilibrium distribution revealed that 99b comprised 77$ of the mixture. Variable temperature studies in CD2CI2-CF2CI2 (Table XVIl) demonstrated the same increasing partiality to tautomer ^ (77$ at + 32. 6° and 92 $ a t -61. 9 °) as the tem perature was lowered as was seen for sulfoxide 99. In this case,

AH° = 1.72 ± 0.07 kcal/mol and AS° = 3.16 ± 0.25 eu.

Sulfoxide 97 was found to display the greatest bias toward tautomer b of a ll of the pentamethylene semibullvalenes. In the CDC13 1H NMR

4 rJ

5 1 b a

spectrum at +^0°, the permanently vinylic protons H3 and H7, which appear as a doublet (£3^4 = £6,7 = ^-5 Hz) at 6 5-18, are located at higher field than the fluxional protons H4 and H6 which appear as doublet of doublets (£3,4 = £6,7 = ^*5 Hz and J4j5 = J5j6 = 7*5 Hz) at

5.22 (Figure Ik). The remaining protons appear as a set of six (Hi, H5, 90 and (3) in the multiplet at 2.75-3. 40 and as a set of four (a) in the complex pattern at 2.45-2.75- Based on the above data, tautomer 9Jb_ ms found to be present at the 89/0 level. The sparing solubility of £7 in

CD2C12 limited the investigation of the temperature dependence of the spectrum to a range of +32.5 to -29-3°- Since the distribution of iso mers at the low temperature was heavily biased (98$), this was also the lim it for reliable calculations. (There may still be a good deal of systematic error introduced when the equilibrium is so heavily shifted.)

Further calculations revealed &H° =3-35 ± 0.22 kcal/mol and AS° =

6 .1 7 ± 0 .8 2 eu.

For ease of comparison, the data obtained from the 1H UMR studies of the equilibria are summarized in Tables XVIII and XIX.

Table XVIII. 1H MR Data (100 MHz, CDC13), Computed Equilibrium Con stants (Keq), and Gibbs Free Energy Values (AG ) for 52^ 6 1 , 6 2 , and 63 (4 0 ° ).

B racket' Chem ical Mol fraction Keq AG° substituent s h i f t , a 6 iso m er b c a l/m o l

0 (94) 3 .0 0 0 .2 0 4 .0 0 -863

S (g6 ) 3 -5 7 0 .3 8 1.63 -305

NCH2Fh (95) 3 .7 5 0 .4 3 1.33 -1 7 6

CH2 (54) 3 .7 5 0 .4 3 1.33 -176

s^n-SO (98) 4 .5 7 0 .6 8 0 .4 7 470

so2 (99) 4 .7 6 0.7 5 0.33 684 anti-SO (97) 5.22 0.89 0.12 1300 aProtons H4 and H6 relative to internal TMS; ^The method employed is not considered to provide accuracy levels of better than 1# at best. 91

Table XIX. Summary o f Thermodynamic D ata f o r S em ib u llv alen es 5 4 , §4, ’ §L an(3 22..

oa oa ob B racket AH AS AS substituent k c a l/m o l eu kcal/mole eu

0 ^ -0 .5 9 1 .1 - 1.06 0 .4 s (96) 0.19 1.0 O.56 3 .1

NCH2Ph (95) 0.43 1 .4 O.58 2 .9

CH2 (5*0 1.13 3 -9 0.70 2 .9

syn-SO ( 98 ) 1.43 3 .2

S02 (99) 1.72 3 .2 anti-SO (9T) 3.35 6 .2 aIn CD£C12 ~CF2C12-TMS (1 :1 :1 ) o r CD2C12 -TMS; bIn CS2 -TOS.

As in the trimethylene series, the 13C NMR spectra of the pentamethylene semibullvalenes parallel the 1H MMR spectra in reflecting the equilibrium distribution of the tautomeric forms. With the use of the non-fluxional semibullvalene resonances given earlier as model values, assignments were made for 54, and §4.-99; The relevant data are sum marized below in Table XX. In all cases, the SFORD spectrum was utilized to confirm to spectral assignments (see Experimental Section).

It is interesting to note the trends in this series. First, and most obviously, is the increase in the C4 and C6 resonance as the vinylic character increases. As before, use of the non-fluxional semi bullvalene carbon resonance values allows calculation of the tautomer ' 92

Table XX. Summary of 13C MR Data (22.6 Hz, CDC13, ambient temperature) for Semibullvalenes Jj4, 94, 98 and 22.*

Chem ical S h i f t s , ppm from IMS Compound Ci Cg,Ce C3,C7 C4 ,C6 C5 a P

94 59-49a 129.06 120.22 54.68 4 8 .8 3 a 3^.13 74.06

9 6 59-62a 118.54 121.83 70.46 5 4 .2 2 a 3 1 . 6 2 ° 33.88b

95° 5 8 . 8i a 118.64 121.45 80 .2 1 52..77a 2 9 .5 1 5^.7^

121.4 H i 5 9 -6a 1 0 9 .1 8 2 .2 52. 4a d d 98 53-36a 9 0 .8 6 122.47 97-93 57-24a 20.39 51-7^

99 5 4 .l2 a 8 9 .6 7 123.01 99-92 57-30a 24.49 56. l l

91 54.6oa 72.40' 124.09 118. 59 61. 24a 19.26 49.10

sl b These values may be interchanged; These values may be interchanged; CAdditional resonances include: 6l.35> benzyl; 127-06, 128.25, 129-11, and l40.0, aryl; 4*, p, and Y carbons appear at 28.0, 29-0, and 30.0.

distribution (see Table XXl). The excellent agreement of these values provides further substantiation for the ordering determined by 1H MR.

A second trend is observed in the values for C5 which by progres sively increasing reflect the change from primarily cyclopropyl to predominantly bis-allylic character. An opposite trend is seen for C15 although here forces arising from the bracketing ring may moderate the trend somewhat. In particular, steric compression or electronic inter actions may cause the slight deviations noted.

Variations in the 6 carbon (a to the heteroatom x) shifts are primarily a reflection of the electronegativity of the heteroatom X. T a b le XXI. Comparison of Equilibrium D istributions Calculated by 1H and ------13C NMR ( CDCI3 a t 4 0 ° ).

Mol fraction Mol fraction Semibullvalene b (13C KMR) b (1H EMR)

94 (X = - 0 -) 0 .l 4 0 .1 2

96 (x = -s-) 0.32 0.38

95 (x = >NCH2Ph) 0.42 0.43

54 (X = -CH2 - ) 0.45 0.43

98 (X = syn >S0) 0 .6 2 0 .6 6

99 (X = > so2 ) 0 .6 4 0 .7 4

97 (X = anti >S0) 0.85 O.89

These values are in excellent agreement with those reported by Lambert and coworkers for six-membered ring saturated heterocycles.94

Particularly interesting are the shifts of the (3 carbons (i.e., a t o th e

SO) in sulfoxides 97 and 98. In the axial isomer (J97), the carbon bonded to the SO functionality appears at 49.10 ppm while the corresponding

carbon in the equatorial isomer (98) is at lower field, 51-74 ppm. This is exactly what is observed in the six-membered ring sulfoxides.95 Here the carbon a to the axial sulfoxide is at 45.1 ppm while the carbon a to the equatorial sulfoxide is at 52.1 ppm. This supplements the previous arguments regarding the sulfoxide configuration.

Although the a carbons and carbons 2 and 8 are progressively more removed from the heteroatom X, the shift of each is s till slightly effected by the electronegativity of the heteroatomie unit. The earlier study of six-membered ring heterocycles by 13C NMR is again instruc tive.94 The " full effect ” of X can be obtained by subtracting the

Table XXII. Adjusted 13C NMR Shifts of Carbons a, 2 , and 8.

a a X Ap a corr Ay* Gs)Cb c o rr

0 -0 .5 33.8 -3-3 1 2 5 .8

S + 0.5 32.1 -2 .k 1 1 6 .1 O N"\ 1 NR - 1 .5 C 28.3d • 1 1 5 .3 d ch2 0 .0 e 0 .0 1 0 9 .1

SOeq -k.k 16.0 -3 .0 8 7 .9

S02 - 2 .6 21. 9 ' -3 A 8 6 .3

SOax -1 2 .2 7-1 -3 .0 6 9 .h aj3 carbon in (CH 2 )5X-27*7 ppm (cyclohexane resonance); ^y carbon in (CH2 ) 5X - 2 7 .7 ppm; CR = -CH3 ; °R = -CH 2 Eh; e2 8 .0 , 2 9 -0 , or 3 0 .0 ppm.

shift of the carbons in cyclohexane (27.7 ppm) from the shift of the p or V carbon in the appropriate heterocycle (see Table XXIl). These values can then be used to correct the shifts of cn and C2 and Ca so that the influence of the heteroatom is removed. The resulting values

(Table XXII) demonstrate that the ordering of these shifts still follows the trend of the equilibrium. Two exceptions are sulfoxides 97 and 98 •where the positioning of the oxygen atom and the lone pair of electrons exerts a particularly large effect.

Lastly, it should be noted that the shift of C 3 and C7 rem ains fairly constant for both the tri- and pentamethylene series. In all cases, the C3 and C7 resonances are between 120.2 and 12k. 1 ppm. This is in agreement with the assumptions that the semibulIvalene framework remains relatively unperturbed and the heteroatom is sufficiently dis tant so that the shifts of carbons 3 , 4 , 6 , and J are uneffected.

The in itial assumption in the pentamethylene series was that the additional methylene group would provide sufficient additional insulation from the effects of the electronegativity of the heteroatom so that the equilibrium would be largely dependent upon the subtle steric influences of the heteroatom. The slight upfield shifts in the Y carbons

(2.h t o 3 .k ppm upfield) and the relatively constant shift of carbons

3 and 7 suggest that this is largely the case.

With this assumption as a working basis, the results of the penta methylene semibullvalenes may be analyzed. It will be recalled that attachment of the alkyl chain to an sp 2 carbon is thermodynamically favored in the absence of steric effects. In this regard, ether is seen to be nearly ideal. At +k0°, gjj-a dominates the equilibrium to the e x te n t o f 86- 88$ and is favored by an enthalpy difference of 0 .5 9 o r

1 .0 6 kcal/mol (depending on the solvent system used for the measurement).

This enthalpy difference begins to approach the 1-2 kcal/mol expected for the attachment of two alkyl groups with no steric complications. As the

C-X bond length increases 0 < N < C < SO 2 (see Table VIIl), the size of the bracketing ring increases and non-bonded interactions typical of medium-sized rings become more important. For this reason, the bonding preference is overcome and tautomer b (containing the smaller bracketing ring) gains favor because of enthalpy considerations.

Tautomer a on the other hand is uniformly favored by entropy fac tors since the larger bracketing ring should have the greater degree of freedom. As the series progresses, the entropy term increases in magnitude perhaps reflecting the increased flexibility of the ring.

The above explanation adequately orders 0 < NR < CH2 < SO2 but does not explain the ordering of sulfide 96 and sulfoxides 97 and 98. It

■will be recalled that the sulfur analogue in the trimethylene series was also slightly anomalous. As before, it would appear that the bond angle in sulfides causes the effective size of the ring to approach that of the amine or the hydrocarbon. The change in C-X-C bond angle might even force the bracketing ring to adopt a different conformation(s) and thus exert unique influences upon the equilibrium. Also, it is perhaps noteworthy that both the nitrogen and carbon X groups possess substituents which may encounter vicinal interactions. As mentioned earlier, non-bonded interactions would be expected to be slightly less in tautomer b and perhaps this is reflected in the enthalpy terms for

95 and 5!+.

The sulfoxides (97 and 98) are significantly different from the other members of the series in many ways. Sulfoxides possess extremely long C-X bonds (see Table VIIl). In this respect, they parallel sulfides and sulfones; however, the C-X-C bond angle is much less in sulfoxides (98°) than in sulfones (107°) or even sulfides (105°). As a result, the effective ring size should be less than the sulfide (96). 97

Sulfoxides differ from the other members of the series in one other very important aspect: the sulfoxide linkage is a highly polarized ps-pd hybridized double bond (>S^+ = 0^~) and may possess as much as one-third of a full charge . 95 In no other case is there charge polar ization of this magnitude.

The result of these differences may be that comparisons of the

sulfoxides with the rest of the series may not be valid. For example, the significant change in the bond angle may cause the sulfoxides to

adopt an entirely different conformation(s). Very little is known

about the conformations of seven- and eight-membered rings because of the ambiguities in the system, i . e . , the multiple conformations

available, the complications introduced by pseudorotation, and the

complex proton interactions in the 1H NMR spectra. It is clear, however, that several conformations are energetically similar.

Hendrickson 96 has calculated for example, that the chair conformation

in cycloheptene (iJ+Oc) was only 1.19 kcal/mol more stable than the boat

co n fo rm atio n (3J+0b) and o n ly 1 .7 kcal/mol more stable than the tw ist- boat conformation (lji-Otb). Others 97 have found that dimethylcyclo-

14 0 c 140b 140tb

heptanones, for example, adopt the twist-boat form preferentially. If

the angle change from nearly tetrahedral (throughout the series) to

98° (in the sulfoxides) were to introduce a major conformational change, 98 the resulting entirely new set of steric interactions would not be expected to be comparable with the rest of the series. The different ratios of tautomers for 97 and 98 could, then, be explicable in terras of steric requirements upon the oxygen atom.

It is difficult, however, to assess the effect of the postulated change in conformation. Throughout this study, subtle changes in molecular parameters have been reflected by substantial changes in the equilibrium distribution of tautomers. It is entirely possible, therefore, that the proposed change of conformation sufficiently explains the observed data. It might be that the second major difference, charge polarization in the sulfoxides, also plays a role. In this regard either a through-bond or through-space interaction could be postulated to accommodate the increased ratio of tautomer b. An explanation based on through bond transmission is not very attractive since, as noted before, the Y carbon experiences no exceptional shifting. A through-space stabilizing interaction may have some merit, however.

The highest occupied molecular orbital (HOMO) in semibullvalene was shown by Hoffman and Stohrer24a to be S 4 while the lowest unoccupied molecular orbital (LUMO) was shown to be S 5 (see below). Interaction of the HOMO with a lone pair of electrons on the heteroatom in the

LUMO 99

HOMO

bracketing ring would not be productive since both orbitals are filled.

The LUMO could in teract at the if- and 6 positions by accepting electron density, but this -would contribute to a weakening of the 3-k and 6-7 n bonds (perhaps favoring tautomer b). The interaction of a positive charge on the sulfoxide sulfur atom would, however, be able to interact favorably with S 4 and provide a stabilizing interaction. Since overlap with the cyclopropane-like orbitals should be better than the overlap with the olefin-like orbitals, tautomer b would be expected to be favored by this interaction. Interaction with the LUMO would, of course, be unproductive since both orbitals are unfilled. The degree of inter action of the vacant orbital on sulfur would be further dependent upon the orientation of the orbital. Hence, the configuration of the sul foxide atom would govern the degree to which tautomer b was favored. In this regard, the axial isomer 97 apparently allows better overlap.

The major objection to the explanation based upon a stabilizing through space interaction is that the amount of actual charge on the sulfur atom is relatively small. Further, and perhaps most importantly, the charged site is a good distance away from the site of proposed interaction. Thus, the magnitude of such an interaction might be quite small or non-existant. In short, the degree to which steric or electronic factors operate in causing the equilibrium displacements observed is unassessable. It is for that matter entirely possible that some other unmentioned factor is the real cause for the relative ordering of the sulfoxides. Clearly, the wide variety of 2,8-pentamethylene semibullvalenes has proven interesting in a number of respects. First, the u tility of 13C NMR in assessing the equilibrium displacements in semibullvalenes has been well documented. Further, the assignment of 13C shifts of heterocycles has' been slightly extended. Most interesting, however, are the subtle responses of the semibullvalene moiety to the varied substituents. In this regard, some insight has been gained into the interplay of ground state equilibrium fo rc e s . F in a lly , some e x p la n a tio n s have been advanced w hich may stimulate further investigations. In short, the study of bracketing effects in annulated semibullvalenes has proven indeed rewarding. 101

F ig u re k. Hie 100 MHz 1H NMR spectrum of 2, 8-trimethylenesemibullvalene 52 in CDC13 at 5°° Hz sweep width.

F ig u re 5. The 100 MHz ^-H NMR spectrum of oxasemibullvalene 6l in CS2 a t 500 Hz sweep width.

F ig u re 6. The 100 MHz 1H NMR spectrum of thiasemihullvalene 6^ in CS2 a t 500 Hz sweep w id th .

F ig u re 7. The 100 MHz ^-H NMR spectrum of azasemibullvalene 62 in CDC1 3 a t 1000 Hz sweep width.

F ig u re 8. The 100 MHz 1H NMR spectrum of azasemibullvalene 25. in CDC1 3 a t 1000 Hz sweep width.

F ig u re 9- The 100 MHz 1H NMR spectrum of oxasemibullvalene 9t in CS 2 a t 500 Hz sweep width.

F ig u re 10. The 100 MHz ^H NMR spectrum of thiasemihullvalene 2*2. in CS 2 a t 500 Hz sweep w id th .

F ig u re 11. The 100 MHz 1H NMR spectrum of 2, 8-pentamethylenesemibull valene in CDCI 3 at 5°0 Hz sweep width.

F ig u re 12. The 100 MHz 1H NT® spectrum of sulfoxide § 8_ in CD2C12 a t 500 Hz sweep width.

F ig u re 15- The 100 MHz 1H NMR spectrum of sulfone 22 in CDC1 3 at 500 Hz sweep width.

F ig u re it. The 100 MHz iH NMR spectrum of sulfoxide 2£ in CD 2C12 a t 500 Hz sweep width. 102

F ig u re 14-

F ig u re 5

Q > —

5 0 4.0 30 2.0 1.08

F ig u re 6 105

NChLPh

F ig u re 7

. 8.0 7.0 6.0 5.0 4.0 3.0 2.0 l.OS F ig u re 8 10k

F ig u re 9

F ig u re 10

5.0 4.0 3.0 2.0 1.0 8 Figure 11 105

H,,H,

Figure 12

Figure 13

5.0 4.0 3.0 2.0 8 F ig u re 1^ RESULTS AND DISCUSSION

Part III. Synthesis and Evaluation of Potentially

Homoarornatic M olecules.

As discussed in the Introduction, the claim that pentaene 5598 'was the first example of a neutral homoarornatic molecule39 met with immediate criticism .45 Paque.tte, et a l . , suggested that the 1H NMR

5 5 b 5 5 e spectrum of 55 (Figure 15) could be indicative of the presence of a diamagnetic ring current and hence, 55-might be better represented by a homoarornatic structure such as 55c, d, or e. Vogel and coworkers45a contested this interpretation, claiming that the spectrum appeared to be that of a molecule more analogous to cycloheptatrienes 5J,911(1 5§_ than to aromatic bridged annulenes 59 and 60. This argument, however, begs

5 8 59 60

106 107

the real issue of how much interaction exists between the orbitals of

carbons 2 and 3 and how much stabilization such an interaction provides.

Clearly, although there might be substantial interaction, it is unlikely

that a neutral homoarornatic molecule would experience the driving force

for delocalization observed for aromatic molecules or for positively or

negatively charged homoarornatic species.

Unfortunately, the 13C NMR and ultraviolet spectra of 55 provide

only substantiation of the gross structure and cannot be used to assess

the degree of delocalization. The most clearcut evidence that 55 is

10-jt homoarornatic comes from the photoelectron spectrum,46 which is

remarkably sim ilar to that of naphthalene. Further, the model which provides the best calculated results is one in which a ll nonbridgehead

atoms are confined to a plane allowing maximum tt o v e rla p .

Thus, while the 1H NMR and photoelectron spectra may argue for a

delocalized structure, there is not an overwhelming accumulation of

evidence which supports this conclusion. Further, there is some

question as to the exact interpretation of the 1H NMR spectrum. Clearly, more information is required so that impartial judgments can be made.

In itial efforts were directed to the synthesis of a suitable model

for pentaene 55_ in its localized form. A good choice for such a model

appeared to be 2,8(k,6)-divinylsemibullvalene lk l (see Scheme XIl).

Because of the rotational freedom anticipated for the vinyl groups,

semibullvalene lk l should not be able to experience significant

delocalization.

Dimesylate 108 appeared to be a suitable precursor for lkjL,

Particularly attractive in this respect was a method developed by Erker 108 and Roth89 who found that excess potassium t-butoxide in dimethylsulfoxide with a trace of water would simultaneously effect the elimination of a sulfonate ester and hydrolysis of a urazole ring.

When this elimination and hydrolysis procedure was applied to 108, oxidation of the resulting intermediate with either eopper(ll) or air surprisingly provided not lh l hut ££!

Although several reaction pathways can he postulated, it is clear that at some point two hydrogens must he oxidatively removed to form the additional double bond. Tiro conclusions may he drawn from this fact: either the oxidation has a very low energy of activation or the product is considerably more stable than the starting material.

Further, it might he possible that this additional stability is due, in part, to the proposed homoaromaticity of 55.

In order to better assess these possibilities, divinyl derivative

2h2 was desired. If lh2 was available, it might be possible to prepare semibullvalene iAl under controlled conditions. The subsequent reactivity of lAlj e.g. , a rapid Cope rearrangement to iM which could then be oxidized to might provide mechanistic insight. Secondly,

1^2 might be induced to undergo a Cope rearrangement to lVj. which could then provide lM- directly. Further, jM would serve as another good model for localized _§£. In short, divinyl derivative Jlj+2 was viewed as a potentially very valuable compound and was the next synthetic goal.

Three different methods were tried in hopes of obtaining JjA2j unfortunately a ll three were unsuccessful. Treatment of 108 with excess

DBU in refluxing THF99 over 53 hr gave only a complex mixture which did not contain the usual >CHN< signals in the 1H NMR spectrum. Activated 109 Br Br Scheme XII

1 4 6 1 4 5

A N ^ O X ‘CBr.

0S02CH3 t-BuOK [o ] OS05CH3 > I / / DMSO- H20

108 5 5 X V

1 4 2 1 4 3 1 4 4

t F P " *

1 4 1 110 aluminum oxide has often been used successfully to effect elimination of mesylates j100 however, treatment of dimesylate 108 with specially ac tivated alumina gave only starting material and oxepane 101 in a 1:1 • ratio. It is ironic that 101, so difficult to synthesize intentionally, would be the only product. The alumina apparently cleaves an oxygen- sulfur bond. The resulting alkoxide then effects an intramolecular

108 1 0 1 displacement of the other sulfonate. Lastly, the synthesis of l42 was attempted via a bis-selenoxide.101 The difficult step in this process was expected to be the syn elimination of the alkyl phenyl selenoxide to form the terminal olefins. Since the use of o-nitrophenylselen- oxides is reported to improve the yield of these reactions and drastically reduce the reaction time,1018, dimesylate 108 was treated with two equivalents of o-nitrophenylselenide anion in absolute ethanol and tetrahydrofuran. The 1H NMR spectrum of an aliquot of the reaction mixture indicated that bis-selenide l4j_ had been formed since the trip let at 6 4.37 (J = 6.5 Hz, 4, -CH2OSO2CH2) and the singlet at 2.97

(6, CH3SO3-) were both absent and a new trip let had appeared at 3-03

(J = 7.0 Hz, 4, -CH2Se-). Also, the aryl region was now a multiplet at 8.20 (2h) and a singlet at 7«50 (Uh). Unfortunately, oxidation of the selenide and subsequent selenoxide decomposition gave rise I l l to a complex mixture which could not he purified. Since none of these unique methods was successful, attempts to synthesize lk 2 w ere abandoned.

It is interesting to note in passing that semibullvalene 1^8_is not readily oxidized to the benzo analogue of 55 (1^-9) - When a solution of

1 4 8 1 4 9 ljj-8 in CDC13 was exposed to a stream of oxygen, only products of benzylic peroxidation and/or substrate decomposition were observed.102

Since small amounts of pentaene 55. were available from the elimination, hydrolysis, and oxidation of dimesylate 108_, a brief examination of the relative reactivity of the various vinylic sites was made. If dibromocarbene were to add to the 6,7 double bond as indicated, ring opening of cyclopropyl dibromide 1^ followed by removal of the bromine atoms would give 1^6., a potentially bishomoaromatic molecule.

As expected, however, dibromocarbene addition was not selective and multiple products were obtained although the reaction was unexpectedly slu g g is h .

Another approach which might be expected to yield definitive information about the nature of 55 would be the preparation of a number of derivatives in which the bridging ring has been modified. Hopefully, one of these derivatives would be very sim ilar to 55 5 yet crystalline so that the distances between carbons 2 and 3 and carbons ka. and 8a 112 could be determined by x-ray crystal structure. These distances are crucial since these carbons must be close enough to interact if the molecule is to be 10-jr or 6-71 homoarornatic. Further, if these derivatives are chosen properly, then, whether they are crystalline or not, meaningful information about the character of the system could be ascertained. One such derivative 1^6 has already been alluded to. This homologue, in addition to being interesting in its own right, could be viewed as the precursor to the cationic homologue of 55 5 i. e ., 150.

1 4 1 5 0

Another useful derivative would be the tropone-like homologue _151.

Should homoarornaticity be of importance, the delocalized representation

151a could prove to be the stable form. Also the tropolone-like

1 5 1 15 1a derivative 152 could provide valuable insight. Here any one of three classical tropolone-like structures (152 a-c) could be envisioned or the delocalized form _l£2d could be possible. Further, protonation of 152 could give a cationic homologue of In short, the preparation 113

.0,

15 2a 15 2b 1 5 2 c

OH H OH 1 5 2

1 5 2 d 1 5 3

of suitable derivatives, only a scant few of which have been mentioned,

appeared to be a very rewarding objective.

Since a great number of derivatives would be desired, the most

efficacious synthetic method would, based upon past experience, require

a suitable common intermediate. Dialdehyde 15^ was considered to be a prime candidate for this role since condensation of 15^ with a variety

of substrates could be envisioned as giving rise to a host of desirable

derivatives (15 5) •

CHO y-N ;ho + x

1 5 4 1 5 5 Ilk

Although dialdehyde l^jt is a worthy synthetic goal, all of the attempts to synthesize it, as outlined in Scheme XIII, have been unsuc cessful. Ideally, diene 8ja might be treated with excess ozone fol lowed by a reductive workup to give 15^ directly. The most straightforward method involved low temperature (-78°) ozonolysis of 87a in a methanol-dichloromethane (2:l) solution followed by reduction by dimethylsulfide.103 Unfortunately, as Pappas and coworkers pointed out later,104 dimethylsulfide begins to reduce hydroperoxides between -20 and -10°. In this case, the proximity of the reaction sites allows cyclization of the reaction intermediates, e ^ ., lgj, to form 1,2- dioxanes, e^g., 158, which are inert to further reduction. This type

Ph Ph ks/.O / ( / /

1 5 7 1 5 8 of reaction, believed to be acid catalyzed, would be expected to occur readily as a result of the sm all amount of hydrochloric acid which is produced as a by-product from decomposition of the dichloromethane.104

Since a chlorinated solvent is necessary to dissolve 87a at low temperatures, another modification was attempted.

The work of Pappas, et_al. ,104 also demonstrated that triphenylphosphine reduced hydroperoxides below -50°. Application of this method produced some positive results as there was evidence of aldehyde-like 115

Scheme XIII

OsO, NalO, 8 9 a

1) 0 ; .CHO CHO

I^DMSO 1 NCS A* N aH C0 3

PhS“

1 5 6 92 116

resonances in the 6 9-10 region of the 1H NMR spectrum of the crude product. Also, triphenylphosphine oxide, the expected by-product, was

isolated. In fact, this proved to be a severe problem as it was

impossible to separate the product from the phosphorus compounds.

Further, the phosphorus compounds interfered with subsequent reactions

o f 1£4.

Reduction of the ozonolysis intermediate(s) with zinc and acetic

acid105 sim ilarly appeared to give a low yield (< 30$) of aldehyde l|?j+, but again it was impossible to purify 154 sufficiently to perform

subsequent reactions. It was also not possible to isolate a hydrazone

derivative from this mixture.

One additional attempt was made to obtain 13ft- from the cleavage of

diene 8ja. Treatment of 8ja with a catalytic amount of osmium tetroxide

in the presence of excess sodium metaperiodate106 was expected to give

154. However, this reaction also gave a number of products and purification and/or further reaction of 154_ was impossible.

Since a variety of methods based upon the cleavage of 8ja had proven unsuccessful, this synthetic approach was discontinued in favor of methods which relied on the oxidation of previously isolated inter mediates. The first such reaction attenpted was the Collins oxidation107 of diol 89a. Unfortunately, the only product isolated

(60$ yield) was apparently lactone 137 which was characterized by its

IR (2900, 1775, 1710, 1400, and 1275 cm-i) and 1H NMR C (6, CDC13)

7.15-7.45 (m, 5, aryl), 5-45 (m, 1, >CHN<), 5-15 (m, 1, >CHN<), 4.40 » ” (s, 2, -CH20-), 2.60 (s, 1, cyclopropyl), and 2.0-2.4 (m, 3, cyclopropyl)] spectra. Since this product could well be expected for a 117 Ph

1 5 7 variety of sim ilar oxidations of 89a, this approach was also abandoned.

An equivalent method of oxidation would be a modified Kornblum oxidation on dimesylate through the agency of sodium bicarbonate and dimethylsulfoxide.108 Workup of this reaction, however, gave a low yield of a multicomponent mixture which was not characterized.

One last attempt involved a method for aldehyde synthesis which had shown some promise in these laboratories.109 Basic to this method was the observation that hydrolysis of an o'-chloro sulfide gives alde hydes in good yield. Accordingly, the bis-thiophenyl ether 156 was prepared by treatment of dimesylate 92. with excess thiophenoxide in

THF. Thioether 156 was characterized only by 1H NMR [ (6, CDC13)

T.00-7.70 (m, 15, aryl), 5-22 (q, j = 2.0 Hz, 2, >CHN=:), 5.05 (ABq,

Jj^q = llt-.O Hz, AvAB = 10.6 Hz, 1)-, -CH2S-), 1.92 (m, 5» cyclopropyl) and

1.52 (m, 1, cyclopropyl syn to thioethers)] since subsequent chlorination with N-chlorosuccinimide gave a complex mixture of products. Attempted hydrolysis of this crude product was unsuccessful.

Although there is an almost endless number of oxidation or cleavage methods which might give 15^, it was felt that the difficulties encountered with the variety of approaches attempted might be the rule rather than the exception. As a result work in this area was Xl8 discontinued in favor of another synthetic strategy which proved to he rewarding. In this approach (outlined in Schemes XIV-XVl), functionalization of diazasnoutene 68 was intended to lead to a variety of methoxy derivatives of 55. Further, the methoxy substituted double bond should now be the double band most reactive toward electrophiles, e.g., carbenes. Presumably, the gem-dihalocyclopropane derivatives could be transformed to a series of tropone- and tropolone-like molecules, e. g. , 151 and 152. Further, one or more of the derivatives of 55 might be crystalline; certainly all of the proposed derivatives could provide insight into the nature of §5,.

As mentioned earlier, photooxygenation of 68_ had been shown83 to produce syn-allyl alcohol 120 stereospecifically in nearly quantitative yield based on unreacted 68 (see Scheme XIV). Further examination revealed that this photooxygenation—and all others in this study— could be facilitated by the use of rose bengal on polymer support

(Photox™) as the sensitizer.110 Previously, removal of the rose bengal or methylene blue sensitizer required careful purification of the product, often by chromatography. Phot ox™, on the other hand, can be removed by simple filtration. Further, any solvent which swells the polymer support may be used.

A llylic alcohol 120a was easily oxidized to enone 158a in 80$ yield with activated manganese dioxide. Surprisingly, only manganese dioxide prepared by Attenburrow's method69 gave consistent results.

Neither commercial manganese dioxide (either direct from the container or activated by heating under vacuum) nor manganese dioxide on activated carbon111 gave consistent results. An even better method of 119

Scheme XIV

’0 .

. 1 20a, R=Ph 1 5 8 a . R=Ph 6 8 a. R = Ph b, R=CH3 b, R=CH3 * * b, R=CH3

OCH;

162 166

OCH OCH

163 167

1 64. X=H 1 5 1 , X=H 165, X= OH 1 5 2 , X=OH oxidation in this system, is the use of Collins reagent.107 With Collins reagent, alcohol 120b was oxidized to enone 158b in 93 % y ie ld . As before, since enone 158a had been fully characterized by IE, XH NMR, and mass spectra and combustion analysis, methyl congener 158b was identified by spectral sim ilarities to 158a. Particularly noteworthy is the 1H NMR spectrum of 158b in CDC13 which in part contains a doublet a t 6 7.11 ( j AB = 1 0 .0 Hz, 1 , -CH=CHC0-) and a d o u b le t a t 5-7 7 (JAB =

10.0 Hz, 1, -CH=CHC0-). The >CHN< protons are slightly non-equivalent appearing as a doublet at 5*17 (J = ^.0 Hz) and a doublet ip.97 (J =

5.0 Hz). Lastly, the methylene protons comprise an AB quartet at 2.88

(J ab = 18.5 Hz and Avab = 1°*3 Hz).

Since enone 158a was available, a potentially profitable offshoot was explored. The successful use of the m esityl group as a ring current probe112 suggested possible applications in the pentaene system. If a suitably substituted pentaene, for example lb l, could be synthesized, then the existence of a ring current would cause deshielding of the ortho methyl groups, whereas the para methyl group which lies outside the ring current would act as the internal standard. Further, diene

IbO could serve as a model for a localized system for calibration purposes. Synthetic access ms envisioned through 1,2 addition of m esityllithium 113 to enone 1§j3 followed by dehydration of the allylic alcohol (see Scheme XV). Unfortunately, the enone is unstable in strong base and undergoes decomposition rather than addition. Parallel in vestigations114 showed that phenyllithium added to 158a to give 159b in very low yield. This fact, coupled with subsequent results in other systems which indicate that 158a is too hindered to react favorably Scheme XV

.OH Ar Li a

159a. Ar=Mesityl t 6 0 b, Ar=Ph

1 6 1

with mesityllithium, suggested that further investigations were not m erite d .

Enone 158b was smoothly converted to methoxydiene 162 in 90$ yield using a large excess of trimethyl orthoformate. The reaction, catalyzed best by oxalic acid, was very sensitive and gave other unidentified products if lesser amounts of the orthoester were used. Enol ether 162 was characterized by its spectral properties and by elemental analysis.

The 1H NMR spectrum of 162 is particularly unique. The vinyl region consists of a doublet at 6 6.20 [J^p = 9*5 Hz, 1, -CH=CHC(0CH3)=CH-], a doublet of doublets at 5-TO [J = 9- 5 and 2.0 Hz, 1, -CH=CHC(0CH3)=CH-], and a doublet at 5-05 [J = 2.0 Hz, 1, -CH=CHC(0CH3)=CH-]. The 122 cyclopropyl protons are found at 2.37 as a multiplet (ill), at 2.00 as a multiplet (2h), and at 0.63 as a doublet (j = 3»5 Hz, Hi). The upfield doublet, the result of the cyclopropyl proton syn to the cyclohexadiene ring, is located only 0.0b ppm downfield from its location in the unsubstituted diene 87b.

It was thought that introduction of a gem-dihalocyclopropane functionality might be easier at this point since lb2 possesses only two double bonds, one of which is activated toward electrophiles.

Adduct 16^ could then provide access to tropone- and tropolone-like derivatives l64 and 165. Treatment of lo2 with Seyferth's reagent,115 phenyl(tribromomethyl)mercury, in refluxing benzene did not give the expected 163 but rather produced enone 158b in ca 60$ yield. Apparently the cyclohexadiene ring is so congested—on the anti face by the urasole ring and on the syn face by the cyclopropyl proton—that the carbene undergoes ether insertion. The resulting enol ether is apparently

162 hydrolyzed upon workup to return enone 158b. Clearly, the cyclopropyl functionality would have to be introduced at the pentaene stage.

Enol ether 162 was hydrolyzed and oxidized in the usual manner to give the 6-methoxy derivative of pentaene 55* The yield for this sequence was 83$, much better than that normally observed for the preparation of semibullvalenes. Ether l6b was an air sensitive, bright 123 yellow, low melting (23-27°) solid. Spectral properties, particularly the NMR and mass spectra, served to characterize 166. Its UV spectrum

[X^ ane 327 (e k. 0 x 103) and 2^7 nm (5.b x 104)] also compares well with that of [xjfj°ctane 335 (e 1.8 x 103) and 239 nm (2.1 x 104)].

Although 166 was unsuitable for x-ray crystallography, th e 1!! and 13C

NMR provided valuable insight into the nature of these pentaenes. For purposes of clarity, these spectra w ill be discussed together with the spectra of the 5-methoxy derivative 168 later in this section.

Enol ether 166 was valuable in its own right, but it was also viewed as the precursor to tropone- and tropolone-like systems 15_1 and

1§2. The steric hindrance on the anti face that seemingly diverted the carbene reaction of precursor 162 was now absent. It was also hoped that the methoxy substituent would provide sufficient enrichment so that the 6,7 double bond would be preferentially attacked by a carbene.

Treatment of 166 with dibromocarbene generated either from Seyferth's reagent115 or from bromoform-t-but oxide surprisingly gave little or no reaction. Ether 166 was recovered essentially intact, only slight decomposition being evidenced. It will also be recalled that reacted only slowly with dibromocarbene. It is not known whether this absence of reactivity is due to the homoaromatic delocalization of 55_and 166; certainly, this lack of reactivity is not that expected of normal pentaenes! In any event, this pathway was unsuccessful.

The next synthetic objective was the other positional isomer of

166, the 5-methoxy derivative 168. By analogy to the previous synthesis, the key intermediate was expected to be enone 16JJ (see Scheme XVI).

Since l6j9 is structurally related to enone 158b, the approach immediately suggested was a ketone transposition sequence. 124

Seheme XVI R

87a. R = Ph 121a. R=Ph 1 73 b, R=CH b, R=CH3

1 6 9 C r 0 *2py (68b) 3

N v^O

H20 2 N2H<

K2C03 H 0 Ac

1 7 0 17 2 1 5 8 b

OCH HC(0CH3); 1 6 9 2) MnO T sO H 168 1 75 125

With this objective in mind, enone 158b was treated with alkaline hydrogen peroxide in ethanol116 to give epoxyketone 170 in 75$ yield.

Although the stereochemistry was hot rigorously proven, the epoxide would be expected to be anti to the cyclopropane ring by analogy to previous reactions. Spectral data and satisfactory combustion analysis served to prove the gross structure. The 1H NMR spectrum of 170 in

CDC13 was particularly helpful in this regard. The bridgehead (>CHN'<) protons are non-equivalent and appear as doublets (j = 3.0 Hz) at

6 5.IT and k. 87. The methine proton o' to the ketone is also a doublet

(j = k.5 Hz) at 3-63 while the methine proton P to the ketone is located with the N-methyl signal as a broad singlet at 3.13* The methylene protons are also non-equivalent, as expected, and appear as an AB quartet (j^g = 15.0 Hz and Av^jb = 26.0 Hz) at 2.88. The cyclopropyl proton syn to the cyclohexanone ring would be expected to be shielded

« (the corresponding proton in IJjL appears at I.36117) if the epoxide

1.36 1 7 1 ring were syn to the cyclopropane ring; however, this is not the case as a ll four cyclopropyl protons comprise a m ultiplet at 2.00-2.20, The epoxide is, thus, tentatively assigned as anti. 126

Treatment of 170 with hydrazine hydrate and acetic acid118 in

refluxing dioxane gave the expected allyl alcohol (172) in low yield

(30$ based on recovered 170). The allyl alcohol was not characterized

hut was oxidized immediately with Collins reagent.107 Although trans

posed enone 169 was obtained in 91$ yield, the low yield on the previous

step made this method unattractive.

An alternative method appeared to he transformation of epoxide 173

to enone lbg. Epoxide 1J3 was expected to he available in turn from

the treatment of endoperoxide 121 with an alkyl phosphite.119 Endo-

peroxide 121b was formed in quantitative yield from diene 87b by photo

oxygenation83 using Photox®^ as sensitizer.110 Because 121b is thermally

unstable, characterization was made by the 1H NMR spectrum in CDC13

[(6, CDCI3) 6.12 (t, J = 4.0 Hz, 2, vinyl), 5-10 (t, J = 3.0 Hz, 2,

>CHN<), 4.93 (t, J = 4.0 Hz, 2, allyl), 3.00 (s, 3, methyl), 1.80-2.40

(m, 3S cyclopropyl), and 1.58 (d, J = 4.0 Hz, 1, cyclopropyl syn to

cyclohexene ring) ]. Since thermal rearrangement of 121a has been shown

to give 174a, the structure of which was proven by x-ray crystal

lography,83 the endoperoxide must be anti to the cyclopropane ring.

1 74a Thus, the shielding of the cyclopropyl proton syn to the cyclohexadiene unit must be due to the olefin moiety. Unfortunately, treatment of 121b with either triphenylphosphine102 or trimethylphosphite gave only endoperoxide 174b and at most a trace of 173 and this approach was deemed im practical.

An alternative method was the direct allylic oxidation of diazasnoutene 68b. In this regard, selenium dioxide or Collins reagent107 seemed to be the reagents of choice. Treatment of 68a with selenium dioxide in dimethoxyethane117 surprisingly did not give the phenyl congener of either allylic alcohol 172 or enone 169, but rather gave diene 87a, presumably via dehydration of the intermediate alcohol.

The chromium trioxide-pyridine complex appeared to be only slightly better, giving 169 in low yields.117 However, reinvestigation of the reaction revealed that with careful processing of the inorganic salts

(see Experimental Section) enone 169 • could be obtained in 43$ yield based on recovered 68a. Since this reaction could be carried out on large scale, it became the method of choice. Enone I69 was fully characterized by spectral methods and elemental analysis. The XH NMR of 169 is particularly helpful in determining its structure. The vinylic proton (3 to the ketone, coupled to the adjacent methylene protons, appears as a doublet of triplets (j = 4.0 and 10.0 Hz) at

6 6.6l while the vinylic proton <2 to the ketone is a doublet of triplets

(j = 2.0 and 10.0 Hz) at 5-95* The bridgehead protons (>CHN<) are non-equivalent and appear as doublets (j = 4.0 Hz) at 5*70 and 4.9T-

The methyl group is a singlet at 3-07 while the methylene protons appear as a doublet of doublets (j = 2.0 and 4.0 Hz) at 2.87. In this case, the enone functionality deshields the syn cyclopropyl proton which appears as a doublet (j = 4.0 Hz) at 2.4-2 while the remaining cyclo propyl protons constitute a multiplet at 1.90-2.40. 128

Enone l6g_ was smoothly transformed to enol ether (68%) with excess trimethyl orthoformate and tosyl acid as a catalyst. The

structural assignment was verified by combustion analysis and spectral data. The 1H KMR of 1J5, in CDC13 is again very distinctive. The vinylic region is now an ABX pattern. The AB portion, at 6 5.UO-6.10, is very similar in appearance to the AB portion of an ABX pattern with

AvAb = 10 Hz’ Jab = 10-0 Hz> Jax = 1 ,0 Hz> and is x = ® Hz.72 This assignment nearly parallels the values expected for this system. The non-equivalent bridgehead protons (>CHPT<) appear as a m ultiplet at 5-60 and as a m ultiplet combined with the X proton of the ABX system at

1*-.9°-5*20. The 0-methyl and N-methyl groups comprise singlets at 3 .6 7 and 3-00; respectively. The cyclopropyl protons are also non-equivalent, appearing as a multiplet at 2.20-2.60 (1H), a multiplet at 1.80-2.20

(2H), and a doublet (j = ^.0 Hz) at 0.90. The cyclopropyl proton syn to the cyclohexadiene ring is not as shielded in this case (O. 9 0) a s i t is in positional isomer l62 (O.63) or in diene &£b (0. 57). This could be ascribed to the fact that the methoxy substituent is now closer to the proton and exerts a small deshielding effect.

The 5-me'thoxy derivative of pentaene 5£ was obtained in 72% yield from the hydrolysis and oxidation of enol ether ljjj,. Ether 168 was a bright yellow, extremely air-sensitive, low melting (^ 5- 50°) solid and was unfortunately not suitable for x-ray crystal structure determination.

The IR spectrum (v1®1, l6l0, 1235, 1205, 1155 , 795, and 7^5 cm"1), the UV max spectrum [\isooctane ^ g (e p#8 x 103), 262 (3-7 x 104), and 238 nm max (2.9 x 104)], and the accurate mass are helpful in assigning the structure of 168. The UV spectrum of 168 compares particularly well 129

with those of 6-methoxy derivative 166 and parent pentaene J55. The ^-H

and 13C NMR spectra provide considerable insight into the nature of 168

and are discussed later. Since attempted carbene additions to l66 and

52 had proven unproductive, a similar attempt was not made in this case.

Although _55, 166, and 168 form a complete series of compounds, the

5,8-dimsthoxy derivative 176, which should be readily available from

endoperoxide 12Tb, would be a worthwhile addition since it might be

more stable and perhaps more crystalline than 168.

As outlined in Scheme XVII, the base induced transformation of 121b

to eneketol 17£ was undertaken as a possible route to the key inter mediate l8l . Previous results117 had indicated that basic aluminum

oxide120’121 would effect the desired reaction in moderate yield1, however, as it commonly the case with this reagent, the reactions varied from batch to batch and could not now be repeated. As an alternative, treatment of 121b with triethylamine119a was slightly

successful and a small amount of 177 was observed; however, as before, thermal rearrangement to bisepoxide 17^-b was the dominant course of the

r e a c tio n .

Another approach would be catalytic hydrogenation121’122 of 121b to give diol 178 which could be oxidized to the intermediate diketone

l8l. Platinum catalyzed hydrogenation of 121b at a slight positive hydrogen pressure gave a good recovery of product; but, by 1H NMR analysis, overreduction had occurred. Further, treatment of the crude product with Collins reagent107 gave no reaction.

With this avenue of approach frustrated, one last method was attempted. Reduction, of 121b with sodium borohydride121 gave a 130

Scheme XVII

1 8 2 183

N s^O J \ o

1 21 a, R = Ph NaBH b,R=CH 3

CHs Nv^O OH OH

179 OH OH 178 DDQ | or CrO ‘ f 2 py

OCH

OCH 1 76 181 180 quantitative recovery of enediol 17§,* The structure of 1J9. was assigned on the basis of IR, mass, and 1H NMR spectra. The XH NMR

spectrum of 179 in CDC13 is particularly definitive: 6 6.18 (dd,

J = 3-0 and I4-.0 Hz, 2, vinyl), 5-18 (dd, J = 2.0 and 3.0 Hz, 2, >CHN<),

5-00 (dd, J = 3-0 and 4.0 Hz, 2, >CH0H<), 3-10 (br s, 2, -oh), 3-05

(s, 3, methyl), and 1.80-2.20 (m, 4, cyclopropyl). Attempted catalytic hydrogenation of 17)9 did not give 178, but rather returned a complex mixture as the result of overreduction. Surprisingly, enediol 179 was extremely resistant to .oxidation to diketone l80. Treatment of 179. with Collins reagent107 gave no reaction while reaction with 2,3- dichloro-4,5-dicyanoquinone123 gave no recognizable products. Since 176 was only desirable and not absolutely necessary, efforts in this area were terminated.

One other possibly interesting off-shoot was explored, however.

Since 174b was readily available, it would be interesting to see if

semibullvalene 182 could be obtained with the thought that the epoxide rings might be induced to open to give the potentially homoaromatic

(l4it) diether 1&3_. The usual hydrolysis route, with hydroxide in

•refluxing 2-propanol, was considered to be too vigorous for the epoxide functionalities, so the milder t_-butoxide-DMSO-water method89 was attempted. Subsequent addition of activated manganese dioxide69 did not produce the expected gas evolution, however, and no identifiable products were isolated. Since even this milder hydrolysis procedure was apparently too severe for the epoxide functionality, this avenue was abandoned. 132

XH and 13C NMR Spectra. With the successful synthesis of methoxy derivatives l66 and 168, attention was directed to a quantitative evaluation of the degree of homoaromaticity present in these pentaenes.

For purposes of comparison, the 1H NMR spectrum of pentaene £5. i s shown in Figure 15. Protons 5-8 could he designated as AA'BB', but since it

5 5 was expected that J5}6 = J j}q, the pattern should simplify to an A2B2.

Since protons 1-2+ are insulated from the rest of the system, analysis of the spectrum is reasonably straightforward and is in agreement with the information gained by spin decoupling.45a’ 50 At highest field (1.76 6) is the doublet due to Hsh (J2a,sb = 7-^ Hz). Next is the doublet of trip lets from H2a (J>a,sb = 7-1*- Hz, J2,2a = £2a,3 =2.1+ Hz) at 3.76.

Spin decoupling of H2a verifies the assignment of Hsh an<3 reveals that the doublet of doublets at 5*97 is due to H2 and H3 (J^}2a = £ga,3 =

2.1+ Hz and J l j 2 = £3,4 = 5*3 Hz). Irradiation of H2a also sharpens the m ultiplet at 6.30-6.1+0 by removing long range coupling with Hi and H4 and thus establishes their location. The A2B2 portion is deceptively simple, but apparent couplings can be discerned. The lowest field doublet of doublets at 6.61+ i s seen to be from H6 and H7 with £6,7 =

1 1 .0 Hz and £5 j6 = £6 ,7 = 6.2+ Hz. The rem ain in g p ro to n s H 5 and Hs are part of the m ultiplet at 6.3O-6.I+O. 133

With these assignments in mind, the spectrum of 6-methoxy derivative 166 may he analyzed. Although 166 does not contain the plane of symmetry which is present in 55, many of the equivalent

2 3 ab

1 6 6

couplings, i.e ., Jij2 = J3,4, should still he present and the values of many of the coupling constants should he similar to those seen in ^ for the corresponding protons. Hie upfield doublet (6 2.03) reflects the absence of symmetry as it is broadened slightly. This signal is clearly from Hsh with J2a,ab = 7*5 Hz. The singlet at 3-68 is due to the methoxyl substituent. The signal at 3*90 from H2a reveals the slight non-equivalence of H2 and H3 since it is a well split multiplet compared to the clean doublet of triplets observed for H2a in 5J5. Spin decoupling of the peak at 3*9° serves to confirm the assignments of

H2a and Hsb and to locate H2 and H3. H2 is the doublet of doublets

•(j.2,2a = 2.7 Hz and J ij2a = 5*2 Hz) at 5-93 while H3 appears as a doublet of doublets- (Jaa,3 = 2 .7 Hz and £3,4 =5*2 Hz) at 6.11.

Irradiation at 3-90 also removed the long range coupling of H2a with Hi and H4. Thus, Hi is found to correspond to the doublet of doublets at lowest field (6.^3) with J i,2 = 5*2 Hz and J i,2a =2.7 Hz while H4 is the doublet of doublets (J2a,3 = 2 .7 Hz and =5*2 Hz) at 6.39* 13^

Of the remaining vinylic protons, H5 is expected to he only slightly coupled and is, in fact, a broad singlet at 6.2k. The other two protons should be strongly coupled with each other and should also exhibit small couplings to H5. Thus, H7 is the doublet of doublets

(J7 js = 6.5 Hz and J s j7 =1-5 Hz) a t 6.0k while Hs is the doublet of d o u b le ts (J7,8 = 6.5 Hz and J s 5s = 1*5 Hz) a t 6 .JO. The r e l a t i v e a s signments of some pairs of protons, e. g. , H7 and Hs, were made on the assumption that the major differentiating factor w ill be the resonance contribution of the methoxyl group which would be expected to shield protons in conjugation with the oxygen group as illustrated by resonance form A. As shown in the semibullvalene .study (Part II), the

A inductive effects of a substituent fall off rapidly with distance.

The spectrum of the third member of the series (l68) is also fairly complex as shown in Figure 16. The upfield doublet (J2a,ab =

OCH3

1 6 8 7*5 Hz) at 1.85 6 is slightly broadened as a reflection of the lack of symmetry of 168 and the resulting non-equivalent couplings. The methoxyl singlet is located at 3-6^, nearly the same position as in the spectrum of 166. The multiplet due to H2a is ^ess well defined than in

166. In this case, H2a is located at 3*75 with J^a,ab - 7*5 Hz and

J2,2a Jsa,3 ^ Hz. Hpin decoupling at 3*75 again verifies the assignments of H2a and Hat end serves to locate H2 and H3. The doublet of doublets = 5>5 Hz and J2a,3 = 2.0 Hz) at 6.10 is due to H3 while the doublet of doublets (J ij2 =5-5 Hz and J2a,3 = 2.0 Hz) a t

6.1(-3 is the result of H2. The remainder of the spectrum is fairly straightforward after it is realized that protons 6, J, and 8 should comprise an ABX pattern. The X portion (Hs) is clearly visible as four nearly equal intensity lines at 5.83. The AB portion (Hy and Hg) is superimposed upon another signal ( H 4 ) at 6.5O-6.9O, but still appears very sim ilar to an ABX pattern with Av^g = 5-0 Hz, Jab = 8,° ’

J-py = b. 2 Hz, and = 1.8 Hz. That leaves Hi as a doublet (Jij2 =

5.5 Hz) at 6.^3* As before, arguments regarding the substituent effects of the methoxyl group were used to make some of the above assignments.

The'salient feature of these three spectra is their great sim ilar ity. This is particularly striking when the data are viewed col lectively (see Tables XXIII and XXIV). The sim ilarity of the coupling constants would imply that the molecular framework including the bonding pattern is much the same in all three compounds. The sim ilarity of the chemical shifts is somewhat perplexing although it is particularly difficult to sort out the various diamagnetic and paramagnetic effects.

It is, therefore, nearly impossible to make definitive statements 136

Table XXIII. Summary of 1H NMR Dataa—Chemical Shifts of Jjj£, 1 66 and 168.

P ro to n 55 166 168

1 6 . 3 0 - 6 .4 0 6.43 6.25

2 5-97 5-93 6.43

2a 5 -7 6 3 .9 0 5-75

5 5-97 6 .1 1 6.10

k 6 . 3 0 - 6 .4 0 6 .3 9 6. 50-6 .9 0

5 6 .3 0 -6 .4 0 6 .2 4 —

6 6 .6 4 — 5.83

7 6 .6 4 6.04 6.50-6.90

8 6.30-6.40 6.30 6 . 50- 6 .9 0

8b 1 .7 6 2.03 1.85 aIn CDC13 (100 MHz) at 40° referenced to internal TMS (6).

regarding the presence or absence of a ring current on the basis of the

XH NMR spectra of these three- compounds.

The in itially reported39*1 l3C NMR of 55, contained an incorrect assignment of the quaternary carbons 4a and 8a as pointed out in an independent investigation.45*1 The average values of these studies

(except for 4a and 8a) are reported in Table XXV along with the values determined for 166 and 168. 137

.Table XXIV. Summary of XH NMR Data3,—Coupling Constants for 166, and 168.

J , cps 55 166 168

1 ,2 5-3 5-2 5 -5

2 ,2 a 2 .h 2 .7 2 .0

2 a , 8b 7 -^ 7-5 7 -5

5 ,6 6.1^ 6.5b '

5 ,7 1-5

5 ,8 1 .5

6 ,7 1 1 .0

a I n CDCI 3 (1 0 0 MHz) a t 40°^ b J 7 j 8 .

The SPORD spectrum of 166 and the model values for 55, may be used

to assign the resonances in the 13C NMR of 166 in CDC13. The methyl

and quaternary carbons are easily distinguished in the SFORD spectrum.

The quaternary carbon bound to the methoxyl group must be the signal at

lowest field (158.7). The other two quaternary signals are assigned

as shown on the basis of a slight shielding of Cga which might be

expected from a resonance contribution of the substituent. The re-

maing signals are all due to carbons bearing one hydrogen. The two

signals at higher field are assigned as C2a and C8b on the basis of the

model values. Of the remaining vinylic carbons, the signal at highest

field must be C7, the carbon which should experience the greatest

shielding due to the substituent resonance. By analogy to 55, C2 and Table XXV. Summary o f 13C NMR Data (22.6 MHz, CDC13, ambient temperature) for 166, and 168.

Compound Ci c2 C2a c3 C4 C4a C5 Cs Cy Ca Csa Cab

51 1 3 3 .7 ° 11 5 .0 55-2 1 1 5 .0 133. 7° 136 A 131. oc 127.5 127.5 131. 0C 136 A i+6-3

l6 6 a 1 3 6 .7 ° l l l . l d 55-6 113. 9 d 132. 8e 139-9° 131. oe 1 5 8 .7 1 0 ^ .2 1 3 1 .1 ° 133-7° 1+6.3

l6 8 b 1 3 1 .1 ° 12 3 .1 5 5 .1 11 3 .8 1 3 6 .1 ° 12 1 .6 A 5 .7 128. 6° 129-7° 130. 1° A l.O ^3-3 a55-1—methyl-, b60. 0—methyl-, "^Interchangeable assignments; ^Interchangeable assignments; eInterchangeable assignments. 139

C3 must tie at 111.1 and 113.9 P P m , respectively. These assignments are also based on the slight shielding expected for C2. The remaining vinylic resonances are possibly interchangeable, but are assigned as indicated to achieve the best agreement with the values for 55..

The assignments for 168 are made using many of the same arguments.

The methyl carbon is again distinctive as are the quaternary carbons

(C2a, C8a, and C5). Carbons 2a and 8b are nearly unchanged from their locations in and Po6. Two vinyl carbons are at slightly higher field and are, by analogy to 55. and 166, assigned as C2 and C3. The remaining signals are due to Ci, C4, Cs> C7, and C3 and are inter changeable for a ll practical purposes.

It is generally agreed124 that the shielding constant a f o r a particular nucleus is approximated by the sum of three terms:

CT = Ofl + a p + o' where cr^ is the diamagnetic contribution, Up the paramagnetic con tribution, and o ' the contribution from neighboring groups/nuclei.

The paramagnetic term is the dominant factor for most 13C shifts.125

Since this term primarily reflects charge polarization, variation in bond order, and average excitation energy,124 the use of 13C chemical shifts as a probe of the electronic nature of aromatic systems has received wide use. In particular, excellent correlation of chemical s h i f t s w ith Hammet o constants has been achieved in 2- and 3-substituted pyridines.126a’^ Others1260’^’e have found excellent correlations for substituted benzenes such that, barring excessive steric interactions, etc. , the substituent effects in polysubstituted systems are additive. This is not necessarily the case of heterocyclic aromatic compounds where there exists a higher degree of bond fixation.124^ The shifts of furan (1%) and pyrrole (185) are noted below as examples. The.

1 09.6ppm 108.2

1 4 2.6 11 8 .5 0 N

1 8 4 1 8 5

situation is complicated further when substituents are introduced at the 3-position as in thiophene and 3 -methoxythiophene (1§£) as

shown be low. Here the effect of the methoxyl substituent is trans

1 8 6 1 8 7 1 8 7a

mitted unegually. Carbon-2 is shielded by 27-5 ppm while the other adjacent carbon (4) is only shielded by 6.8 ppm. This is nearly what would be expected in the absence of aromatic delocalization and if the substituent were acting only via resonance form 18.73.

It is in this context that the 13C NMR data for 55.? l66j and 168 must be viewed. The only appreciable substituent effects seen in 166 are in the shift of C7 which is shielded by 23-3 ppm and in the shift

of C6 which is deshielded by 31*2 ppm. In 168, a similar result is

observed as C4a is shielded by it-. 8 ppm and C5 is deshielded by 1^-7 ppm HM

It would appear then that the only substituent effects noted are for the carbon directly attached to the substituent and for the q? carbon which can undergo a resonance interaction with the oxygen lone pairs.

It is difficult to interpret the real meaning of such data. It would seem that the other a carbon should e x p e rie n c e some s h ie ld in g i f the bonding were truly delocalized. However, this must be tempered with the fact that the generally recognized aromatic thiophene derivative

1&5 exhibits a substituent effect largely in only one direction. It cannot, therefore, be unequivocally stated whether or not there is delocalization in 55, 166., or l68.

In summary, this study has explored some possible routes to new potentially homoaromatic molecules. In all cases, diazasnoutene 68

served as an adaptable synthetic intermediate which allowed the exploration of a wide variety of paths with a minimum of duplicative effort. Two of these methods provided positional isomers of the methoxy substituted pentaene 5£. Lastly, these derivatives have brought forth new information which will hopefully help assess the degree of neutral homoaromaticity that and its derivatives possess. lh2

Figure 15• Hie 100 MHz 1H NMR spectrum of pentaene in CDCI3 at 500 Hz sweep width.

Figure 16. The 100 MHz 1H NMR spectrum of pentaene 168 in CDCI3 at 500 Hz sweep width.

Figure 17- The 100 MHz 1H NMR spectrum of pentaene 166 in CDCI3 at 500 Hz sweep width. 1^3

F ig u re 15

OCH 2a

Heb

- J k

F ig u r e l 6

OCH

7.0 6.0 5.0 4.0 3.0 2.0 8 F igure 17 EXPERIMENTAL

Nuclear magnetic resonance (NMR) spectra were recorded using

Varian T-60, A-60A, HA-100 and Jeolco MH-100 instruments for proton

spectra, and a Bruker 90 spectrometer for Fourier transform spectra

(both carbon and proton). Apparent splitting are reported to ±0.5 Hz

in all cases and chemical shifts are referenced to internal tetramethylsilane. Perkin-Elmer 137 and h6 j spectrometers were used to record infrared spectra, while ultraviolet and visible spectra were

obtained using a Cary 14 spectrophotometer. Mass spectra were obtained

with an AEI-MS9 instrument at an ionizing potential of 70 eV. Unless

otherwise noted, melting points were determined in open capillary tubes

with a Thomas-Hoover apparatus or a Mel-temp hot block (mp > 225°) and

are corrected, while boiling points are uncorrected. Elemental

analyses of crystalline solids were performed by the Scandinavian

M icroanalytical Laboratory, Herlev, Denmark, or by Chemalytics, Inc.,

Tempe, Arizona. Semibullvalenes were too air, acid, and/or thermally

unstable for combustion analysis. Lastly, a Welsbach ozonator was used

for ozone generation.

Hexamethylphosphoramide (HMPA) was d istilled from calcium hydride

and stored over activated molecular sieves prior to use as was dimethyl

sulfoxide (DMSO). Acetone was distilled from potassium permanganate

and ethyl acetate was distilled from anhydrous potassium carbonate.

Dioxane, ether, and tetrahydrofuran were dried by distillation from Ibk 1^5 sodium-benzophenone ketyl. Pyridine v/as dried and stored over potas sium hydroxide pellets. Triethylamine was purified and dried by successive distillations from phenyl isocyanate, phthalic anhydride, and potassium hydroxide and v/as stored over barium oxide. Dichloro- methane was stirred with concentrated sulfuric acid, rinsed with v/ater and 10$ sodium carbonate solution, dried over calcium chloride, and distilled from barium oxide prior to use. Pentane was stirred with concentrated sulfuric acid followed by aqueous potassium permanganate solution, rinsed with water, dried, and distilled from anhydrous potas sium carbonate. Reagent grade benzene v/as dried over sodium wire.

Chromium trioxide v/as dried overnight at 80° under vacuum. Unless otherwise stated, solutions were dried over anhydrous magnesium ...n a.. O U J . J L Uuc •

Aa ,^-Hexalin-9,10-dicarboxylic Anhydride (jl). A mixture of kjk.k g

(2.92 mmol) of acetylene dicarboxylic 0 acid monopotassium salt and 3 Si o f 3 N

sulfuric acid v/as mechanically stirred

until the salt dissolved. The aqueous

solution was thoroughly extracted with

71 ether (10 x 250 ml) and the combined organic layers were washed with brine (150 ml), dried, filtered, and evaporated to dryness. Trituration with pentane, followed by further evaporation, gave a white solid which was dried under vacuum to yield

325 g (98$, lit52 79$) of acetylenedicarboxylic acid, mp 173-175°

( lit127 mp 175-176°). Freshly prepared acetylenedicarboxylic acid (325-0 g, 2.85 mol),

2 g of hydroquinone, and 5°0 ml of glyme were added to a 7 X. autoclave cylinder previously cooled in a Dry Ice-acetone bath and 1.5 I o f co n densed 1,3-butadiene was added. The mixture, securely enclosed in a rocking autoclave, was heated with shaking for 12 hr at 1^0°. Three batches of the resulting product were combined and, after a ll low boiling m aterial had been removed by rotary evaporation and distillation at 25 mm, the volatile products were distilled, bp above lkO° at 2-5 mm*

The solid distillate was melted on a steam cone, gently poured into 5 & of ether, and mechanically stirred several times with 10$ sodium carbonate solution. The aqueous layer vras acidified to give the un wanted monoadduct, l,^-cyclohexadiene-l,2-dicarboxylie acid, which was removed by filtration and air dried.128 When acidification of tlie aqueous layer no longer produced solid, the ether layer was washed with saturated salt solution (300 ml), dried, filtered, and evaporated to dryness. The solid residue was recrystallized from ether-petroleum ether to yield 39° g (22$, lit52 21$) of anhydride ^71, mp 97-103° CHPl (lit129mp 102-103°); 3 1855, 1770, and 980 cm-1; XH NMR (6, CDC13)

5.98 (quint, J = 1.7 Hz, 4, vinyl), 2-72 (downfield d of ABq,

Jab =15*5 Hz, further split into quartets J = 1.7 Hz, k, allyl) and

2 .2 0 ( u p fie ld d o f ABq, = 1 5 -5 Hz, a l l y l ) . cis-9,10-Bis(hydroxymethyl)-A2,6-hexalin (,72a). A solution of 110.0 g

(O.5U mol) of 71 in 500 ml of dry tetrahydrofuran was added dropwise with stirring to a suspension of 26.6^+ g (O.7O mol) of lithium aluminum hydride in 750 ml of tetrahydrofuran in a 5 A three-necked flask fitted 1*1-7

with a mechanical stirrer, reflux con

denser, and dropping funnel. Following

addition, the grey mixture was heated

at reflux for hO hr while maintaining

anhydrous c o n d itio n s . The r e a c tio n

72a. R = H mixture, cooled to 0°, was quenched by careful addition of water to give a white, viscous suspension. With continued cooling, 10$ sulfuric acid was added u n til a pH 1-2 m s reached. The solution m s diluted to 12 & with water and then cooled in ice for 30 “in. The fluffy white precipitate m s removed by fil tration, washed with water, and a ir dried to give 99-90 g (91$ > l i t 50

80$) of hexalin diol mp lVf-1500 (lit52 indicates that the compound exists in several polymorphic forms with melting points ranging from

125 to 176°); v S l3 3^0 and 2890 cm"1* 1H NMR (6, CDC13) 5-55 (m, ^ vinyl), 3.60 (s, -CHgO-), and 2.03 (br s, 8, allyl). c is -9,10 - Bi s (methane sulfonyloxyme th y l) - A 2 > 6-hexal in (72b). To a

mechanically stirred solution of 230 g OR (2.00 mol) of methane sulfonyl chloride

in *4-00 ml of dry pyridine m s added,

at -5 to -10°, a solution of 125 g OR (0 . mol) of diol 72a in 500 ml of dry

-72 b , R = S02CH3 p y rid in e . The anhy d ro u s r e a c tio n mixture ms stirred at ~5 to -10° for 2 hr and at 0° for 2 hr. While a temperature below -5° m s maintained, water was carefully added to destroy the excess methanesulfonyl chloride. The mixture was gently poured into enough ice-cold 10$ hydrochloric acid to make the solution distinctly acidic and was allowed to stand at 0° for 1 hr. The resulting tan precipitate was removed by filtration, washed with water, and air dried to give 22k.h g (99-6$, lit50 99-5$) of dimesylate 72bj mp 113.5-115° (lit52 mp 115-116°); 3020 , 2940, 1650, 1450, 1425,

1340, 1175, and 1080 cm"1; XH NMR (6, CDC13) 5-52 (br s, 4, vinyl),

4.20 (s, 4, -CH20-), 2.98 (s, 6, methyl), and 2.07 (br s, 8, allyl).

12-Thia[4.4.3]propella-3,8-diene (£5.). A mechanically stirred slurry

of 400 g (I.67 mol) of crushed sodium

sulfide nonahydrate (rinsed with ethanol

prior to use) in 1.4 £ of anhydrous

HMPA was heated with collection of the

aqueous a z e o tro p e (max bp 125° / l 5 mm)

which was later discarded. The resulting blue-green mixture was cooled to 40° and 226 g (0.644 mol) of dimesylate 72b was added, prior to heating at 120° for 19 hr under anhydrous conditions. The brown suspension was cooled, diluted with

3 £ of water, and extracted with ether (5 x 500 ml). The combined extracts were washed thoroughly with water (5 x 2 1) and brine, dried, filtered, and evaporated. (The aqueous phases were saved for HMPA recovery.) Chromatography on silica gel with petroleum ether elution provided 121.55 g (98-5$, lit50 98.8$) of sulfide 73 as a waxy solid, rap 55-56° (lit50 mp 55-5-56.5°); ^max 3 2820 and 1660 cm"1; XH NMR

(6, CDCI3) 5-45 (m, 4, vinyl), 2.76 (s, 4, -CH^S-), and 2.08 (m, 8, a l l y l ) . ik9

ll-Chloro-12-tM a[^.U.3lpropella-3,8-diene 12,12-Dioxide C&). A

mixture of 100 g (O.52 mol) of sulfide

73, J0 .8 g (O.53 m ol) o f N -c h lo ro su c -

cinimide, and 1.5 1 of carbon tetra SO. chloride was magnetically stirred and

heated at reflux under nitrogen for 74 20 hr. The reaction mixture was cooled to ambient and a quantitative yield of suecinimide was removed by filtration. The filtrate was concentrated in vacuo and the residual oil was taken up in ether (500 ml). The mechanically stirred solution was cooled to 0° and a standardized ethereal solution containing

I .05 mol of monoperphthalic acid55 was added dropwise. The solution was allowe d to warm to ambient temperature and was stirred overnight under anhydrous conditions. The mixture m s filtered to remove precipitated phthalic acid and the resulting solution was washed vrith 0.5 I sodium hydroxide solution, water, and brine prior to drying. After filtration, evaporation of solvent gave a nearly quantitative yield of ehloro- CHCI3 sulfone Jit, mp 110-112° (lit50 mp 111-113°)•, v. - 1. max 1320 and 1120 cm' ^ flMR (6, CDCI3) 5-59 (m, vinyl), 5-05 (s, 1, >CHCl), 3-16 (ABq,

JAB = 13.0 Hz, Avab = 15.2 Hz, 2, -CHgS02-), and 1.59-2.82 (m, 8, a l l y l ) .

[if.4.2]Propella-3,8,ll-triene (X5,)* To an ice-cold, mechanically stirred solution of 13^.^ g (O.52 mol) of jk i n 1. 5 & of dry tetra hydrofuran was added incrementally over 5 min JOh g (2.71 mol) of commercial powdered potassium t-butoxide. The tan slurry was allowed 150

to warm to room temperature and was

stirred and heated at reflux for 2.5 hr

under a nitrogen atmosphere. The dark

mixture was cooled, carefully diluted

75 w ith 5 & cold water, and thoroughly extracted with pentane (8 x 200 ml).

The combined organic layers were washed with water and brine, dried, and filtered. Concentration of the solution in vacuo (no heat) provided a viscous oil which, after vacuum distillation, afforded 65.7 g

(80$, lit50 72/°) of ,75 as a clear liquid, bp 65-67°/^.5 mm [ lit49 h e a t bp 75 °/5 mm]; ^max 5000, 2890, 2800, and 1650 cm"1; ^ NMR (6, CDC13) 5.65-5-88 (br m with sharp s at 5-68, 6, vinyl), and 1.96-2.12 (m, 8, a l l y l ) .

[it-, it-. 2]Propella-2,i|-,8^11-tetraene (77). Pyridinium hydrobromide

1 3 o perbromide (106. 5 g, 0.53 mol) was

added in one portion to a mechanically

stirred, cooled (0°) solution of 51-0 g

(0.312 mol) of triene 75, in 1 I o f

carbon tetrachloride-acetic acid (l:l), 77 and the resulting mixture was stirred under nitrogen at 0° for it- hr. After dilution with 2 & of water, the layers were separated. The aqueous phase was then extracted with carbon tetrachloride (2 x it-00 ml). The combined organic layers were rinsed with water (2 x 500 ml), 0.5 N sodium hydroxide solution

(2 x 250 ml), and brine before drying and filtration. Evaporation of 151 solvent in vacuo (no heat) left a quantitative yield of dibromide as a o CHC1 white, waxy solid, mp 53-56 •, 3 3°30, 3000, 2910, 2815, lMi-5, lltJO, 1155, and 990 cm"1; XH NMR (6, CDC13 ) 5-7-6 .k (m, 1+, vinyl),

4.5-5-1 (m, 2, -CHBr-), and 1.7-2.8 (m, 8, ally l and -CH^CHBr-).

Without further purification, the dibromide v/as taken up in 1.7 4 of HMPA. Anhydrous lithium chloride (132.5 g, 3*12 mol) and anhydrous lithium carbonate (209-0 g, 3*12 mol) were added and the resulting o slurry was mechanically stirred and heated at 100 under nitrogen for

2k hr. The mixture was cooled and diluted with 2 I of water and the product was removed by continuous extraction with petroleum ether

(1+0-60°). (The aqueous phase was saved for HMPA recycling. ) The petroleum ether extracts were combined, washed with water (5x) and brine, dried, and filtered. Concentration of the solution in vacuo

(no heat) gave a dark oil which, after vacuum distillation, provided

37.1 g (76.3$, lit50 66.0f0) of tetraene 7J131 as a light yellow oil, bp 55-62°/l. 5 mm-, 3010, 2880, 2790, 1070, 875, and 760 cm"1-,

4ax°Ctan6 268 2260250 (^60), and 223 nm (2320H 1H NMR (6,

CDCI3) 5.11-2-5.97 (m, 8, vinyl) and 1.88-2.07 (m, k, a l l y l ) .

5,j8-pihydro-N-phenyl-l,li-:ll-a,8a-diethenophthala.zine-2,3jlH ,ll-H)-

dicarboximide (78a)* A magnetically stirred solution of 22.3 g of crude tetraene 77 in a minimum volume of pentane was cooled to 0° and

a solution of N-phenyltriazolinedione132 in a minimum volume of ethyl

acetate was added dropwise until a pink color persisted. The slurry

was allowed to warm to room temperature whereupon the white solid j 8a

was removed by filtration. The yellow filtrate was evaporated to 152

dryness to give a solid residue which

was recrystallized from dichloro- ° \L r - n ^ / n 1 \ t / methane-petroleum ether (65-110°). The purified 78a was removed by filtration. o 78 a. R=Ph The mother liquor was concentrated and then chromatographed on silica gel with petroleum ether (65-110°) elution to give 6.37 g of triene 75i> elution with chloroform-ether (l:l) gave additional adduct j 8a. The combined yield of 78a v/as 33- ° S ( 9 7 based upon unreacted triene 73., lit50

75-*$) as a white solid, mp 252-255° (from dichloromethane-petroleum ether) (lit50 mp 253-255°); v^ 13 1760, 1720, 1690, 1490, and 1400 cm-1; 1H BMR (6, CDC13) 7-43 (s, 5, aryl), 6.28 (t, J = 3.5 Hz, 2, vinyl protons adjacent to hetero ring), 5-78 (m, 4, cyclobutene and ey c lo h e x e n e), 4 .7 3 ( t , J = 4 .0 Hz, 2, >CHN<), and 2 .2 0 -2 .7 0 (m, 4 , a l l y l ) .

5,8-Dihydro-N-methyl-1,4:4a,8a-diethenophthalazine-2,3(1H,4h) - dicarboximide (j8b)« In a manner strictly analogous to the preparation

of adduct 7.8a, 37-07 g (0.27 mol) of

tetraene JT7 in 100 ml of pentane was

reacted with excess N-methyltriazoline-

dione in ethyl acetate to give 4.0 g of

starting triene 75. and 58.0 g (99-7$) 2&b, R=CH, of adduct 78b, mp 177-179° (from acetone- KBr ether); 1770, 1700, 1450, 1390, and 1190 cm-1; XH NMR (6, CDCI3)

6.17 (q, J =3.5 and 3-5 Hz, 2, vinyl protons adjacent to hetero ring), 153

5*7^-5-85 (m with s at 5-77, cyclobutyl and cyclohexyl), ^.58 (t,

J = 3-5 Hz, 2, >CHN<), 3-00 (s, 3, methyl), and 2.22-2.14-O (m, b, a lly l) - , c a lc d m/e_ 269.11639, found 269.11689.

2,2a,6,9-Tetrahydro-N-phenyl-l,5a,2,5-ethanediylidene-5aH-cyclobuta-

[ci]phthalazine-3,^[lH,5H]-dicarboximide (&la). A solution of 1.2 g

(3.43 mmol) of triene J)3a in ^00 ml of

oxygen-free acetone-benzene (l:l) was

irradiated for 1.25 hr with a ^50 watt

Hanovia lamp through a 2 mm Corex filte r

while nitrogen was bubbled through the

solution. The solvent was removed in

Vacuo to give a yellow solid which was combined with earlier runs. Crystallization from dichloromethane-hexane gave some pure cube 8la. The mother liquor was concentrated and chro matographed on silica gel packed with chloroform:, ether elution gave additional 8la. The average yield of 8la per run was 0.9^- g (78%, lit50 bbf>), mp 18k.0-186.0° (from acetone)-, 1755, 17OO, 1500, lk00,

1195, 1170, and 1125 cm-1; 1H 1R (6, CDC13) 7-3-7-7 (m, 5, aryl),

5 .9 2 ( t , J = 3-0 Hz, 2, v in y l.) ^ .7 8 (dd, J = 2. 5 and 3.5 Hz, 2 , >CHN<),

3*57 (m, 2, methine), 2.86 (t, J = 3-° Hz, 2, methine), and 2.32

(d, J = 2.5 Hz, b, a l l y l ) .

Further elution with chloroform provided variable amounts of a material which lacked vinylic signals and is presumed to be a dimer 133 v ^ 13 1755, 1690, 1500, ibZO, and 1300 cm-1:, ^ NMR (6, CDC13) 7-^0 (m,

10, a r y l ) , I+.58 (m, >CHN<), 3 -5 5 (®, b, methine), 3*0 (m, 8, methine) and 2 . 3 0 (s, If, cyclohexyl). 154

2,2a-Dihydro-N-phenyl-l,5a,2,5-ethanediylidene-5aH-cyclobuta[d]- phthalazine-3,4(lH, 5H)-dicarboximide (83). To a cooled (-78°) solution

of O.5O g (1.51 mmol) of cube 8la in Ph 100 ml of dichloromethane was added

dropwise with stirring and protection

from m o istu re 4 .9 ml (1 .6 2 mmol) of a

O.33 M solution of bromine in carbon

t e tr a c h lo r id e . The s o lu tio n was s ti r r e d 8 3 for 15 min at -78°, and was allowed to warm to room te m p e ra tu re . After 30 min, the solvent was removed in vacuo to provide dibromide 82 in quantitative yield as a white solid:, v S l3 3010, 1760, 1695, 1505, and 1420 cm"1; XH NMR (6, CDC13) 7-25-

7.65 (m, 5 , a r y l ) , 4 .7 2 (m, 2 , >CKN<), 4-55 (m, 2, >CHBr), 3 -35 -3-80

(m, 2, methine), 2.80-3*25 (m, 4, methylene), and 2.38 (m, 2, methine).

Crude cube dibromide was dissolved in 125 ml of dry tetrahydro furan and stirred under nitrogen overnight with 2.00 g (13-2 mmol) of

1,5-diazabicyclo [5-4-0]undec-5-ene. The excess base was quenched by the addition of 50 ml of water. After stirring for 30 min, the solution was extracted with dichloromethane (2 x 75 ml). The combined extracts were washed with water (2 x 150 ml), 10$ hydrochloric acid (150 ml),

10$ sodium carbonate solution (150 ml), and brine before drying, filtration, and evaporation to dryness (no heat). The crude product contained starting dibromide 82, cycloreversion product 84, and the desired, therm ally unstable diene 83^ 1H NMR (6, CDC13) 7-1-7-5 (m, 55 a r y l ) , 5 .1 -5 -7 (AA'BB', 4 , d ie n e ) , 4 .6 0 (m, 2 , >CHN<), and 3-3"3-8

(m, 4, methine)-, 1H NMR [6, benzene-ds-py-d5(l:l)] 7-94 (d, J = 8.0 Hz, 155

2, aryl), 7-4-7.7 (nb 5? aryl), 5.1*5-5.82 (aa'bb", 1*, diene), 4.98

( t , J =3.0 Hz, 2, >CHN<), and 3 . 6O-3 . 7O (m, 4 , m eth in e).

N-Ehenyl-2,5-benzo-3,4-diazabicyclo[4.2.0]oet -7-ene-3,4- dicarboximide (84). Diene 83 was recrystallized three times from

2-propanol to give analytically pure _84 Ph as white needles, mp 218.1-218.4°; GHC1 vmax 3 1765 ’ 1710 ’ 1505» and ^ cm" 1» 1H NMR ( 6 , CDCI3 ) 7 -1 0 -7 -5 0 (m, 9 ,

aryl), 5-74 (s, 2, vinyl), 5-49 (d,

J = 4 .0 Hz, 2 , >CHN<), and 3-5 6 (d ,

J = 4.0 Hz, 2, methine)-, ^ NMR [6, benzene-ds-py-d5 (l:l)] 7-42-7-60 (m, 9? aryl), 5-80 (m, 4, vinyl and

>CHN<), and 3-80 (d, J = 4.0 Hz, 2, methine); calcd m/e_329-1164, found 329. H 69.

Anal. Calcd for C2oHi5N302: C, 72-94; H, 4.59i N, 12.76.

Found: C, 7 2 -6 6 ; H, 4-79; N, 12.80.

Benzocyclooctatetraene (.85). A mixture of 430.3 mg (1-30 mmol) of 84

and 850 mg (15-0 mmol) of potassium

/ \ ✓v hydroxide in 20 ml of 2-propanol was

i | I C - ) J stirred and heated at reflux under

nitrogen for 2 hr. The reaction mix

ture was cooled (0°), acidified (pH 2) 8 5 v /ith 1 Ofi hydrochloric acid solution,

s t i r r e d 5 m in, and b a s if ie d (pH 8 ) v/ith 3 N ammonium hydroxide s o lu tio n .

Pentane (15 ml) and 2.60 g of manganese dioxide on carbon were added 156 and caused the immediate evolution of gas. The black suspension was

stirred at 0° for $0 min and at ambient temperature for 1 hr. The

organic layer was decanted and the aqueous solution was washed with

pentane. The combined extracts were washed with water (2 x 50 ml) and

brine (50 ml), dried, and filtered. Excess solvent v/as removed by

d istillatio n through a 100 mm vacuum-jacketed Claisen-Vigreux head.

Preparative vpc (6' x 4" % SF-96 on 60/80 mesh Chromosorb G at 120°,

and 60 ml/min, t^ -t = 11 min) gave 85 mg (b2fo) of benzocyclooctatetraene

1H NMR (6, CDCI3) 6.8-7.2 (AA'BB', 4, aryl), 6.21 (ABq, JAB = 12.0

Hz, Avab = 52.6 Hz, 4, vinyl), and 5-80 (s, 2, vinyl) identical with an

authentic sample.60

2,2a,6,9-Tetrahydro-W -methyl-l,5a,2,5-ethanediylidene-5aH-cyclobuta[d]- phthalazine-3,4[lH,5Hl-dicarboximide (8lb). A solution of 1.00 g

(3.72 mmol) of adduct 78b in 400 ml of

deoxygenated acetonitrile-acetone (4:1)

was irradiated for 45 min through a

1 mm Corex filte r with a 450 watt

Hanovia lamp while nitrogen was bubbled

through the solution. After removal of 81b, RrCH3 solvent in vacuo, the residual brown oil

was chromatographed on silica gel packed with chloroform. Ether elution

gave 0.48 g of starting triene 75 contaminated with some photo by

product (s) and then O.5O g of cube 8lb. Recrystallization from acetone-

petroleum ether (65-110°) gave pure cube 8lb as a white solid,

mp 158.0-161.0°; ^50, 1685, 147Q, 1450, 1400, and 1270 cm-1; ^ NMR (6, CDCI3) 5-86 (t, J = 2.5 Hz, 2, vinyl), k.63 (dd, J = 2. 5 and

3 .5 Hz, 2, >CHN<), 3*50 (m, 2 , m e th in e ), 3 -0 7 ( s , 3> m eth y l),

2.83 (t, J = 3-0 Hz, 2, methine) and 2.28 (d, J = 2. 5 Hz, allyl); calcd m/e_269.116^, found 269.1169.

A secondary photoproduct of mass 538, presumably a dimer, was

isolated with further elution in amounts which varied from run to run

but was not characterized further.

la, 2 ,3,6,7,7a-Hexahydro-N-phenyl-l, 2a, 6a-metheno-lH-eyclopropa[b] -

naphthalene-2,7-biimine-8,9-dicarboximide (68a). A solution of 6.)+0 g

(19-3 mmol) of cube 8la, and 102.7 g

(600 mmol) of reagent grade silver

nitrate in 1^00 ml of dioxane and 350 ml

of water was stirred and heated at

reflux under nitrogen and in the dark

for if8 hr. The grey solution was 68 a . R = P h concentrated in vacuo then diluted with

200 ml of chloroform and 200 ml of wa ter. The layers were separated and

the aqueous layer was rinsed thoroughly with chloroform x 150 ml).134

The combined organic layers were rinsed with water (4 x 200 ml) and

brine, dried, filtered, and evaporated to give a greater than

quantitative yield. Recrystallization from 2-propanol gave 6.08 g

(95$, lit50 89$) of pure snoutene 68a, mp 1^8.0-1^9-5° (lit50 mp lW .0-

150.0°); v2Sl3 1770, 1700, 1500, lta ), 1290, and 1130 cm-1; 1H NMR (6, UlcwC CDCI3) 7-25-7.67 (m, 5, aryl), 5-50 (m, 2, vinyl), ^.88 (t, J = 2.0 Hz,

2 , >CHN<), 2 .5 2 (m, a l l y l ) , and 1 .9 0 -2 .1 2 (m, c y c lo p ro p y l). la , 2,3,6,7 j 7a-Hexahydro-N-methyl-l,2a,6a-metheno-lH~cyclopropa[b]- naphthalene-2,7-biimine-8,9-dicarboximide (68b). As in the preparation

of snoutene 68a, a solution of 9 .1 0 g

(0.03^ mol) of cubyl derivative 8lb and

l*)-0.0 g (O.82U mol) of reagent grade

silver nitrate in 2 .k I of d io x an e -

water (5:1) was stirred and heated at

reflux in the dark under nitrogen for

68 b. R=CH 3 53 br. A similar work-up afforded

8.9^ g (98.3$) of snoutene 68b, mp I69.O-I7O.50 (from acetone-pentane)-, v 1755, 1685, lif-55, and 1390 cm"1; 1H WMR (6, CDC13) 5-^8 (t, max J = 1.5 Hz, 2, vinyl), *+.78 (t, J = 2.5 Hz, 2, >CHN<), 3-05 (s, 3, m e th y l), 2 .b6 ( s , k, allyl), and 1.96 (m, U, cyclopropyl); calcd m/e_

269.116^, found 269.1169. la ,2,7,7a-Tetrahydro-N-phe nyl-1,2a,6a-methe no-IH-c yclopropa[b]- naphthalene-2,7-biimine-8,9-dicarboximide (87a). To a magnetically

stirred solution of 5*0^ g (15*0 mmol)

of snoutene 68a in $00 ml of dichloro

methane at -78° was added dropwise a

solution of bromine (2.*J-3 g, 15-2 mmol)

in *1-5 ml of carbon tetrachloride. The

solution was then stirred for 15 min 8 7 a . R = Ph and allowed to warm to ambient tem perature while anhydrous conditions were maintained. The clear yellow solution was evaporated in vacuo to leave a quantitative yield of 159 dibromide as a white, frothy solid; 1H NMR (6, CDC13) 7* *1-0 (m, 55 aryl),

*+•95 (m, 2 , >CHN<), il-.33 (m, 2 , >CHBr), and 1.9 3 -5 * 1 0 (m, 8 , c y c lo p ro p y l and methylene).

A magnetically stirred solution of this dibromide in *1-00 ml of anhydrous tetrahydrofuran under nitrogen was treated with excess l,5-diazabicyclo[5**+.0]undec-5-ene (DBU) (11.6 g, 75*0 mmol) and heated at 50° for 12 hr. Hydrolysis of the unreacted base was accomplished by adding excess water and stirring for 3° min. The solution was con centrated in vacuo, then diluted with water and dichloromethane. The layers were separated and the aqueous phase was extracted with dichloro methane (2 x 100 ml). The combined organic layers were washed with water, 5$ hydrochloric acid solution (3 x), 10$ sodium carbonate solution, anu brine. After drying, filtration, and solvent removal, there was obtained *)-. 98 g (99$) of diene 87a. Chromatography on silica with chloroform elution gave pure 87a as a white solid, mp 182.5-183*0° rnirn .. (from acetone-ether); 3 1770, 1700, 1505, and 1*1-15 cm" ; ^ NMR (6,

CDCI3 ) 7.37 (m, 5? aryl), 5.75-6.*1-0 (AA'bb' pattern, *1-, olefinie),

5*20 (dd, J = 2.5 and 3.0 Hz, 2, >CHN<), 1.92-2.60 (m, 3, cyclopropyl), and O.56 (d, J = *1- Hz, 1, cyclopropyl syn to cyclohexadiene ring);

13C NMR (ppm, CDCI3) 155*36 (N-C=0), 131*63, 129*16, 128.19, 125.78,

123.05, and 122.66 (aryl and diene), 60.*1-5 (>CHN<), *1-7*8*1- (quaternary cyclopropyl), and 28.lj-1, 25.7*+, and 25*35 (remaining cyclopropyl); c a lc d m/e_ 329*ll6*+, found 329*117*+*

A n a l. C alcd f o r C2oHi5N302 : C, 72*9*+; H, *+*59; N, 1 2 .76.

Pound: C, 72-91; H, *+.6 3 ; N, 12.75* la,2,7,7a-Tetrahydro-N-methyl-l',2a,6a-metheno-lH-cyclopropa[b] naphthalene-2,7-biimine-8,9-dicarboximide ( 87b). In a manner strictly

parallel to the synthesis of _87a, 2.00 g

(7*44 mmol) of snoutene 68b v/as treated

with one equivalent of bromine.

Following an analagous workup, the

corresponding dibromide v/as reacted at

50° with 5-84 g (37>1 mmol) of

-87b, R = CH 3 l,5-diazabicyclo[S.4.0']undec-5-ene. A similar isolation procedure gave I.7O g (86%) of diene 87b as a white solid, mp 187.0-188.5° (from chloroform-ether); 1770, 1710, 1465,

1395, 1230, 825, 800, 755, 700, 670, and 520 cm"1; 1H NMR (6, CDC13)

5.75-6.30 (AA'BB' pattern, 4, vinyl), 5-03 (dd, J - 2.0 and 3*5 Hz

>CM<), 2.92 (s, 3, methyl), 2.20-2.62 (m, 1 cyclopropyl), 1.90-2.15

(m, 2, cyclopropyl), and O.57 (d, J = 4.0 Hz, 1, cyclopropyl syn to cyclohexadiene)•, calcd m/e_ 267.IOO8, found 267-1014.

Hexahydro-1,6-his(hydroxymethyl)-N-phenyl-1,2,3-metheno-lH-4,5- diazacycloprop[cd]indene-4,5-dicarboximide (8ga)'. Into a l l three-

necked round bottom flask equipped

with a stopper, a condenser with gas

outlet, and a bubbler inlet was placed

3.43 g (10.4 mmol) of 87a in 500 ml of

chloroform. This solution was cooled

in ice while a stream of ozone was

8 9 a . R = H introduced until a slight blue color 161 persisted. Excess ozone was purged with oxygen (10 min), ethanol

(80 ml) was added to aid solubility, and a solution of sodium borohydride (4.00 g, 106 mmol) dissolved in 20$ aqueous ethanol was added dropwise to the cooled reaction mixture. After 12 hr, 10$ hydrochloric acid was added (to pH 4), the layers were separated, and the organic layer was washed with 10$ hydrochloric acid and saturated sodium bicarbonate solutions. These washings were reextracted with chloroform and the combined organic phases were washed with brine, dried, and evaporated to yield 2.95 g (84$) of white crystalline diol 89a:

3420, 1750, and 1685 cm-1; 1H HMR (6, py-d5) 7*1-7*4 (m, 5, aryl), 5*55

(s, 2, -OH), 5-50 (m, 2, >CHN<), 3-90 (s, 4, -CHgOH), and 1.80-2.10 (m,

4, cyclopropyl).

A 25O mg (0.74 mmol) sample of this

impure diol was dissolved in pyridine

(25 ml) and treated with acetic

anhydride (465 mg, 4.36 mmol). The

solution was stirred under nitrogen

. - overnight and evaporated to dryness in 8 9b, R = COCH3 vacuo. There was obtained 190 mg (6l$) of diacetate 8gb, mp 170.2-170.5° (from 2-propanol); 1740, 1700,

1405, 1230, and 1215 cm"1; XH MR (6, CDC13) 7-50 (m, 5, aryl) 5*15 (m,

2 , >CHN<), 4 .2 0 (ABq, J = 12.5 Hz, AvAB = 26-9 Hz, 4 , -CHgO-), 2 .1 7 (m,

4, cyclopropyl), and 1.94 (s, 6, methyl); calcd m/e 451.1743, found

451.1750.

Anal. Calcd for C22H2iN306: C, 62.4l; H, 5-00; N, 9*92.

Found: C, 6 2 .2 6 ; H, 4 .9 6 ; N, 9 . 8 7 . Hexahydro-N-phenyl-2H,4H-cyclopropa[3' ,4']pentaleno[l' ,6': l ,3,2]cyclopropa[l,2-£]furan-l,5-biimine-6,7-dicarboximide (,90). Method A. A

solution of 75 mg (0.22 mmol) of Ph N s p ° unpurified 89a in 15 ml of freshly dis oK'V tilled pyridine was heated at reflux

for 2 hr with 54.1 mg (0.274 mmol) of

p-toluene sulf onyl chloride. The

solution was cooled, acidified to pH 2

with 5$ hydrochloric acid, and extracted with chloroform (5 x 25 ml). The usual workup afforded a tan-colored oil (> 100$), thin layer chromatography of which indicated two major products, one of which was subsequently shown to be _g0.

Ifethpd. B. A solution containing 200 mg (0.59 mmol) of impure 8§a and 112.5 mg (0.612 mmol) of p-toluenesulfonyl chloride in 25 ml of anhydrous pyridine was stirred overnight at 25°. Removal of solvent in vacuo left a solid residue which was taken up in anhydrous tetra hydrofuran (25 ml). Sodium hydride (36.1 mg, 0. 756 mmol) was added in portions and the mixture was stirred at 25° overnight. Water was carefully added and the mixture was extracted with chloroform

(5 x 50 ml). Workup afforded a brown send-solid which when chromato graphed on F lorisil gave 162 mg (85.5$) o f ,5 mp 219.5-220.0° (from

2-propanol); 2860, 2840, 1765, 16955 and 1410 cm-1; 1H NMR (6,

CDCI3 ) 7 .2 5 -7 .6 0 (m, 5, a r y l ) , 5-15 (m, 2, >CHN<), 3-88 (ABq, J = 9-0

Hz, Av = 18.3 Hz, 4, -CH^O-), and 1.9-2.2 (m, 4, cyclopropyl); calcd m/e 321. 1 1 1 3 , found 3 2 1 . 1 1 1 8 . 165

Anal. Calcd for Ci8Hi5N303: C, 67.28; H, 71* N, 1 3 . 08.

Found: C,. 67. 1 3 ; H, 4 .7 1 ; N, 12.73-

Hexahydro-1, 6-bi s (hydr oxymethyl) -N-phenyl-1,2,3 -me the no -1H-4,5 -diazacycloprop[cd]indene-4,5-dicarboximide Dimethanesulfonate (jg2)- To a

solution of impure 89a (280 mg,

N ^O 0.825 mmol) in 25 ml of dichloromethane was added under nitrogen 250 mg of 0 'XsV t purified triethylamine followed by OS02CH3 cooling to -10° and dropwise intro o so 2ch 3 duction of methane sulf onyl chloride 92 (209 mg, 1.82 mmol). After 15 min, the organic phase m s washed with ice water, cold 5 1° hydrochloric acid, cold saturated sodium bicarbonate solution, and brine. From the dried dichloromethane solution there m s isolated 399 mg (87^) of .§2 as a white solid, mp 214.7-215.0° (from tetrahydrofuran-ether-hexane); p tip i T 3 1770, 1700, 1500, 1400, 1350, 1175, 1040, and 945 cm" ; NMR UicwC (6, CDCI3) 7-2-7-5 (m, 5, aryl), 5-15 (m, 2, >CHN<), 4.30 (s, 4,

-CH20-), 3-00 (s, 6, methyl), and 1.8-2.3 (m, 4, cyclopropyl); calcd m/e 495.077, found 495-076.

Anal. Calcd for C20H2iN303S2: C, 48.48; H, 4.27; N, 8.48.

Found: C, 48.29; H, 4.33; N, 8.21.

Hexahydro-3-benzyl-N-methyl-2H,4H-cyclopropa[31,4' ]pentaleno[l' ,6 ': l,3,2]eyclopropa[l,2-c]pyrrole-l,5-biimine-6,7-dicarboximide (gjaj- A solution of 570 mg (1.15 mmol) of dimesylate _§2 and 400 mg (3-74 mmol) of freshly distilled benzylamine in 75 ml of acetonitrile ms stirred i6k

at ambient temperature under nitrogen

for 111- hr. The solution was evaporated

to give a yellow o il which, after

chromatography on silica gel (elution

with ether), yielded 296 mg (63 $) o f 93a

as a white solid, mp 158.3-158.5° (from

93_a \-NC!l2Ph tetrahydrofuran-ether); 3°3°}

2920, 2800, 1 7 6 5, 17^5, 1700, 1505, and llHO cm-1; 1H HMR (6, CDC13)

7.45 and 7.20 (m, 10, aryl), 5*02 (m, 2, >CHN<), 3-53 (s, 2, benzyl),

2-59 (ABq, J^b = 9'k Hz> AvAB = 3 7 *h Hz, U, -CH2N<), 2.^3 (m, 1, cyclopropyl syn to nitrogen ring), and 1.95 (m, 3? cyclopropyl); calcd m/e klO.Tjk3, found lHO.1747.

Anal. Calcd for C25H22N402: C, 73-15} H, 5-^0; N, 13-65*

Found: C, 72.9^5 H, 5-^8; N, 13-53-

N-Benzyl-2a,7b-dihydro-5H,7H-pentaleno[l' ,6': l,3,2]cyclopropa[l,2-e]- pyrollidine (62b) H-benzyl-2,It-,5a,5b,5c,5d-hexahydrocyclopropa[3,^]- pentaleno[l,6-cd]piperidine (62a).135 A stirred mixture of 32I1-.^ mg

(0.792 mmol) of 93a,

316.8 mg (7-92 mmol)

of sodium hydroxide, NCH,Ph and 20 ml of 2-

propanol m s refluxed a b _ for 1 hour under 62 argon. The milky

yellow solution was cooled to 0°, acidified to pH 6 with 50$ aqueous acetic acid, hasified

(pH 9) with JH ammonium hydroxide, and diluted with 20 ml of pentane.

To this clear, yellow solution was added, in one portion, 688 mg.

(7-92 mmol) of activated manganese dioxide69 and stirring was continued at ambient temperature for 9 hr. After filtration, the clear, yellow filtrate was diluted with pentane, rinsed with water (b x 10 ml), and dried by shaking with sodium chloride-sodium sulfate. Excess solvent was carefully distilled utilizing argon bubbling. The concentrated solution was transferred to a sublimator, and the remaining solvent was distilled at high vacuum. Sublimation (40-95°, 1.2 x 10"3 torr) gave 100 mg (54$) of 62^ as low melting (ca. 10°) white solid;

3055, 30^5, 2945, 2890 , 2800, 2765, 1735, 7*1-0, 715, and 695 cm"1;

X - r tane 215 (e 6700), 231 (e 5000), 241 sh (e 2800), and 252 sh nm

(e 1600); XE MR (6, CDCI3) 7-32 (m, 5, aryl), 5-53 (

= 5-0 Hz, J4>5 = J5j6 = 2.0 Hz, 2, H4 and Hs), 5-29 (d, £3,4 = £e,7

=5*0 Hz, 2, H3 and H7), 3*75 (s, 2, benzyl), 3.26 (dt, J ij5 = 6.5 Hz,

£4,5 = £ 5 ,6 = 2.0 Hz, 1, Hg), 3.18 (ABq, = 9-0 Hz, AvAb = *1-9.6 Hz,

* * -, -CHg-N<), and 2.98 (d, £ 1 , 5 = 6.5 Hz, 1, Hi). Spin decoupling:

Double irradition of the signal at 6 5*53 collapsed the peak at 5*29 to a singlet and the peak at 3-26 to a doublet, £1^5 = 6.5 Hz. Saturation at 3.26 gave a doublet at 5*53, £3,4 = £6,7 = 5*0 Hz. 13C MR (ppm,

CDCI3) 138.98, 128.68, 128.14, and 126.90 (d, aryl), 123.18 (d, C4 and

Cs), 121.88 (d, C3 and C7), 61.51 (s, C2 and C8), 59-35 and 53.90 (t, benzylic and a to semibullvalene framework), and 53-85 and 52.66 (d, Ci and C5); calcd m/e 235-1381, found 235-1355- 166

Hexahydro -1 ,6-bi s (2 -hydroxyethyl) -N-phenyl-1 , 2,3 -metheno-ILH-1!, 5-diazacycloprop[cdTindene-4,5-dicarboximide (100a). Following the procedure

for the ozonolysis of 87a, 2.00 g (6.05

mmol) of- snoutene j58a in 250 ml of chlo N v/.O roform was reacted with excess ozone at

0°. The mixture, diluted with 40 ml of

absolute ethanol, was reduced with

2.29 g (60.5 mmol) of sodium borohydride

1.Q°3’ ^-ll as ^ef0re> a sim ilar work-up afforded

2.17 g (97*856) of frothy solid diol 100a; vKBr 3450, 2925, 1750, 1690,

1500, 1410, 1265, and 1020 cm"1; 1H NMR (6, CDC13) 7-43 (m, 5, aryl),

5 .0 8 ( t , J = 2 .5 Hz, 2 , >CHN<), 3 .7 2 ( t , J = 6 .5 Hz, 4, -CH20H), 3 -0 0

(concentration dependent) (s, 2, -OH), and 1-50-2.20 (m, 8, cyclopropyl and cyclopropyl carbinyl).

Crude diol (113 mg, 0.3I mmol) was

dissolved in 12 ml of dry pyridine and N ^ O treated with excess acetic anhydride

(628 mg, 6 .1 6 mmol). The s o lu tio n was

, ~ o r ' R stirred overnight at 25 under

anhydrous conditions, then evaporated

IQOb, R=C0CH3 ^.0 dryness in vacuo. Thick layer

chromatography (ether elution) followed by recrystallization from

tetrahydrofuran-ether gave 39 mg (3 Ofo) of white solid diacetate 100b,

mp 112.4-113.3°; 1740, 1695,-1495, 1405, and 1235 cm"1; aH HMR (6,

CDCI3 ) 7*45 (m, 5, a r y l ) , 5 .O5 (m, 2 , >CHN<), 4 .0 0 -4 .5 0 (m, 4 ,

-CH202CCH3), 2.03 (s, 6, CH3C02-), and 1.65-2.15 (m, 8, cyclopropyl and 167 cyclopropyl carbinyl); calcd m/e_ 451-1743, found 451.1750.

Anal. Calcd for C24H25N306: C, 63.85; H, 5-58; N, 9-31.

Found: C, 63 -7 4 ; H, 5-64; N, 9-28.

Hexahydro-1,6-bis(2-hydroxyethyl)-N-phenyl-1,2,3-metheno-lH-4,5-diazacycloprop[cd]indene-4,5-dicarboximide-6-tosylate (102a). To a solution

of 420 g (l.l4 mmol) of crude diol 100a

Ph in 25 ml of anhydrous tetrahydrofuran

I ( was added 216 mg (l.l4 mmol) of tosyl

chloride (recrystallized from chloro-

form-petroleum ether prior to use) and

1.4 ml (9.36 mmol) of purified

102a. X=OTs triethylamine. The resulting solution was stirred at room temperature under nitrogen for 5 hr and then heated at reflux for 14 hr. The mixture was cooled, concentrated in vacuo, and diluted with 75 ml of chloroform. The organic layer was rinsed with 10$ hydrochloric acid, water, and brine prior to drying, fil tration, and evaporation. There was obtained 523 mg (88$) of monotosylate 102a as a tan oil; v ^ t 1770, 1700, 1495, 1410, 1360, 1 1 9 0 ,

1175, and 750 cm"1; 1H NMR (8, CDC13 ) J.15-7-83 (m, 9, aryl), 5.10 (t,

J = 2.5 Hz, 1, XHN<), 4.88 (t, J = 2.5 Hz, 1, >CHN<), 4.18 (t, J = 6.0

Hz, 2, -CHpOTs), 3.73 (t, J = 6.5 Hz, 2, -CHgOH), 2.58 (s, 1, -OH),

2.38 (s, 3, CJ^-Aryl), and 1.75-2.20 (m, 8, cyclopropyl and cyclopropylcarbinyl).

Hexahydro-1-(2-hydroxyethyl)-6-(2-iodoethyl)-N-phenyl-1,2,3-metheno-

lH-4,5-diazacycloprop[cd1inde ne-4,5-diearboximide (102b). The crude 168

monotosylate (523 mg, 1.00 mmol) was

dissolved in 50 ml of purified acetone N - ^ 0 and stirred with sodium iodide (280 mg,

I .87 mmol) u n d er a n itro g e n b la n k e t fo r

1 hr at ambient temperature and for 18 hr

at reflux. The flocculent, white 1_Q2b, X= I precipitate was filtered from the warm solution and the filtrate was evaporated to dryness. A chloroform solution of the filtrate residue was rinsed with water, 10$ sodium thiosulfate solution, water, and brine before drying. Filtration and evaporation of solvent gave 3^8 mg (73$) °f crude iodohydrin 102b.

Chromatography on silica gel with ether elution gave 31° mg of pure 102b . KBi* as a frothy, white solid; vmay 3*<30, 1755? 1895? 1500, 1^-10, and 1260 cm"1; 1H NMR (6, CDC13) 7-1)-8 (m, 5, aryl), 5-15 (t, J = 2.5 Hz, 1,

>CHN<), 5-02 ( t , J = 2 .5 Hz, 1 , >CM<), 3*78 ( t , J = 6 .0 Hz, 2 , -CgpOH),

3.29 (t, J = 8.0 Hz, 2, -CH2I), 2.95 (br s, 1, -OH), and 1.60-2.20 (m, 8, cyclopropyl and cyclopropylcarbinyl).

Decahydro-H-phenylcyclopropa[3 ',b ']pentaleno[l ',6 1,3,2lcycloprop -

[l,2-d]oxepin-l,7-hiimine-8,9-dicarboximide (101.). The purified

Ph iodohydrin 102b (310 mg, 0. 65 mmol) was

dissolved in 50 ml of anhydrous tetra-

hydrofuran and stirred under nitrogen

with freshly prepared silver oxide136

(309 mg, 1*35 mmol) overnight at room

1 0 1 temperature and for 2b hr at reflux. The brown suspension was cooled and the solid was removed by filtration.

The yellow filtrate was evaporated to dryness and the residue m s taken up in chloroform. The organic solution was rinsed with water (2x) and brine, dried, filtered, and evaporated to dryness. Prepative thin layer chromatography with chloroform elution provided 130 mg (57$) of KBr oxepane 101, mp l8^<3-l8t. 7° (from chlorof orm-cyclohexane) •, vfflax 2890,

1770, 1700, 1505, lk95, 1^10, 1295, 1275, 1135, 1115, IHO, 1070, 1005

765, and 710 cm"1; 1H NMR (6, CDC13, 100 MHz) 7.25-7.65 (m, 5, aryl),

U.8 5 ( t , J = 2 .5 Hz, 2, >CHN<), 3-85 (dtp, J = 1 3 .0 , 5 .0 , and 3 -0 Hz, 2 ,

>CH0-), 3-lil-'(dq, J = 13.0, 8.0, and 2.0 Hz, 2, >CH0-), and 1.87-2.50

(m, 8, cyclopropyl and cyclopropylcarbinyl); calcd m/e 3^9-1^+26, found

3 ^9. 1^33 -

Anal. Calcd for C2oHi9N303: C, 68.75; H, 5-^8; N, 12.03.

Found: C, 68.1fl; H, 5-^2; N, 12.01.

1,2, ^, 5,6a, 6b, 7b, 7c-Octahydroc yclopropa[3, b-]pentaleno[1,6-de]oxo c in

(jgljb) ~ 2a,5,6,8,9,9b-Hexahydropentaleno[l'’,6 /:l,3,2]cycloprop-

[l,2-d]oxepin (j?(fa). 135 A mixture of b-60.8 mg. (1-32 mmol) of oxepane 101

7^1 mg (13.2 mmol) •

of powdered potas

sium hydroxide, and

was mechanically a b 9 4 stirred and heated at reflux under

argon for 1 hr in a 250 ml three-necked round bottom flask fitted with a septum-covered gas inlet with stopcock and a condenser topped with a gas inlet. The milky, yellow solution was cooled to 0°, acidified (pH 6) with 50$ aqueous acetic acid, stirred for 10 min, basified (pH 9) with 50 $ aqueous ammonium hydroxide, and diluted with 50 ml of pentane. To the clear yellow solution was added in one portion 1.16 (13*2 mmol) of

Attenburrow manganese dioxide69 which triggered immediate gas evolution.

Stirring was continued under argon at ambient temperature for 9 to, whereupon the mixture was filtered through Celite to give a clear yellow solution which was extracted with pentane (5 x 25 ml). The combined pentane extracts were rinsed with water (5 x 25 ml) and dried over anhydrous sodium sulfate. The solution was filtered and concen trated by careful distillation with argon bubbling. The viscous con centrate was transferred to a sublimator and the remaining solvent was distilled. Two sublimations (k0-80°/l. 5 x 10“4 torr) gave 158.9 mg

(60.1$) of semibullvalene <&, mp ^2-55°; 5055, 2920, 2860, and

1110 cm"1; \ ^ octane 212.5 (e 9-0 x 103), 23O.O ( k.J x 103), and

250.0 nm (2.8 x 103); 1H NMR (6, CDC13, 100 MHz) 5-20 (d, J 3 j4 = J S j7

= 1.5 Hz, 2, H3 and H7), ^.03 (dt, J = 11.6 and 5-0 Hz, 2, >CH0-),

3.35 (d, Jlsr = 7-0 Hz, 1, Hi), 3-19 (dt, J = 6.5 and 11.6 Hz, 2,

XJH0-), 2.9k (t, Ji}5 = 7-0 Hz, J 4 j5 = J 5 j 6 = 5-5 Hz, 1, H5) 2.87 (dd with additional splitting, ^3^4 = J.6,7 = !• 5 Hz, J 4 j5 = J s 56 = 7*0 Hz,

2, H4 and H6), 2.50 (m, *1-, -CH2CH0-). Spin decoupling: saturation of the signal at 5 5-20 collapsed the signal centered at 2.87 to a doublet

J4 5 = J5j6 = 7-0 Hz. Conversely, double irradiation of the 2.87

signal simplified the 5*20 signal to a singlet. NMR (ppm, CDC13) 171

129.06 (s, C2 and CB), 120.22 (d, C3 and C7), 7^.06 (t, -CH20-), 59-^9

(d, Ci or C5), 5^-68 (d, C4 and C6), ^8.83 (d, C5 or Ci), and 3*M3

(t, -CH2CH20-); calcd m/e Ijk. lOkk-, found 17*1-. 10^9.

Hexahydro-1,6-bis (2-hydroxyethyl) -h-phenyl -1,2,3 -metheno-lH-*f, 5 - diazacycloprop[cd]indene-*t-, 5-dicarboximide Dimethylsulfonate (108).

Me thane sulfonylchloride (l. Ot g,

If. 10 mmol) was added dropw ise t o a N ^ O stirred solution of crude diol 100a

OMs (1.50 g, If. 10 mmol) and purified triethylamine (1.2^ g, 12.3 mmol) in

100 ml of dichloromethane at -10° under 108 a nitrogen blanket. After 15 min, the organic phase was washed with ice water, cold 5$ hydrochloric acid, cold saturated sodium bicarbonate solution, and brine. The solution was dried, filtered, and evaporated to give a quantitative yield of dimesylate 108 as a tan semi-solid; 1755, 1&95, 1500, 1*4-10, 1350

1170 , 955 , 800, 765, 7*1-5, and 710 cm-1; 1H NMR (6, CDC13) 7-^5 (m, 5, a r y l ) , 5-0 1 ( t , J = 2. 5 Hz, 2 , >CHN<), *t--37 ( t , J = 6 .5 Hz, It-,

-CI^OsSCHs), 2.97 (s, 6, -O3SCH3), and 1.75-2.10 (m, 8, cyclopropyl and cyclopropylcarbinyl). it—Benzyldecahydro-W-phenyl-1,7-hiimino-2H-cyclopropa[3 ',h ’’Ipentaleno-

[ l',6':l,3,2]cycloprop[l,2-d]azepine-8,9-dicarboximide (lOgj. A solution of 523 mg (1.00 mmol) of crude dimesylate 108 and 3*1-8 mg (3.25 mmol) of freshly distilled benzylamine in 75 ml of acetonitrile was stirred at ambient temperature under nitrogen for hk hr. The solution was filtered 172

and evaporated to dryness to give a tan Ph oil which, after chromatography on

silica gel (elution with 5°$ ether-

petroleum ether), afforded 266 mg

(6l$) of amine 109 as a white solid,

mp l84.5-l85.it-0 (from chloroform-ether); 1 0 9

1770 ’ 1705» 1505» lk95’ lIt05’ 11285 and 788 cm-1; 1H NMR (6, CDC13) 7*47 (m, 5, aryl), 7.25 (s, 5, aryl),

4.78 (t, J = 2.0 Hz, 2, >Cffl<), 5-52 (s, 2, benzyl), and 1.70-2.90 (m,

12, methylene and cyclopropyl); calcd m/e_ 458.2055, found 4j8.P060.

Anal. Calcd for C27H2sN4.02: C, 73-95; H, 5-98; N, 12-78.

Found: c, 73-51; H, 5-95; N, 12.78.

5 - Benzyl-2,5,4,5, 6a , 6b, 7b, 7c -octahydro -lH-cyc lopropa[3,4(]pentaleno-

[ l,6-de]azocine (9£b) a 7-Benzyl-2a,6,7,8,9>9h-hexahydro-5H-pentaleno—

[ l ', 6 ': l,3,2]cycloprop[l,2-d]azepine (§5a).135 The hydrolysis and

oxidation of 455 Jng

(1.04 mmol) of 109 was

carried out in the NCH2Ph same manner as 101

a using 346 mg 9 5 (8 .6 6 mmol) o f

powdered sodium hydroxide, 30 ml of 2-propanol, 900 mg (8.66 mmol) of activated man-

69 ganese dioxide, and 30 ml of pentane to give, after sublimation

(40-60°/l.2 x 10"3 to rr), 145 mg (53$) of semibullvalene ^ a s a low 173 melting (ca 10-15°) light yellow solid; 3 ^0 , 2930, 2815 , 2790,

1730, 1605, 1530, 1500, 1455, 12*2*5, 1375, 1350, 1315, 1220, 1180, 1150,

1115, 1028, 772, 750, 730, and 698 cm-1; xj|°0ctane 230 (e 7-5 x 103),

2k0 (5.1 x 103), and 255 nm (2.5 x 103); XH NMR (5, CDC13 , 100 MHz)

7.17 (s, 5, aryl), 5-00 (d, J3j4 = Js,7 = 5-0 Hz, 2, H3 and Hr ), 3-75

(m, 2, H4 and H6), 3-53 (s, 2, benzylic), and 2.10-3.20 (m, 10, methylene, Hi, and H5). Spin decoupling: double irradiation at 6 5*00 collapsed the multiplet at 3-75 to a doublet, J4,5 = 1.5,6 = 5-0 Hz, while irradiation at 3*75 simplified the doublet at 5*00 to a singlet.

H5 could not be specifically located. 13C NMR (ppm, CDC13) 129-11,

128.25, and 127.06 (d, aryl), 123-23 (s, aryl quaternary), 121.2*5 (d,

C3 and Cy) , 118.62* ( s , C2 and Ca ) , 80.12 (d , C4 and C6 ), 6 1 .3 5 ( t , benzylic), 58.8l (d, Ci or C5), 5^-71}- (t, -CH2CH2N-), 52.77 (<3, C5 or

Ci), and 29.51 (t, -CH2N-); calcd m/e 263.1672*, found 263.1677.

Decahydro-N-phenylcyclopropa[3 /,2* ']pentaleno[l ', 6 ': 1,3,2]c yclopropa-

[l,2-d]thiepin-l,7-biimine-8,9-dicarboximide (110). To 250 ml of

refluxing ethanol was added simul

taneously a solution of 1.97 g

(3*75 mmol) of crude dimesylate 108 in

60 ml of N, N-dimethylacetamide137 and a

solution of 0.90 g (3-75 mmol) of sodium

sulfide nonahydrate in ethanol-N,N- 1 0 dimethylacetamide (l:l). The addition, carried out under nitrogen, was complete after 8 hr and the resulting suspension was stirred and heated overnight at reflux. After the 1 7 ^ solvent had been removed via bulb-to-bulb distillation at high vacuum

(c a 1 mm) with minimum heat, the reaction mixture was chromatographed on 75 g of silica gel (chloroform-ether (l:l) elution) to yield 0 .8 8 g

(Ski) of sulfide 13.0, trip 209.5-210.5° (from chloroform-ether)-, vKBr 2920, IflclX 1750, 1690, 1^95} 1^55, 1^10, 1295, 1125, 770, 755, and 697 cm"1;

^•H HMB (6 , CDCI3 ) 7 -^7 (m, 5, a r y l ) , if. 89 (m, 2 , >CHN<), 2 .1 0 -3 -5 0 (m,

5 , methylene and cyclopropyl syn to thiepane ring), and 2 .0 0 (m, 3 , cyclopropyl); calcd m/e_ 365-1198, found 365- 1203 .

Anal. Calcd for C2oHi9]!fe02S: C, 65-73; H, 5-24; N, 11-50-

Found: C, 65,50; H, 5-26; N, 11-50. l , 2,lf,5,6a ,6b,7b,7c-Octahydrocyclopropa[3,^]pentaleno[l,6-de_]thiocin

(96b) ~ 2a,5,6,8,9,9b-Hexahydropentaleno[l'',6':l,3,2]cyclopropa-

[l,2-d]thiepin (96a).135 In a fashion strictly parallel to the prepa

ration of 599-2 mg

( 1 .6 2 mmol) o f s u lf id e

110 was hydrolyzed

w ith 65O.O mg ( l 6 . 0

a b mmol) of powdered

9 6 sodium hydroxide in 30

ml of 2-propanol and

oxidized with 1.^1 g (l6.2 mmol) of Attenburrow manganese dioxide. 69

Sublimation (75-l00°/l.6 x 10 “3 mm) afforded 185 mg (60$) of semibull- o KBr . valene g6 as a low melting, light yellow solid, mp 23-25 ; 301+5,

29*30, 2920, 2880, 2855, 2825, 1735, 1605, 1530, 1500, ll)-55, 1^5, 1^30,

1335, 1315, 1300, 1280, 1255, 12*35, 1220, 1175, 1110, 930, 915, 875, 814-5, 8^0, 800, 755, 7^0 , 730 , 690 , 635 , 625, and 600 cm"1; x J|Joctane

225 (e 1.7 x 103), 235 (1.5 x 103), and 2^0 nm (1.2 x 103); XH NMR (6,

CDC13, 100 MHz) 5.13 (d, J3j4 = J6j7 =3-0 Hz, 2, H3 and H7), 3-57

(dd, J3j4 = JSj7 = 3.0 Hz, J4j5 = J5j6 = 5.9 Hz, 2, Ei and He), and

2.IO-3.35 (m, 10, methylene, Hi, and H5). Spin decoupling: Double

irradiation at 6 5*13 caused the peak at 3*57 to collapse to a doublet,

J.4,5 = J5,6 = 5*9 Hz. Saturation at 3*57 simplified the signal at

5.13 to a singlet, but H5 could not be located exactly. 13C NMR (ppm,

CDC13) 121.83 (d, C3 and C7), II8.5I4 (s, C2 and Cs ), 70. 14-6 (d, C4 and

Ce), 59-62 (d, Ci or C5), 5^-22 (d, C5 or Ci), and 33-88 and 31-62 (t,

-CH2S- and -CH2CH2S-); calcd m/e_ 190.08l6, found 190.0818.

Decahydro-H-phenylcyclopropa[3']pentaleno[l ', 6 ': l,3,2]cyclopropa-

[l,2-d]thiepin-l,7-biimine-8,9-dicarboximide 4-Dioxide (ill)- To a

stirred, cold (0°) solution of 365 mg

(1.00 mmol) of thiepane 110 in 25 ml

of chloroform was added dropwise a

solution of 346 mg (2.00 mmol) of so. m-chloroperbenzoic acid in 5 ml of

chloroform. The solution was stirred 111 with protection from moisture for 5 hr

after warming to ambient temperature. The reaction mixture was diluted

to 75 ml with chloroform, rinsed with saturated sodium bicarbonate

solution, dried, filtered, and evaporated to dryness to give, after

recrystallization from chloroform-ether, 330-7 mg (83$) of sulfone 111

as a white solid, mp 338-5-339-5 (dec, sealed tube); VmS 1755, 17°0, 176

1495, 1455, 1415, 1528, 1282, 1152, 1106, 762, and 698 cnT1; 1H NMR (6,

CDCI3 , 90 MHz) 7.51 (m, 5, a r y l ) , 4 .9 5 ( t , J = 5*1 Hz, 2 , >CHN<), 5 -7 °

( t , J = 1 3 .0 Hz, .2 , -CHS02 - ) , 2 .8 2 -3 -1 5 (m, 2 , -CHS02- ) , 2 .6 0 (d ,

J = 6.0 Hz, cyclopropyl syn to thiepane ring), and 1.87-2.17 (m, 7,

cyclopropyl and cyclopropylcarbinyl).

Anal. Calcd for C20H19N3O4S: C, 60.44; H, 4.82; N, 10.57.

Found: C, 60.02; H, 4.88; N, 10.4l.

l,2,4,5,6a,6b,Tb,7c-Octahydrocyclopropa[5,4]pentaleno[l,6-de3tliiocin-

5,5-Dioxide (99b) ^ 2a,5,6,8,9,9b-Hexahydropentaleno[l',6 1,3,2]-

cyclopropa[l,2-d]thiepin 7,7-Dioxide (92a)-135 In a way exactly analogous

to the preparation

of J?4_, the hydroly

sis and oxidation of SO 2 SO. su lfo n e 111 (495 mg, a b 1.24 mmol) was car ried out using 496 mg 9 9 (12.4 mmol) of powdered sodium hydroxide, 5° ml of 2-propanol, 1.08 g (12.4 mmol) of

activated manganese dioxide,69 and 5° ml of pentane. The product was, however, only slightly soluble in pentane so, after the filtration of the reaction mixture, the aqueous solution was shaken with ether

(5 x 25 ml). The combined extracts were ’trashed with water (5 x 25 ml) and dried over anhydrous sodium sulfate. Filtration and rotary

evaporation (no heat) left a suspension of the product in 2-propanol.

Trituration with pentane gave 116 mg (42/,) of semibullvalene 22, as a 1 7 7

o K!Rr light yellow air sensitive solid, dec > 120 (sealed tube); vmy 2950,

2925, 1^08, 1398, 1352, 1315, 1275, 1130, 1112, 8^5, 808, 759, 73^,

J+80 , and kk-5 cm-1; x j^ octane 225 (e 1.9 x 103), 235 (l. 5 x 103), and

2^5 ran (1.3 x 103)*, ^ NMR ( 6, CDC13, 100 MHz) 5-19 (d, J 3 j4 = J 6j7

=3-8 Hz, 2, H3 and H7), ^*76 (t, J 3?4 = Je,7 =3-8 Hz, J5jS = J 4j5

= 3.5 Hz, 2, H4 and Hs), 3-30 (m, 1, Hs), 3*20-3.25 (m, If-, -CH2S02-),

3 .0 0 (m, 1, H i), and 2 .7 5 -2 -9 0 and 2 .2 0 -2 -5 0 (m, it-, -CH2CH2S02 ).

Spin decoupling: saturation at 6 5.19 simplified the triplet at b .j6 to a doublet, J4j5 = J5j6 =3*5 Hz. Double irradition at 4.76 caused the signal at 5*19 to collapse to a singlet and the multiplet at 3*30 to simplify to a doublet, Ji,s =3*5 Hz. The location of H 5 was confirmed by saturation at 3*3° which collapsed the signal at ^.76 to a doublet, J3j4 = J.6,7 =3*8 Hz. Hi must therefore resonate at 3*00*

13C NMR (ppm, CDC13) 123.01 (d, C3 and C7), 99*92 (d, C4 and C6),

89.67 (s, C2 and C8), 57*30 (d, Ci or C5), 56.11 (t, CH2S02), 5h. 12 (d ,

C5 or C i), and 2h.k9 (t, CH2CH2S02); calcd m/e 222.071^, found 222.0718.

Decahydro-W-phenylcyclopropa[3 ']pentaleno[l ', 6 1,3,2]cyclopropa-

[l,2-d]thiepin-l,7-biimine-8,9-dicarboximide W -0xide (jd^) and

Decahydro-N-phenylcyclopropa[-3 '>b ']pentaleno[l',6 l,3,2]cyclopropa-

[l,2-d]thiepin-l,7-biim ine-8,9-dicarboxim ide ^(3-Oxide (.13^).

A* Oxidation with m-Chloroperbenzoic Acid. To a stirred, cold

(0°) solution of 365 mg (1.00 mmol) of thiepane 110 in 25 ml of chloroform was added dropwise with protection from moisture a solution of 173 mg (1.00 mmol) of m-chloroperbenzoic acid in 5 ml of chloroform.

The solution was slowly warmed to room temperature and was stirred for 178

an additional 5 hr. Ph P 'h k ^ o Following dilution j r t & 0^,'f with 75 ml of.chloro

form, the mixture was

rinsed with saturated

sodium bicarbonate 11 2 11 3 solution, dried with sodium chloride-anhydrous sodium sulfate, filtered, and evaporated to dryness to return 297 mg (78$) of product. Proton NMR (CDC13) with the addition of europium shift reagent revealed the presence of two isomers whose bridgehead signals (>CHN<) integrated to an 80:20 ratio. The isomers were separated by preparative thin layer chromatography which required five elutions with 10J, acetone-chloroform.

The major isomer jL12, Rf = O.55, was recrystallized from chloroform- petroleum ether (65-HO0) to give a white solid, mp 292.5-295*5° (dec); v, 1755, 1700, 1495, 1415, 1278, 1258, 1150, 1068, 1027, 1015, 785,

770, 765, 717, and 710 cm"1; ^ NMR (6, CDC13) 7*50 (s, 5, aryl), 4.95

( t , J = 2 .5 Hz, 2 , >CHN<), 5*00 (m, 4 , -CH2S 0 -), 2.45 (m, 4, cyclopropylcarbinyl), and 2.08 (m, 4, cyclopropyl).

Anal. Calcd for C2oHi9N303S: C, 62.97*. H, 5-02; N, 11.02.

Found: C, 62.52; H, 5*02; N, 10.91.

The minor isomer 115, Rf = 0.24, was found to interconvert to 112 upon heating and was, therefore, not purified further; 1755, 1700,

1500, 1495, 1410, 1125, 1020, and 765 cm-1-, XH NMR (6, CDC13) 7*50 (s,

5, aryl), 4.95 (t, J = 2. 5 Hz, 2, >GHN<), 2.20-5.60 (m, 6, -CH2S0 and

-CHCH2SO-), 2.05 (m, 4, cyclopropyl), and 1.20-2.00 (m, 2, >CHCH2S0-). 179

B. Oxidation with Sodium Metaperiodate. To a solution of 365 mg

(1.00 mmol) of sulfide 110 in 50 ml of ice-cold methanol m s added dropwise a solution of 213-9 S (1-00 mmol) of sodium metaperiodate in

10 ml of water. The solution was stirred overnight at ambient temperature, solid was then filtered off, and the solvent was removed in vacuo to yield 330 mg (87$) of product. U tilization of proton NMR with lanthanide shift reagent revealed the presence of two isomers, as before, in an 80:20 ratio.

Thermal Interconversion of Sulfoxides 112 and 113,. A solution of care fully recrystallized (dichloromethane-acetone-petroleum ether and low heat) sulfoxide 113, in deuterioehloroform-tetramethylsilane contained in a sealed NMR tube was heated in a constant temperature bath at 70-0° and was periodically monitored. Interconversion m s immediately ap parent. After 2 hr, the thermodynamic equilibrium ratio was nearly established. There was no apparent change in the appearance of spectra after 6 hr and at this time the spectrum had an appearance nearly identical to those of the crude reaction mixtures from oxidation of sulfide 110.

Recrystallization of isom erically pure 113, from chloroform, acetone, or ethyl acetate also affected the interconversion.

Chemical Epimerization of Sulfoxide 112. To a solution of sulfoxide 112

(190 mg, 0.5 mmol) in 10 ml of dichloromethane under argon was added

6 k.k mg (0.5 mmol) of trimethyloxonium fluoroborate. The Meerwein's reagent gradually dissolved and after 30 min a white, flocculent precipitate appeared. Filtration gave 200.5 mg of white solid; l8 o

v!2) 1755, 1700, l^O , 11*20, llHO, 1275, 1120, IO75, 1020, and 780 cm-1. JUcUv The solid was suspended in 3° ml of water and 5*5 ml of 0.1 N sodium

hydroxide was added to cause complete solution. This solution became

cloudy with the appearance of a white precipitate and, after stirring

for 30 min, the mixture was extracted with dichloromethane (1* x 10 ml).

The combined organic extracts were rinsed with water (10 ml) and brine

(10 ml), dried over anhydrous sodium sulfate, filtered, and evaporated

to afford 176.2 mg (92.7$) of sulfoxide 113, identical to the minor

products from oxidation of sulfide 110.

Photooxygenation of Sulfide 110. A solution of 50 mg (0.l4 mmol) of

sulfide 110_ and 25 mg of rose bengal107 in 10 ml of anhydrous methanol was irradiated for 35 min with a 600 watt Sylvania DYS tungsten-halogen

projector lamp as a stream of oxygen was bubbled through the solution.

(The lamp was air and water cooled as usual.) The solution was evapo

rated to dryness with a minimum of heat. The crude mixture contained

sulfone 111 and sulfoxide JLIJ by thin layer chromatography and by NMR.

Preparative chromatography gave 10.7 mg (2 Cffo) of sulfone 111 and 15-0 mg

(29$) of sulfoxide (The low yields were due to the difficulty of

extracting the compounds from the adsorbent.) Both compounds were

identical to authentic samples.

Identical solutions of 25 mg of sulfide IIP and 25 mg of Photox,

25 mg of sulfoxide 112 and 25 mg of Photox, and 25 mg of sulfoxide llj^

and 25 mg of Photox each in 10 ml of dry methanol were irradiated (for

up to 2.5 hr) as before and monitored by thin layer chromatography. It

was found that as sulfone 3A1 appeared from the oxidation of sulfide 110 l 8 l at most a trace of sulfone was evident in the sulfoxide oxidations.

Secondly, no interconversion of sulfoxides was noted.

Photooxygenation .of Sulfide 117. A solution of 55 nig (0.20 mmol) of sulfide 11786 and 4.7 mg of rose bengal107 in 5 ml of dry methanol was irradiated as above for 13 min while a stream of oxygen was passed through the solution and then the solution was evaporated to dryness with minimum heat. By thin layer chromatography and NMR, there were only two products in an approximately equal ratio, sulfone 118 and syn-sulfoxide ll6. Preparative tic with 30$ acetone-chloroform gave

24.4 mg (40JS) of sulfone 118 at Rf = 0.42; XH NMR (6, CDC13) 5-11 (t,

J = 2 .5 Hz, 2, >CM <), 3 .3 8 (ABq, = 15 Hz, AvAB = 4 l Hz, 4 ,

-CH2S02-), 3-04 (s, 3, methyl), and 1.91-2.44 (br m, 4, cyclopropyl).

At Rf = 0.l4 was found 12.6 mg (21.7$) of syn-sulfoxide 116; 1H NMR

(6, CDCI3) 5-07 (dd, J = 3.8 and 2-5 Hz, 2, >CM<), 3-22 (br s, 4,

-CH2S0-), 3-07 (s, 3j methyl), 2.84 (d, J = 4.5 Hz, 1, cyclopropyl syn to sulfoxide ring), and 1.81-2.40 (br m, 3> cyclopropyl).

Both components were identical to authentic samples. 86 Their low solubility caused low recovery of material from the adsorbent.

O x id atio n o f S u lfid e 122,.

A. By Singlet Oxygen. In accordance with the method outlined above, 4 l.l mg (0.15 mmol) of sulfide 122, in 5 ml of dry methanol was photooxygenated for 5 min using 2*3 mg of rose bengal107 as sensitizer.

The crude residue contained (thin layer chromatographic analysis) starting m aterial 122,, sulfone 12£, and both sulfoxide isomers 12J, and

124. Preparative thin layer chromatography (elution with 10$ methanol- 182 ethyl acetate) gave 5-9 mg of starting material (Rp = 0.72), 15-9 mg

(40$) of sulfone 12£_ (R^ = 0.40), identical to an authentic sample,86

6.7 mg (l8$) of one sulfoxide (Rf = 0.125), and 15-7 mg (36.4$) of the other sulfoxide isomer (Rf = 0.06). The sulfone ms further purified by passage through a charcoal-Celite column to give a white solid; KBr vmax 1760> l69°> 1395J 1390, 1310, 1300, and 1140 cm-i;

1H NMR (6, CDCI3) 4.95 (dd, J = 3-0 and 5-0 Hz, 2, >CHN<), 3-65 (m, 2, methine), 3*53 (br s, 6, methine and methylene), and J .0 8 (s, 3» methyl); calcd m/e 307-0626, found 307.06319.

The sulfoxides were prepared independently (see below).

B. By m-Chloroperbenzoic Acid. To a cooled (0°) solution of

41.1 mg (O.15 mmol) of sulfide 122 in 2 ml of dichloromethane m s added dropwise a solution of 25.8 mg (0.l4 mmol) of m-chloroperbenzoic acid

in 1 ml of dichloromethane. The resulting solution ms allowed to warm to ambient temperature and was stirred in a dry atmosphere overnight.

The solution m s diluted to 20 ml with dichloromethane and rinsed with

5 ml of 10$ sodium bicarbonate solution, dried over anhydrous sodium

sulfate, filtered, and evaporated to dryness to give 40-3 mg (92.3$) of white solid. Chromatography of the residue gave l6.6 mg of one sul

foxide (Rf(10$ methanol-ethyl acetate) = 0.125) and 7-7 mg of the other

sulfoxide (Rf = 0.06). The sulfoxides were combined with like isomers

generated by method A and were chromatographed on charcoal-Celite.

Recrystallization of the more rapidly eluted isomer (Rf = O.125) from

dichloromethane-ether gave a white solid, mp > 205° (dec); I76O,

1690, 1460, 1395, 1270, 1220, 1030, 1005, and 695 cm"1; 1H NMR (6, 183

CDC13 ) 5-02 (d d , J = 2 .0 and *f.O Hz, >CHN<), 3 -6 7 (m, 2 , m eth in e ),

3.37 (t, J = 3*0 Hz, 2, methine), 3 .1 3 (ABq, = lif.5 Hz, AvAB = 3 3 . If

Hz, If, methylene), and 3 .O5 (s, 3, methyl); calcd m/e 291.0677, found

291. 0682.

The other isomer (Rj> = 0.66), a white solid, was obtained in lesser quantity; 1755, 1690, 1^75, 1280, 1270, and 7^5 cm"1; 1H NMR (6,

CDCI3 ) k.kj) (d d , J = 2 .5 and If.0 Hz, 2 , >CHN<), 3-73 (m, 2, m e th in e ),

3-15 (ABq, JAB = 13*5 Hz, AvAB = 27-9 Hz, If, methylene), 3.08 (t,

J = 3*0 Hz, 2, methine), and 3-00 (s, 3, methyl); calcd m/e 291.06775 found 291.0682.

Hexahydro-N-methyl-1 ,5a, 2 , 5 - ethane diylidene - 5 aH-c yclobut a [d Iphthalaz ine -

3,MlH,5H)-dicarboximide (127,). A solution of 250 mg (O.93 mmol) of

diazabishomocubane 8la. in 10 ml of

CH anhydrous tetrahydrofuran was added to

a prereduced suspension of 20 mg of

platinum oxide in 20 ml of dry tetra

hydrofuran under a hydrogen atmosphere.

A slight positive pressure of hydrogen 1 2 7 was maintained until uptake ceased

(ca k5 min). The suspension was filtered through Celite and the resulting solution was evaporated to dryness. The residue was tr it urated with ether, and evaporated to dryness to give a quantitative yield of 127 as a white solid, mp 122.5-123.0° (from dichloromethane- cyclohexane); 2980, 29k0 , 2915, 2895, 287O, 2835, 17 60 , 1 7 0 5 , lVf5, 1390, and 1255 cm"1; ^ NMR (6, CDC13) If.If8 (dd, J = 2.0 and 1 &

1*.0 Hz, 2, >CHN<), 3*43 (to m, 2, methine), 3.O7 (s, 3, methyl), 2-90

(t, J = 2. 5 Hz, 2, methine), and I .63 (m, 8, methylene); calcd m/e

271.1320, found 271.3321*.

Anal. Calcd for C15Hi7N302: C, 66.40; H, 6-32; N, 15-49-

Found: C, 66.16; H, 6.34; N, 15>38.

1,2,4,5,6a, 6b,7h,7c -0ctahydrocyclopropa[3,4]penta.leno [ l,6-de]thiocin

3ci’-Qxide (9Xb) ~ 2a, 5,6,8,9,9h-Hexahydropentaleno[l ', 6 l,3,2]cyclopropa[l,2-d]thiepin 7^-Oxide (97a).135 A mechanically stirred mixture

o f 494 mg (1 .3 0 mmol)

of pure sulfoxide JL12,

730 mg (13.0 mmol)

of powdered potassium

a 9 7 fi hydroxide, and 30 ml.

of 2-propanol was

heated at reflux for

45 min under argon in a 250 ml three-necked round bottom flask fitted with a septum covered gas inlet with stopcock and a condenser topped with a gas inlet. The dark solution was cooled to 0°, acidified (pH 6) with 50^ aqueous acetic acid,-stirred for 10 min, basified (pH 9) with

5056 aqueous ammonium hydroxide, and diluted with 50 ml of ether.

Activated manganese dioxide69 (1.13 g, 13*0 mmol) was added in one portion and the chocolate colored suspension was stirred under argon for 9 hr. The mixture was filtered through Celite and the aqueous phase was extracted with ether (3 x 25 ml). The combined ethereal solutions were rinsed with water (5 x 25 ml), dried over anhydrous 185 sodium sulfate, filtered, and evaporated to dryness. Recrystallization from ether-petroleum ether (65-HO0) gave b$ mg (l6$) of semibullvalene

2[as a tan solid, mp > 120° (dec, sealed tube); 2950, 1695, ib^O,

1395, 15^0, 1315, 1280, 1260, 1110, lOkO, 1015, T95, 750, and 7^0 cm'1-,

XH NMR ( 6, CDCI3, 100 MHz) 5.22 (dd, J4}5 = J5jS = 7.5 Hz, J3 ) 4 = J s , 7 =

4.5 Hz, 2, H4 and Hs), 5-18 (d, £34 = le }7 = b.5 Hz, 2, H3 and H7),

2 .7 5 -3 -^ 0 (m, 6 , -CH2 SO-, Hi an d H s ) , and 2.1f5-2.75 (m, b, -CH2CH2SO-).

Spin decoupling: saturation at 6 5*22 simplified the multiplet at

2.75-3-^0, thus confirming the location of H5. 13C NMR (ppm, CDC13)

12b. 09 (d, C3 and C7), 118.59 (d, C4 and Cs), 72-1+0 (s, C2 and C8),

61.2^ (d, C5 or Ci), 5^-60 (d, Ci or C5), 10.10 (t, -CH2S0), and 19-20

(t, -CH2CH2SO); calcd m/e 206.0765, found 206.0768.

1,2, U, 5, 6a , 6b, 7b, 7c -Octahydrocyclopropa [3, l|-]pent aleno [ 1,6-de ~]thioc in

3P-0xide (98b) ^ 2a,5,6,8,9,9b-Hexahydropentaleno[l ',6 l,3,2]cyclopropa[l,2-d]thiepin 7P-0xide (,98a) .135 A suspension of 657 nig (1.72 mmol)

of a mixture of

sulfoxides 112 and

113 (cei 1: l) and

\ | t O 1-93 g (3b.b mmol)

of powdered potassium a 9 8 hydroxide in 50 ml of

2-p ro p a n o l was mechanically stirred and heated at reflux under argon for kO min in a

250 ml three-necked round bottom flask fitted with a septum covered gas inlet with stopcock and a condenser topped with a gas inlet.

The dark mixture was cooled to 0°, acidified (pH 6) with 50$ 186 aqueous acetic acid, stirred 10 min, basified (pH 9) with 50% aqueous ammonium hydroxide, and diluted with 100 ml of ether. To the clear, brown solution m s added, in one portion, 1.49 g (17-2 mmol) of

Attenburrow manganese dioxide.69 The coffee colored mixture m s stirred under argon at ambient temperature for 12 hr, then filtered through Celite to give a clear, yellow solution. Water was added

(200 ml) and the layers were separated. The aqueous phase m s washed with ether (5 x 50 ml) and the combined ether extracts were washed with water (5 x 30 nil)* The solution was dried over anhydrous sodium sulfate, filtered, and evaporated to give a brown oil which contained some product. Extraction of the aqueous solutions with dichloromethane

(5 x 50 ml) and sim ilar processing of the extracts provided additional product as a tan solid. Hie crude 98 was dissolved in a minimum volume of dichloromethane and triturated with ether to give 3° mg of brown impurities. Recrystallization of the mother liquor from dichloro methane -petroleum ether (65-110°) provided 127 mg (36%) of semibullvane

98 as a tan solid, mp > 170° (dec, sealed tube); v*?*! 3°30, 2920, 1430, — - IllCbA. 1335, 1020, 925, 820, and 750 cm"i; XH NMR (6, CDC13, 100 MHz) 5-20 (d,

J.3,4 = J e ,7 = 5-5 Hz, 2, H3 and H7), 4.85 (t, J3j4 = J.6,7 = 5 -5 Hz, 2 ,

H4 and Hs ) 2.4-4.0 (m, 8, Hi, H5, -CH2S0, and -CHCH2S0), and 1.50-2.20

(m, 2, -CHCH2S0); 13C NMR (ppm, CDC13) 122.1*7 (d, C3 and C7), 97-93 (d,

C4 and C6), 90.86 (s, C2 and C8), 57-24 (d, C5 or Ci), 53-56 (d, Ci or

C5), 51-47 (t, -CH2S0), and 20.39 (t, CH2CH2S0); calcd m/e 206.0765, found 206.0768.

2a,8b-Dihydrocyclopent[cd]azulene (5)5.) •

A. Oxidation by Copper(ll). To a stirred slurry of 450 mg 187

(too mmol) of potassium t-butoxide in

10 ml of anhydrous dimethylsulfoxide

and 0.1 ml of water was added under

argon 200 mg (O.38 mmol) of dimesylate

108. The m ix tu re im m ediately tu rn e d

dark brown and, after stirring for

30 min at ambient temperature, the mixture was poured into 80 ml of ice water. To this solution was added a previously prepared solution of 2.0 g of copper(ll) chloride in 80 ml of water and 1 ml of concentrated hydrochloric acid buffered to pH 3 with sodium acetate. The expected copper complex did not form but rather gas evolution ms evidenced. The mixture ms made basic (pH 9) by the addition of potassium hydroxide. Pentane (20 ml) m s added and the mixture was stirred for 30 min at room temperature. The pentane layer was drawn off and the aqueous solution m s rinsed with pentane

(5 x 20 ml). The organic extracts were combined, rinsed with water

(5 x 20 ml), and dried over anhydrous sodium sulfate. The solution ms filtered and excess solvent m s removed by careful distillation under argon. Preparative vpc (2* x % " 6 f0 Q,F1 on 60/80 Chromosorb G at

115° with a flow of 50 cc/min, tret = 11 min) gave ca to mg of .55. identical by mass spectral and NMR data to a known sample.

B. Air Oxidation. The hydrolysis ms carried out as above on the same scale j however, after the reaction mixture was poured into ice water, the aqueous phase was extracted with pentane (25 ml) and ether

(k x 25 ml). No precautions were taken to exclude air. The combined organic solutions were rinsed with water (5 x 20 ml) and brine, dried over anhydrous sodium sulfate, and filtered. Evaporation of solvent

in vacuo (no heat) returned a brown viscous oil which exhibited an NMR

spectrum identical to that of 5J5*

Attempted Addition^ of JDibromocarbene to 5.5.* To a cooled (0°) mixture

of pentaene _55_ (ca i|0 mg, 0.26 mmol) and 121 mg (1.08 mmol) of potas

sium t-butoxide in 5 ml of pentane under argon was added 65.8 mg

(0.26 mmol) of bromoform in 2 ml of pentane. The slurry became

progressively darker as it was stirred under argon for 17 hr. The

mixture was poured into water and extracted with ether (5 x 20 ml). The

ethereal solutions were combined, rinsed with water (5 x 10 ml) and

brine, dried over anhydrous sodium sulfate, filtered, and evaporated

to dryness (no heat) to give a brown oil. The NMR spectrum indicated

formation of a multicomponent mixture as did the m ss spectrum which

contains a cluster of peaks at the desired 326 as well as major peaks

at 314-1, 355, ^29, and 503.

Attempted Preparation of 1A2.

A. From Treatment of with Activated Alumina. Aluminum oxide

(1*4. 8 g of Brinkman 90 active neutral activity i) was added to a 100 ml

three-necked round bottom flask equipped with two stoppers and a gas

inlet with stopcock. The alumina was activated by heating at 300-550°

for 1I4- hr under vacuum (ca 15 ram). The flask was cooled under vacuum

and then opened under argon. Dichloromethane (20 ml) was added and to

this magnetically stirred slurry was added 1.00 g (1.91 mmol) of

dimesylate 108 in 15 ml of dichloromethane. The resulting mixture was

stirred under argon for 2l+ hr. Filtration of the mixture and 189 evaporation of solvent gave a white frothy solid which by MMR was a

5O-5O mixture of starting m aterial 108 and oxepane 101.

B. From Treatment of 108 with DBU. To a solution of 200 mg

(O.38. mmol) of dimesylate 108 in 25 ml of anhydrous tetrahydrofuran was added 3^7 mg (2.28 mmol) of l,5~diazabicyclo[5* b.O^undecene (DBU).

Tlie solution was stirred and heated at reflux for 53 hr, then diluted • with water and extracted with chloroform (5 x 25 ml). The combined extracts were rinsed with 10% hydrochloric acid (2 x), water, and brine, dried, filtered , and evaporated to dryness. The NMR spectrum of the crude product did not contain any >CM< signals and was otherwise non-descript.

C. Via a Bis-selenoxide. To a magnetically stirred, cooled (0°) suspension of 227 mg (1.00 mmol) of o-nitrophenyl selenocyanate in 5 ml of absolute ethanol in a 25 ml three-necked round bottom flask fitted with two stoppers and a condenser with an argon inlet was added portionvri.se b2.0 mg (1.10 mmol) of sodium borohydride. Vigorous gas evolution was noted as the solution turned dark red. After 10 min, a solution of 262 mg (0.50 mmol) of dimesylate 108 in 10 ml of anhydrous tetrahydrofuran was added dropvri.se. The mixture was allowed to warm to ambient temperature and stirred for 11 hr. A solution of 0A5 ml

(k. 5 mmol) of 30% hydrogen peroxide and 2^2 mg ( 2 .b mmol) of triethylamine was added dropwise. The resulting mixture was stirred under argon overnight before dilution with kOO ml of water and extraction with chloroform (5 x 5° ml). The chloroform layers were combined, rinsed with water, saturated sodium carbonate solution, and brine. The solution was dried, filtered, and evaporated to dryness. Chromato graphy of the crude residue on 8 g of Florisil gave four components, one of which appeared by KMR and mass spectral data to be the desired product 1^2 . Efforts to isolate 108^ proved unsuccessful, however.

1,2,5 ? 6,6a, 6b, 6c ,6d-Octahydro-3P-hydroxy-Ii-phenylbenzo[l,3]cyclopropa-

[ l ,2,3-cd]cyclopropa[gh1pentaleno-l,6-biimine-7,8-dicarboximide (120a).

A solution of 1.20 g (3-70 mmol) of di-

azasn o u ten e 68a_ and 0 .1 0 g o f ro se b en -

g a l107 in A00 ml o f anhydrous re a g e n t

grade methanol was irradiated for 21 hr

with a Sylvania DYV tungsten halogen

projector lamp while a slow stream of 1 2 0 a oxygen was bubbled through the solution.

(The light was positioned in the immersion well in the usual manner and was cooled by a stream of air in addition to normal water cooling. )

The reaction mixture was transferred to a 1 i round bottom flask and cooled in ice. Following portionwise addition of 1.00 g (26.A mmol) of sodium borohydride, the solution was stirred at room temperature for 30 min, 10 ml of b IT potassium hydroxide solution was added, and the resultant mixture was concentrated in vacuo. ter (150 ml) and chloroform (100 ml) were added to the residue, t ayers were separated, and the aqueous phase was rinsed with c - ■'form (A x 100 ml).

The combined extracts were washed with water and bi dried, filtered, and evaporated to dryness to give a nearly quantitative yield of a pink, frothy solid. Chromatography on 2k g of silica gel gave 0.A8 g o f 191 unreacted 68a with chloroform elution and 0.73 g (95$ hased on recovered

68§J of allylic alcohol 120a vdth acetone -chloroform (l:l) elution.

Recrystallization from chloroform-pentane provided pure 120a as a white solid, mp 206-208°; 1760, 1695, 1500, 1410, 1135, and 1110 cm-1; UlclX 1H RMR (6, CDCI3) 7.22-7.55 (m, 5, aryl), 5-88 (dd, = 10.0 and

2.0 Hz, -CH=CHCH0H-), 5-46 (d, J = 10.0 Hz, 1, -CH=CHCH20H-), 4.90-

5.18 (m, 2, XHN<), 3.70 (m, 1, >CH0H), 3-25 (d, J = 6.0 Hz, 1, -OH), and 1.15-2.60 (m, 6, methylene and cyclopropyl); calcd m/e 347-1270, found 347.1275.

Anal. Calcd for C 2 OHi7N30 3 : C, 6 9 .1 5 , H, 4.93=, N, 12.10.

Found: C, 69.01; H, 5-05; N, 11.82. l ,2,3,6,6a,6h,6c,6d-Octahydro-3P-hydroxy-N-methylhenzo[l,3]cyclopropa-

[l,2,3-cd]cyclopropa[ghlpentalene-l,6-biimine-7,8-dicarboximide (120b).

In a process which duplicated the R preparation of 120a, 2.00 g (7*44 mmol)

of diazasnoutene 68a was photooxygenated

H using 0.10 g of rose bengal107 in 400 ml OH of anhydrous methanol followed by 3*00 g

1 20 b, R=CH 3 (80.0 mmol) of sodium borohydride.

After workup of the reaction, chromato

graphy on 30 g of silica gave 0.43 g of unreacted 68b with chloroform elution and 1.56 g (94$ based on recovered

£8b) of allylic alcohol 120b with 1:1 acetone-chloroform elution.

Recrystallization from chloroform-ether gave pure 120b. as a white solid, mp 194.0-195.0°; vjjjj 3410, 1755, 1685, and 1465 cnT1; xH NMR (6, CDC13) 192

5-90 (dd, J = 2.0 and 10.0 Hz, 1, -CH=CHCH0H-) 5-^8 (d, = 10.0 Hz,

1 , -CH=CHCH0H-) , 4.95 (to t , J = 4 .0 Hz, 2, >CHH<), 3 .8 - 4 .1 (m, 1 ,

CHOH), 3.03 (s, 3, methyl), and I.3-2.7 (m, 7j OH, methylene, and cyclopropyl); calcd m/e_ 285.1113 , found 285.1119*

A n al. Calcd f o r C15H i5W303: C, 6 3 . 16; H, 5-26; N, 14.74.

Found: C, 63.13; H, 5-29; N, 14.64. l,2,3,6,6a,6b,6c,6d-Octahydro-N-phenyl-3-oxohenzo[l,33cyclopropa-

[l,2,3-cd]cyclopropa[ghlpentalene-l,6-biimine-7,8-dicarboximide (jj^8a)-

A suspension of O.67 g (1.94 mmol) of

alcohol !20a and 4.0 g (46 mmol) of

Attenburrow manganese dioxide69 in

125 ml of dichloromethane was stirred

under anhydrous conditions at ambient

temperature for 18 hr. Filtration of 158a. R=Ph the mixture through Celite and evaporation of solvent gave 0.54 g (80 ,6ffo) of crude enone 158a- Fil tration through alumina and recrystallization from chloroform-pentane afforded pure 158a as a v/hite solid, mp 214-215°; 1770, 1710, 1500, — IuiLX. 1410, 1280, and 1135 cm"1; 1H MR (6, CDC13, 100 MHz) 7-28 (br s, 5, aryl), 6.98 (d, = 9-0 Hz, 1, -CH=CHC0-), 5-68 (d, = 9-0 Hz,

1, -CH=CHC0-), 5-12 (d, J = 4.0 Hz, 1, >CHN<), 4-96 (d, J = 4.0 Hz, 1,

>CHN<), 2.84 (ABq, = l 8 .0 Hz, AvAB = 12.6 H z), and 2 .0 8 (m, 4 , cyclopropyl); calcd m/e 345.1113} found 345*1117.

Anal. Calcd for C20H15N3O3: C, 69.55; H, 4-38; N, 12.17

Found: C, 69*16; H, 4 .4 l ; N, 11.93- 193 l,2,3?6,6a,6b,6c,6d-0ctahydro-N-methyl-3-oxobenzo[l,3]cyclopropa[l, 2,3- cdOcyclopropa[gh]pentalene-l,6-biimine-7,8-dicarboximide (158b)* To a

mechanically stirred, nitrogen .

blanketed solution of 11.10 g (11.4 ml,

l4 l mmol) of anhydrous pyridine in

400 ml of dichloromethane was added

portionwise 7*00 g (70.0 mmol) of

powdered dry chromium trioxide. After • 158 b, R=CH 3 ca 30 min, a solution of 3*62 g

(12.7 mmol) of alcohol 120b in 100 ml of dichloromethane was added dr op wise to the deep burgundy solution. A black tar formed immediately; after 45 min, the mixture was filtered through Celite. The dark organic solution was decolorized by rinsing with 10$ sodium hydroxide solution

(4 x 50 ml). Further washing with 5$ hydrochloric acid, 10$ sodium bicarbonate solution, and brine, followed by drying, filtration, and evaporation gave 3*34 g (93$) of enone 15&b as a white solid, mp I96.O-

197*5° (from dichloromethane-ether); 1765, 1695? 1460, and 1395 cm"1; 1H NMR (6, CDC13) 7*11 (d, J^b = 10.0 Hz, 1, -CH=CHC0-), 5-77 (d,

JAg = 10.0 Hz, 1, -CH=CHC0-), 5*17 (tr d, J = 4.0 Hz, 1, >CHN<), 4.97

(b r d , J = 5*0 Hz, 1 , >CHN<), 3*05 ( s , 3 , m eth y l), 2 .8 8 (ABq,

J^-g = 18.5 Hz, = 10.3 Hz, 2, methylene), and 1-95-2.40 (m, 4, cyclopropyl); calcd m/e^ 283.0957? found 283.0961. l,6,6a,6b,6c,6d-Hexahydro-3-methoxy-N-methylbenzo[l,3]cyclopropa-

[l,2,3-cd)lcyclopropa[gh]pentalene-l,6-biimine-7?8-dicarboximide (l62_)*

A solution of 2.00 g (7.O5 mmol) of enone 158b, 10.0 g (94.3 mmol) of trimethylorthoformate, and 1.00 g (110 mmol) of oxalic acid in 160 ml of 1,2-dichloroethane-methanol (1:1) was stirred and refluxed under nitrogen for 18 hr. The cooled solution v/as evaporated to dryness.

The white residue was taken up in chloroform (200 ml) and washed with

10$ sodium bicarbonate solution and brine. After drying over anhydrous

sodium sulfate and filtration, the solution was evaporated to give

2.75 g of oil. Chromatography on silica gel with ether elution

furnished 1.87 g (90$) of methoxydiene 162 as a v/hite solid, mp 189-5-

192.0° (sealed tube) (from ethyl acetate-ether); 1760, 1 7 1 0 , 1695>

1645, 1460, 1395, 1235, and 800 cm"1; XH HMR (6, CD013) 6.20 (d,

= 9-5 Hz, 1, -CH=CHC0CH3<), 5-70 (dd, J = 9-5 and 2.0 Hz, 1,

-CH=CHC0CH3) 5-05 (d, J = 2.0H z, -CEKJOCHq-), 5-00 (m, 2, >CHN<), 3-53

(s, 3, -OCH3), 2.95 (s, 5, -NCH3), 2.37 (m, 1, cyclopropyl), 2.00 (m, 2,

cyclopropyl), and O.63 (d, J = 3*5 Hz, 1, cyclopropyl syn to cyclohexadiene ring).

Anal. Calcd for Ci6Hi5N303: C, 64. 64; H, 5* 09; N, 14.13-

Pound: C, 6 k.6 6 -, H, 5-10; N, 14.27.

Attempted Carbene Addition to 162• A mixture of 297 mg (1.00 mmol) of

methoxydiene 162 and 529*4 mg (l.OO mmol) of phenyl(tribromomethyl)

mercury in 3 0 ml of anhydrous benzene v/as stirred and heated at reflux

in an inert atmosphere for 4 hr. The solution was allowed to cool

(precipitate formation) and evaporated to dryness. Chromatography on

20 g of silica gel with benzene elution gave 231.1 mg of product which

was > 90$ enone 158b by mass spectral and HMR data.

2a,8b-Dihydro-6-methoxycyclopent[cd]]azulene (166). With strict

adherence to the procedure outlined for the preparation of 94, 594 mg (2.00 mmol) of methoxydiene l62 was

hydrolyzed with 800 mg (20.0 mmol) of

powdered sodium hydroxide and ^0 ml of

2-propanol and oxidized using 1.7^ S

■J6 6 (20.0 mmol) of activated manganese

dioxide69 and 75 ml of pentane. Workup

as before gave a sublimed (^0-50°/

3. 5 x 10-4 mm) yield of 303*5 mg (82.5 1°) of ether 166 as a bright yellow solid mp 23-27°i 2910, 1625, 1510, 1^55, 1^00, 1255, 1220, 11^5,

1020, 850, 830, and 805 cm-1:, x jjj°ctane 327 (e ^.0 x 103) and 2^7 nm

(5*^ X 104); ^ HMR (6, CDOls, 100 MHz) 6.1{-3 (dd, J ijS = 5*2 Hz,

J i,2a =2.7 Hz, 1, Hi), 6.39 (dd, J3>4 = 5*2 Hz, J2a,4 =2.7 Hz, 1,

H4), 6.30 (dd, J7j8 = 6.5 Hz, J5)8 = 1.5 Hz, 1, Ha), 6.2>i (br s, 1, Hs),

6.11 (dd, J2a,3 =2.7 Hz, J3j4 = 5*2 Hz, 1, Hs), 6.0U- (dd, J7?a = 6.5

Hz, J5)7 = 1.5 Hz, 1, H7), 5*95 (da, J2,2a =2-7 Hz, J i>2 = 5*2 Hz, 1,

H2), 3*90 (m, 1, Hsa), 3*68 (s, 3, -0CH3) and 2.03 (d, J2aj8b = 7*5 Hz,

1, Hst)). Spin decoupling: saturation at 3*90 collapsed the peak at

2.03 to a broad singlet, the peaks at 5*95 and 6.11 to doublets, J i}2 =

Js ,4 = 5*2 Hz, and the peaks at 6*39 and 6.^3 to doublets with the same coupling constants. Double irradiation at 5*95 or 6.11 simplifies the signal at 3*00 to a broad doublet, J2ajsb =7*5 Hz-, 13C NMR (ppm,

CDC13) 158.65 (s, C6), 139.85 (s, 04a), 136.66 (d, Ci), 133*67 (s, C8a),

132.81 (d, C4), 131.05 (d, Ca or C5), 130-95 (d, C5 or C8), 113-87 (d,

C3 ) , 111-lh (d, C2), 10l|.l8 (d, C7), 55.63 (d, C2a), 55*1^ (q, methyl), and ^6.32 (d, C8t>); calcd n^/e 18^.0888, found l81(-.0891. 196

Attempted Addition of Dibromocarbene to 166.

A. Bromoform-t-Butoxide Method. Under an argon atmosphere,

185 mg (1.65 mmol) of possium t-butoxide was added to a yellow solution of 152 mg (0.825 mmol) of ether 166 in 30 ml of anhydrous benzene. To the brown slurry was added dropwise over 30 min 795 M.1 of a solution of 1 ml of bromoform in 10 ml of benzene (0.825 mmol of bromoform).

A slightly exothermic reaction occurred as the mixture turned darker.

The slurry was stirred overnight at ambient temperature, then water was carefully introduced. Following the addition of benzene (50 ml), the layers were separated and the organic phase was washed with water

(3 x 50 ml), dried over anhydrous sodium sulfate, filtered, and evaporated to dryness to give a greater than quantitative yield. The dark green, oily residue contained (> 90$) starting material as evidenced by NMR.

B. Seyferth's Reagent. A solution of 152 mg (0.825 mmol) of ether 166 and 437 (0.825 mmol) of phenyl(tribromethyl) mercury in 10 ml of dry benzene was stirred and heated at reflux under argon for 10 hr.

The solution was cooled to ambient temperature and the precipitate removed by filtration. Evaporation to dryness gave > 100$ recovery of a dark green oil. By NMR analysis, this substance was largely starting l66. No evidence of adduct formation was found.

Preparation of Intermediate Endoperoxide 121b. A cooled (-10 to -15°, water-methanol bath) mixture of 267 mg (1.00 mmol) of norcaradiene 87b and 100 mg of Photox107 in 200 ml of dichloromethane was irradiated for 2 hr with a Sylvania DYS tungsten halogen projector lamp while a slow stream of oxygen was bubbled

through the solution. (Hie lamp was

positioned in the immersion w ell'in the

usual manner and was cooled by a stream

of air in addition to the normal water

— cooling. ) Hie Photox was removed by 121b. R=CH3 filtration and the solution m s evaporated to dryness (no heat) to return a quantitative yield of the therm ally unstable endoperoxide 121b, 1H HMR (6, CDC13) 6.12 (t, J =

4 .0 Hz , 2 , v i n y l ) , 5-10 ( t , J = 3 -° Hz, 2 , >CHN<), 4.93 ( t , J = 4 .0 Hz,

2, allyl), 3-00 (s, 3) methyl), 1.80-2.40 (m, 3, cyclopropyl), and

I .58 (d, J = 4.0 Hz, 1, cyclopropyl syn to endoperoxide ring).

2 ,3 ,4 ,5 -Bi sepoxyde c ahydro -N-phenylbe nzo [ 1,3 ] eye lopr opa [ 1,2,3 - cd] - cyclopropa[gh1pentalene-l,6-biimine-7,8-dicarboximide (174b). A

solution of 100 mg (0.28 mmol) of

endoperoxide 121b (from the photo

oxygenation of diene 8jb) in 8 ml of

1,2 dichloroethane was heated at reflux

for 1.5 hr. Hie solvent m s evaporated

and the residue m s recrystallized from 1 7 4 b. R=CH3 1,2-dichloroethane to give 80 mg (80$>) of bisepoxide 174b, mp 290-295°; ' S 01 1765’ and 1700 cm-1;

1H NMR (6, py-ds, 90 MHz) 7.97 (dd, J = 8.0 and 1.5 Hz, 2, aryl), 7.21-

7 .4 6 (m, 3 , a r y l ) , 5-35 ( t , J = 2-5 Hz, 2 , >CHN<), 3 -12-3-39 (m, 4 , epoxide methines), and 1.94-3.17 (m, 4, cyclopropyl). Anal. Calcd for C20H15N3O4: C, 6 6 . hj', H, 3 .l 8 ; N, 1 1 .63.

Found: C, 66.0 8 ; H, 3 .2 3 ; N, 11.39-

2,3,3,5-Bisepoxydecahydro-N-methylbenzo[l,3]cyclopropa[l,2,3-cd]cyclopropa[gh]pentalene-l,6-biimine-7,8-dicarboximide (lj|+a). From the

R thermolysis of endoperoxide 121a., there r ^ o was obtained bisepoxide 3j3b whose 1H 3 HMR was identical to that of bisepoxide 0 . lj3b except for the absence of aryl signal and the addition of a peak at a . 1 74 R=Ph 3.O3 6 (s, 3, methyl).

Attempted Synthesis of YQ. In the manner described above for the p re p a ra tio n o f 121b , 267 mg (l.O O mmol) o f n o rc a ra d ie n e 8j b was p h o to

oxygenated in 100 ml of dichloromethane. The sensitizer was removed

and l38 |il (155 mg, 1.25 mmol) of freshly distilled trimethylphosphite was added. The solution was stirred for 1.5 hr at 0° and 3 hr at

ambient temperature. Evaporation of solvent and chromatography of the

residue on silica gave 65.3 mg of starting diene contaminated with

phosphates with ether elution, 55>5 mg of uncharacterized phosphorous

compounds with chloroform elution, 121 mg of bisepoxide r£ 3 b w ith 1 :1

chloroform-methanol elution, and 131-5 mg of a tan frothy solid, a

multicomponent mixture by HMR, with methanol elution.

4,5 -Epoxydecahydro-N-raethyl-3 -oxobenzo[l,3 ]cyelopropa[l, 2,3 -cd]cyclo-

propa[gh]pentalene-l,6-biimine-7,8-dlcarboximide (iJO.). To a stirred

solution of 283 mg (1.00 mmol) of enone 158a in 100 ml of absolute 199

ethanol was added, in one portion, a

solution of 69 mg (0.5 mmol) of potas

sium carbonate and 102 mg (3.00 mmol)

of hydrogen peroxide in 5 ml of water.

The solution darkened slightly upon

' addition, became cloudy, then cleared. 170 A f te r h hr, the mixture was concentrated to ca ■§ volume and diluted with k00 ml of water. The aqueous solution was rinsed with chloroform (5 x 50 ml). The combined organic layers were washed with saturated sodium bicarbonate solution and brine, dried over anhydrous sodium sulfate, filtered, and evaporated to dryness to yield 22k. 7 mg (75*2$>) of epoxyketone 170. Recrystalliza tion from acetone-chloroform then ethyl acetate gave pure 170 as a white solid, mp > 262° (dec); 1765, 1755, 1710, 1690, lk6o, lkOO,

1255, 12^5, 970, 760, and 7^5 cm"1; ^ HMR (6, CDCI3) 5-17 (d,

J = 3*0 Hz, 1, >CHH<), 4 .8 7 ( d , J = 3 .0 Hz, 1 , >CHN<), 3-63 (d ,

J = k. 5 Hz, 1, methine a to ketone), 3*13 (hr s, k, methyl and methine

P to ketone), 2.88 (ABq, J^g = 15*0 Hz, Av^g = 26.0 Hz, 2, methylene), and 2.00-2.20 (m, k, cyclopropyl); calcd m/e_ 299*0906, found 299*0910.

Anal. Calcd for C15H13N3O4: C, 6 0 .2 0 ; H, k .38; H, lk .O k .

Found: C, 60.17; H, k. 50; N, 13*89* l,2,5,6,6a,6b,6c,6d-Octahydro-N-methyl-2-oxobenzo[l,3)cyclopropa-

[l,2,3-cd]cyclopropa[gh]pentalene-l,6-biimine-7,8-dicarboximide (l69_).

A* From Epoxyketone 170* A solution of 22k mg (0.75 mmol) of epoxyketone 170, 1 ml of hydrazine hydrate, and k drops of glacial 200

acetic acid in 15 ml of dry dioxane

was heated on a steam cone for 0.5 hr.

The mixture was concentrated nearly to

dryness and diluted with 5 % hydrochloric

acid. The aqueous solution was

extracted with chloroform (5 x 20 ml)

1 6 9 and the combined organic layers were

rinsed with saturated sodium bicarbonate solution and brine, dried,

filtered, and evaporated. The tan oily residue was purified by

preparative thin layer chromatography to give b7«5 mg of starting 1J0

and 50-0 mg (3C$ based on recovered 170) of allylic alcohol rj2.

This product was not characterized but was immediately oxidized

with Collin's reagent. To a nitrogen-blanketed solution of 185 mg

(2.3^ mmol) of dry pyridine in 10 ml of dichloromethane was added with

vigorous stirring 106 mg (1.06 mmol) of dry chromium trioxide. After

20 min, crude 172 in a minimum volume of dichloromethane was added

dropwise to the burgundy solution. The solution darkened with the

formation of black tars and, after 15 min, the mixture was filtered

through Celite. The organic solution was rinsed with 10 % sodium

hydroxide solution (2x), 10$ hydrochloric acid, saturated sodium

.bicarbonate solution, and brine. Drying, filtration, and evaporation to

dryness provided ^5*3 mg (91.2$1) of enone 163, identical to that

synthesized by method B.

B. From Diazasnoutene 68b. To a mechanically stirred solution of

7-12 g (225 mmol) of pyridine in 300 ml of dichloromethane in a 5 00 ml three-necked Morton flask fitted with a stopper and a condenser topped with an argon inlet was added portiomri.se 11.2 g (112 mmol) of dry chromium trioxide. The Collin's reagent ms stirred for 15 min, then transferred under argon via a glass adapter to a solution of 2.00 g

(7.44 mmol) of diazasnoutene 68b in 10 ml of dichloromethane in an identically equipped 2 i, three-necked Morton flask. The flask was stoppered and the mixture was vigorously stirred for 12 hr at ambient temperature. A second addition of Collin's reagent [from 7*40 g

(74.0 mmol) of chromium trioxide and 11.7 g (l48 mmol) of pyridine in

200 ml of dichloromethane] was made and the mixture was stirred for an additional 12 hr. The mixture was filtered through glass wool and the solids were thoroughly rinsed with dichloromethane. The combined organic solutions were rinsed with 10$ sodium hydroxide solution

(2 x 200 ml), 10$ hydrochloric acid (150 ml), saturated sodium bicarbonate solution (150 ml), and brine (150 ml), dried, filtered, and evaporated to give ca 1 g of brown oil. The inorganic solids were dissolved in base and the resulting green flocculant solid was removed by filtration. The solids were thoroughly washed with dichloromethane and the aqueous solution was rinsed with dichloromethane (4 x 150 ml).

The combined organic solutions were processed as before to give an additional 1 g of tan oil. Chromatography of the combined residue on

80 g of silica gel gave, with chloroform elution, 420 mg of unreacted

68b and 720 mg (43*4$ based on recovered 68b.) of enone 163,. Re

crystallization from ethyl acetate gave pure 163, as a light yellow

solid, mp 224.0-226.5° (dec sealed tube); 1760, 1700, 1660, l455>

and 1395 cm-1; XH MMR (6, CDCI3) 6.61 (dt, J = 4.0 and 10.0 Hz, 1, -CH=CHCO-), 5-95 (at, J = 2.0 and 10.0 Hz, 1, -CH=CHC0-), 5-70 (d, j = b.o Hz, l, >CM<), b.97 (d , J = b.o Hz, l, >chn<), 3.07 (s, 3, methyl), 2.87 (dd, J = 2.0 and b.O Hz, 2, methylene), 2.^+2 (d, J = ^.0

Hz, cyclopropyl syn to cyclohexenone), and 1.9°-2.*l-0 (m, 3 5 cyclo propyl); calcd m/e 283.0957} found 283.0961.

Anal. Calcd for C15H13N3O3: C, 63.60; H, b.62\ N, l 4 . 83.

Found: C, 63. H, b.69; N, 1 ^ .8 l.

1 ,6, 6a , 6b, 6c , 6d-Hexahydro -2-methoxy-K-methylhenzo [ 1,3 ]c yc lopropa-

[l,2,3-cd]cyclopropa[gh]pentalene-l,6-biimine-7}8-dicarboximide (173.)*

In a process vdiich mirrored the c h 3 preparation of enolether 162, 2^3 • 8 mg

/I (0.86 mmol) of enone l 6|5 was converted

to methoxydiene 175 hy use of ^0 mg of

p-toluene sulfonic acid, 915 mg (8.60

mmol) of trimethylorthoformate, and 1 75 10 ml of 1:1 methanol-1,2-dichloro-

methane in 68.3$ yield. Chromatography on 5 g of silica gel with

chloroform elution followed by ethyl acetate-ether recrystallization

gave analytically pure 175 as a white solid, mp 156.5-158.0°; vmax 1768,

1705, 1^65, 1395, 1260, 12U0, 790, 75O, and 725 cm-1; HMR (6, CDC13)

5.I1O-6.IO (AB portion of ABX, Av^g = 10 Hz, = 10.0 Hz, J^x = 1 Hz,

IgX = 8 Hz, 2, -CH=CH-CH=C0CH3-), 5-60 (m, 1, >CHN<), k. 9 0- 5-20 (x

p o r tio n o f ABX, 1, -CH=C0CH3- and m, 1 , >CHN<), 3 -6 7 ( s , 3 , -0CH3 ) ,

3.00 (s, 3, >NCg3), 2.20-2.60 (m, 1, cyclopropyl), 1.80-2.20 (m, 2,

cyclopropyl), and 0.90 (d, J = U.O Hz, 1, cyclopropyl syn to 203 cyclohexadiene); calcd m/e_ 29T•1113 , found 297*1119*

Anal. Calcd for C16Hi5N302: C, 6 k. 6k-, H, 5*09; N, 1^.13*

Found: C, 61k 50* H, 5*12; N, i lk 07. .

2a,8b-Pihydro-3-methoxycyclopent[cd]azulene (l68). In a manner fully-

duplicating the procedure for the

preparation of 9!b the hydrolysis and

o x id a tio n o f 396 mg ( 1*33 mmol) o f

ether 175, was carried out using 53^- mg

(13*3 mmol) of activated manganese

- dioxide,107 and 75 ml of pentane to give,

after sublimation (36 0[ 7 x 10"4 mm),

172.^ mg (72.1$) of 168 as a bright yellow solid, mp ^5-50 ; vmY 1610,

1235, 1205, 1155, 795, and 7^5 cm"1* 3^ (e 1.8 x 103),

262 (3.7 x 104), and 238 nm (2 .9 x 104 j* XH JJMR (6, CDC13 , 100 MHz)

6. 50-6.90 (m, AB portion of ABX sim ilar to Av^g = 5*0 Hz, J^g = 8.0 Hz,

= 1*. 2 Hz, = 1.8 Hz, 3, Hr, Hs , and H4 ), 6.1)3 (d d , J l j2 = 5 .5 Hz, jo 2a =2.0 Hz, 1, H2), 6.25 (d with additional splitting, Jx 2 = 5*5

Hz, 1, Hx), 6.10 (dd, J3,4 = 5*5 Hz, J2a,3 = 2.0 Hz, 1, H3), 5*85 (o.»

X portion of ABX, 1, H6), 3*75 (dt with additional splitting,

J.2a ,8b = 7*5 Hz, J2,2a = J.2a,3 =2.0 Hz, 1, H2a), 3*6^ (s, 3, methyl), and I.85 (br d, J2a,sb = 7*5 Hz, 1, Hab)* Spin decoupling: saturation at 6 3*75 collapsed the signal at 1.85 to a broad singlet and the peaJcs at 6.10 and 6.^3 to doublets, J2j2a = J>a,3 =5*5 Hz. Conversely,

double irradiation at 6.10 or 6.^3 simplified the m ultiplet at 3*75‘»

13C HMR (ppm, CDCI 3 ) 1^5.67 (s, Cs ), 1^1.04 (s, C8a), 136.07, 131.06, 150.09, 129-66, and 128.58 (d, C4, Ci, C8, C7, and Cs), 125.08 (d, C2

or C3), 121.5T (s, C^g,), 113*80 (d, C3 or C2), 60.08 (q, CHs), 55*07

(d, C2a or 08b)> and ^5*31 (3, C8b or C2a)', calcd m/e 18^.0888, found

184.092.

Attempted Hydrolysis of Bisepoxide 17j£b* To a stirred slurry of 1.12 g

(10.0 mmol) of potassium t-butoxide in 35 ml of dry dimethylsulfoxide

and 0.15 ml of water was added 299 mg (l.OO mmol) of bisepoxide lT^b-

The mixture immediately darkened. After being stirred for 30 min at

ambient temperature, the mixture was poured into ^00 ml of water and

extracted with ether. The ether extracts were combined and diluted

with 50 ml of water. Activated manganese dioxide107 (869 mg, 10.0 mmol) was added and the mixture was stirred overnight under argon. (No gas

evolution v/as noted. ) As a precaution, the aqueous layer v/as diluted with 100 ml of ether and an additional 869 mg (10.0 mmol) of manganese

dioxide v/as added. Again, no gas evolution was noted, but the mixture v/as also stirred under argon overnight. For both mixtures, the sus pension v/as filtered and the resulting solution v/as extracted with

ether. The ether extracts v/ere washed with water and brine, dried over anhydrous sodium sulfate, filtered, and evaporated. No residue

remained from either mixture.

l , 2, 5, 6, 6a , 6b , 6c ,6d-Octahydro- 2 , 5-bishydroxy-N-methylbenzo[l, 3 ] -

cyclopropa[l, 2,3 -cd']cyc lopropa [gh]pentalene-l, 6-biimine-7,8-

dicarboximide (179). A solution of endoperoxide 121b (from singlet

oxygen oxidation of 133 mg (0.50 mmol) of diene &7b) v/as prepared as

described previously. The solution v/as filtered and transferred to a 205

500 ml round bottom flask, cooled to ' CH3 -10°, and reduced with 100 mg (2.6b

J f mmol) o f sodium b o ro h y d rid e. The

m ix tu re was s t i r r e d a t -10° f o r 30 min

th e n 1 ml o f 10$ ammonium c h lo rid e

s o lu tio n was added. The r e a c tio n 179 0H mixture v/as diluted with 100 ml of water and the aqueous layer was washed with chloroform (100 ml). The combined chloroform extracts were rinsed with water and brine, dried, filtered, and evaporated to dryness to give a quantitative yield of

179; vjgpG 3500, 336O, 17^5, and 1670 cm'1; 1H HMR (6, CDC13) 6.18

(dd, J = 3.0 and 4.0 Hz, 2, vinyl), 5-18 (dd, J = 2.0 and 3*0 Hz, 2,

>CHN<), 5-00 (d d , J = 3 .0 and 4 .0 H z, 2 , >CH0H), 3-10 (b r s , 2, -OH),

3.O5 (s, 3, methyl), and 1.80-2.20 (m, 4, cyclopropyl). The N-phenyl isomer has calcd m/e 363-1219, found 363-1224.

Attempted Catalytic Hydrogenation of Diol 179- The crude diol 179. was taken up in 100 ml of methanol-ethyl acetate (l:l) and 5 ml of dichloromethane. The solution v/as added to a stirred, prereduced suspension of 80 mg of platinum oxide in 5 ml of ethyl acetate. When hydrogen uptake ceased, the catalyst v/as removed by filtration and the solution was evaporated to dryness to give a white frothy solid which (HMR analysis) v/as seen to be a complex mixture on the basis of multiplets from 6 1.0 to 5-0-

Attempted Oxidation of Diol 179.

A. With Collins Reagent. Treatment of 150 mg (O.5O mmol) of 206 diol 179 with a large excess of Collins reagent [from 1.00 g of dry chromium trioxide, 1.7^- g (22.0 mmol) of pyridine, and 50 ml of

dichloromethane prepared as in the oxidation of .158b] gave no reaction after 21 hr of stirring under nitrogen.

B. With DDQ,. A solution of 150 mg (0.50 mmol) of diol 179 and

227 mg (1.00 mmol) of 2,3-dichloro-^,5-dicyanoquinone (DDQ) in 75 ml

of dry dioxane was stirred overnight under argon. The solution was

evaporated to dryness and chromatographed on silica gel hut no

recognizable products were eluted.

Attempted Catalytic Hydrogenation of Endoperoxide 121b. Endoperoxide 121b

(from oxidation of 267 mg (l.OO mmol) of diene 8jb) m s dissolved in

ethyl acetate (120 ml) and added to prereduced catalyst (80 mg of

platinum oxide in 30 ml of ethyl acetate) in an atmospheric hydro

genation apparatus. After being stirred for 2 hr at -10°, the mixture

was allowed to warm to ambient temperature and stirred for 3 hr. The

catalyst m s removed by filtration and the solution was evaporated to

g iv e 27b.5 mg of frothy, white solid. Ho olefinic signals were present

in the HMR spectrum hut the 6 1.0-2.2 region was much too intense.

Further, attempted Collins oxidation of the crude product, using

300 mg (3.00 mmol) of dry chromium trioxide, and ^98 mg (6.3 mmol) of

pyridine, returned only starting material.

Attempted Preparation of Eneketol 177.

A. By Treatment of Endoperoxide 121b with Basic Alumina. A

solution of endoperoxide 121b (from oxidation of I .50 g (5-62 mmol) of diene &fb) in 30 ml of dichloromethane m s added to a slurry of 47 g of alumina (Woelm Basic, Activity i) in 30 ml of dichloromethane. The mixture was stirred at -10° overnight and the slurry was poured onto a chromatography column. Elution with 1:1 chloroform-methanol gave 1.50 g of bisepoxide 174b and only traces of the desired 17X’

B. By Treatment of Endoperoxide 121b with Triethylamine. To a cooled (0°) solution of endoperoxide 121b (from 267 mg of diene 87b) in

15 ml of dichloromethane in a 25 ml round bottom flask was added 1.0 ml of triethylamine. The solution turned brown and solid was deposited.

After being stirred for 1.5 hr, the mixture was poured into 10$ hydrochloric acid. The aqueous phase m s washed with chloroform

(2 x 50 ml) and the combined extracts were rinsed with 10$ hydro chloric acid, water, and brine, dried, filtered, and evaporated to give a tan solid. By NMR analysis, the product was seen to be a mixture containing mostly bisepoxide 174b and some desired product 177- APPENDIX

Variable Temperature 1H NMR Data, Computed Equilibrium

Constants,and Gibbs Free Energy Values for Semi- bullvalenes 9^_, 9^., §6^, 9Tj an^ 98.

208 209 i a* Table IX. Variable Temperature H KMR Data, Computed Equilibrium Constants (Keq) and Gibbs Free Energy Values (AG°) for the Fluxional System ^ 94a-

Temp Chem ical Mol Seq AG° s h i f t 13 f r a c tio n (94b -* 94a) °C H43 He (94a/94b) k cal/m o l

+32.3 2 .7 3 0.120 7 -3 1.21

+ 3 -7 2 .6 8 0.105 8 .6 1.18

- 2 9 .6 2.61*. 0.092 9 .8 1.13

-4 4 .4 2 .5 7 0.071 13. I 1.17

-5 6 .3 2 .5 6 0.068 1 3 .8 1.13

- 7 2 .8 2 .5 2 0.055 1 7 -1 1.13

- 88.lt 2. Iff 0.040 2 4 .0 1.17

-9 9 -9 2 .4 6 0.037 2 6 .1 1.13

v a100 MHz in CS2-TMS:, referenced to internal TMS and corrected so that the H3 and H7 resonance remained constant at 5*04 S. 210

Table ,X. Variable Temperature H NMR Data,a Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9hp ~

Temp Chem ical Mol & q AG° s h i f t b f r a c ti o n (9*£b -• §!*a) °C H*, H6 9kb c a l/m o l

+32-3 2.92 0.1 7 8 1*. 60 928

+ 3 -7 2.90 0.172 1*. 80 865

- 2k. 6 2. 81* O .I5I* 5 .5 0 81*3

-kk.k 2-77 0.132 6 .5 6 856

- 5 6 .3 2.76 0.129 6 . 7 *1- 823

- 7 2 .8 2.72 0.117 7 .5 5 806

- 8 8 . k 2.67 0.1 0 2 8 .8 5 802

-9 9 -9 2.66 O.098 9 .1 6 761*

a100 MHz in 1:1:1 CD2C12-CF2C12-TMS; ^referenced to internal TMS and corrected so that the CHDC12 resonance remained constant at 5*32 6. 211

Table XI. V ariable Temperature 1H NMR D a ta ,3" Computed E q u ilib riu m Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9£b r.s 9£a*

Temp Chem ical Mol Seq AG° s h i f t *0 f r a c ti o n (2 & ^ 2 § a) °C ' H4, He 9£i> c a l/m o l

+34.2 3 -5 4 0 .3 7 1 .7 1 327

- 3 .4 3 -66 o . 4 i 1 .4 6 204

- 9-7 3 .7 0 0 .4 2 1-39 172

-3 4 .3 3 -79 0 .4 5 1.24 103

- 51.2 3 .8 3 0 .4 6 1 .1 8 71

-5 1 -7 3 .8 7 0 .4 7 1.12 53

-5 6 .7 3.88 0.47 1.12 45

-6 6 .0 3 -9 4 0 .4 9 1 .0 4 15

-7 5 . ^ 3 .9 6 0 .5 0 1.00 0

-85.6 4.o4 0.52 0 .9 1 -3 4

a100 MHz in CS2-TMS; ^relative to internal TMS and corrected so that the Aryl-CHg- resonance remained constant at 3-51 &• 212

Table XII. Variable Temperature 1H NMR Data,a Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9^b t* 9£a*

Temp Chem ical Mol & q ag° sh ift'* 5 f r a c tio n (2a-: 95a) °C H4, H6 2a (9 £ a /§ a ) c a l/m o l

-3 ^ -3 if. 19 0.57 O.7 6 130

-5 1 -T 4 .2 7 0.59 0 .6 8 170

-5 6 .7 k.29 0 .60 0 .6 7 173

- 6 6 .0 b.32 0.61 o.6if l8k

-75*^ If. 35 0.62 0 .6 2 188

- 8 5 .6 U.3 9 • 0.63 0 .5 8 203

a100 MHz in 1:1:1 CD2C12-CF2C12-T1'J1S^ ^relative to internal TMS and corrected so that the CKDCI 2 resonance remained constant at 5*32 6. TaM ejQ II. Variable Temperature 1H NMR Data,a Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9§b =* 96a.

Temp Chem ical Mol *eq AG° s h i f t 13 f r a c tio n (96b -» 9oa) °C H4, H6 96b (96a /3 £b) c a l/m o l

+33 0 5 3 .kb 0. 3 l|- 1.95 -1+12

+33 -^ 3 - 1+6 0.35 1.90 -393 +30.0 3.5O O.36 1 .8 0 -356

-2.6 3 ° 53 O.37 1-73 -297

-12.1+ 3-58 O.38 1 .6 2 -251

-1 6 .6 3 ° 59 O.39 1 .5 8 -2l+l

-2 6 .8 3 .6 1 0.39 1 .5 8 -219

-3 3 .0 3 .6 5 O.llO 1.1+8 -189

-3k.k 3 .6 6 O .lll 1. 1+5 - 1 8 0

- in . 8 3 .6 8 O .lll 1. 1+5 -161+

-Mi. 3 3 .7 0 0.1+2 1-39 -1 5 0

-5 ^ .6 3 * 7l+ 0. 1+3 I .3 1 - 1 2 1

-5 5 .0 3-75 0.1+3 1.31 - 1 1 6

-6 3 .1 3*79 0. 1+5 1 .2 2 - 90

-6 7 .8 3 .8 1 0. 1+5 1 .2 2 - 78

- 79.5 3 .8 6 0 . 1+7 1.11+ - 50

- 92.5 3 .9 6 0.50 1 .0 0 0

a100 MHz in CS2-TMS; ^relative to internal TMS. 2ib

Table _XIV. Variable Temperature 1H NMR Data,a Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 96b ~ £6a.

Temp Chem ical Mol & q AG° s h i f t b f r a c ti o n (96b -* 9§a) °C H*, Hs 96b (9 6 a /§ 6b) c a l/m o l

-3 3 .0 3 .8 5 0 .4 6 1 .1 5 -68

- i n . 8 3 .8 8 0 .4 7 1 .1 1 -48

-5 5 -0 3 .9 0 0 .4 8 1 .0 8 -35

- 6 7 .8 3 .9 2 0 .4 9 I .05 -23

-7 9 -5 3 .9 3 0 .4 9 I .05 -17

a100 MHz in 1:1:1 CD2C12-CF2C12-TMS‘, ^referenced to internal TMS and cox-recLed so that the CKDC12 resonance rem ined constant at 5*32 6. 215

Table XV. Variable Temperature 1H NMR Data,a Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System 9Tb ~ 97a.

Temp C hem ical Mol AG° «eq s h i f t b f r a c ti o n °C H4, H6 97b k cal/m o le

-•32.5 5-33 O.92 O.O87 l.k9

+16.9 5 .3 8 0 .9 ^ 0.069 1.5k

+ 5 -0 5 -1i-3 0-95 0.052 I .65

-1 7 .1 5 .^ 9 0 .9 7 0.0 5 2 1-75

-2 9 .3 5-52 O.98 0.0 2 2 1.85

a 100 MHz in CD2C12-TMS; referenced to internal TMS and corrected so that the CHDC12 resonance remained constant at 5-32 G. 216

Table XVI. Variable Temperature XH MR Data,3, Computed Equilibrium C onstants (Keq), and Gibbs F ree Energy Values (AG°) f o r the Fluxional System § 8b i* 98a-

Temp Mol Chemical 5>q AG° shiffcb f r a c ti o n (§8b - 98a) °C ' H4 , Hs 28b (98a/98b) c a l/m o l

-0 0 .9 k. 5b 0 .6 8 O.h-8 lAlj-

+2 0 .0 lu68 0 .7 2 0.39 5^9

0 .0 U.76 0 .7 5 0 .5b 586

- 1 5 .0 h-.8ii- 0 .7 7 0.30 618

-3 0 .0 4 .9 0 0 .7 9 0.27 633

-kb. 0 5.02 0 .8 3 0.21 711

-5 7 -0 5.10 0 .8 5 0 .1 8 737

-7 3 -0 5.22 0 .8 9 0.13 812

•u ®90 MHz in CD^Cla-TMSj relative to internal TMS corrected so that the CHDC12 resonance remained constant at 5-32 6. 217

Table XVII. Variable Temperature 1H WMR Data,9, Computed Equilibrium Constants (Keq), and Gibbs Free Energy Values (AG°) for the Fluxional System £§b r-; 99a.

Temp Chem ical Mol ^ q AG° s h i f t '*3 f r a c tio n (2 2 b -* 2§a) °C H4, Hs 2 2 S3 (99b/99a) c a l/m o l

+32 .6 4 .8 6 0 .7 8 • 29 778

+2 6 .0 k.9 0 0 .7 9 • 27 779

+14.8 1+.96 0 .8 1 .24 818

+ 6 .6 4 .9 8 0 .8 1 . 2k 817

+ 3 .3 5 .0 2 0.83 .2 1 859

+ 1 .5 5 .0 2 0.83 .23. 853

- 1 .5 5 .0 4 0.83 .2 1 844

- 9 -1 5 .0 8 o .8 4 •19 882

- 1 1 .2 5 .0 8 0 . 8k .19 875

- 1 5 .2 5.13 0 .8 6 •17 923

- 18.3 5.13 0 .8 5 •iT . 912

- 2 0 .8 5 .1 6 0 .8 7 • 15 943

-2 7 .9 5 .2 0 0 .8 8 .1 4 962

•4 5 .5 5-27 O.90 .1 1 1000 • 0 -6 1 .9 5. 3^ 0 .9 2 00 1045

a100 MHz in 1:1:1 CD2C12-CF2C12-TMS; ^referenced to internal TMS and corrected so that the CHDC12 resonance remained constant at 5*32 &• 218

REFERENCES

1. A. C. Cope and E. M. Hancock, J. Amer. Chem. Soc. , 60, 2903 (1938).

2. (a) L. Claisen and E. Tietze, Chem. Ber. , jj8, 275 (1925 )?> (b) L. Claisen, ibid., 45_, 3157 (1912); (c) C. D. Hurd and M. A. Pollack, J . Amer. Chem. S o c ., 60, 1905 (1938).

3. (a) A. C. Cope and E. M. Hardy, ibid. , 62, 44l (1940); (b) A. C. Cope, K. E. Hoyle, and D. Heyl, ibid. , 63^, 1843 (l94l); (c) A. C. Cope, C. M. Hoffman, and E. M. Hardy, ibid. , 1852 (194-1); (d) D. E. Whyte and A. C. Cope, ibid. , 66, 1684 "(19W ).

4. W. von E. Doering, V. G. Toscano, and G. H. Beasley, Tetrahedron, 27, 5299 (1971).

5- (a) W. von E. Doering and W. R. Roth, ibid. , 18, 67 (1962); (b) W. von E. Doering and W. R. Roth, ibid. , 19, 715~Tl963) (c) W. von E. Doering and W. R. Roth, Angew. Chem. , Intern. Ed. Engl. , 2, 115 (1963).

6. For reviews see: (a) E. Vogel, ibid., 2, 1 (1963), (b) S. J. Rhoads, in P. de Mayo, " Molecular Rearrangements, " Vol. 1, Interscience, New York, N. Y., 19&3, 884; (c) M. Ranciu, Stud. Ceret. Chim., 15_, 603 (l967)i (d) H* M. Frey and R. Walsh, Chem. Rev. , 69, 1^3 (I9&9)i (e) K. F u k u i and H. T ujim oro, Mech. Mol. M g r. , 2, 117 (1969), (f) E. W interfeldt, Fortschr. Chem. Forsch., 16, 75 (197°), (g) L. A. P a q u e tte , Angew. Chem. , I n te r n . Ed. E n g l., 10, 11 (1971)*

7* R* B. Woodward and R. Hoffmann, " The Conservation of Orbital Symmetry," Academic Press, New York, N. Y., 1969-

8. It is geometrically impossible for rigidly held trans-1,2-divinyl group to undergo reaction.

9 . (a ) E. V o g el, K. -H. O tt, and K. G ajek, J u s tu s L ie b ig s Ann. Chem. , 644, 172 (1961); (b) J. M. Brown, B. T. Golding, and J. J. Stofko, J . Chem. S o c ., Chem. Commun. , 319 (1973 )i (c) M. P. Schneider and J. Rebell, ibid., 283 (1975J.

10. (a) J. S. McKennis, L. Brener, J. S. Ward, and R. P ettit, J. Amer. Chem. Soc. , §2., 4957 ( I 9 7 l ) ‘» (B) L. A. P a q u e tte , R. F. D avis, a n d D. R. Jam es, T etra h e d ro n L e tt. , 1615 (1974). 219

11. For example see: (a) P. S. Wharton, Y* Sumi, and R. A. Kretchmer, J. Org. Chem., 22j 234 (1965):, (h) P. S. Wharton and R. A. Kretchmer, ibid., 33s 4258 (1968); (c) W. Grimme, J. Amer. Chem. Soc., 94, 2525T1972).

12. (a) G. Schroder, Angevr. Chem. , Intern. Ed. Engl. , 2, 48l (1963); Chem. Ber., 97, 3140 (1964); (b) M. Saunders, Tetrahedron L ett., 1699 (19&3);~T c ) R. Merenyi, J. F. M. Oth, and G. Schroder, Chem. Ber. , 97, 3150 (1964); (d) A. Allerhand and H. S. Gutousky, J . Amer. Chem. Soc. , 87, 4092 (1965).

13. For reviews see: (a) G. Schroder and J. F. M. Oth, Angew. Chem. Intern. Ed. Engl., 6, 4l4 (1967); (b) M. Banciu, Stud. Ceret. Chim. , 18, 547 (1970).

14. W. von E. Doering, B. M. Ferrier, E. T. Fossel, J. H. Hartenstein, M. Jones, Jr., G. Klump, R. M. Rubin, and M. Saunders, Tetrahedron, 23., 3943 (1967).

15. (a) U. Biethan, H. Klusacek, and H. Musso, Angew. Chem., Intern. Ed. Engl., 6, 176 (1967); (b) H. Tsuruta, K. Kurabayashi, and T. Mukai, Tetrahedron L ett., 3775 (1967)-

16. (a) L. G. Greifenstein, J. B. Lambert, M. J. Broadhurst, and L. A. Paquette, J ■ Org. Chem., 22> 1210 (1973); (h) L. A. Paquette and M. J. Broadhurst, ibid., 38, 1893 (1973)*

17. (a) L. A. Paquette and T. J. Barton, J. Amer. Chem. Soc. , 89, 5480 (1967); (b) P. Wegener, Tetrahedron L ett., 4983 (19&7); (c) L. A. Paquette, J. R. Malpass, G. R. Krow, T. J. Barton, J. Amer. Chem. Soc., 91, 5296 (1969); (d) G. R. Krow, J. Reilly, D. M. Fam, J. Marakowski, and K. C. Ramey, ibid. , 98, 3636 (1976).

18. (a) A. G. A n a sta ssio u and B. Y. -H. Chao, J . Chem. S o c., Chem. Commun., 277 (1972); (b) L. A. Paquette, U. Jacobsson, and M. Oku, J. Chem. Soc., Chem. Commun. , 115 (1975); (c) A. G. Anastassiou, E. Reichmanic, and J. C. Wetzel, Tetrahedron L ett., 1651 (1975); (d) A. G- Anastassiou, A. E. Winston, and E. Reichmanis, J. Chem. Soc., Chem. Commun. , 779 (1973)*

19- H. Nakanishi and 0. Yamamoto, Tetrahedron Lett. , 1803 (1974).

20. H. Gunther and J. Ulmen, Tetrahedron, 22, 378l (1974).

21. G. Schroder, J. F. M. Oth, and R. Merenyi, Angew. Chem. , Intern. Ed. Engl. , 4, 752 (1965).

22. R. Bicker, H. Kessler, and W. Ott, Chem. Ber. , 3151 (1975)-

23. J. B. Lambert, Tetrahedron Lett., 1901 (1963). 220

2b . (a) R. Hoffman and W. D. Stohrer, J. Amer. Chem. Soc., 92.5 69^1 (l97l)j (b) M. J. S. Dewar and W. W. Sehoeller, ibid. , 92.3 1^81 (1971)9 (c) M. J. S. Dev/ar and D. H- Loo, i b i d . , 95, 7201 (1971)-

25- For reviews on homoaromaticity see: (a) P. M. Warner, " Recent Developments in Homoaromaticity and Antihomoaromaticity, 11 in press, personal communication to Dr. L. A. Paquette:, (b) S. Winstein, Chem. Soc. , Special Publication, Ho. 21, 5 (1967); (c) S. W. Staley and D. W. Reichard, J. Amer. Chem. Soc., 91, 3998 (1969); (d) P. Alberg, D. L. Harris, and S. Winstein, ibid. , 92, h-^4-5^- (1970); (e) S. Winstein, Quart■ Rev. , 23, 1^1 (19^97; 7f) W. J. LeNoble, "H ighlights of Organic Chemistry," Marcel Dekker, Inc., Hev/ York, N. Y., 197^, PP 355-367.

26. For examples see: (a) H.-P. Loffler and G. Schroder, Angew. Chem., Intern. Ed. Engl. , 7 , 756 (1968); (b) J. F. M. Oth, E. Ivkchens, H. Rdttele, and G. Schroder, Justus Liebigs Ann. Chem., 7^2., (1971); (c) A. Busch and H. M. R. Hoffmann, Tetrahedron Lett. , 2379 (1976); (d) J. C. Barborak, S- Chari, and P. von R. Schleyer, J. Amer. Chem. Soc. , 92.5 5275 (1971); (e) 0. Krow, J. Reilly, D. M. Fam, J . M arakowski, and K. C. Ramey, ib id . , 9 8 , 3636 (1976); (f) J. F. M. Oth, R. Merenyi, H. Rottele, and G. Schroder, T e tra h e d ro n L e tt. , 39^-1 (1 9 6 8 ).

27* K. Hamski, R. Molojcic, S. Borcic, and D. E. Sunko, J. Amer. Chem. Soc. , 92, 653^- (1970) and references cited therein.

28. (a) R. Hoffmann, Tetrahedron Lett., 2907 (1970 )•, (b) H. G unther, ibid., 5173 (19 /07

29* For examples see: (a) F.-G. ICLarner, ibid., 19 (197*0; (b) H. Gunther, B. D. Tunggal, M. Regitz, H. Scherer, and T. Keller, Angev/. Chem. , I n te r n . Ed. E n g l ., 10, 563 (1 9 7 1 ); (c) H. Gunther and T. K e lle r , Chem. Ber. , 3231 (19707" (!) 0. M a ie r, Angev/. Chem., Intern. Ed. Engl., 6, b02 ( 1967).

30. (a) H. E. Zimmerman and G. L- Grunewald, J. Amer. Chem. Soc. , 88_, 183 (1966); (b) H. E. Zimmerman, R. W. Binkley, R. S- Givens, G. L. Grunewald, and M. A. Sherwin, ibid., 91, 3316 (1969).

31* (a) H. E. Zimmerman and H. Iv/amura, ibid., 92., V763 (1968):, (b) H. E. Zimmerman and H. Iv/amura, ib id ., 9£, 2015 (1970).

32. (a) J. Meinwald and D. Schmidt, ibid., 91, 5877 (1969); (b) J. Meinwald and H. T su ru ta , i b i d . , 91, 5877 (1 9 6 9 ); (c) H. E. Zimmerman, J. D. Roberts, and J. Schantl, ibid. , gl, 5878 (1969).

33* R* Mhlherbe, Helv. Chim. Acta, 96, 28^5 (1973 7

3 ^. L. A. P a q u e tte , J . Amer. Chem. S o c ., 92, 5765 (197°). 221

35* R- A sk an i, T etra h e d ro n L e tt. , 33^-9 (1970)*

36. A. K. Cheng, P. A. L. Anet, J. Mioduski, and J. Meinwald, J. Amer. Chem. S o c ., 9 6 , 2887 (197^)•

37* (a) D. R. James, G. H. Birnberg, and L. A. Paquette, J. Chem. Soc. , Chem. Commun., 722 (197^)» (h) D. R. James, G. H. Birnberg, and L. A. P a q u e tte , J . Amer. Chem. Soc. , 7^65 (197*0; (<0 L. A. Paquette, D. R. James, and G. H. Birnberg, ibid. , 96, 7**-5*J- (197*0; (d) L. A. Paquette, W- E. Volz, M. E. Beno, and G. G. Christoph, ibid.3 2L 2562 (1975).

38. L. Libit and R. Hoffmann, ib id ., 96.5 1370 (197*0-

39* (a) L. A. Paquette, R. E. Wingard, Jr., R. K. Russell, ibid., 4739 (1972); (b) E. Wenkert, E. W. Hagaman, L. A. Paquette, R. E. W ingard, J r . , and R. K. R u s s e ll, J . Chem. S o e ., Chem. Comm-, 135 (1972); (c) R. K. Russell, L. A. Paquette, L. G. Greifenstein, and J. B. Lambert, Tetrahedron L ett., 2855 (1973); (d) R. E. Wingard, Jr. R. K. Russell, and L. A. Paquette, J. Amer. Chem. Soc. , §6, 7*1-7*+ (197*0-

*1-0. (a) R. C. Haddon, V. R. Haddon, and L. M. Jackson, Fortschr. Chem. Forsch., 16, 103 (1971); (b) M. Barfield, D. M. Grant, and D. Ikenberry, J. Amer. Chem. Soc., 97, 6958 (1975) and references cited therein; (c) T. G. Schmalz, T. D. Gierke, P. Beak, and W. H. Flygare, Tetrahedron L ett., 2885 (197*+); (d) R. Benassi, P. Lazzeretti, and F. Taddei, J. Phys. Chem., 79, 8*4-8 (1975); (e) D. Lewis and D. Peters, " Facts and Theories of Aromaticity, " Crane Russak, New York, N. Y. , 1975; (f) H. J. Dauben, Jr., J. D. Wilson, and J. L. Laity, in " Non-benzenoid Aromatics," Vol. 2, Ed. by J. P. Snyder, Academic Press, New York, N. Y., 1971, P* l£>7; (g) E. D. Bergmann and B. Pullman, " Aromaticity, Pseudo- arom aticity, and Anti-aromaticity, " Israel Academy of Sciences and H u m an ities, Je ru sa le m , 1971*

*+1. (a) J. L. von Rosenberg, J. E. Mahler, and R. Petit, J. Amer. Chem. Soc., 8*+, 28^4-2 (1962)-, (b) N. C. Deno, Progr. Phys. Org. Chem. , 2, 129 (19&0; ( c ) S. Winstein, C. G. Kreiter, and J. I. Brauman, J. Amer. Chem. Soc., 88, 2047 (1966). ii-2. W. von E. Doering, G. Laber, R. Vonderwahl, N. F. Chamberlain, and R. B. W illia m s, i b i d . , jOA, 5*+*+8 (1956).

1+3. (a) F. A. L. Anet, ibid. , 86, *+58 (196*0; (b) F. R. Jensen and L. A. Smith, ibid. ,~B6, 95o"~(196*+).

*+*+. For example see: (a) E. Vogel, Pure and Appl. Chem. , 20, 237 (1969); (b) F. A. L. Anet and G. E. Schenck, J. Amer. Chem. Soc., 22., 556 (1971); (c) I. Pikulik and R. F. Childs, Can. J. Chem., 22.5 l8l8 (1975)« (d) G. Maier, Angew. Chem., Intern. Ed. Engl., 6, *402 (1967). 222

45- (a) E. Vogel, U. H. Bririker, K. Nachtkamp, J. Wassen, and K. Mullen, ibid., 12 , 758 (1973); (8) H. Gunther, H. Schmickler, U. H. Brinker, K. Nachtkamp, J. Wassen, and E. Vogel, ibid. , 12_, 760 (1973).

46. G. P. C e a sa r, J . G reen, L. A. P a q u e tte , and R. E. W ingard, J r . , Tetrahedron L ett., 1721 (1973)* i+7* R. K. Russell, Fh.D. Thesis, The Ohio State University, 1975-

48. L. A. Paquette, R. K. Russell, and R. L. Burson, J. Amer. Chem. S o c ., 2 7 , 6l2h (1975).

4-9. J. C. Philips, Ph.D. Thesis, The Ohio State University, 1969.

50. R. E. W ingard, J r . , Ph.D. T h e s is , The Ohio S ta te U n iv e rs ity , 1971*

51. (a) L. A. P aq u ette and J . C. P h i li p s , J . Chem. S o c., Chem. Commun., 680 (1969)’, (b) L. A. Paquette, R. E. Wingard, Jr., J. C. Philips, G. L. Thompson, L. K. Read, and J. Clardy, J. Amer. Chem. Soc., 23.j 4508 (1 9 7 1 ); (c) L. A. P a q u e tte , J . C. P h ilip s , and R. E. Wingard, J r., ibid. , 92.3 ^5l6 (1971)-

52. W. B. Scott, Ph.D. Thesis, University of B ritish Columbia, 1965- Available through the Catalog Division, National Library of Canada, Ottawa (Canadian Thesis on Microfilm No. 483H (b) W. B- Scott and R. E. Pincock, J. Org. Chem. , 32, 3374 (1967).

53. " Background Information on Hexamethylphosphoric Triamide, " Office of Occupational Health Surveillance and Biometrics, National Institute for Occupational Safety and Health, October 24, 1975*

54. Aqueous solutions of HMPA are extracted with chloroform (4x) and • the combined extracts are concentrated to remove solvent and water. D istillation from calcium hydride gives usable HMPA. If the HMPA is to be used in a sulfide reaction, a second distillation from sodium sulfide is advised. Further, HMPA recovered from sulfide reactions should be kept separate from other recovered HMPA.

55* E. E. R oyals and L. L. H a r r e ll, J . Amer. Chem. S o c ., 77, 3405 (1955).

56. (a) L. A. Paquette, "Mechanisms of Molecular M igrations," Vol. I, B. S. Thyagarajan, Ed., Interscience Publ. , Inc., 1968, pp 121-156:, (b) L. A. Paquette, Accts. Chem. Res., 1, 209 (1968).

57* The addition of powdered glass allows the reaction to proceed at a lower temperature (60° vs 90°). Also, the use of lithium fluoride instead of lithium chloride seems to give a cleaner reaction. P. F. King, private communication. 223

58. The cyclohutene double bond has been shown to be less reactive than the cyclohexyl double bonds. See L. A. Paquette and G. L. Thompson, J. Amer. Chem. Soc. , 7118 (1972).

59* D. C. Liotta, unpublished results.

60. M- J. Kukla, unpublished results.

61. (a ) H. E. O 'N eal and S. W. Benson, J . Phys. Chem. , J2, 1866 (1968)-, (b) H. E. O'Neal and S. W- Benson, Int. J. Chem. Kinetics, 2, ^23 (I970)i (c) R- Srinivasan, ibid., 1, 133 (19&9)*

62. M. J. Carmody, Ph.D. Thesis, The Ohio State University, 1976.

63- S. V. Ley, private communication.

6 k. H. Gunther and T. K e lle r, Chem. Ber. , 3231 (1970).

65. Adapted from a procedure provided by Professor Berson. See J. A. Berson and S. S. Olin, J. Amer. Chem. Soc., 915 777 (1969).

66. (a) D. D. Reynolds and W. 0. Kenyon, ibid. , 72, 1593 (1950); (t>) A. W. Schrecker and J. L. Hartwell, ibid. , 7£, *+32 (1955); (c) S. Wolff, A. B. Smith, III, and W. C. Agosta, J. Org. Chem., 32.J 1607 / •» Ar»l. \

67. (a ) R. M. O tte n b r ite and P. V. A lsto n , i b id . , 39.? 1115 (197*0; (b) R. M. Ottenbrite and P. V. Alston, ibid. , 3X-- 3360 (1972); (c) R. M. Ottenbrite and P. V. Alston, J. Heterocycl. Chem., 10, 785 (1973).

68. R. K. Crossland and K. L. Servis, J. Org. Chem. , 33.9 3195 (1970).

69. Mn02 prepared by the Attenburrow method has proven to be best suited for these oxidations: J. Attenburrow, A. F. B. Cameron, J. H. Chapman, R. M. Evans, B. A. Hems, A. B. A. Jansen, and T. Walker, J . Chem. Soc. , 109*1- (1952).

70. R. C. Schnur and E. E. van Ihmelen, J. Amer. Chem. Soc., 97, k6h (1975).

71. (a) J. Meinwald and H. Nozaki, ibid., 80, 3132 (1958); (b) L. A. Paquette and R. W. Begland, ibid. , 5159 (1968).

72. (a) F. A. Bovey, "Nuclear Magnetic Resonance Spectroscopy," Academic Press, New York, N. Y., 1969; (b) R. M. Silverstein and G. C. Bassler, " Spectrometric Identification of Organic Compounds," 2nd Ed., John Wiley and Sons, Inc., New York, N. Y., 1967. 22k

75. (a) P. Friedman and P. Allen, Jr., J. Org. Chem., 50, 780 (1965); (b) A. Muller, E. Funder-Fritzsche, W. Konar, and E. Rintersbacher- Wlasak, Montash, 8k,'-1206 (1953)*

7k. C- R. Johnson and D. McCants, Jr., J. Amer. Chem. Soc., 87, 1109 (1965)', (b) C. R. Johnson and D. McCants, Jr., ibid., 86 , 2935 ( 196k ).

75- C. R. Johnson and D. McCants, Jr., J. Amer. Chem. Soc., 87, 5k05 (1965).

76. For recent reviews of criteria for configurational assignment of cyclic sulfoxides, see (a) R. Lett and A. Marquet, Tetrahedron, 3379 (197k); (b) R. R. Fraser, T. Durst,_ M. R. McClory, R. Viau, and Y. Y. Wigfield, Int. J. Sulfur Chem. , Part A, 1, 133 (1971).

77* K. W. Buck, A. B. Foster, W. D. Pardoe, M. H. Qadir, and J. M. Webber, J. Chem. Soc., Chem. Commun., 357 (1966).

78. (a) W. 0. Siegl and C. R. Johnson, Tetrahedron, 27, 5kl (1971)’, (b) C. R. Johnson and W. 0. Siegal, J. Amer. Chem. Soc., £1, 2796 (197°)j (c) R- M. Dodson, E. H. Jancis, and G. Klose, J. Org. Chem., 3J5, 2520 (1970); (d) E. Jonnson, Ark. Kemmi., 26, 557 (l967)» (e) E. Jonnson and S. Holmquist, ib id ., 29, $01 (1968).

79* J* B. Lam bert and R. G. Keske, J . Org. Chem., JJL, 3^29 (1966).

80. (a) W. Amann and G. Kresze, Tetrahedron Lett■, k909 (1968)-, (b) E. T. Strom, B. S. Snowden, J r., and P. A. Toldan, J. Chem. Soc., Chem. Commun., 5° (1969)» (c) R. D. G. Cooper, P. V- DeMarco, J. C. Cheng, and N. D. Jo n e s, J . Amer. Chem. Soc. , ^ I j lk08 (197°).

81. (a) S. Ohtsuru, K. Tori, and M. Fukuyama, Tetrahedron L ett., 2877 (1 9 7 0 )j (b ) R. R. F r a s e r and Y. Y. W ig fie ld , J . Chem. Soc. , Chem. Commun., lk71 (1970).

82. S. Vazquez and J. Castrillon, Spectrochim. Acta, Part A, ^OA, 2021 (197k).

85. (a) J. B. Lambert, D. A. Wetzel, H-n. Sun, and K. K. Lilianstrom, J . Amer. Chem. Soc. , 9 8 , 3778 (1976)', (b ) J . R. Wiseman, H. 0. Krabbenhoft, and B. R. Anderson, J. Org. Chem., kl, 1518 (1976).

8k. (a) L. A. Paquette, C. C. Liao, D. C. Liotta, and W. E. Fristad, J. Amer. Chem. Soc. , submitted; (b) L. A. Paquette, D. C. Liotta, C. C. Liao, T. G. Wallis, N. Eickman, J. Clardy, and R. Gleiter, ibid., submitted; (c) L. A. Paquette, D. C. Liotta, and A. D. Baker, Tetrahedron Lett. , submitted.

85. C. Batich, E. Heilbronner, C. B. Quinn, and J. R. Wiseman, Helv. Chim. Acta, 52? (1976) and references cited therein. 225

86. (a) J. A. Baltrop, P. M. Hayes, and M. Calvin, J. Amer.. Chem. Soe., j6 , k^>k8 (I95^)j (b) G. 0. Schenck and C. H. Krauch, Chem. Ber. , 9

87. Sulfides lTf and 122 and authentic samples of 118, 11£, 120.5 and 122, were kindly provided by Dr. R. K. Russell.

88. R. G le ite r , p r iv a te com m unication.

89- D. C. Liotta and A. D. Baker, private communication.

90. Gerhard Erker, Ph.D. Thesis, Ruhr-Universitat Bochum, 1973* Thanks are extended to R. K. Russell and Professor W. R. Roth for sup plying this procedure.

91. (a) E. Vogel, W. Wiedemann, H. Kiefer, and W. F. Harrison, Tetrahedron Lett. , 673 (1963 )', (b) E. Vogel, W. Maier, and J. Eimer, ibid., 655 (1966).

92. L. A. Paquette, D. E. Kuhla, J. H. Barrett, and R. J. Haluska, J. Org. Chem., 2its 2866 (1969).

93* (a ) L. A. P a q u e tte and J . C. P h i li p s , J . Amer. Chem. S o c ., j20, 3898 (1968)-, (b) L. A. Paquette, T. Kahihana, J. F. Hansen, and J. C. Philips, ibid., 152 (1971)*

9^. See footnote 27 of reference 39* The chlorinated solvent systems (CDCI3 -TMS, CD2C12 -TMS, and CD2C12 -CF2C12 -TMS) p ro v id e com parable data while C3CI4-TMS and CS2-TMS form another self-consistent set of solvents.

95- J* G. Pritchard and P. C. Lauterbur, J. Amer. Chem. Soc., 83, 2105 (1961).

96. S. Glazer, R. Knorr, C. Ganter, and J. D. Roberts, ibid., 6026 (1972).

97- N. L. Allinger. M. T. Tribble, and M. A. M iller, Tetrahedron, 28, 1173 (1972).

98. Since is definitely not a semibullvalene, the Chemical Abstracts name of 2a,8b-dihydrocylopent[cd]azulene w ill be used. The num bering scheme compatible with this name is indicated. The as sistance of Dr. Kurt L. Loening of Chemical Abstracts Service in naming th is and several other compounds is gratefully acknowledged.

99* H. Oediger, F. Moller, and K. Eiter, Synthesis, 591 (1972). 226 100. (a) G. H- Posner, R. J. Johnson, and M. J. Whalen, J. Chem. Soc., Chem. Commun. , 28l (1972); (h) C. Mercier, P. Soucy, W. Rosen, and P. Deslongchamps, S y n th etic Commun., 5., l 6 l (1975)*

101. (a) K. B. Sharpless and M. W. Young, J. Org. Chem. , 9*4-7 (1975); (b) K. B. Sharpless and R. P. Lauer, J. Amer. Chem. Soc. , 9£., 2699 (1973)-

1 0 2 . T. G. W allis, private communication.

105. J. J. Pappas, W- P. Keaveney, E. Gancher, and M. Berger, Tetrahedron Lett. , *4-275 (1976). loh. J. J. Pappas, W. P. Keaveney, M. Berger, and R. V. Rush, J. Org. Chem., 55, 787 (1968).

105. J . L. W arn ell and R. L. S h rin e r, J . Amer. Chem. Soc. , 79, 5160 (1957)-

106. (a) S. H* Graham and A. J. S. Williams, J. Chem. Soc.(c), 655 (1966)j (b) R. Pappo, D. S. Allen, Jr., R. U. Lemieux, and W. S- Johnson, J. Org. Chem., 21, *4-78 (1956).

107. (a) R. Ratcliff and R. Rodehorst, ibid., 5,5, ^0°° (1970); (b) J. C. Collins, W. W. Hess, and F. J. Frank, Tetrahedron Lett. , 5365 (1968).

108. A. P. Johnson and A. Pelter, J . Chem. Soc. , 520 (196*4-).

109. R. A. Snow and L. A. F a q u e tte , unp u b lish ed r e s u l ts .

110. The use of rose bengal on polymer support (Photox^ available from Hydron Laboratories, Inc.) allows easy recovery and reuse of the sensitizer by filtration of the suspension prior to workup. This modification is the preferred method.

111. L. F. Fieser and M. Fieser, " Reagents For Organic Synthesis," Vol. I, John Wiley and Sons, Inc., New York, N-Y., 1967, P 937*

112. (a) B. Bock, K. Flatau, H. Junge, M. Kuhr, and H. Musso, Angew. Chem. , Intern. Ed. Engl., 10, 225 (1971); (h) G. Hafelinger, R. G. Weissenhorn, F. Hack, and G. Westermayer, Angew. Chem. , I n te r n . Ed. E n g l., 11, 725 (1972).

115. (a ) H. Hock and F. E rn s t, Chem. Ber. , 92 , 2736 (1 959); (b ) G. Fraenkel, S. Dayagi, and S. Kobayashi, J. Chem.- Phys., 72, 953 (1968). i l k C. N. Shih, unpublished results.

115. D. Seyferth and R. L. Lambert, Jr., J . Organometal. Chem. , 16, 21 (1969). 227 ll6. L. F. Fieser and M. Fieser, " Reagents for Organic Synthesis," Vol. I, John Wiley and Sons, Inc., New.York, N. Y., 19673 P H66; (b) R. L. Wasson and H. 0. House, Organic Synthesis, Coll. Vol. IV, 552 (1965).

117- C. C. Liao, private communication.

118. . (a) P. S. Wharton and D. H. Bohlen, J. Org. Chem., 26, 3615 (l96l); (b) H. E. Johnson and D. G. Crosby, ibid. , 27., 2205~fl962); (c) L. F. Fieser and M. Fieser, " Reagents for Organic Synthesis, " Vol. I., John Wiley and Sons, Inc., New York, N. Y. , 19 67, p ^39-

119- (a) P. A. Bums, C. S. Foote, and S. Mazur, J. Org. Chem., Hi, 899 (197*8)*, (b) G. 0. Pierson and 0. A. Runquist, ibid. , 5H, 3^5^ (1969); (c) L. Horner and W. Jurgeleit, Justus Liebigs Ann. Chem., 591, 138 (1955)-

120. M. Mousseron-Canet, J. -C. Mani, and J.-P. Dalle, Bull. Soc. Chim. F r., 608 (1967).

121. K. Gollnick and G. 0. Schenck in " 1,H Cycloaddition Reactions," J. Hamer, Ed., Academic Press, New York, N.Y., 1967 3 P 255*

122. 0. Wallach, Justus Liebigs Ann. Chem. , 392.5 H9 (1912).

123- L. F. Fieser and M. Fieser, "Reagents for Organic Synthesis," Vol. I, John Wiley and Sons, Inc., New York, N. Y., 1967, P 215-

12H. (a) A. Saika and C. P. Slichter, J. Chem. Phys., 222, 26 (195H); (b) G. C. Levy and G* L. Nelson, " Carbon-13 Nuclear Magnetic Resonance for Organic Chemists," W iley-Interscience, New York, N.Y., 1972.

125. M. Karplus and J. A. Pople, J. Chem. Phys. , 38, 2803 (1963).

126. (a) H. L. Retcofsky and F. R. McDonald, Tetrahedron Lett. , 2575 (1968); (b) H. L. Retcofsky and C. E. G riffin, ibid. , 1975 (1966); (c) J. Mason, J. Chem. Soc.(A), IO38 (1971); (d) H. Spiesecke and W. G. Schneider, J. Chem. Phys., 3J5, 731 (1961); (e) G. J. Ray, R. J. Kurland, and A. K. Colter, Tetrahedron, 27, 735 (l97l)»

127. T. W. Abbott, R. T.. Arnold, and R. B. Thompson, Organic Synthesis, Coll. Vol. II, 10 (19H3).

128. The combined yield of monoadduct was 30° S (28^).

129* K. Alder and K. H. Backendorf, Chem. Ber. , 713 2199 (1938).

130. L. F. Fieser and M. Fieser, "Reagents for Organic Synthesis," Vol. I, John Wiley and Sons, Inc., New York, N. Y., 1967 j PP 966- 970. 228

1 3 1 . Modest amounts of starting triene were also present and could be recovered following the next reaction.

132. J. Stickler and W. Pirkle, J. Org. Chem. , JJL, 3^ ^ (1966).

133- parent ion could not be detected in the mass spectrum.

134. . The silver was recovered from the aqueous phase as silver chloride following the addition of excess salt (WaCl).

135* All solutions and solvents were flushed with argon prior to use and exposure to the atmosphere was minimized.

136. Silver oxide was prepared by stirring equimolar amounts of silver nitrate and sodium hydroxide (0.1 N solutions). - The resulting brown solid was removed by filtration and was rinsed several times with water, acetone, and ether. The powdery solid ms dried under house vacuum in the dark.

1 3 7 . N,N-Dimethylacetamide is toxic by dermal contact and chronic inhalation has been shown to cause liver damage. J. A. Riddick and W. B. Bunger, "Organic Solvents," Vol. II, 3rd Ed., W ile y -In te rs c ie n c e , pp 8M1- 8^5*