INFORMATION TO USERS

This was produced from a copy of a document sent to us for microfilming. 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 material submitted.

The following explanation of techniques is provided to help you understand markings or notations 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 through an image and duplicating adjacent pages to assure you o f complete continuity.

2. When an image on the Him is obliterated with a round black mark it is an indication that the film inspector noticed either blurred copy because of movement during exposure, or duplicate copy. Unless we meant to delete copyrighted materials that should not have been Aimed, you will And a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., is part of the material being photo­ graphed the photographer has followed a deAnite method in “sectioning” the material. It is customary to begin Aiming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the Arst row and continuing on until complete.

4. For any illustrations that cannot be reproduced satisfactorily by xerography, photographic prints can be purchased at additional cost and tipped into your xerographic copy. Requests can be made to our Dissertations Customer Services Department.

5. Some pages in any document may have indistinct print. In all cases we have Aimed the best available copy.

University Micrdrilms International 300 N. ZEEB ROAD. ANN ARBOR, Mt 48106 18 BEDFORD ROW. LONDON WC1R 4EJ, ENGLAND 7922533

MUTHARD, JEAN LOUISE SYNTHESIS AND CHEMISTRY OF TOP.OLOG 1C ALLY UNUSUAL POLYQUINANES.

THE OHIO STATE UNIVERSITY, PH.D., 1979

University Microfilms International 3 0 0 n. z e e b r o a d, a n n a r b o r, mi ab io g SYNTHESIS AND CHEMISTRY OF TOPOLOGICALLY

UNUSUAL FOLYQUINANES

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Jean Louise Muthard, B.S.

*****

The Ohio State University

1979

Reading Committee: Approved By Dr. Leo A, Paquette Dr. John A. Secrist, III Dr. John S. Swenton Q u a g . C S Adviser Department of Chemistry To my family ACKNOWLEDGMENTS

The author is grateful to Professor Paquette for his guidance as a research adviser and his boundless enthusiasm as a chemist. She is extremely indebted to the many graduate students, postdoctoral fellows, and friends who have been supportive on both the emotional and intellectual levels during her time at The Ohio State

University. Special thanks are extended to Dr. Morey

Osborn and Dr. Michael Geckle for their unselfish assistance in obtaining many of the carbon magnetic resonance spectra reported here. The chemical contri­ butions of Dr. Tadeusz Cynkowski, Dr. Otto Schallner,

Dr. Robert Snow, Dr. Matthew Wyvratt, Dr. Richard

Bartetzko, Professor Rolf Gleiter, and Professor Gary

Christoph are gratefully acknowledged.

Though the existence of Matthew Wyvratt has at times made the last few years more difficult, his encouragement and understanding have been invaluable in the completion of this work. The author expresses the deepest appreciation to Matt, her husband.

iii VITA

May 25, 1953 ...... Born, Atlanta, Georgia

1974 ...... B.S., high honors, University of Florida, Gainesville, Fla.

1 9 7 4 - 1 9 7 5 ...... University Fellow, The Ohio State University, Columbus, OH

1975 - 1977...... Research Assistant, Department of Chemistry, The Ohio State University, Columbus, Ohio

1977 - 1978 ...... Dissertation Fellow, The Ohio State University, Columbus, OH

1978 - 1979 ...... Research Assistant, Department of Chemistry, The Ohio State University, Columbus, OH

PUBLICATIONS

"C^g-Hexaquinacene, 11 L.A. Paquette, R.A. Snow, J.L. Muthard, and T. Cynkowski, J. Am. Chem. Soc., 100, 1600 (1978).

"Long Range Electronic Transmission in Conformationally Rigid ar-Diketones, " R. Bartetzko, R. Gleiter, «T- L. Muthard, and L.A. Paquette, J. Am. Chem. Soc., 100, 5589 (1978).

"Quantitative Assessment of pp-cr Overlap in a Topologically Convex Triene. Electronic and Crystal Structure Analysis of C 1 6 -Hexaquinacene." G.G. Christoph, J.L. Muthard, L.A. Paquette, M.C. Bohm, and R. Gleiter, J. Am. Chem. Soc., 100, 7782 (1978).

FIELD OF STUDY

Major Field: Organic Chemistry

iv TABIiE OF CONTENTS

Page DEDICATION ...... ii

ACKNOWLEDGMENTS. . ■ ...... iii

VITA ...... iv

LIST OF TABLES ...... vi

LIST OF FIGURES ...... vii

LIST OF SCHEMES ...... ix

INTRODUCTION ...... 1

RESULTS AND DISCUSSION ......

Part I. The Electronic Structure and Chemistry of Some Conformationally Rigid (y-Dfketones...... 8

Part II. Synthesis and Chemistry of a C 2 Symmetric Bislactone ...... 43

Part III. C^g-Hexaquinacene...... 61

EXPERIMENTAL ...... 108

APPENDIXES

A. Selected Proton Magnetic Resonance Spectra 158

B. Carbon Magnetic Resonance Spectra of Triols 115 and 1 1 6 ...... 170

REFERENCES ...... 172

v LIST OF TABLES

Table Page

1. Comparison Between Measured Vertical Ionization Potentials# Iy j, of 28, 29, 22, and 30_ and Calculated Orbital Energies (e^) 24

2. Measured and Calculated Electronic Absorption Spectra of 28, 29, 22, and 30_...... 28

3. Emission Data of Compounds 28, 29, 22, and 3<3 32

1 3 4. C NMR Spectral Parameters for Hexaquinacene ...... 77

5. pp—o Overlap Integrals Estimated Using Slater O r b i t a l s ...... 83 LIST OF FIGURES

Figure Page

1 . PE Spectra of 28, 29, 22, and 3(3 Between 8 and 16 e V ...... 23

2. Correlation Between the First Bands of the PE Spectra of 22, 29, 28, and 30^ Based on MINDO/3 Calculations ...... 26

3. Comparison Between the First Bands in the Electronic Absorption Spectra of 28, 29, 22, and 30^ in Methylene Chloride...... 30

4. Comparison of the Orbital Energies for 28 without (left) and with (right) TT* - 17* Interaction ...... 32

5. Orbital Alignments of 1,4-Diketones .... 44

6 . Orbital Alignment Present in 5j5...... 46

7. Orbital Alignments Present in 1,4-Diketones 78 (A) and 77 (B) ...... 69

8 . X-Ray Crystal Structure of C --Hexaquinacene ( 2 4 ) ...... 80

9. View of C.g-Hexaquinacene (24) Down the Three-fold axis ...... 81

10. The He(I) Photoelectron Spectrum of C -- Hexaquinacene (24)...... 84

11. Orbital Energy Plot Showing the Variations in e(Il) and ai(ll) Levels as a Function of Distance Between Interacting II Bonds . . . 8 6

12. Correlation of the Ionization Energies of 24, 95, and 96...... 87

vii FIGURES (CONT'D)

Figure Page

13. X-Ray Crystal Structure of C .. ,-Hexaquinane ( 9 4 ) ...... 8 8

14. Depiction of the Intramolecular Hydrogen- Hydrogen Distances in 9 4 ...... 89

15. Vacuum Pyrolysis Apparatus ...... 98

viii LIST OF SCHEMES

Scheme Page I. Synthesis of Diketones 22, 29, and 30^ . . 20

II. Synthesis of Biscyclopentenone 55_ .... 45

III. Reductive Fragmentation of Enone S5 . . . 47

IV. Attempted Ring Contraction of Bislactone 6 0 ...... 53

V. Attempted Synthesis of 24 from 84 .... 72

VI. Thermolysis of Triquinacene (95) ...... 97

VII. Hydroboration of C^-Hexaquinacene (24). . 102

ix INTRODUCTION

The chemistry of six-membered ring annulation

reactions has been thoroughly studied due to the early

interest in steroid chemistry. The construction of

polyquinanes, fused five-membered ring compounds/ has

received increased attention only over the past decade.

The development of synthetic methodology in this area

has been motivated, in part, by the recognition of the

.medicinal importance of prostaglandins and the isolation

of natural products possessing polyfused cyclopentanoid

structures. Interest has also grown in the construction

of higher polyquinanes with an all syn-cis stereochemistry

which lends them a sphere-like topology. The optimal

arrangement of fused cyclopentane rings in the spherical

pentagonal dodecahedrane (1 ) has not yet been realized.

1

1 The majority of the early attention in the polyquinane field was directed toward the synthesis and chemistry of the diquinane, bicyclo[3.3.0]octane ( 2 ).

The parent hydrocarbon and its functionalized derivatives

2 were prepared by utilizing such methods as the Dieckmann cyclization, vinyl cyclopropane rearrangements/ and transannular chemistry of cyclooctadiene and cyclooctane derivatives.

Among the triquinanes are the carbon skeletons of several natural products including hirsutic acid C (3)/^

2 3 the coriolin 4/ and hirsutine (5). Eaton and coworkers 4 have synthesized a cis, syn, cis stereoisomer (6 ) of this ring system. The structure of the sesquiterpene/ retigeranic acid (7) is based in part on the tricyclo-

[6 .3.0.0^'8jundecane ring system. Triketone 8 , 8 an example of this ring system/ has been prepared by Weiss and coworkers via the condensation of 4,5-dioxopentanoic acid with dimethyl 3-ketoglutarate and subsequent acid- promoted cyclization. A molecule with interesting topology is triquinacene (9) / which was initially 7 synthesized by Woodward and coworkers in 1964 and has

8 — 11 been the object of several subsequent syntheses. ” The

00H 7 8

physical properties and chemical behavior of 9^ provide a precedent for those of some of the molecules discussed in this dissertation. Pinacolic coupling of appropriately functionalized triquinacene units afforded in several 12 steps the bivalvane 10^. The possibility of intra­ molecular reactions of highly functionalized bivalvanes to afford 1 is still under investigation. The starting point for many of the compounds

synthesized in this dissertation possesses a tetraquinane

carbon skeleton which has received considerable attention.

Domino Diels-Alder cyclo addition of 9,10-dihydrofulvalene with dimethyl acetylenedicarboxylate is known to afford 13 14 diesters 11 and 12. ' Reductive cleavage of the

central carbon-carbon bond in 1 1 gave the functionalized 1 4 tetraquinane 13. The parent hydrocarbon 14 has been

COOCH ■COOCH

COOCH

COOCH 11

15 synthesized by Pukunaga and Clement. Weiss and

coworkers have isolated the tetracyclic triketone 15 in :o o c h

;o o c h 13 5 low yield from the condensation of glyoxal with

1 6 dimethyl 3-ketoglutarate followed by decarboxylation.

The analogous unsaturated hydrocarbon 16^ was prepared by 17 Srinivasan via pyrolysis of the photoproduct derived from benzene and bicyclo[3.2.0]hepta-2,6 -diene. With the

15 16 synthesis of tetracyclo[5.5.1.0^' ^]tridecane-

1 8 2/ 6 / 8 ,12-tetraketone (17)/ Weiss and coworkers have further established the condensation of dimethyl 3- » ketoglutarate with 1 ,2 -dicarbonyl compounds as a general method for the synthesis of polycyclic cyclopentanoid systems. The trivial name of "staurane" has been

17 proposed for this symmetrical carbon skeleton. A fourth 19 basic tetraquinane framework 18 has been prepared by 6

Eaton, and coworkers by the symmetrical annulation of two

five-membered rings onto cis-bicvclo[ 3 .3.0]octane-2,8 -

dione. Compound 18 serves as the precursor for penta- 19 quinane 19, which again possesses convex topology.

18 19

The introduction of one additional carbon into the 19 framework of 19_ results in peristylane 20, a hexa- 20,60 quinane with five-fold symmetry. A C^Q-hexaquinane 21 has been prepared by Paquette and coworkers and a few

COOCH

20 21 higher polyquinanes have been synthesized as researchers

attempt to construct the closed spherical array of

cyclopentane rings known as dodecahedrane. 7

The compounds discussed in this dissertation all possess fused five-membered rings. A series of diketones with carbocyclic framework 2 2 ^ were synthesized and their electronic structure examined. An unusual Il-mediated

COOCH

COOCH

22 23 24

1 ,6 -dicarbonyl reduction of compound 23^ was discovered and the further chemistry of the reduction products was explored. The basket-shaped triene, C^^-hexaquinacene

(24), has been prepared and its properties and chemical behavior have also been examined.

Some of the results contained in this dissertation had their inception in the work of Dr. Tadeusz Cynkowski,

Dr. Otto Schallner, Dr. Robert Snow, and Dr. Matthew

Wyvratt. Dr. Richard Bartetzko and Professor Rolf Gleiter measured the emission and photoelectron spectra of the diketones and performed the corresponding molecular orbital calculations. To these individuals the author expresses her gratitude. RESULTS AND DISCUSSION

Part I. The Electronic Structure and Chemistry of Some

Conformationallv Rigid g-Diketones

In recent years, chemists have ceased to consider the electronic structure of organic molecules as consisting of a u bond framework on which nonconjugated functional groups or multiple bonds possess orbitals localized within each such group. Recognition of the electronic interaction of nonconjugated II electrons within a molecule has prompted considerable theoretical and experimental interest. This interaction can take the form of through-space orbital overlap and/or through-bond or hyperconjugative mechanisms. 2 1 The extent of these orbital interactions can be determined by the use of 22 photoelectron spectroscopy, which allows the direct measurement and the assignment of the important ionization potentials of a molecule.

Bloomfield and Moser studied the spectral properties of the [4.4.2]propella-ll,12-diones 25-27 and found them to possess n -• II* spectral bands at unusually long 23 wavelength in the visible region. Diene 25, which is 9

25 26 27

pink in tlie crystalline state and in solution, absorbs

at X_ 537.5 run with e 71.7 in cyclohexane, while max compound 27 is yellow as a solid and in solution with

X 461 nm and e 73. The dihydro derivative 26 forms max pink crystals but affords orange solutions with two

absorption maxima at X „ 460-464 (e 38.8) and 532-535 nm ^ max (e 32). These spectra remain unchanged in a variety of

solvents. Compound 26 also yields an infrared spectrum

(in carbon tetrachloride solution) in which the carbonyl

absorptions appear to be a composite of the carbonyl

spectra of 25^ and 2 7 .

The explanation of these observations put forward

by Bloomfield and Moser was that the high visible maximum seen for 25^ resulted from resonance stabilization

of the excited state due to interspatial interaction

between the six-membered ring double bond and the dione.

The two absorption maxima observed for 26_ might result

from different interconverting conformational orientations

of the molecule (see 26a and 26b). Spatial interaction 10

26b

of the II orbitals would be maximized in conformation 26b leading to a similar X__„ as that observed for 25. In TTlaX *' 26ei/ this interaction would be significantly less# resulting in an absorption maximum like that of 27. A recent determination of the crystal structure of 25^ shows that the molecule has the two six-membered rings in the boat form as in 25a and deviates slightly from ^ v 24 symmetry.

25 a

25 A molecular orbital (MO) calculation/ based on models for 25-27 consisting of the dione II system interacting with two appropriately situated ethylenes 11 minus the cr bond framework, has indicated that saturated

27 should absorb at longer wavelengths than diene 25.

Inclusion of the a valence electrons in the calculations yielded the experimentally observed order of transitions.

Therefore, through-bond interactions predominate over the through-space options in these compounds.

A recent study of the photoelectron spectra of compounds 25-27 has clarified the electronic structure 26 of these molecules still further. The data were 27 interpreted using Koopmans 1 theorem (equation 1) which

equates the negative value of the orbital energy (ej) to the vertical ionization potential (Iv j) measured from the photoelectron spectrum. In these molecules the low-energy molecular orbitals which must be considered are the n+ and n_ pair of the dicarbonyl and the TI+ and

II pair of the ethylene moieties plus some cyclobutane a orbitals. On the basis of previous studies, a through-bond interaction exists between the 2p orbitals of or-dicarbonyls which results in an n+/n splitting of 28 about 2 eV with n at lower energy than n+. However, 12

29 the n+/n splitting should be small. The positions of the n, (8.65 eV) and n (10.4 eV) bands observed in the "T* —• spectrum of 27_ remain essentially the same in 2j5 and 26.

The spectrum of 2(5 contains one new band at 9.50 eV (IT) and that of 25^ contains two new bands at 9.35 and 10.00 eV (n_/n+). According to McGlynn and coworkers, inter­ action between the n, and IT, orbitals, which have the T T same symmetry, is most likely to occur.

From the photoelectron spectrum of 26, no isomerism is apparent in the gas phase for this molecule, even though the visible and infrared spectra of its solutions indicate the presence of conformational isomers. As the temperature of a solution of 26^ is decreased, the 25-like absorption component decreases until at 77°K the remaining absorption is like that of 27. This process has been shown to be totally reversible.

The photoelectron spectral data for 25-27 lead to the conclusion that 27^ should absorb at the longest wave­ length (have lowest energy n+ -♦ IT+ * transition) which is contrary to experiment. An explanation for the energy differences of these compounds may lie in changes in the

CO/CO dihedral angle from the ground to the first excited 30 state. Variations in this dihedral angle greatly affect the energy of the II+ * carbonyl antibonding M O 1 s, with the n+ -» II+ transition predicted to be of lowest energy when the dicarbonyl system is in a coplanar conformation (dihedral angle of 0 ° or 180°)

Compounds 25_ and 27_ have different absorption band shapes with the band for 2J5 being rather sharp with weak vibrational fine structure. At 7 7 ° K , the vibrational fine structure is more pronounced and the 0 - 0 26 progression is the most intense peak. The visible spectrum of 2?, however, shows a Gaussian band shape.

Based on these absorption band shapes, McGlynn and coworkers have concluded that for 25 the CO/CO dihedral angle is the same in both the ground and excited states and that it is probably 0°. Large changes in the dihedral angle must exist between the ground and excited states * of 27, thereby increasing the n+ -* Il+ transition energy and causing the absorption maximum to occur at shorter wavelengths.

The uncertainty of the conformation of 25-27 in solution makes it difficult to draw any firm conclusions about the above case. Thus, the availability of «- diketones 28, 29, 22, and 30 prompted us to undertake a study of their photoelectron and electronic spectra 14

28 29

22 30 for comparison. These diketones possess similar, but conformationally rigid molecular structures to those of

25-27; thus the geometrical relationships of the double bonds {or cyclopropane rings) to the -COCO- moieties can be estimated with a high degree of certainty.

The carbocyclic framework of diketones 28-30, which consists of four fused cyclopentane rings, was assembled via a domino Diels-Alder cycloaddition, under 13 14 conditions worked out by Paquette and Wyvratt. '

Sodium cyclopentadienide in tetrahydrofuran was 15

l< THF -78°

31 b

C H 3OOC- C= C- COO CH3

H COOCH

COOCH COOCH

COOCH3

COOCH3 COOCH, COOCH3

11 12 oxidatively coupled by addition of iodine at low temperatures to afford 9,10-dihydrofulvalene 31 whicb was immediately reacted in situ with dimethyl acetylene- dicarboxylate. The desired diester 11 was obtained from this reaction in 12% yield via the pincer Diels-Alder reaction shown in path a. Diester 12 was isolated in 5% 16 yield from the domino Diels-Alder reaction (path b) of

31 and the dienophile. Two procedures were utilized for isolation of the desired diester 11. Flash distillation, by slow addition of the crude reaction mixture to an evacuated flask at 2 0 0 °, afforded a dark green distillate which was subjected to controlled saponification. Diester 12 reacts faster than the more sterically hindered 1 1 , and thus by using mild conditions and a short reaction time diester 1 1 ^ could be isolated free of 12. The second method involved selective saponification of the crude reaction mixture and produced diester of not quite as high purity as the first procedure.

The necessary cleavage of the central carbon-carbon bond between the two ester functions in 1 1 ^ was accomplished by overnight heating in toluene with tri- 14 methylchlorosilane and a sodium dispersion. The bis ketene acetal 32_ formed was protonated under conditions of kinetic control by slow addition to methanol to afford a mixture of diesters from which the major endo, endo isomer 33^ could be isolated in 70% yield by fractional crystallization from hexane. Equilibration 17

Si (CH3I3 (J00CH3 COOCH3 COOCH3 Na, C0*0CH3 J J - ^ 7 (CH3 )3 SiCl, CH 3 OH toluene/ A

SiICH3 ]3

11 32 33

of 3j3 to its less sterically hindered epimeric diesters occurred with a catalytic amount of sodium methoxide in methanol at room temperature.

To form the desired diketone an intramolecular acyloin condensation needed to be carried out between the two ester functions in 33. An initial attempt using sodium-potassium alloy in refluxing benzene only caused epimerization to the exo, exo epimer of starting material.

Once 3J3 epimeri zed/ as it would in the presence of any base, an acyloin reaction between the esters became impossible. Conducting the reaction in the presence of 31 32 trimethylchlorosilane, however/ removed any excess base which was present or formed during the course of the reaction. A solution of 33 in toluene was heated at 18 reflux with trimethylchlorosilane and a sodium dispersion for 13 hr and the bis silyl enol ether 34 was hydrolyzed by treatment with methanol at gentle reflux. The gold oil was identified as a-hydroxy ketone 3J5 on the basis of its NMR spectrum which has a complex olefinic region and the proton a to the hydroxyl group appears as a doublet at 8 4.05. Crude 35 was oxidized with dimethyl

QOOCH3 Na,

(CH3J 3 SiCl,

toluene/ L

33 FeCl HCl CH OH 34 ether

OH DMSO

35 28

33 sulfoxide (DMSO) and acetic anhydride at room temperature to give diketone 2J3 in 50% yield from 33^ after column chromatography. The high degree of symmetry

1 13 for 2J3 is verified by its simple H and C NMR spectra

(five lines). The conversion to 2 8 could also be carried 19

out by direct oxidation of .34 with three equivalents of anhydrous ferric chloride and a trace of hydrochloric 32 acid in refluxing ether. This method afforded diketone

28 as a yellow solid in 65% overall yield from 33^ and furnished a much cleaner product than the dimethyl- sulfoxide-acetic anhydride procedure, which suffered from the DMSO removal problem. However, the ferric chloride oxidation was sensitive to excessive moisture and would sometimes fail to produce the deep green reaction mixtures which denoted a successful oxidation.

In these cases, tv-hydroxy ketone 3J5 would be recovered.

To obtain the dihydro derivative 29, diketone 2!3 was reduced with diimide in methanol, carefully following the reaction progress by thin layer chromatography. A mixture of diketones 28, 29, and 22^ was obtained which could be separated by preparative thin layer chromato­ graphy to afford a 26% yield (based on 213 consumed) of

29 as yellow plates. Catalytic hydrogenation of 33 with

10% palladium on charcoal at 50 psig of hydrogen gave a 98% yield of 36, which was converted by subsequent acyloin cyclization and direct oxidation to tetrahydro derivative 22 in 55% yield. Another route to this diketone involved hydrogenation of Diels-Alder product 20

Scheme I

11 EtZnI

28

33 37

1) Na,(CH 3 )3 SiCl, H 2 '

Pd-C toluene, A * V 2) CH3OH

COOCH3 COOCH3 co'och3 COOCH

36 38

1) Na, (CH3 ) 3 SiCl, 1) Na, (CH3 )3 SiCl, toluene, A toluene, A 2) PeCl3, HC1, 2) FeCl3, HCl, ether, A ether, A

22 30 21

1 1 followed "by sodium and trimethylchlorosilane promoted reductive cleavage of the central carbon-carbon bond to afford 3j5 in 57% yield. It was difficult to separate pure 36_ by fractional crystallization from the epimeric mixture of diesters produced by this reductive cleavage, so hydrogenation of 3J3 was the preferred method.

Simmons-Smith cyclopropanation of 1JL afforded almost exclusively the exo,exo bdscyclopropyl diester 37, which was readily purified by column chromatography and recrystallization from hexane. The C 2 V symmetry of this molecule is substantiated by the presence of seven lines 13 in its C NMR spectrum. The reductive cleavage of 37, again with sodium and trimethylchlorosilane, proceeded cleanly and in good yield {90%) to 38. Breaking the central carbon-carbon bond changes the geometry of the molecule so that the cyclopentane rings are no longer puckered, but each one is almost planar. This affected the NMR shifts of the cyclopropyl methylene protons, which in compound 37_ both lie within the shielding cone of the five-membered ring and thus are all observed from

38 22

6 0.17 to -0.33. In 38, "however, only one of the cyclopropyl protons is shielded (6 -0.12 to -0.42) while H„ appears in the normal shift range for cyclo- propyl protons (6 1.02 to 0.55). Biscyclopropyl diketone

30 was then prepared by the same procedure as before in

70% yield.

The photoelectron (PE) spectra of diketones 28, 29, 34 22, and 30 were recorded by Gleiter and Bartetzko and are shown in Figure 1. Table 1 contains the measured vertical ionization potentials, together With the results of molecular orbital calculations. Once again, the data were interpreted on the basis of Koopmans' theorem

(equation 1 ) .

Comparison of these PE data with those of related molecules allows predictions to be made concerning the location of different ionization bands. As discussed previously, a split of about 1.5 eV is expected between the two lone pair combinations (n± ) of the dicarbonyl OR ^ R moiety with n+ above n . ' An estimation of 9.5 eV was made for the II bond energy, based on the ionization

3 6 potential of norbornene (8.97 eV) and the inductive effect of the C2°2 ^unc't:^on (0.4-0.5 eV).^^a"c Through- bond interaction of the II-orbital electrons with the U1 U1 H < < oc OC H z z 3 3 O o o o

a 7 a o to 11 1 2 ' a 14 15 16 6 7 a o 10. 11 12 13 14 15 16 I. P. (eV) l.P.(eV)

iu ® @

a 7 8 9 10 11 12 13 14 15 IB a 7 8 9 10 11 12 13 14 15 16 I.P.(eV) I.P.leV) NJ 0J

Figure 1. PE spectra of 28, 29, 22, and 30 between 8 and 16 eV. 24

TABLE 1

COMPARISON BETWEEN MEASURED VERTICAL IONIZATION

POTENTIALS, Iv J# OP 28, 29, 22, AND 3£ AND

CALCULATED ORBITAL ENERGIES (eV)

Compound Band I Assignment “GJ “Gj V, J (MINDO/3) (CNDO/S)

28 1 8.85 8.69 9.28 altn+> 2 9.15 bl(nJ 9.25 9.42

3 1 0 . 0 a (II ) 9.61 9.86

4 10.3 b 2 (n_) 10.41 11.43

29 1 8.80 a' (n+ ) 8.77 9.33

2 9.45 a' (IT) 9.51 9.72 (£S} 3 10.4 a" (n ) 10.37 11.39

2 2 1 8.82 8.85 9.43 al tn+ ) 2 10.26 V n-> 9.87 11.32 30. 1 8 . 6 a,(n^) 8.70 9.37 1 + 2 9.6 9.56 10.55 2 v ^ W 3 1 0 . 0 a2(*2J 9.75 11.00 4 10.14 9.99 11.14 al (*3> 5 10.7 b2(nJ 10.23 11.32 6 10.74 1 2 . 0 0 10.7 J 25

cr frame would result in a split between the II_ and n+ 21 37 linear combinations ' shown below.

7T_ 7T+

On the basis of previous studies of through-bond inter­ action across four cr bonds in [4.4.2jpropella-3, 8 ,11-triene ng 3o (39) and diene 40, this split has been estimated by

Gleiter to be about 0.5—1.0 eV. Xt has also been shown previously that the introduction of nonconjugated double bonds into a mono- or polycyclic monoene shifts the

39 40 center of gravity ~e (IT) to lower energies by -0.1 to -0.2 29 36 eV per added double bond. ' Thus, in the case of 213 the bands due to ionization from II and n+ should be between 9 and 10 eV.

In the region from 9.6 eV on, the PE spectrum of 30 shows strongly overlapping bands. It is felt that the orbital energies of the four linear combinations of the

Walsh orbitals together with that of the n_ combination lie in this region. 26

The results obtained from MUSIDO/3 calculations'^ agree quite well with the experimental values. A comparison of the first bands in the PE spectra of

28-30 (see Figure 2) indicates a relatively constant split between n and n+ in all the diketones. For compound 213 there is a large split between II and II+ which/ on the basis of calculations/ is due to a strong ;>*rV*' through-bond interaction between the II_ linear combination and the a frame.

Figure 2. Correlation between the first bands of the PE

spectra of 22/ 29, 28/ and 30^ based on MINDO/3

calcuations. Strongly overlapping bands are

indicated by the shaded areas. 27

Diketones 28-30 are all yellow crystalline solids,

the intensity of the color decreasing in the order:

30 > 22 > 29^ > 28. The absorption bands measured for

these compounds are listed in Table 2 together with the 40 results of a CNDO/S calculation as performed by Gleiter.

On the basis of the calculations, the first absorption *if maximum results from a HOMO (n ) to LUMO {II, ) transition. T T The latter arises from linear combination of the two IT*

orbitals located at the carbonyl groups. Changing the

solvent from methylene chloride to ethanol causes this

band to shift to lower wavelengths.

Figure 3 compares the visible absorption bands of

the series of diketones and the Gaussian band shape is

readily apparent. A bathochromic shift is observed in

the absorption maxima of 28, 29, and 22 which is the r opposite of the trend for the maxima of 25_ to 2.1. Several

esqplanations for this observed trend in the energy of the * n, - II, transition are possible. As discussed previously, T T differences between CO/CO dihedral angles in the ground

and excited states can greatly affect the transition 30 energy. The energy difference between the maxima of the first bands in the absorption and emission spectra provides a good estimate of the corresponding angle TABLE 2

MEASURED AND CALCULATED ELECTRONIC ABSORPTION SPECTRA OP 28, 29, 22, and 30

Observed Calculated Compound Band \[nm] X [nm] Predominant Configuration lo9 emax

28 A 408.5a 1.20 405.4 n* (CO) - n, (74%); + +

403.7b 1.18 n*— - n “r (2 6 %) B 271.5C 3.14 270.9 n* (CO) - n+ (26%); 246b n* - n+ (74%)

C 228.5C 3.99 212.4 n*(co) - n_(c=c)(1 0 0 %)

A 421.0a 453.3 II* (CO) - n (67%); 29 1.34 T T

1.33 n*(co) - n(c=c) (2 6 %) 417.6b *r B 255.7C 3.54 306.1 II* (CO) - n (40%); T i 27 2b II* (CO) - n_ (44%)

A 1.49 456.7 II* (CO) - n (94%) 22 427.2a + + 424.2h 1.48 Table 2 (continued)

Observed Calculated Compound Band X[nm] log e \[nm] ^ max Predominant Conf.

30 A 412.0a 1.23 456.0 11* (CO) - n (90%) + + 407.4b 1 . 2 2 B 230.0C 3.61 302.0 II* (CO) - n+ (58%)

II* (CO) - n (41%) T —

Methylene chloride.

bEthanol.

cCyclohexane. 30

0.8 -

29

0.4 - -

0.2 -

4 0 0 5 0 0

Figure 3. Comparison between the first bands in the electronic absorption spectra of 28/ 29/ 22/ and 3(3 in methylene chloride. 31 difference. As shown in Table 3, only small differences exist in the case of our compounds. Also, the ionization potentials of the first band, corresponding to n+, in the PE spectra remain relatively constant through the series of compounds. The remaining possibility is a spatial interaction between b^(II_) of the olefinic * moiety and ) of the C 2 0 2 group. According to the calculations, such an interaction exists, leading to an ★ increased energy separation between n+ and II+ as depicted in Figure 4.

Therefore, the hypsochromic shift observed with the addition of double bonds in the series of conformationally rigid a-diketones 22, 29, and 2t3 is due to a spatial interaction between the orbitals of the olefin and those of the diketone. The PE spectra of these compounds and those of 25 through 71_ show strong similarities, and in both cases indicate through-bond interaction is occurring.

The overriding factor in the opposite shift trend observed for 27, to 25^ is changes in the CO/CO dihedral angle, as mentioned earlier. In 27, the six-membered rings are more flexible and a larger angle change is observed than in the more rigid 25. The bathochromic shift caused by these angle differences of the cyclobutane carbonyls overwhelms the through-space effect.

Another possibility which would decrease the through-space 32

TABLE 3

EMISSION DATA OF COMPOUNDS 28, 29, 22, and 30a'b

~emiss ~abs V* Compound % a x max max ~~ max run cm- 1 cm- 1

28 464.0 21 552 2928 29 465.5 21 482 2271

2 2 473.0 21 142 2267 30 470.0 21 277 2995

Methylene chloride solution.

1. The data are uncorrected.

€ n.

Figure 4. Comparison of the orbital energies for 28

without (left) and with (right) II+ -

interaction. 33

interaction, is the adoption by 25^ of conformations like

25b in solution.

25b

Benzilic acid rearrangement of diketone 28 would produce the molecular framework 41, which could be

converted into compounds of interest for solvolysis

studies. Treatment of similar a-hydroxy carboxylic

base

28

'H

V *

43 42

acids with thionyl chloride in refluxing benzene has been shown to give the corresponding ketones^ (i.e. 42) .

Reduction of 42 should produce alcohol 43, which is of 34

interest because the geometry of the molecule causes the double bond ri-orbitals to cant inward such that appreciable overlap between them and the p-orbital of an apical carbonium ion would be expected. This neighboring group participation should lead to a nonclassical ion such as 44. Thus, a comparative solvolysis rate study of

44 derivatives of £3 with and without double bonds or cyclopropane rings would be most informative as to the extent of this participation.

Several different solvent and base combinations were tried in an attempt to convert 28 to 41. Treatment of

28 with aqueous potassium hydroxide in dioxane at 41 reflux, potassium hydroxide in t-butanol and water at 42 reflux, extremely concentrated aqueous potassium hydroxide at reflux, or sodium methoxide in methanol at 42 reflux only afforded starting 2J3 after acidification.

The double bond p-orbitals may hinder the approach of the reagent to the carbonyl p-orbital, thus preventing any reaction. A photochemical Wolff rearrangement of ar-diazo ketone 45^ would also result in the desired ring 43 contraction to 46a, in this case. Treatment of 28^ with p-toluenesulfonylhydrazine resulted in only partial

0 COOR

28 1) TsNHNHg f

2 ) basic alumina 46 a R H 45 b R — CH^ c R = Et conversion to a mixture of mono- and ditosylhydrazones.

The reaction failed to go to completion even in refluxing methanol. After removal of unreacted tosylhydrazine, 44 the crude product mixture was stirred with basic alumina to afford a semi-solid whose infrared spectrum contained a definite diazo stretch. However, further pursuit of this reaction sequence was abandoned due to the difficulty in forming monotosylhydrazone cleanly. 36

46 The Favorskii rearrangement presented another method of achieving our goal. Treatment of a-hydroxy ketone 3J5 with tosyl chloride in pyridine afforded the corresponding tosylate 4 7 . This conversion required a longer reaction time than expected and thus, further demonstrated the inaccessibility of the (3 face of the molecule to reagents. Treatment of 47 with sodium 45 methoxide in methanol at reflux yielded diketone 2J9 as the major product. As has been observed previously 46 in analogous molecules, elimination of p-toluenesulfinate

0 ■OTsOTs TsCl NaOCH3/CH3OH » 46b pyridine

47 OH

35

S0C1 NaOEt -+> 46c ether, &

48 37

anion following base abstraction of the most acidic proton (a to -OTs) occurred rather than the desired ring contraction. To rule out the possibility of such elimination, or-halo ketones were next considered. Treat­ ment of acyloin 3j5 with phosphorus tribromide and pyridine in benzene failed to give the desired or-bromo ketone. Reaction of 35_ with neat thionyl chloride gave a compound whose crude H NMR spectrum was consistent with 48. Treatment of crude 48^ with sodium ethoxide in refluxing ether resulted in hydrolysis back to 35. From examination of molecular models it can be seen that removal of HA produces an enolate anion whose orbitals are improperly aligned for delocalization of the negative charge over the carbonyl p-orbitals. Rather, upon treatment with base, removal of H_ and formation of the

0

48 resulting stabilized enolate anion probably occurs. In the case of 48, because of the strain caused by cyclopro­ pane formation and the problem with removal, the 38

desired Favorskii rearrangement would have to occur via the semibenzilic mechanism (quasi-Favorskii), involving nucleophilic attack by the base on the carbonyl. Initial formation of the enolate resulting from abstraction would prevent such nucleophilic attack from occurring.

The next route to ketone 42^ tried was the Ruzicka 47 cyclization of a diacid (see equation 2 for an example).

The attempted preparation of the required diendo diacid

r - r r r - ° oH —275 / —1 0 0 - [T^t >=°

55%

49 by treatment of diketone 2!3 with hydrogen peroxide and sodium hydroxide under a variety of conditions^® yielded mainly starting diketone. Conversion of diacid

50 to its bistrimethylsilyl ester with trimethylchloro-

silane and triethylamine in toluene followed by central bond cleavage with sodium and trimethylchlorosilane in refluxing toluene produced an epimeric mixture of cleaved diacids with 49 as the major epimer. Reaction of this diacid mixture on a small scale with barium oxide at 47 200-300 under reduced pressure produced no 42. 39 QOOH COOH 1) E^N, (CH3)3SiCl Na, (CH3 )3SiCl toluene, A QOQH COOH 50 3) CH3OH ■ I l U

COOCH3 1) {CH3 } 3 S 1 I, co'och3 49 42 * wH CHC1- 2) H+

33

Treatment of diendo diester 33_ with anhydrous lithium 49 bromide in dimethylformamide at reflux afforded only unepimerized starting material. However, 3J3 reacted with 50 sodium cyanide in hexamethylphosphoramide at 73° to give largely diexo diacid 51. Trimethylsilyl iodide has

0 0 ft

51 been found to cleave esters to the corresponding acids 51 under mild, neutral conditions. Treatment of 33 with 40

this reagent in refluxing chloroform produced epimeri- cally pure diendo diacid 49^ in 98% yield. The structural assignment was confirmed by comparison of the olefinic region of its NMR spectrum with those of 51^ and the mixture formed by reductive cleavage of the bistrimethyl- silyl ester. The ester cleavage reaction with trimethyl- silyl iodide is very sensitive to the purity of diester

33 used. In cases where 3^ had not been chromatographed on silica gel prior to use, reaction of trimethylsilyl 52 iodide with the double bonds occurred. Attempted cyclization of 49_ by treatment with thionyl chloride at 53 reflux followed by heating with anhydrous methanol gave an epimeric mixture of cleaved diesters as the major product plus a gold oil whose NMR spectrum indicates that some reaction had taken place at the double bonds.

Anhydrous potassium fluoride has been found to serve as a catalyst in the cyclization of adipic acid to cyclo- 54 pentanone and has the advantage of being less basic than barium oxide, whose basic character could cause epimerization to a diacid which is incapable of under­ going intramolecular cyclization. However, pyrolysis of

49 with anhydrous potassium fluoride at 270° did not 41

lead to any reaction other than some decomposition.

Conducting these heterogeneous pyrolyses on a small scale presented difficulties in the isolation of any product by distillation.

Access to the molecular framework of 42 might also be achieved via the Dieckmann reaction of diester 33_ to give 52. A minor product from the reaction of ljL with sodium in liquid ammonia at -78° was tentatively assigned

C00CH3 0 c o 'o c h 3 -COOCH

33 52

EC 1 structure 52 by Wyvratt on the basis of the H NMR spectrum of the product mixture. Repetition of this experiment, quenching with aqueous ammonium chloride solution, and separation of the products by preparative thin layer chromatography did not result in the isolation of any 52. Condensation attempts with sodium methoxide and sodium hydride afforded only epimerization of ,33 and with lithium N-isopropylcyclohexylamide as a base 42

unepimerized 33^ was recovered. In most cases/ the

Dieckmann condensation is unsuccessful if the product 56 cannot form a stable p-keto ester enolate and ketone

52 is incapable of forming such a stabilized enolate.

However, diester 5JS has been successfully cyclized in good yield to £54 using the sodium salt of dimethyl 57 sulfoxide. Treatment of 3J3 with this sodium salt at

90° followed by quenching with aqueous acetic acid afforded a rather insoluble white solid whose infrared

and mass spectra are indicative of an anhydride.

1) Na C H 2 SOCH3, C00CH CH-SOCH-,0°

53 54 RESULTS AND DISCUSSION

Part II. Synthesis and Chemistry of a C^ Symmetric

Bislactone

The stereoelectronic requirements for cleavage of

the central carbon-carbon bond in 1 ,4-diketones have been

investigated. In the case of the reductive cleavage of

a cyclopropyl ring conjugated to a carbonyl function, the

carbon-carbon bond which is broken by reaction with lithium

in liquid ammonia is the one orthogonal to the plane of 53 the carbonyl group (equation 3)„ Likewise, the C 2 -C 3

bond of a 1 ,4-diketone will undergo reductive cleavage if

that bond is orthogonal to the planes of both carbonyl 59 functions. This alignment allows maximum overlap in the

transition state between the carbonyl II-orbitals and the

carbon p-orbitals which develop during the breaking of the

sigma bond. Figure 5 illustrates an orbital alignment which facilitates 1,4-diketone reductive cleavage (A) and one which prevents it (B).

43 44

Figure 5. Orbital alignments of 1,4-diketones. In A

cr-bond reductive cleavage can occur, whereas

in B it will not occur.

The synthesis of biscyclopentenone 55_ was accomplished by Paquette and coworkers in eight steps and 43% overall 60 yield from diester 11 (Scheme II) . In 55 as in 11, two ester groups are situated such that treatment with sodium 14 and trimethylchlorosilane would be expected to result in cleavage of the central carbon-carbon bond. However, sodium reduction of 55^ followed by filtration into anhydrous methanol afforded a compound whose mass spectrum indicates four hydrogen atoms had been incorporated. The infrared spectrum indicates the cyclopentenone and ester moieties are still present while the NMR spectrum possesses two different methyl ester singlets. Consideration of 45

Scheme II

1) KOH, CH3OH 2) KOH/H20 I«/KI

COOCH-* 3) HC1 COOH NaHC03 COOCH- COOH 11 NaOCH3,

c h 3o h

Zn(Cu)

CH OH )H COOCH OH |COOCH COOCH, COOCH, COOCH,

c o o c h3

BF 4“ KOH/DMSO NaOH, 0 ► ♦ 2)HBF A

COOCH3 COOCH- COOCH3

C 00C H3

h C r o 46

compound_J55^ shows that the molecule contains two 1/ 6- dicarbonyl units each of which possesses a double bond. Since the carbon framework of the a ,B-unsaturated carbonyl moieties is rigid and the ester groups can rotate, the orbital alignment of each unit is that depicted

in Figure 6, which is an arrangement conducive to reductive

cleavage of the C^-C,. sigma bond. As shown in Scheme m , the electrons are transferred to the ketone carbonyl groups whose reduction potential is generally lower than that of the saturated carboxylic esters. The silylated tetraenolate

56 thus formed was protonated under conditions of kinetic control to give 57. This structural assignment is based on the lack of symmetry implied by the presence of two 1 13 methyl ester peaks in the H NMR and 22 lines m the C

Figure 6. Orbital alignment present in 55. Scheme III

COOCH 3

COOCH 3

2 e

CH,0

t C H j g H OSi(CH3 )3 0" “OCH 3 CHiO "0 (CH3 )^5iO OSi(CH3 ) 3 OCH 56

H 0 o f - \ "h 0 c h 3ooc c o o c h 3 CH 3 00C C O O C H 3 48

NMR spectrum. The molecule most likely adopts the configuration illustrated with the cyclohexane ring in a skew-boat conformation61 and most large substituents equatorially disposed.

COOCH

COOCH

57

This configuration also places the axial o--carbornethoxy proton within the deshielding cone of the adj acent double bond, thus changing its shift substantially relative to that of its equatorial counterpart, as seen in the 1H NMR spectrum. i Treatment of 55^ with sodium in liquid ammonia followed by quenching with ammonium chloride solution under conditions of thermodynamic control led to highly symmetrical diester *513 in 66% yield. The H NMR spectrum shows one singlet for the methoxy groups while there are 13 11 lines in the C NMR spectrum indicating C symmetry.

The molecule exists in a conformation similar to 57 with 49

58

the ester groups in equatorial positions. Sodium methoxide in methanol at room temperature for a prolonged period completely epimerized 5^7 to 58; conversely/ treatment of

58 with methanolic hydrogen chloride resulted in its conversion to 57_. Careful examination of the ^H NMR spectrum of the crude product from reaction of 5_5 with sodium in liquid ammonia shows that a trace of 57_ is present. Since excess trimethylchlorosilane was not removed from the reaction mixture prior to quenching and formation of 57/ methanolic hydrogen chloride would have been produced in the process. This might explain the exclusive production of 57_ from the reductive cleavage in the presence of trimethylchlorosilane. However/ when excess trimethylsilyl chloride was removed prior'to 62 quenching, only J57 was isolated.

For reasons which will become evident later, the further chemistry of symmetrical 58 was pursued. 50

Controlled catalytic hydrogenation of 5(3 was possible in ethyl acetate solution at -23° and 50 psig with acetic acid-washed 10% palladium on carbon. At room temperature under the same conditions a mixture of products was formed# while no hydrogenation occurred when the catalyst had not been treated-with acid. Filtration of the low temperature hydrogenation mixture and concentration of the ethyl acetate filtrate gave a mixture from which the major product 59_ could be isolated by recrystallization from ethyl acetate. Continued washing of the catalyst with methanol and dichloromethane yielded only additional 5j3

(84% total yield). The mass spectrum of 59 indicates that

10% Pd/C

-23 OCH

58 59 two moles of hydrogen were taken up and both the NMR 13 (one signal for the two methyl esters) and C NMR spectra

(11 lines) are consistent with a symmetrical molecule.

Examination of the configuration of 5J3 shows that reaction on a catalytic surface would encounter the least steric 51

hindrance if it occurred in the manner illustrated above.

The very minor products may result from epimerization of

59 in the presence of the acid-washed catalyst. Thus, hydrogenation produces a molecule which is ideally set up for lactone formation within its two identical cup­ shaped halves. Treatment of 59^ with sodium borohydride in ethanol at room temperature resulted in hydride attack on the ketones from the less sterically hindered convex face followed by nucleophilic attack of the alkoxide anions produced on the ester functions to yield 60^ in 38% yield. Lactone formation is corroborated by the appearance of protons at 6 4.9 in the NMR spectrum and a typical

60 mass spectral fragmentation pattern. The 10 lines 13 exhibited in the C NMR spectrum confirm the symmetry of the molecule. In an attempt to minimize overreduction and increase the yield of 60, the less reactive reducing 6 3 agent sodium cyanoborohydride was tried in methanol at 52

about pH 4, but no reaction occurred, possibly because of solubility problems.

Bislactone 60 is a functionalized dimeric triquinane which upon cleavage of bond a would produce the functiona­ lized d&-bivalvane 61. Besides its unique synthesis, 61_ is of interest because the functionality present could

60 61 assist in the process of joining the two identical halves 12,64 of the molecule to form a dodecahedrane-lihe structure.

Prior to bond cleavage attempts, attention was focused on the transformation of the 6-lactones into cyclopentanones

62 (Scheme IV ). Treatment with methanolic hydrogen chloride or methanolic hydrogen bromide could form 63 which upon conversion to an organometallic reagent would hopefully undergo intramolecular addition to the ester group. Methanolic hydrogen chloride reacted with 60^ to give a new compound whose NMR spectrum exhibits a 53

Scheme IV

HC1 or HBr, 60 62 CH-.OH /

CH-aOOC COOCH. OHC CHO 67 b X = Br

SOC1

HO OH 66

single methyl ester peak/ hut whose mass spectral

fragmentation was inconsistent with structure 63a. The

compound produced on treatment of 60_ with methanolic

hydrogen branide has a promising NMR spectrum but its

mass spectrum does not fit 63b. Further characterization

of these reaction products was not attempted. Using 54

followed by the addition of methanol led to a complex mixture from which no single product could be isolated.

Another route for conversion of bislactone 60^ to biseyelopentanone 62_ involves an initial reduction to bislactol 64. Although treatment of 59_ and 60^ with 66 o diisobutylaluminum hydride in toluene at -70 produced the desired 64/ the method of choice was reduction of !59 with lithium aluminum hydride in tetrahydrofuran at 0°.

Prolonged standing at room temperature or heating of 64_ resulted in sane elimination of water to form a mono- or dihydropyran contaminant 65^ as indicated by the appearance of olefinic resonances in the NMR spectrum. Treatment of 64 with neat thionyl chloride at room temperature gave

65 bischloro ether 66. The NMR spectrum exhibits an absorption at 6 6.13 (-CHC10-) and a broad singlet at

6 4.28 (> CHO). Since 66 hydrolyzed rapidly upon exposure to air/ the crude compound was dissolved in dry 55

benzene and treated with a silver perchlorate solution at room temperature. Column chromatography of the reaction mixture on Plorisil gave no traces of the hoped for ene aldehyde 67. Conversion of 67_ to 62^ would have been attempted using a rhodium (I) complex known to catalyze the addition of aldehyde functions to carbon-carbon double 67 68 bonds or various Lewis acids. At this point, on the 55 basis of the results of other workers and those described above, preparation of 62 was abandoned and the cleavage of bond a joining the two bislactone moieties in 6(3 was pursued.

Bond a in 60_ would be expected to be subject to reductive cleavage with metals, since it is the C^-C^ bond of a 1, 4-dicarbonyl system. Treatment of bislactone 60 with sodium in liquid ammonia at -78° followed by saturated ammonium chloride solution resulted in formation of 6j4 only. In this case, it is possible that the intermediate radical anion abstracted a proton from ammonia since the starting material could not readily donate a proton and no added proton donor was present during the course of the reaction. To minimize the possibility of solvent proton donation, the reduction was carried out in hexamethyl- o 69 phosphoramide-tetrahydrofuran with sodium metal at 0 , 56

but only starting bislactone was isolated. An uncharacteri- zed oil/ whose H NMR spectrum exhibited a very broad absorption at high field, was obtained from treatment of

60 with dispersed sodium and trimethylchlorosilane in refluxing toluene followed by addition to methanol.

Close study of the configuration of 60^ shows that the I,4-dicarbonyl subunit is held such that the transition state for cleavage of bond a is not at all stabilized by orbital overlap as discussed previously (see Figure 5, alignment B). A great deal of strain develops in the molecule when an attempt is made to align the lactone carbonyls properly for reductive cleavage. At high temperatures the flexibility of the molecule might increase, thus permitting the desired cleavage to occur, but bislactone 6(3 was recovered unchanged after treatment for 70 12 hr with aclactivated zinc dust and acetic acid in xylene 4o 5^,71 at 120

Failure to reductively cleave bond a prompted an attempt to sever it by oxidative' means. The first step was generation of a double bond, which could be oxidatively cleaved, in place of bond a. Initial attempts involved formation of the bisenolate of 60_ followed by addition of one equivalent of halogen in the hope that the remaining 57

enolate anion would affect dehalogenation to afford 68-

However, to attain the charge delocalization necessary

60 68

for stabilization of the bisenolate, the 1/4—dihetone

system would need to adopt a planar configuration which places a great deal of strain on the molecule. Thus,

attempts were redirected toward halogenation of a mon-

enolate anion and subsequent dehydrohalogenation. The

lithium alhylamides were attractive as bases for this purpose because they would be easy to remove from the

final reaction mixture and the amine formed could potentially act as the base for the dehydrohalogenation.

The desired enolate anion proved unexpectedly difficult 72 to generate. With lithium diisopropyl amide as the base, experiments conducted at -78° and at 0° produced only

starting material; whereas when the reaction mixture was

allowed to warm to room temperature prior to quenching with halogen a product was isolated, in some cases, whose mass spectrum corresponds to bislactone plus diisopropyl­ amine. It appears that rather than abstracting the 58

enolizable proton, which is not readily accessible because of its hindered location, the base attacked the lactone moiety to afford a compound such as 69. Even use of the highly hindered lithium 2, 2,6,6-tetramethyl- piperidide resulted in some attack on the lactone plus

69 recovery of starting material. The non-nucleophilic triphenylmethide base has been used successfully by

Ourisson and coworkers in the conversion of lactone 70^ to the a'-bromo compound 7JL which upon further treatment

1) 0-aCLi 0 3----- » 0 2) BrCCH^Br CH3

70 71 72 with 1,5-diazabicyclononene (DBN) gave 72^ in 70% overall 73 yield. In our case, the stenc bulkiness of triphenylmethide could present problems in removing the 59

desired proton in 60; "however/ gentle heating of the reaction mixture resulted in enolate formation. Treatment of bislactone 60_ in tetrahydrofuran with slightly more than two equivalents of lithium triphenylmethide was followed by heating the reaction mixture at gentle reflux for 3 hr. Two equivalents of iodine were added at 0° and the mixture was stirred at room temperature for 3 hr.

Workup afforded 68 in yields ranging from 30-60%. The 13 C NMR spectrum is consistent with a symmetrical molecule (10 different carbons) and exhibits both carbonyl and olefin lines, while the infrared and NMR spectra corroborate the continued existence of the 6-lactone moiety. It has not been determined whether 6j3 was formed by monohalogenation of a bisenolate or by halogenation of a monoenolate anion and subsequent dehydrohalogenation by the excess base present in the reaction mixture.

The sometimes erratic yields of the conversion discussed above prompted a search for an alternate method.

Intramolecular coupling of the bisenolate with cupric 74 chloride in dimethylformamide resulted in mainly starting material plus a small amount of chlorinated bislactone. Selenium dioxide oxidation in refluxing 60

dioxane failed# possibly because the two hydrogen atoms

to be removed are trans to one another. No reaction also

occurred with 2, 3-dichloro-5,6-dicyanobenzoquinone (DDQ) 75 in refluxing dioxane. Irradiation of 60_ in refluxing

carbon tetrachloride in the presence of N-bromosuccinimide gave starting material again.

Cleavage of the double bond in 68 was attempted with

ozone in methanol and methylene chloride followed by 76 dimethylsulfide addition. No reaction occurred when the ozonolysis experiment was conducted at -23° or 0°. 77 78 Oxidation with ruthenium tetroxide * also resulted in

isolation of mainly starting unsaturated bislactone.

The limited quantity of 6J3 available precluded further

exploration of oxidative cleavage conditions. RESULTS AND DISCUSSION

79 The basket-shaped molecule C^g-hexaquinacene (24)

Is of considerable synthetic and theoretical interest because of its convex topology and high degree of symmetry/ three mirror planes intersecting a three-fold rotation axis (£3v) . The geometry of 2j4 could allow trishapto coordination of the double bonds to suitable metal atoms which would result in the enclosure of an unusual spherical

24 cavity. The arrangement of the three p-II orbitals in this triene is such that a cyclic six-electron pp-cr homoconjugative interaction within the molecule could be 80 substantial. Of prime importance to the synthetic chemist is the relationship of J24 to the pentagonal dodecahedrane 1/ and the possibility of incorporating the necessary four additional carbon atoms in the proper manner to attain 1.

61 62

The synthesis of 2£ has been accomplished in eight

steps from domino Diels-Alder diacid 50^, which consists

of six five-membered rings and 14 carbon atoms. The

incorporation of the two additional carbon units necessary

for the C^-hexaquinacene framework was accomplished by

the dropwise addition of a tetrahydrofuran solution of

50 to lithium hydride followed by the addition of excess methyllithium. The reaction mixture was quenched by

addition to a cold, stirred solution of dilute hydrochloric

acid. Diketone 73_ was isolated in 85% yield after

chromatography of the crude reaction mixture on silica gel.

The infrared spectrum of 73 indicates the presence of a

| COOH COOH 63

— 1 13 ketonic carbonyl (1682 cm ) and the C NMR spectrum

(6 lines) attests to the symmetry of the molecule. The inverse quenching procedure utilized in the above reaction was very-important in this case, since the addition of acid to the reaction mixture would result in further reaction of the diketone formed with the excess base present to produce the tertiary alcohol 74^ or intra­ molecular aldol condensation product 75. An alternate method for the conversion of 50 to 7J3 involved treatment of the diacid with thionyl chloride and pyridine in benzene followed by reaction of the crude acid chloride with seven equivalents of lithium dimethylcuprate in 81 ether and tetrahydrofuran to afford 73_ in 63% yield.

Now that the requisite 16 carbon atoms were present the key step of our synthesis involved the formation of cyclohexenedione 7<5 which is set up to undergo a [2 + 2] cycloaddition to either of the norbornene double bonds to afford 77. Compound J2. possesses the basic structural framework of 24. Several methods were tried for joining

P /v

76 77 64

the ketone methyl groups of 73^ to form the desired 1,4- cyclohexanedione ring. Success was achieved using a 74 modification of Saegusa's oxidative coupling procedure, in which the bisenolate was generated by addition of a solution of 73^ to two equivalents of lithium diisopropyl - amide in tetrahydrofuran and hexane solution at -78°. The cold bisenolate solution was stirred for six hours prior to transfer via cannula to a solution of anhydrous cupric chloride in dimethylformamide and tetrahydrofuran (5:1) at -78°. After column chromatography on silica gel, compound 78^ was isolated in 53% yield. Inverse addition of the bisenolate to the cupric chloride solution was critical to the success of this coupling reaction, since the reverse procedure resulted in significantly lower yields of 78, plus a substantial amount of 75_. The enhanced acidity of the methylene protons or to the carbonyl groups in 78_ allows proton transfer to unreacted bisenolate from

78 as it forms, thus decreasing the yield of coupled 1 diketone. The H NMR spectrum of 78 is quite simple and exhibits narrow multiplets at 6 6.05 (olefinic protons),

3.59 (allylic protons), and 2.60 (2CH) and a singlet at

1 3 2.33{-CHoC=0). The C NMR spectrum confirms the £2v 65

1) LiN(iPr)2 2 equiv 73 ► +■ 21 2) CuCl2/ s / 7 4 - DMF, THF COCH c o c h 2ci

78 79 symmetry of 78^ with its six different carbon lines: one olefinic, one carbonyl, and four saturated carbons.

In addition to 78, varying amounts of aldol product

75, monochloride 79^ and an incompletely characterized product were isolated from the above reaction. The former two side products were probably formed from mono- enolate, which may have been present prior to the coupling due to incomplete bisenolate formation. Treatment of 73_ with one equivalent of lithium diisopropylamide at -78° afforded 75 in fair yield. The reaction of cupric chloride with carbonyl compounds and their enolates in 82 dimethylformamide has been known to afford a-chloro ketones. Formation of 79^ directly from the bisenolate of

73 is not considered likely, since the remaining enolate would probably displace the chlorine. The above con­ clusions are further substantiated by the decreased yields of 79 and 75 when the time allowed for bisenolate 66

formation was increased. In cases where the n-butyllithium used to form the base contained alkoxides or the reaction mixture was exposed to the cupric chloride in dimethyl - formamide solution for a prolonged period another side product, tentatively assigned structure 80, was isolated.

Condensation of coupled diketone with dimethylformamide apparently occurred.

80

Dehydrogenation of the 1,4-diketone moiety in 7f3 was affected by treatment with selenium dioxide and potassium dihydrogen phosphate in dioxane at the reflux temperature.

Triene dione 7£ was formed in 75% yield after workup and its structure is confirmed by the presence of two different i olefinic resonances in the NMR spectrum, a singlet at

6 6.61 (2 hydrogens) and a multiplet at 6 5.94 (4 hydrogens).

The reaction mixture was heated at reflux for only 1.5 hr since prolonged heating resulted in a lower yield of 76. 67

This observation may be due to a thermal retro Diels-Alder reaction which would convert triene dione 76_ into compound

81. This compound was never actually isolated from the reaction mixture and characterized/ however. Treatment of 78 with DDQ in refluxing dioxane did not result in

81 formation of 76/ but some of the starting dilcetone decomposed.

The [2 + 2] cycloaddition of the enedione moiety in

76 to one of the symmetry-equivalent norbornene double 82 bonds has substantial literature precedent. The trans­ formation was affected in 87% yield by irradiation of dilute solutions of 76_ in dry benzene through Pyrex with

d 13 3500 A light in a Rayonette reactor. The nine line C

NMR spectrum corroborates the symmetry of 77. It was found that the reaction would not go to completion in the

1.5 hr photolysis time unless the sample of 76 used had been washed with a saturated sodium bicarbonate solution 68

during workup of the selenium dioxide dehydrogenation.

Prolonged irradiation resulted in precipitation of a

polymeric solid. Purification of the cage-like 7J7 proved

difficult because of its unstable, highly strained nature,

so the crude photolysis product was immediately subjected

to the subsequent reaction.

Two bonds in 77 must be cleaved to form the tetra- hydrohexaquinacenedione which serves as the precursor to

the desired 24. These two bonds constitute the C^-C^

sigma bonds of the two 1,4-dicarbonyl subunits in 77. As has been discussed previously, the orbital alignment between the carbonyl p-II orbitals and the sigma orbitals of the C 2 ~C3 bond is critical to successful reductive 59 cleavage. For example, the geometry of cyclohexanedione

78 places the two sets of orbitals in an undesirable orthogonal relationship (see A in Figure 7). The theory

is substantiated experimentally, in that treatment of 78 with zinc in acetic acid at reflux resulted in no reaction.

However, the 1,4-dicarbonyl units in 77^ are positioned

such that maximum orbital interaction is possible during reductive cleavage (see B in Figure 7). 69

A B

Figure 7. Orbital alignments present in 1,4-diketones 7j3

(A) and 77 (B).

It had been discovered previously, that the similar

1, 4-diketone 8j2 upon reaction with zinc dust in acetic

acid at room temperature gave 83, the product of cyclo- 71 butane bond cleavage. In our case, treatment of 77

Zn, HOAc

25

82 83 70 with activated zinc dust in acetic acid at room tempera­

ture resulted only in cleavage of the central norbornyl-

type bond to give 84. The structural assignment is 13 corroborated by the off-resonanc e C NMR spectrum which 70

exhibits each of the seven saturated carbon signals as a doublet. If the cyclobutyl bond had been cleaved, a combination of five doublets, one triplet, and one singlet would have been observed for the saturated carbon atoms.

hv Zn HOAc 25°

76 77 84

Zn, HOAc 118°

85

The cleavage of the more highly strained central bond permits the molecule to spread open slightly and the geometric alignment of its remaining 1,4-dihetone unit becomes less ideal than before. Zinc in acetic acid at the reflux temperature (118°) still afforded cleavage to 71

85, however. Treatment of T7 under these conditions followed by column chromatography on silica gel and elution with 30-50% ethyl acetate in hexane gave a 60% 13 yield of 85. The C NMR spectrum of this compound exhibits nine lines and the mass spectrum indicates the addition of two hydrogen atoms. Varying amounts of IQ were also isolated from this reaction by elution of the column with 10% ethyl acetate in hexane. This latter compound probably was formed by zinc-acetic acid reduction of any 7(5 which did not undergo photo cycliz at ion in the previous step.

The final transformation, conversion of the a-methyl-

ene ketone groups to double bonds, proved unexpectedly difficult. Initial attempts by Snow involved cleavage of the 1,4-diketone unit in 8-4 with concomitant double bond formation (Scheme V). Reduction of 84 with dispersed sodium in refluxing toluene followed by inverse addition of the resultant bisenolate 88 to a diethylchlorophosphate solution failed to produce the desired bisenolphosphate 84 87. Treatment of 84 with sodium borohydride in tetra- hydrofuran and methanol afforded the cyclic ketol 88, whose structure was assigned on the basis of mass spectral 72 Scheme V O _(E_tO) 2PCI __

\9 (EtO) OP(OEt), 86

Li,NH \ J V

NaBH4 24

84 /l) Na/NH3 LiAlH4/ / 2) NH Cl 4 THF, A

90 a X Br __ Cl

data and similar behavior of 82^ under the same conditions/ Q C as observed by Eaton. However/ reduction of 84 with

lithium aluminum hydride in tetrahydrofuran at reflux gave diol 89. Attempts to convert 89/ via its bistetra-

hydropyranyl ether, to dibromide 90a or directly to its

corresponding dichloride 90b failed. The desired type

of reductive cleavage has been achieved by Paquette on 73

an analogous 1/ 4-dichloride system using sodium in liquid . 86 ammonia.

Attention was then focused on the conversion of 85^ to C^g-hexaquinacene. Reductive deoxygenation by treat­ ment of 85^ with base followed by capture of the bisenolate with diethylchlorophosphate and subsequent reduction with 84 lithium in diethylamine afforded no 24. Reaction of

85 with tosylhydrazine in methanol produced no correspond­ ing bistosylhydrazone, ruling out the possible use of a

Bamford-Stevens reaction to obtain 24. Starting R5 was recovered unchanged after attempted deoxygenation with zinc and trimethylchlorosilane in refluxing tetrahydro- . 87 furan.

Conversion of 85^ to its endo,endo diol 91 was accomplished in 93% yield with lithium aluminum hydride in tetrahydrofuran at the reflux temperature. Treatment of the dimesylate derived from 9JL with bases such as potassium t-butoxide in dimethyl sulfoxide and lithium

2,2/6,6-tetramethylpiperidide gave only 91. Apparently nucleophilic displacement by attack on sulfur was occurring rather than elimination. Preliminary results

6 ' with the reaction of the dimesylate on an alumina surface were encouraging and prompted further investigation. 74

The bulkier tosylates have been found to eliminate more readily under these heterogeneous conditions than 89 mesylates, so the reaction was explored using ditosylate

92. This compound was obtained in 73% yield from treat*, ment of 9jL with p-toluenesulfonyl chloride in pyridine and was purified by recrystallization from ethyl acetate.

Stirring of a solution of 9J2 in dichloromethane with activated neutral alumina at room temperature or at reflux resulted in very slow and incomplete formation of

24. Changing to basic alumina in dichloromethane at room temperature gave largely C^g-hexaquinacene, but column chromatography of the crude reaction mixture on basic alumina afforded 9j2 and the monotosylate. Heating a suspension of basic alumina coated with 9j2 in refluxing toluene gave J24 plus an uncharacterized minor olefinic product. The cleanest reactions occurred when neutral alumina (activity super I) was stirred with a dichloro­ methane solution of 9£ (10 g of alumina per mmol of 92) followed by removal of the solvent. The alumina residue was suspended in dry toluene under a nitrogen atmosphere and the mixture was heated at reflux for a prolonged period. Column chromatography on neutral alumina and 75

LiAlH 4

TH E1, A 0 85

24 recrystallization from ethyl acetate or acetone afforded

C16-hexaquinacene in 41% yield.

Even with the use of highly activated alumina, some hydrolysis occurred to give an epimeric mixture of alcohols. The use of tetrahydrofuran as a solvent was tried, in the hope that it would block some of the acid sites on the alumina and further minimize hydrolysis.

However, almost all of the tosylate was recovered unchanged. Treatment of 8-quinolinesulfonates with alumina has been found to afford higher yields of 89 elimination products. Thus, in a further attempt to improve the yield of 24, diol 91 was treated with 76

8-quinolinesulfonyl chloride in pyridine, but none of the desired derivative was isolated. It is possible that the cavity of the molecule cannot accommodate two sterically bulky quinolinesulfonyl groups. In some instances, a minor olefinic component, which was not characterized, was present in the crude reaction mixture.

This may be 93, the other possible double bond positional isomer, formed by elimination involving removal of the tertiary proton to afford the more highly substituted olefin.

93

C^g-hexaquinacene crystallizes as colorless needles and melts at 231.5- 234°c with some accompanying subli­ mation, in an oil bath preheated to 226°c. If the melting point is determined by slow heating in the presence of air, the needles become deep red in color but retain their integrity, even at 350°c. Apparently oxidation occurs on the surface of the crystals in air at high temperature, since similar heating of j24 in a sealed tube 77

under argon resulted in regular melting behavior. The infrared spectrum of 24 is that of an olefinic hydro­ carbon, with its strongest bands at 3030, 2936, 2900,

1353, 743, and 720 cm’“^. As expected from its symmetry, 24 possesses a marvelously simple 60 MHz

NMR spectrum with all six vinylic hydrogens appearing as a singlet at 6 5.36 while the remaining hydrogens appear as a narrow multiplet at 3.76-3. 20. The four lines of the 13 13 C NMR spectrum and the C-H coupling constants obtained via gated decoupling are listed in Table 4.

TABLE 4

13C NMR SPECTRAL PARAMETERS FOR C^-HEXAQUINACENE

Carbon 13C resonance *3C-H coupling ppm Hz Hz

a 131.567 2646.484 .160.156 b 60.15 2 1209.960 138.184 c 54.908 1104.492 134.278 d 53.064 1067.382 137.676 78

The fully saturated C^g-hexaquinane 9j4 was prepared in 88% yield by catalytic hydrogenation of 2j4 over 10% palladium on charcoal. This colorless, highly crystalline compound possesses a NMR spectrum which consists of two broad multiplets, 6 3.77-2.17 and 2.07-1.13, and a 13 C NMR spectrum which exhibits four different carbon lines.

94

The physical and electronic structural properties of C^g-hexaquinacene can be reasonably compared with

3 those of triquinacene (95) and cis -1,4,7-cyclononatriene

(96). In each of these three trienes the double bonds

95 96 adopt an arrangement which is potentially suitable for 80 effective pp-<* overlap resulting in homoaromatic 79

90 stabilization. Information about the extent of this

overlap can be obtained from a consideration of

photoelectron spectroscopy studies and X-ray crystal

structure data. A preliminary comparison of the vacuum

ultraviolet spectra shows a single absorption maximum without vibrational fine structure at 192 nm (e 20/000)

for 24 in cyclohexane solution while triquinacene 91 exhibits a maximum at 187 nm (e 13/000) in isooctane 9 2 and cyclopentene at 180 nm (e 10/000).

On the basis of models, the spherical geometry of

24 permits a more ideal in-plane alignment of the p-H orbitals than is found in 95_ or 96. The X-ray crystal

structure of C^6-hexaquinacene is shown in Figures 8 and 93 9 together with the important bond distances and angles.

The molecule deviates slightly from C^v symmetry, but the most important observation is that while the central

cyclopentane rings are planar the peripheral cyclopentene

rings are puckered outwards such that the plane of the

C3A-C4A-C4B-C3B olefin fragment makes a 5.4° dihedral

angle with the plane of the C3A-C2-C3B unit. In o 94 triquinacene (95) this same angle is 2.2 . Apparently, the p-TI orbitals are canted inward at an appreciable Figure 8. X-ray crystal structure of C^g-hexaquinacene (24).

CD o 9. View of C16-hexaquinacene (24) down

■three-fold axis. 82

angle, but 11-11 repulsion causes the molecule to spread

apart increasing the intramolecular distance between the

olefinic carbons. 80 An overlap integral (S^), which is based upon the

internuclear distance (R) between the carbon atoms on which the p orbitals are located and the orientation of the two p orbitals relative to each other and to an internuclear vector, indicates the extent of orbital interpenetration. Calculations give an S value of 0.054 for 24, which is identical to that previously determined on for £5 (see Table 5) . 95 . Photoelectron spectral data for 96 indxcates interaction between its II bonds, with a 0.9 eV energy difference between the ionization bands for the e(n) and a^t^) orbitals. In the case of 95, this split has 96 been found to be only 0.35-0.40 eV. This difference between 9!5 and 96 has been explained on the basis of hyperconjugative effects. Similar interactions occur between the sigma framework and e (II) and a^II) in 96, but in 95^ the interaction of a^(Il) with the sigma framework is stronger than that involving e(II). 93 The photoelectron spectrum of 24 (Figure 10). as determined by Gleiter exhibits a single peak separated 83

TABLE 5

pp-CX OVERLAP INTEGRALS ESTIMATED USING

SLATER ORBITALS

Compound R (A) a 0 (deg)a S

C^g-Hexaquinacene (24) 2.848 33.99 0.054

Triquinacene (95) 2.533 59.18 0.054 3 cis -1,4/7-Cyclonona- triene (96) 2.464 54.18 0.066

Biphenyl (sp2-sp2)'b 1.520 0.0 0.687

^ is the internuclear distance between the carbon atoms and 0 is the angle between the p-orbital axis and the vector K. 1^ Biphenyl is included in this table for reference; the other overlaps are consequently on the order of 10% of normal single bond C-C overlaps. Figure 10. The He(I) photoelectron spectrum of spectrum photoelectron He(I) The 10. Figure CPS 8 9 ^-eauncn (24). C^g-hexaquinacene 10 11 2 I.R(eV) 12 84

85

by about 1 eV from strongly overlapping bands. The

energy difference between e (II) and a^dl) is predicted to

be less than 0.5 eV, because of the rather large

internuclear distance between the II units. The first

band in the PE spectrum is therefore assigned to ioni­

zation from both e (II) and a^(II). Molecular orbital

calculations of both the MINDO/3 and extended Huckel

type predict to be of higher energy than e (II) in

this case, due to the stronger interaction of a^(II) with

the sigma framework. This orbital ordering is a function

of the distance between the II bonds under consideration

(Figure 11). As the correlation of ionization energies

due to 24, 9 5, and 96 in Figure 12 shows, hyperconjugation

with the sigma framework in 24_ predominates over homo­

conjugation. As a chemical consequence of this lack of

homoaromatic character, C^g-hexaquinacene did not react when attempts were made at reduction with potassium in

liquid ammonia or oxidation with CotAz)^ in a flow- 97 through system.

An X-ray crystal structure of C^g-hexaquinane (94) 98 is illustrated in Figures 13 and 14. It can be seen that the molecule adopts a twisted configuration in order 86

12 3 II I -11

E (e v ) I

-12

-13 2 3 4 — A

Figure 11. Orbital energy plot showing the variations

in e(ID and ^(11) levels as a function of

distance between interacting II bonds. Figure 12. Correlation of the ionization energies of

24_, 95, and 96.

to obtain the optimal van der Waals radii between the three hydrogens labelled H(l), H{2), and H(3). The three central cyclopentane rings are each almost planar, while the three peripheral rings adopt half chair conformations.

The availability of dimethyl diketone 73 prompted

interest in the potential for five-membered ring formation between the ketone units and the double bonds. Epoxide formation at the double bonds would provide electrophilic sites for attack by the enolates of the ketone functions 88

Figure 13. X-ray crystal structure of C^g-hexaquinane

(94) . 89

H( I)

H(a

H(2)

Figure 14. Depiction of the intramolecular

hydrogen-hydrogen distances in 94. 90 to form two additional cyclopentanone rings in the manner shown. Further elaboration of 98^ could produce molecules of use in the construction of the dodecahedrane

framework.

mCPBA base OH

tteCHi ]C0CH c o c h 3 COCH, 98

Attempts to form bisepoxide 97 by treatment of a dichloromethane solution of 73^ with two equivalents of m-chloroperbenzoic acid (mCPBA) resulted in isolation of no bisepoxide, after washing with sodium bisulfite and sodium bicarbonate solutions. Analytical thin layer chromatography (tic) of the reaction mixture before and after workup indicated that the product was undergoing further reaction during isolation. If the reaction mixture was placed directly on a silica gel column, without undergoing aqueous extraction, elution with ethyl acetate removed a white solid whose infrared spectrum exhibits two hydroxyl absorption bands and no carbonyl band. The assigned structure 99^ possesses a symmetry axis which is consistent with the eight different carbon 91

OH H

'3’ OH

99

13 13 lines of the C NMR spectrum. The off-resonance C

NMR spectrum contains two singlets, five doublets, and one quartet and one of the quarternary carbons has a chemical shift of 106 ppm, which is consistent with a hemiketal. The mass spectrum is correct for 99 less one molecule of water. The m-chlorobenzoic acid formed during the course of the epoxidation reaction or the acidic sites on the silica gel must have promoted protonation of bisepoxide 97_ and subsequent ring opening via the oxygen of the ketone. This type of reaction has been observed previously in the conversion of 100 to 101 99 under acidic conditions. COCH

100 101 92

Treatment of 7^ with m-chloroperbenzoic acid in a two-phase medium of dichloromethane and 1 N aqueous sodium bicarbonate"1-00 resulted in the successful formation of 97 in 97% yield. Due to the diminished activity of the mCPBA in this biphasic medium, the reaction required a longer time period and excess mCPBA. When 9J7 was treated with four equivalents of the hindered lithium

2,2,6,6-tetramethylpiperidide, mainly starting bisepoxide plus a small amount of an alcohol were isolated. Reaction of 97_ with lithium diisopropylamide at -78° followed by warming of the reaction mixture to room temperature and inverse addition to saturated ammonium chloride solution afforded a crystalline compound whose infrared spectrum exhibits a hydroxyl absorption and only a trace of ketonic carbonyl absorption. The mass spectrum indicates the new compound to be isomeric with 97, but further characterization was precluded because of decomposition upon recrystallization. The structure was tentatively assigned as vinyl ether 102. Consideration of molecular models of 97^ shows that it is impossible in this case to properly align the orbital of the a-carbon atom of the planar enolate ion for the backside attack on the epoxide 93

OH HO

102

desired in an SN2 displacement. Thus, the availability 101 of the lone pairs on oxygen leads to 0-alkylation.

Cleavage of the central norbornyl bond in diketone

73 would allow the molecule to adopt a better orientation

for successful C-alkylation of the ketone enolate. This

carbon-carbon bond cleavage was accomplished by treatment

of 73_ with trimethylchlorosilane and dispersed sodium 14 metal in refluxing toluene. Epimerically pure 103 was obtained in 62% yield after chromatography on silica

gel and recrystallization from ethyl acetate. The C ^

symmetry of 103 is corroborated by the presence of six 13 lines in its C NMR spectrum and the appearance of a

singlet for the olefinic protons m the H NMR spectrum.

Diketone 103 was converted into its bisepoxide 104 in 1) Na. mCPBA

C0CH

104 94

74% yield with mCPBA in the two-phase reaction medium

described previously. This reaction occurred more

slowly than the epoxidation of 73^ to 97. Monitoring of

the reaction by tic showed that the initial transfor­

mation to monoepoxide was very rapid, but the second

epoxidation required considerable time. Treatment of

104 with four equivalents of lithium diisopropylamide

at -78° initially and then at room temperature afforded

mainly starting 104. Another cyclization attempt

involved formation of the bisenolate at -78°, followed

by heating of the reaction mixture at the reflux

temperature of tetrahydrofuran overnight. A large

amount of solid precipitated, but after workup only

starting bisepoxide was isolated.

Although the enolate anion was formed under the

conditions of kinetic control, abstraction of the ct- methine proton may still have occurred to form the more

stable enolate. This thermodynamically-favored enolate

anion would be unable to react with the epoxide ring and upon quenching would revert back to 104, although some

epimerization might take place. However, if the desired enolate did form, the molecular geometry must have again 95

prevented nucleophilic attack on the epoxide. The bisepoxysultone 105 has been prepared by Hales and upon treatment with sodium hydride has successfully cyclized 52 to afford 106. In this case, the anion formed could align itself properly for nucleophilic displacement of the epoxide since it was not constrained as were the enolates of 97 and 104 to maintain a stabilizing orbital interaction with its adjacent carbon.

HO' NaH DMSO

r so S0o so 106

The ability of C^g-hexaquinacene to form metal complexes involving coordination with all three double bonds is of great interest. A useful analogy is the 3 QOb behavior of cis -1/ 5/ 9-cyclododecatriene 107 in the presence of metal complexes. This compound approximates the peripheral framework of the hexaquinacene structure, but itself does not exist in the same type of crown conformation.®^*5 However, complexes of 107 with cuprous 96

/_\ / \

107 108

102 103 triflate and nickel (O) do exist in conformation

108. An initial complexation attempted by treatment of 104 24 with a molybdenum (diglyme) tricarbonyl complex in benzene at room temperature resulted in isolation of only starting C^g-hexaquinacene. No complexation occurred with rhodium 1,5-cyclooctadiene chloride dimer in benzene at room temperature or at 90° for a prolonged period.

No further attempts to form metal complexes were made, due to our limited expertise and equipment in this area.

It is possible that the interior cavity of triene 2_4 is rather inaccessible to approach by metal atoms.

Triquinacene (95), a similar triene whose concave surface is more open than that of 24, is also reluctant to form 105 metal complexes. The use of metal vaporization 1 06 techniques was considered, but early work with cyclo- 107 nonatriene 96 proved unsuccessful.

The next area studied was the thermal stability and reactions of C^g-hexaquinacene (24). Pyrolysis of 97

triquinacene (95) above 600° results in clean conversion 108 to known compounds (Scheme VI). Vacuum pyrolysis of

24 was carried out by slow vaporization through a hot quartz tube and collection of the product on a cold finger. This apparatus (illustrated in Figure 15) minimized the possibility of secondary reactions occurring at the high temperatures used. Up to 700° at a pressure of 0.1 mm of Hg, C^g-hexaquinacene was recovered unchanged, but at 800° a yellow oil appeared together with starting 24. The NMR spectrum of the product

Scheme VI

600 650 4>

750 700°

750° 98

■ 5.75

Quartzn

u

Figure 15. Vacuum pyrolysis apparatus 99

exhibits a broad aromatic absorption and its mass spectrum indicates loss of two hydrogen atoms from 24.

By analytical thin layer chromatography it appears that more than one new compound is present. Repeated attempts to achieve a clean/ high percentage conversion of hexaquinacene to product by slightly altering the temperature and pressure conditions were disappointing.

Apparently, at the high temperatures necessary to achieve the initial reaction, secondary reactions or total destruction were unavoidable. The experiments were also conducted close to the temperature limits of the apparatus' used. Exact identification of the pyrolysis products was therefore not pursued.

Photolysis of a triene system whose double bonds lie in close proximity, as in 24, should lead to unique molecules. Irradiation of a dilute pentane solution of triquinacene (95) for 30-50 hr resulted in about 45% conversion to a mixture of eight products, the two major ones of which were 109 (26%) and 110 (49%).^^ When a _3 1 x 10 M solution of C^g-hexaquinacene in purified pentane was irradiated through Vycor with a medium 100

95

109 110 pressure 450 W Hanovia light source for 33.5 hr under nitrogen a new compound was produced in 22% yield together with a 32% recovery of unchanged 24. The components of the reaction mixture were separated by preparative thin layer chromatography on silica gel

(hexane elution). The photoproduct is a waxy solid whose

NMR spectrum exhibits a single olefin peak at 6 5.63

(2H) plus a broad multiplet between fi 3.57 and 2.13 (14H).

Its mass spectrum shows it to be isomeric with 24 and only nine lines, one with an olefinic chemical shift, 13 are present in the C NMR spectrum. Thus, the photo­ product was assigned structure 111 which would arise

24 111 from an intramolecular [2+2] cycloaddition and possess a plane of symmetry. Further purification of 111 was 101

effected by column chromatography on 10% silver nitrate- silica gel with 20% ether in hexane followed by sublimation.

Monitoring the progress of the photolysis by tic indicated that the formation of 111 occurred extremely slowly after the first 12 hr of irradiation. If a pentane solution of 2j4 was irradiated through quartz with a 200 W Hanovia lamp, all of the starting material was consumed, but in addition to 111 at least two minor products were present together with a large amount of undesirable material. In an unsuccessful attempt to promote [2+2] cycloaddition, 5% acetone in pentane was used as a solvent. At the present time, no evidence has been found for the formation of 112, the product of intramolecular [2+2+2] cycloaddition.

112

Homologation of the C^6-hexaquinacene structure by reaction of a borane reagent with the double bonds presents a possible means of enclosing the molecular 102

cavity (Scheme VII) . Similar treatment of trans, trans, trans-1, 5,9-cyclododecatriene 113 with borane and tri- ethylamine in diglyme at high temperatures resulted in 110 the formation of 114. Treatment of 24_ with disiamyl- borane in tetrahydrofuran followed by workup with 111 aqueous sodium hydroxide and hydrogen peroxide afforded a mixture of triols 115 and 116 in 70% yield.

BH 3 .N(C2 H 5 ) 3 ,,. » diglyme, A

113 114

Scheme VII

1) (Hr) 2BH, THF 2) NaOH, H20 ■ohL HO""' 3) H 20 2 ’""OH

24 115 116

117 103

13 The C NMR spectrum of this mixture contains 13 lines, even though the symmetrical triol 116 contains six different carbons and the unsymmetrical triol 115 contains 16 different carbons. The extended hydrocarbon framework probably results in only small differences in carbon shift between 115 and 116.

Previous work on the hydroboration of triquinacene

(95) with diborane followed by oxidative workup afforded 112 a mixture of triols 118 and 119 in 71% yield. Unlike

115 and 116, these triols were extremely soluble in water, which helped in their purification, and could be separated by high pressure liquid chromatography on silica gel 13 eluting with acetone. The C NMR spectrum of the mixture of 118 and 119 exhibits 14 resonance lines. If one permits the assumption that a rough estimate of the relative amounts of the two isomers present can be

1 ) b h 3 - t h f

95 118 119 104

obtained by a comparison of the areas of lines for carbons in similar environments, hydroboration with diborane produced a mixture consisting of 37% of the

symmetrical triol 119 and 63% of the unsymmetrical triol

118. The use of disiamylborane produced a triol mixture containing 41% of 119. Statistically a 75% to 25% ratio of 118 to 119 would be expected.

Similar analysis of the triol mixture obtained from hydroboration of C^g-hexaquinacene showed it to contain

28% of the symmetrical triol 116. This nearly statistical ratio is surprising, since the double bonds in 2^4 lie closer together than those in 95_ and thus hydroboration with a bulky reagent should favor the symmetrical isomer

116. The addition of borane to olefins is reversible 113 upon heating, so that heating the intermediate borane formed in this case should afford the less sterically hindered intermediate borane which upon oxidative workup gives 116. In practice, treatment of 24 in diglyme with 9-borabicyclononane followed by heating at

148° and oxidative workup did not significantly alter 13 the isomer ratio as observed by C NMR.

Analytical tic of the mixture of 115 and 116 was discouraging for separation by chromatographic means. 105

Treatment of the triol mixture with benzoyl chloride and pyridine afforded a mixture of benzoates in 94% yield which could be separated by preparative tic on silica gel with 15% ether in dichloromethane elution. After chromatography, 19% of the mixture was the symmetrical tribenzoate 120 (R^ = 0.78) which is a crystalline solid 13 with 10 lines in its C NMR spectrum. Unsymmetrical tribenzoate 121 (R^ = 0.85, 81%) was isolated as a semi- 13 solid and its C NMR spectrum exhibits 23 lines as

0CO',,lt

expected. The NMR spectra of 120 and 121 are quite similar, with that of 120 exhibiting greater fine structure. Treatment of the benzoates with potassium hydroxide in methanol and water at the reflux temperature afforded the separated triols.

Attempts to convert the triols to their corresponding triketones were unsuccessful. Triol 116 reacted with

Jones reagent to afford a mixture, which decomposed when 106 an attempt was made to separate it by preparative tic and no triketone 117 was isolated. Oxidation of 115 with

Jones reagent produced a solid whose infrared spectrum indicates that hydroxyl functions are still present.

Even prolonged oxidation did not remove the hydroxyl absorption band from the infrared spectrum. The desired triketone 121 may have formed, but in the presence of aqueous acid undergone intramolecular cyclization to afford the hemiketal 122. Since the starting alcohol

115 has its hydroxyl groups situated in the exo orientation,

121 is the necessary precursor to 122. No further work was done on this reaction sequence because of the extremely low yield of the desired triol 116 coupled with the value of its precursor 24.

•"‘•OH

0 OH 121 122

Many facets of the chemistry of C^g-hexaquinacene remain to be explored. Homologation of its diketone precursor 85 in the hope of preparing dodecahedrane is 107

already under active investigation. A considerable amount of success has already been realized in this area.

The solvolysis of suitable derivatives of 2j4 would provide a chemical assessment of the interaction between the olefinic moieties and a carbonium ion. More work on the chemistry of this unusual basket-shaped molecule will doubtless be forthcoming.

» EXPERIMENTAL

Melting points were determined in open capillaries

with a Thomas-Hoover apparatus and are uncorrected

unless otherwise stated. Proton magnetic resonance

spectra were recorded with Varian A-60A, Varian EM-360,

Varian HA-100, and Bruker HX-90 spectrometers, while

carbon magnetic resonance spectra were obtained with the

Bruker 90 instrument. Apparent splittings are given in

all cases. The splitting patterns, which were observed

in the single-frequency off-resonance decoupled (SFORD)

carbon magnetic resonance spectra, are designated as

s, quarternary carbon; d, methine carbon; t, methylene

carbon; q, methyl carbon. Infrared spectra were

determined on a Perkin-Elmer Model 467 instrument. Mass

spectra were recorded on an AEI-MS9 spectrometer at an

ionization potential of 70 eV. Elemental analyses were

performed by the Scandinavian Microanalytical Laboratory,

Herlev, Denmark.

In experiments requiring dry solvents, dioxane,

ether, and tetrahydrofuran were distilled from sodium-

benzophenone. Hexamethylphosphoramide, dimethyl

108 109

sulfoxide, dimethyl formamide, toluene, and benzene were

distilled from calcium hydride. Dichloromethane was

distilled from phosphorous pentoxide, methanol from magnesium methylate, and pyridine from potassium hydroxide. Organic solutions were dried over anhydrous

sodium sulfate unless otherwise stated.

2a,3,3a,5a.6,6a,6b,6c-Octahydro-3,6-ethanodicvclopenta

f cd,qh]pentalene-7,8-dione (28) . Trimethylchlorosilane

(20 ml, 154 mmol) was added to a

sodium dispersion (3.28 g, 143

mg-at) in dry toluene (100 ml) in

a 500-ml, three-necked flask

equipped with nitrogen inlet,

condenser, Hershberg stirrer, and

addition funnel. A solution of 33^ (1.00 g, 3.64 mmol)

in 20 ml of dry toluene was added dropwise to the stirred

reaction mixture which was then heated at reflux for

13 hr. The excess sodium was separated by filtration through a pad of Celite which was washed with toluene,

and the filtrate was concentrated. The residue dissolved

in dry toluene (25 ml) was added dropwise under nitrogen to a stirred solution of anhydrous ferric chloride

(1.78 g, 11.0 mmol) in 50 ml of dry ether containing

10 drops of concentrated hydrochloric acid. The resulting 110

dark green mixture was gently refluxed for 1 hr, cooled, and treated with 100 ml of saturated aqueous ammonium sulfate solution. The layers were separated and the aqueous phase was extracted with methylene chloride

(3 x 100 ml). The combined organic layers were washed with dilute sodium bicarbonate (80 ml) and saturated sodium chloride solutions (80 ml), and then dried.

Concentration afforded a gold solid which was purified by chromatography on silica gel (elution with 20% ether in hexane) to yield 0.50 g (65%) of 28. Recrystallization from ethyl acetate-hexane gave yellow needles, mp 246°

(dec); 1720, 1710, 1696, and 1691 cm-1 ? h i NMR (6, jTIcIX CDC13) 5.65 (d, J = 1 Hz, 4H) , 3.83-3.23 (br m, 6H) , and

3.13-2.70 (br m, 2H); 13C NMR (ppm, CDC13) 195.96, 132.36,

57.29, 56.38, and 51.34? m/e calcd 212.0837, obs 212.0840. i Anal. Calcd for c^4H i2°2: C/ ^9.22; H, 5.70.

Found: C, 79.24; H, 5.9 4.

2a,3,3a,5a,6,6a,6b,6c-Octahvdro-8-hydroxv-3,6-ethano- dicyclopenta[cd,gh]pentalene-7-one (35). a sodium

dispersion (785 mg, 34.1 mg-at)

was formed in 25 ml of dry toluene

in a 100-ml three-necked flask Ill

equipped with nitrogen inlet, condenser, Hershberg

stirrer, and addition funnel. The Hershberg stirrer was replaced with a magnetic stirring bar and trimethyl- chlorosilane (5.54 ml, 43.6 mmol) added through the condenser. A solution of (239.2 mg, 0.872 mmol) in

8 ml of dry toluene was added dropwise and the resultant mixture was heated at reflux for 16 hr. The excess sodium was removed by filtration through a Celite pad and the filtrate was concentated. Dry methanol (20 ml) was added dropwise under nitrogen to the residual yellow oil and the resulting solution was heated at reflux for

1 hr. After removal of the methanol in vacuo, the residue was dissolved in 30 ml of methylene chloride and washed with water (2 x 30 ml). The combined water layers were extracted with additional methylene chloride (30 ml) and the organic layers were washed with saturated sodium chloride solution (30 ml). After drying, concentration gave 223.7 mg (> 100%) of 35_ as a gold oil, NMR (6,

CDC13) 6.05-5.35 (m, 4H) , 4.05 (d, J = 3 Hz,' 1H), 3.75-

2.80 (m,8H), and 2.37 (br s, 1H).

2a,3,3a# 5a,6, 6a/6b,6c-0ctahydro-3,6-ethanodicvclopenta-

[cd,qh]pentalene-7,8-dione (28). Crude 35^ (0.872 mmol) 112

was dissolved in dry dimethyl sulfoxide (3.10 ml, 43.7 mmol) under nitrogen and acetic anhydride (2.06 ml,

21.8 mmol) added via syringe. The reaction mixture was stirred at room temperature for 16 hr then poured into

30 ml of ice-water. This was extracted with ether

(4 x 20 ml) and the combined organic layers were washed with 10% sodium bicarbonate solution (2 x 30 ml), water

(4 x 30 ml), and saturated sodium chloride solution

(30 ml). After drying, concentration gave 209.1 mg of yellow needles which were purified by chromatography on silica gel (elution with 20-40% ether in hexane) to afford 92 mg (50% based on 33) Qf 28.

Dimethyl Octahydro-3, 4, 7-metheno-lH-cyclopenta[a]pentalene-

7, 8 (7aH)-dicarboxylate (133) . solution of 11^ (1.00 g,

3.67 mmole) in 60 ml of ethyl

Q00CH3 acetate (distilled from potassium COOCH3 carbonate) in a Parr hydrogenation

bottle was degassed with nitrogen

and 10% palladium on carbon (200

mg) was added. The mixture was hydrogenated on the Parr apparatus at 50 psig for 16 hr.

The catalyst was removed by filtration through a pad of

Celite and rinsed with 50 ml of methylene chloride. 113

Concentration of the filtrate afforded 1.08 g of

colorless solid which was recrystallized from hexane to

give 0.9 3 g (92%) of 1 2 3 as white needles, mp 139-139.5°

(corr); 1720 cm- 1 ; h i NMR (6, CDC10) 3.67(s,6H), ITiaX J 2.47(m, 6H), and 1.68(m,8H); l3C NMR (ppm, CDC13) 172.33,

60.37, 58.00, 51.15, 47.53, and 22.55; m/e calcd

276.1361, obs 276.1366.

Anal. Calcd for <-^g^20<34: ^9.54; 7.30. Found: C, 69.36; H, 7.20.

Dimethyl Decahydro-2,4,5-metheno-lH-cyclopropa[a]cyclo- propa[3,4]cyclopenta[l,2-e]pentalene-2,6 (laH)-dicarboxylate

(37 ). a 1M solution of ethylzinc iodide in ether (27.0

ml, 27.0 mmol) in a 100-ml three-

COOCHg necked flask equipped with nitrogen

inlet, condenser, rubber septum,

and magnetic stirring bar was

treated via syringe with methylene

iodide (2.2 ml, 27.0 mmol). The resulting solution was heated at reflux for 30 min prior to dropwise addition of 11^ (500 mg, 1.84 mmol) in dry ether (5 ml) at room temperature. The reaction mixture was heated at reflux for 18 hr, then poured into cold I saturated ammonium chloride solution (70 ml). The ether

layer was separated and the aqueous phase was extracted 114

with ether (40 ml). The combined ether layers were washed with saturated ammonium chloride solution, dried,

and concentrated to give 710.3 mg of oily yellow solid.

Chromatography on silica gel (elution with 20% ether in

hexane) afforded 523.4 mg of white solid which was

recrystallized from hexane to yield 394.7 mg (72%) of

37, mp 123-123.5° (corr); vKBr 1741 and 1722 cm-1? 1H — max NMR (6, CDC13) 3.48 (s, 6H), 2.43(m, 4H), 1.57(m, 2H),

1.25-0.88(br m, 4H), and 0.17-(-0.33)(br m, 4H)? 13C NMR

(ppm, CDC13) 172.40, 64.52, 57.34, 51.22, 35.34, 10.49,

and 2.38? m/e calcd 300.1361, obs 300.1367.

Anal. Calcd for c iqH 20°4i C/ 6.71. Pound: C, 71.84? H, 6.62.

Dimethyl Dodecahvdrodicyclopenta fcd.ahlpentalene-3,6-

dicarboxylate (36) A fine sodium dispersion (821 m

35.7 mg-at) was formed in 25 ml of

C00CH3 dry toluene in a 100-ml, three­ c o |o c h 3 ' necked flask equipped with nitrogen

inlet, condenser, Hershberg

stirrer, and addition funnel.

Trimethylchlorosilane (6.0 ml,

47.3 mmol) was added through the condenser and the

reaction mixture was stirred for 10 min prior to the 115

addition of a solution of 123 (250 mg, 0.911 mmol) in

5 ml of dry toluene over a 30 min period. The resulting mixture was heated at reflux for 8 hr, then the excess

sodium was separated by filtration through a Celite pad,

followed by a toluene rinse. The filtrate was concen­

trated in vacuo and the residue, as a solution in 5 ml of

dry toluene, was added dropwise to dry methanol (25 ml) with stirring under nitrogen. After completion of the

addition the reaction mixture was stirred for 30 min before 100 ml of water was added. The layers were

separated and the water layer was extracted with ether

(3 x 50 ml). The combined organic layers were washed with water (1 x 25 ml) and saturated sodium chloride

solution (1 x 25 ml) prior to drying. Concentration gave 351.6 mg of yellow oil which was chromatographed on silica gel (elution with 2 0 % ether in hexane) to yield 156.6 mg (62%) of 36. Recrystallization from hexane afforded white crystals, mp 94.5-95.5° (corr)? vKBr 1734 and 1190 cm-1; 1H NMR (6 , C D C l J 3.67(s,6H), max 3

3.47-2.53 (br m, 8 H), and 2.10-1.58(br m, 8 H) ; l3C NMR

(ppm, CDClg) 173.95, 56.61, 54.47, 51.02, 46.36, and

30.49? m/e calcd 278.1518, obs 278.1524. 116

Anal. Calcd for C]_6**22°4: ^9.04? H, 7.97. Pound: C, 69.11? H, 7.94.

This same diester (36) Was also prepared by catalytic hydrogenation of 33^ (50 mg, 0.182 mmol) in ethyl acetate (5 ml) containing 10% palladium on carbon

(20 mg) at 50 psig of hydrogen. After 13 hr the catalyst was removed by filtration through a Celite pad and the filtrate was concentrated to yield 49.7 mg

(98%) of 36.

Dimethyl Tetradecahydrobiscyclopropa[4,5]cyclopenta-

[l,2,3-cd: 1 1,21,31-qh]pentalene-2,4-dicarboxylate (38).

A fine sodium dispersion (602.6

COOCH 3 m g ' 2 6 , 2 was formed in C00CH 20 ml of dry toluene in a 50 ml,

three-necked flask equipped with

nitrogen inlet, condenser,

Hershberg stirrer, and addition funnel. Trimethylchlorosilane (4.5 ml, 34.8 mmol) was added through the condenser and the resulting mixture was stirred for 1 0 min prior to the dropwise addition of

37 (200 mg, 0.67 mmole) in 4 ml of dry toluene. After completion of the addition, the mixture was heated at 117

reflux for 7 hr and then the sodium was removed by- filtration through a Celite pad, followed by a toluene rinse. The filtrate was concentrated and the resulting light yellow oil in 5 ml of toluene was added dropwise to dry methanol (20 ml) under nitrogen. The reaction mixture was stirred for 30 min, then diluted with 80 ml of water. The product was extracted with ether ( 3 x 40 ml) and the combined organic layers were washed with water ( 1 x 2 0 ml) and saturated sodium chloride solution

(1 x 20 ml) before drying. Concentration gave 283.7 mg of light yellow oil which was chromatographed on silica gel (elution with 2 0 % ether in hexane) to afford 181.1 mg

(89.6%) of 38, Recrystallization from hexane gave white crystals, mp 128-129° (corr) ; 1735 c m ” '1'; NMR IrlclX

(6 , CDC13) 3.63(s, 6 H), 3.13-2.35(br m, 8 H), 1.58-1.22

(m, 4H), 1.02-0.55(br m, 2H), and -0.12-(-0.42)(m, 2H)?

13C NMR (ppm, CDC13) 173.67, 60.26, 55.11, 51.42, 50.55,

24.53, and 17.73; m/e calcd 302.1518, obs 302.1522.

Anal. Calcd for ci8**22^4: C/ 71.50; H, 7.34. Found: C, 71.51; H, 7.43. 118

Dodecahydro-3,6-ethanodicyclopenta[cd,gh]pentalene-7,8 - dione (22). To a sodium dispersion (324.3 mg, 14.1

mg-at) in 1 2 ml of dry toluene in

a 50 ml three-necked flask

equipped with a nitrogen inlet,

Hershberg stirrer, condenser,

and addition funnel was added

trimethylchlorosilane (2.28 ml,

18.0 mmol) . A solution of 36^

(100 mg, 0.360 mmol) in 2 ml of dry toluene was added dropwise and the stirred reaction mixture was heated at reflux for 13.5 hr. The excess sodium was removed by filtration through Celite and the filtrate was concentrated.

A solution of the residue in 2 ml of dry toluene was added dropwise under nitrogen to a stirred solution of anhydrous ferric chloride (175.2 mg, 1.08 mmol) in 5 ml of dry ether containing 6 drops of concentrated hydro­ chloric acid. The resulting dark green mixture was heated at reflux for 1 hr, cooled, and treated with 1 0 ml of saturated aqueous ammonium sulfate solution. The aqueous phase was washed with methylene chloride (3 x 10 ml) and the combined organic layers were washed with 119

dilute sodium bicarbonate solution ( 1 0 ml) and saturated sodium chloride solution (10 ml). After drying, concentration afforded 150.1 mg of oily yellow solid which was chromatographed on silica gel (elution with

20% ether in hexane) to yield 42.3 mg (55%) of 22 -

Recrystallization from hexane gave yellow needles, mp 181-182.5° (corr); 1701 cm-1 ? h i NMR (5, CDCl_) l U Q A O

2.98(br m, 8 H) , and 2.17-1.37(br m, 8 H) ; ^"3C NMR (ppm,

CDC13) 201.11, 60.50, 56.20, 51.89, and 28.34; m/e calcd

216.1150, obs 216.1154.

Anal. Calcd for C.^H^gC>2 : C, 77.74; H, 7.46.

Found: C, 77.62; H, 7.56.

Tetradecahydro-2, 4-ethanobiscyclopropa[4, 5]cyclopenta-

[1,2,3-cd:1 1,2',3 1-qh]pentalene-5, 6 -dione (30) . a sodium

dispersion (299 mg, 13.0 mg-at)

was formed in 1 2 ml of dry toluene

in a 50-ml, three-necked flask

equipped with nitrogen inlet,

condenser, Hershberg stirrer, and

addition funnel. Trimethyl- chlorosilane (2 . 1 1 ml, 16.6 mmol) was added through the condenser and the resulting mixture was stirred for

1 0 min prior to the dropwise addition of 38 ( 1 0 0 mg, 120

0.331 mmol) in 2 ml of dry toluene. After the reaction mixture was heated at reflux for 12 hr., filtration through a Celite pad removed the excess sodium and the filtrate was concentrated. A solution of the residue in

2 ml of dry toluene was added dropwise under nitrogen to a stirred solution of anhydrous ferric chloride

(161.1 mg, 0.993 mmol) in 5 ml of dry ether containing

6 drops of concentrated hydrochloric acid. The resulting dark green mixture was gently heated at reflux for 1 hr, cooled, and 10 ml of saturated aqueous ammonium sulfate solution was added. The aqueous layer was extracted with methylene chloride (3 x 10 ml) and the combined organic layers were washed with dilute sodium bicarbonate

(10 ml) and saturated sodium chloride solutions (10 ml).

After drying, concentration afforded 131.4 mg of yellow solid which was purified by chromatography on silica gel

(elution with 20-40% ether in hexane) to give 55.9 mg

(70%) of 30. Recrystallization from ethyl acetate gave o KR t" yellow crystals, mp 260.5-262 (corr); v 1720 and 1 1 max 1703 cm ; 1H NMR (6, CDCl3) 3.35-2.60 (br m, 8H) , 1.25-

0.95 {m, 4H) , and 0.65-(-0.02) (m,4H); 13C NMR (ppm, CDCl3)

198.14, 60.08, 52.13, 48.97, 19.12, and 10.01; m/e calcd

240.1150, obs 240.1155. 121

Anal. Calcd for ^i6^16^2: ^9.97; 6*71. Found: C, 79.79? H, 6.75.

1/ 2/ 2a^ 3/ 3a^ 5a, 6, 6a/ 6b, 6c-Decahydro-3, 6-ethanodicyclo-

pentaCcd, qh]pentalene-7,8-dione ( 29) . A solution of _28

(400 mg, 1.89 mmol) in 40 ml of

dry methanol containing potassium

azodicarboxylate (399 mg# 2.05

mmol) was treated dropwise under

nitrogen at -78° with glacial

acetic acid (400 jj.1, 7.00 mmol).

After completion of the addition,

the mixture was stirred for 30 min at -78° and for 3 hr at

room temperature. Additional potassium azodicarboxylate

(160 mg, 0.82 mmol) and glacial acetic acid (6 drops) were added at 0° and the mixture was maintained at this

temperature for 12 hr then diluted with 50 ml of water

and extracted with methylene chloride (2 x 50 ml). The

combined organic layers were washed with 10% aqueous

sodium bicarbonate solution (50 ml) and water (50 ml) prior to drying and concentration. Preparative thin

layer chromatography of the residue (457 mg) on silica gel (four elutions with 30% ethyl acetate in hexane 122

afforded 34.4 mg of _22 (r ^ = 0.23)/ 40.3 mg of starting

28 (r ^ = 0.39)/ and 104.3 mg (26%) of the desired ^

(R_£ = 0.31). Recrystallization from hexane yielded yellow plates, mp 182.5-184° (corr); 2941, 1720,

1710, 1704, and 1685 cm"1; 1H NMR (6, CDCl3) 5.64(d,

J = 1 Hz, 2H), 3.51(m, 2H), 3.09{m, 6H), and 2.38-1.13

(br m, 4H); 13C NMR (ppm, CDC13) 197.79, 130.89, 58.50,

57.63, 55.68, 53.99, 53.11, and 27.09; m/e calcd

214.0994, observed 214.0997.

Anal. Calcd. for c i4**i4°2: C ' 78.48; H, 6.58.

Found: C, 78.79; H, 6.78.

2a, 3, 3a, 5a, 6, 6a, 6b, 6c-0ctahydrodicyclopenta[cd,_gh]- pentalene-3, 6-exo, exo-dicarboxylic acid ( 5_3) . To a

solution of 33^ (50 mg, 0.182

H mmol) in 4 ml of dry hexamethyl-

phosphoramide was added sodium

cyanide (142 mg, 2.9 mmol)

under a nitrogen atmosphere.

The reaction mixture was heated

at 73° with stirring for 25 hr and the resulting cloudy yellow solution was poured into

40 ml of water and extracted with ether (10 ml) to 123

remove unreacted 33^ The water layer was acidified with

10% hydrochloric acid, extracted with ether (2 x 10 ml),

and the combined ether layers were washed with water

(1 x 60 ml, 1 x 30 ml) and brine (30 m l ) . After drying,

concentration afforded 37.2 mg (83%) of ^ which was

recrystallized from ethyl acetate/hexane, mp 215-220° dec;

VKB5 1730, 1703, and 1697 cm"1; 1H NMR (6, d^-DMSO) 5.60- max ~~t) 5.17(m, 4H), 3.52-3.20(m, 6H) , and 2.85-2.70(m, 2H);

m/e calcd 246.0892, obs 246.0897.

2a, 3, 3a, 5a, 6, 6a, 6b, 6c-0ctahydrodicvclopenta[cd,qhDpentalene-

3,6-endo, endo-dicarboxylic acid (49). Diester 33^ (15.30 g,

55.77 mmol, filtered through a

Q00H short column of silica gel with C00H 20% ethyl acetate in hexane)

was dissolved in 250 ml of dry

chloroform (passed down a

column of grade I alumina prior

to distillation from phosphorus pentoxide) in a flame-dried flask protected from light, and trimethylsilyl iodide (1.42 ml, 10.92 mmol, distilled

from copper bronze and stored over copper) was added via

syringe. The pale pink solution was heated at 70° for 124

70 hr under nitrogen and then concentrated to afford a

tan solid. The solid was dissolved in 300 ml of ether

and was extracted with 10% aqueous sodium hydroxide

solution (3 x 100 ml) and washed with brine. The combined

aqueous extracts were acidified with concentrated hydro­

chloric acid (90 ml) and then were extracted with 1:1

ethyl acetate-ether (1 x 500 ml, 2 x 250 ml). The combined organic layers were washed with water and brine prior to drying. Concentration afforded 13.51 g (98%) of 49^ which was recrystallized from ethyl acetate, mp 251-253.5° dec;

3046, 2956, 2931, 2885, 1708, 1409, 1266, 1217, 758, l U O A and 745 c m “ \- NMR (6, d^-acetone) 11.83(br s, 2H),

5.62(s, 4 H ) , and 3.58-2.88 (m, 8H) ,* m/e calcd. 246.0892, obs. 246.0897.

Dimethyl 1, 2, 3, 3"j3, 4

3, 6-dioxodicvclopent[b, hi -as-i ndacene- 4B, 5(3-dicarboxyl ate

(57). A sodium dispersion (300 mg, 13.0 mg-at) was formed

H in 10 ml of dry toluene contained in a 50-ml three-necked flask fitted

/H 0 with a condenser, nitrogen inlet, 'COOCH and Hershberg stirrer. Upon cooling to room temperature without stirring, trimethylchlorosilane

(3.5 ml, 26.9 mmol) was added through the condenser 125

followed by dropwise addition of a solution of ^5 (100 mg,

0.36 mmol) in 1 ml of dry toluene. The resulting yellow

mixture was heated at reflux with stirring for 4 hr,

cooled to room temperature, and filtered through Celite

into 15 ml of anhydrous methanol. After concentration of

the filtrate under reduced pressure, the residue was

dissolved in dichloromethane (50 ml), washed with aqueous

sodium bicarbonate solution, and dried. Concentration

afforded 70 mg (50.6%) of 57 t which was recrystallized

from ethyl acetate, mp 215° dec; 1740, 1695, and

1645 cm-1; 1H NMR (6, CDC13) 3.74 (s, 3H), 3.53(s, 3H),

3.64-2.91(br m, 5H), and 2.82-2.09(br m, 13H); l3C NMR

(ppm, CDC13) 203.27, 202.68, 184.12, 183.67, 173.04,

172.33, 149.02, 148.08, 52.23, 51.65, 45.73, 44.33 (2C),

41.73, 40.98, 40.82, 40.04, 37.77, 37.60, 37.05, 25.90,

and 25.74; m/e calcd 384.1573, obs 384.1577.

Anal. Calcd for c/ 68.73; H, 6.29.

Found: C, 68.54; H, 6.24.

Dimethyl l,2,3,3l,p,4g,5B#5lg,6,7,8",9,9ltt,9,,p, 10-Tetra- decahydro-3, 6-dioxodicvclooentCb,h]-as-indacene-4 a, 5|3- dicarboxylate (58). Freshly cut sodium metal (1.86 g, H 0.0809 g-at) was dissolved in 450 ml

of ammonia (dried over and

C H 30 0 C COOCH j distilled from sodium) under 126

nitrogen in a 1-4 three-necJced flask fitted with Hershberg

stirrer, addition funnel, nitrogen inlet, and dry-ice

condenser. A solution of (0.93 g, 2.45 mmol) in 60 ml

of dry tetrahydrofuran was added dropwise at -78° during

1.5 hr. The reaction mixture was stirred for an additional hour and then allowed to come to room temperature with

evaporation of the ammonia during 2.5 hr. Saturated

aqueous ammonium chloride solution (179 ml) was added

dropwise to the residue. The resulting yellow solution was extracted with methylene chloride (5 x 100 ml) and the

combined organic layers were washed with brine (100 ml) prior to drying. Concentration gave 1.34 g of a white

foam, which was recrystallized from ethyl acetate to

afford 619 mg (66%) of as a colorless solid, mp 248° dec; 1740, 1690, and 1640 cm-1, 1H NMR (6, CDC1-., IUcLX O 100 MHz) 3.67(s, 6H), 3.33-2.97(m, 2H), and 2.93-2.28(m,

16H); 13C NMR (ppm, CDClg) 202.23, 184.35, 173.82, 149.84,

52.26, 45.57 (2C), 40.43, 39.52, 36.92, and 25.94; m/e

calcd 384.1573, obs 384.1573, obs 384.1577.

Anal. Calcd for C22H24°6: C/ H' 6-29* Found: C, 68.51; H, 6.34. 127

Base Promoted Epimerization of j57. Formation of _58. To a solution of _57 (39.0 mg, 0.102 mmol) in 15 ml of dry methanol was added 1 ml of a dilute solution of sodium methoxide in methanol. The reaction solution was stirred at room temperature under nitrogen for 6 days prior to removal of the solvent under reduced pressure. The residue was dissolved in methylene chloride (5 ml) and washed with water (4 ml) and brine (4 ml). After drying, concen­ tration afforded 30 mg (78%) of 58.

Acid Catalyzed Epimerization of 58. Formation of 57.

Hydrogen chloride gas was bubbled through a solution of

5J3 (20 mg, 0.052 mmol) in 2 ml of dry methanol for 10 min.

The resulting yellow solution was stirred for 10 hr under nitrogen then concentrated under reduced pressure. The residue was treated with saturated sodium bicarbonate solution (10 ml) and extracted with dichloromethane.

After drying, the organic phase was evaporated to give

15 mg of _57 contaminated with approximately 10% of 58.

Dimethyl Octadecahydro-3,6-dioxodicvclopent[b,h]-as- indacene-4,5-dicarboxylate (59_) . A solution of ^8^ (500 mg,

1.30 mmol) in 200 ml of ethyl acetate

(distilled from potassium carbonate)

in a hydrogenation bottle was CH300C COOCH3 deoxygenated by bubbling nitrogen 128 through it. Then 166.5 mg of 10% palladium on carbon

(treated with glacial acetic acid in water to pH 5, then washed thoroughly with water and dried overnight at 0.1 mm) was added. The magnetically stirred mixture was connected to a Parr apparatus, cooled to -23° in a dry ice- carbon tetrachloride slush bath, and the hydrogenation was begun at 50 psig. After 10 hr, the reaction mixture was disconnected from the hydrogenator and allowed to come to room temperature while bubbling nitrogen through it. The catalyst was separated by filtration through Celite and washed with a small amount of ethyl acetate. Further washings with methanol (50 ml) and methylene chloride

(200 ml) were collected in a separate flask. Concentration of the ethyl acetate wash followed by recrystallization afforded which was combined with the solid obtained by concentration of the latter two washes to give 422 mg

(83.6%) of 59, mp 246° dec? 1740 and 1730 cm-1; 1H max NMR (6, CDC13) 3.66(s , 6H), 3.25(m, 2H), 3.01-1.61(br m,

18H), and 1.35-0.91(br m, 2H); l3C NMR (ppm, CDC13) 220.82,

174.60, 54.41, 51.81, 44.46, 43.75, 42.90, 41.08, 37.44,

36.99, and 25.29; m/e calcd 388.1886, obs 388.1891.

Anal. Calcd for C22H28°6: C' H, 7.26. Found: C, 67.70; H, 7.26. 129

O ct adecahydro-3,6-dioxacyclopenta[cd]pentaleno[2,1,6-hi11-

acephenanthrylene-4, 5-dione (60). A suspension of 59

(200 mg, 0.515 mmol) in 138 ml of

absolute ethanol was gently heated

under nitrogen until complete

dissolution was achieved and allowed

to return to room temperature.

Sodium borohydride (80.5 mg, 2.13

mmol) was added in one portion and the mixture was

stirred for 16 hr prior to the slow addition of 25 ml of

10% hydrochloric acid. The ethanol was removed jin vacuo

and the residue was diluted with water (75 ml) and extracted

with dichloromethane (4 x 75 ml). The combined organic

layers were washed with brine (75 ml)/ dried, and concen­

trated to yield 214 mg of white solid, which was recrystal­

lized from ethyl acetate to give 64.7 mg (38.3%) of 60, mp 283° dec; vKB^ 1755 and 1735 cm"1; 1H MMR (5, CDC1_) max j 4.90 (q, J = 8.5 Hz, 2H), and 3.90-1.38(br m, 22H); l3C

NMR (ppm, CDC13) 172.23, 82.77, 44.73, 42.20, 42.07, 41.94,

37.64, 37.26, 32.57, and 27.44; m/e calcd 328.1674, obs

328.1679.

Anal. Calcd for c2oH24°4: 73*14* H' 7-37- Pound: C, 73.37; H, 7.19. 130

1,2,2a,6a/7,8,8a. 8b,8c,9,9a,9b,10,10a*10b, lOc-Hexadecahvdro-

3,6~dioxacyclopenta[cd1pentaleno[2,1/6-hi i1acephenanthrvlene-

4,5-dione (68) . t o a solution of triphenylmethane (244.3

mg, 1.0 mmol/ purified by solid

distillation followed by recrystal­

lization from anhydrous methanol

and drying 35 hr under high vacuum)

in 1 ml of dry tetrahydrofuran was added dropwise 654 p.1 of 1.53 M n-butyllithium (1.0 mmol) at 0°. This solution was stirred for 1 hr at room temperature prior to addition to a solution of (50 mg,

0.152 mmol) in 5 ml of dry tetrahydrofuran in a flame-dried,

25-ml three-necked flask fitted with a nitrogen inlet, condenser, rubber septum, and magnetic stirring bar. The base (800 g,l, 0.48 mmol) was added dropwise at 0° and the resulting red solution was stirred for 45 min at room temperature then heated at a gentle reflux for 2.5 hr. A solution of freshly sublimed iodine (116 mg, 0.46 mmol) in 2 ml of dry tetrahydrofuran was rapidly added at 0°.

The reddish-brown solution was stirred for 3 hr at room temperature then washed with saturated sodium thiosulfate solution (3 x 10 ml). The combined aqueous layers were 131

extracted with ether (2 x 10 ml), and the organic phase was washed with sodium thiosulfate (10 ml), water (10 ml),

and brine (10 ml). The organic phases were dried and

concentrated to leave an oil which was purified by silica

gel chromatography. Elution with hexane removed

triphenylmethane and its iodide? an uncharacterized yellow oil was eluted with 20% ether in hexane, and elution with 20-50% ethyl acetate in hexane gave 30 mg (60%) of

68. Recrystallization from ethyl acetate afforded a white solid, dec 298° (sealed tube, bath preheated to

260°)? vKB5 2952, 2879, 1741, and 1720 cm- 1 ? NMR (6, max CDC13) 4.79(m, 2H) and 3.07-0.92 (series of m, 20H);

13C NMR (ppm, CDC13) 166.19, 131.72, 81.71, 43.60,

42.06, 39.96, 38.81, 37.66, 31.13, and 27.34? m/e calcd

326.1518, obs 326.1523.

Eicosahydro-3,6-dioxacyclopenta[cdlpentaleno[2,1,6-hi i]- acephenanthrylene-4, 5-diol (64). Compound 60_ (200 mg,

0.515 mmol) was partially dissolved

with gentle heating in 20 ml of dry

tetrahydrofuran and lithium

HO OH aluminum hydride (97.6 mg, 2.57

mmol) was added in small portions 132

to the stirred suspension at 0° under nitrogen. After

1.75 hr the reaction mixture was allowed to come to room

temperature over a 15 min period prior to quenching with

10% ammonium chloride solution (10 ml) . The lithium salts were removed by filtration through Celite and washed with

dichloromethane. After dilution with water (80 ml), the

filtrate was extracted with dichloromethane (4 x 40 ml)

and the combined organic layers were washed with water

(40 ml) and saturated sodium chloride solution (40 ml).

After drying# concentration afforded 206.9 mg of white

solid which was recrystallized from ethyl acetate to give

100.7 mg (62%) of white crystalline 6 4 # mp 208-210° dec;

Vmax 3380' 1055/ 1027# 992# and 980 cm-1; m/e calcd for

M+-H20 314.1882# obs 314.1887.

7# 8-Diacetvl-3a,3b# 4# 6a# 7# 7a-hexahvdro-3# 4# 7-metheno-3H-

cyclopenta[a]pentalene (73)„ A. Direct Methylation of 50.

Lithium hydride (1.10 g, 138.4 mmol)

was suspended in 200 ml of dry

tetrahydrofuran in a flame-dried#

1 - 1 three-necked flask equipped COCHj with magnetic stirring bar# nitrogen

inlet, rubber septum# and addition funnel. A solution of

diacid 5(3 (10.0 g# 40.9 mmol) in 250 ml of dry tetra­ hydrofuran was added dropwise via cannula to the stirred 133

reaction mixture. Stirring at room temperature was

continued for 45 min? the mixture was cooled to 0° and

methyllithium (1.6 M in ether, 180 ml, 288 mmol) was

added rapidly dropwise. After completion of the addition,

the reaction mixture was stirred at room temperature over­

night prior to transfer via cannula to a cold (0°),

stirred solution consisting of 1 A of water, 300 ml of

10% hydrochloric acid, and 300 ml of saturated ammonium

chloride solution in a 5 - A three-necked flask. This mixture was extracted with ether (1 x 500 ml, 3 x 200 ml)

and the combined organic layers were washed with water

(1 A), saturated sodium bicarbonate solution (200 ml), and water (2 x 400 ml). After drying, concentration afforded

8.68 g of 7^_ which was purified by column chromatography on silica gel (elution with 20% ethyl acetate in hexane) to give 8.31 g (85%) of 73. Recrystallization from hexane yielded long white needles of 73, mp 113°; 1682, Lj~ r UlcUrC 1353, and 1273 on"1; 1H NMR (ppm, CDC13> 6.10(m, 4H),

3.31.(m, 4H) , 2.58 (m, 2H), and 2.02 (s, 6H) ? 13C NMR (ppm,

CDC13) 207.24, 132.96, 77.19, 65.12, 59.47, and 28.95? m/e calcd 240.1150, obs 240.1155.

Anal. Calcd for C16H16C>2; C, 79.97? H, 6.71.

Found: C, 79.91? H, 6.82. 134

B. Cuprate Addition to the Diacid Chloride. Thionyl

chloride {1.7 ml) was added to a solution of ^ (0.98 g,

4.0 mmol) in benzene (40 ml) and pyridine (5 ml). The mixture was stirred at room temperature under nitrogen

for 3 hr and heated at reflux for 30 min. The solvent was evaporated# the residue was again dissolved in

benzene (20 ml)/ and solvent was removed one more time.

In a separate flask/ lithium dimethylcuprate was prepared by addition of methyllithium (1.86 M in ether/

30.44 ml/ 56 mmol) to cuprous iodide (5.12 g, 28 mmol) in

40 ml of anhydrous ether at -5° with stirring. This mixture was cooled to -78° and the diacid chloride in

ether (30 ml) and tetrahydrofuran (10 ml) was added drop- wise. After 2 hr at -78°, absolute methanol (4 ml) was

added slowly. The reaction mixture was allowed to warm to room temperature and poured into saturated ammonium chloride solution (300 ml) . The ether layer was separated

and the aqueous phase was extracted with ether (3 x 75 ml).

The combined ethereal extracts were washed with water

(100 ml), dried, and concentrated to afford a yellow oil.

Crystallization from ether-hexane afforded 0.60 g (63%) of 73 135

2, 6-Diacetyldecahydro-2, 4, 5-metheno-2H-oxireno[3 1,4']- cyclopentaCl1 / 21:4, 5]pentaleno[l, 2-b]oxirene (97).

Diketone ^ (200 mg, 0.832 mmol)

was dissolved in 20 ml of methylene

chloride and 10 ml of 1 M sodium

bicarbonate solution was added. To

this rapidly stirred mixture was added m-chloroperbenzoic acid (328 mg, 1.8 mmole) in small portions. After 49 hr, during which time the reaction was monitored by thin layer chromatography

(silica gel, elution with 50% ethyl acetate in hexane,

= 0.25) and small amounts of m-chloroperbenzoic acid were added as needed, the layers were separated and the organic phase was washed with 10% sodium bicarbonate solution (10 ml), 10% sodium bicarbonate solution (10 ml) together with 10% sodium bisulfite solution (10 ml) (2x),

10% sodium bicarbonate solution (10 ml), and water (2 x

10 ml). After drying, concentration afforded 219.6 mg

(97%) of ^97 which was recrystallized from ethyl acetate/ hexane, mp 167.5-169° (corr); v KBr 1685, 1680, 1350, and ^ max 844 cm-1; 1H NMR (6, CDC13) 3.44(m, 4H), 2.97(m, 4H),

2.19(s, 6H), and 2.13 (m, 2H); l3C NMR (ppm, CDC13) 205.73,

70.64, 57.65, 49.37, 35.80, and 29.58; m/e calcd 272.1048, obs 272.1054. 136

Anal. Calcd for C16H1604: c/ 70.57; H, 5.92.

Found; C, 70.73; H, 5.97.

Decahydro-1, 9-dimethy1-1H-6, 4,5,7b-(epoxyethanylylidyne)- cyclopenta[4,5]pentaleno[l, 6-bc]furan-l,3, 7, 9-tetrol (99).

To a solution of 73^ (100 mg, 0.416

OH mmol) in 2 ml of methylene chloride

.OH at 0° was added dropwise under a nitrogen atmosphere a solution of

m-chloroperbenzoic acid (168 mg,

0.830 mmol) in 5 ml of methylene chloride. The reaction mixture was stirred at 0° for 45 min and then at room temperature for six hr. The mixture was concentrated and placed on a silica gel column using the solvent-coating technique. Elution with 50% ethyl acetate in hexane removed m-chlorobenzoic acid and a minor product and ethyl acetate elution afforded 95.8 mg (75%) of 99, which was recrystallized from ethyl acetate, mp 271-274° dec;

Vmax 3410' 3352' 1390, 1383, 1109, 987, and 862 cm-1; XH

NMR (6, dg-pyridine) 6.38(br s, 4H), 4.55-4.27(m, 4H),

3.43-3.12(m, 4H), 2.87(m, 2H), and 1.6(s, 6H); 13C NMR

(ppm, dg-pyridine) 106.20(s), 84.08(d), 75.60(d), 70.23

(s), 61.57(d), 55.82(d), 53.04(d), and 22.57(q); m / e calcd for M+-H2G 290.1154, obs 290.1160. i 137

3, 6-Diacetyl-2a, 3# 3a, 5a, 6, 6a/ 6b/ 6c-octahvdrodicvclopenta-

[cd,qhlpentalene (103) . A sodium dispersion (375 mg# 16.3

mg-at) was formed in 10 ml of dry

toluene contained in a 50-ml three­

necked flask equipped with a

COCH condenser, nitrogen inlet, and C0CH-. Hershberg stirrer. Upon cooling to

room temperature, trimethyl-

1 « ci' 1 ana f O 0 ml , 17.3 mmole) was added through the

condenser and the mixture was stirred for 15 min. A

solution of 7J3 (100 mg, 0.416 mmole) in 2 ml of dry toluene

was added and the resulting mixture was heated at reflux

with stirring for 16 hr. The reaction mixture was allowed

to cool to room temperature and filtered through Celite.

The filtrate was concentrated and the residue was dissolved

in a small amount of toluene then added dropwise under a

nitrogen atmosphere to 10 ml of dry methanol. The

resulting gold solution was stirred for 30 min, concen­

trated, and the residue was dissolved in 10 ml of

dichloromethane and 10 ml of water. The aqueous phase

was extracted with dichloromethane (2 x 7 ml) and the

combined organic layers were washed with water (7 ml), 138

dried, and concentrated to afford 116.6 mg of gold solid. The crude product was purified by preparative thin layer chromatography on silica gel (elution with 40% ethyl acetate in hexanes, three times), to afford 20.7 mg

(20%) of exo,endo epimer 124 (R^ = 0.42), mp 179-182°;

NMR (8, CDC13) 5.50(q, J = 4.5Hz, 4H), 3.63-2.83(m, 8H), and 2.23(m, 6H); m/e calcd 242.1307, obs 242.1313, and

62.1 mg (62%) of the desired diendo epimer 103 (Rf =0.36).

Recrystallization from ethyl acetate gave pure 103, mp

194-199° (dec, bath preheated to 169°); NMR (6, CDC13)

5.63(s, 4H), 3.67-2.90(m, 8H), and 2.22(s, 6H); 13C NMR

(ppm, CDC13) 207.79, 133.87, 58.02, 52.68, 51.95, and

29.86; m/e calcd 242.1307, obs 242.1313.

Anal. Calcd for ci6H18°2: C' ^9.31; 7.49. Found: C, 79.22; H, 7.43.

2, 4-Diacetyldodecahydropentaleno[l", 6" : 3, 4, 5; 3(l, 4" : 3 1, 41 , 51!- dicyclopenta[1,2-b:1 1,21-b1]bisoxirene (104). Diketone 103

(100 mg, 0.41 mmol) was dissolved in

6 ml of dichloromethane and 2 ml of

0.5M sodium bicarbonate solution

was added. To this stirred mixture

was added m-chloroperbenzoic acid

(177 mg, 80%, 0.82 mmol) in portions. After 2 days, during which time the reaction was monitored by thin 139

layer chromatography (silica gel# elution with 50% ethyl acetate in hexane, R^ = 0.18) and small amounts of m-chloroperbenzoic acid were added as needed, the layers were separated and the organic layer was washed with saturated sodium bicarbonate solution (10 ml) together with 10% sodium bisulfite solution (10 ml) (2x). The combined aqueous phases were washed with dichloromethane

(10 ml) and the combined organic layers were washed with saturated sodium bicarbonate solution (10 ml) and water

(10 ml). After drying, concentration yielded 83.7 mg

(74%) of 104, which was further purified by recrystalli­ zation from ethyl acetate/hexane to afford white plates, mp 166-170°; vKBr 1701, 1696, 1368, and 813 cm"1; 1H NMR e max '

(6, CDC13) 3.52 (s , 4H) , 3.32-2.67(m, 8H), and 2.47-2.00

(m, 8H) ; 13C NMR (ppm, CDC13) 206.64, 61.72, 58.86, 57.29,

47.46, and 29.61; m/e calcd 274.1205, obs 274.1213.

Anal. Calcd for C' 70.05; H, 6.61.

Found: C, 69.93; H, 6.84.

3,3a,3b,4,6a,7a-Hexahydro-l,3,4,7-[l]pentanyl[5]ylidyne-

7H-cyclopenta[a]pentalene-9,12-dione (78) . in a flame-

dried, 500-ml three-necked flask

equipped with a nitrogen inlet,

magnetic stirring bar, rubber 140 septum, and addition funnel was placed diisopropylamine

(6.92 g, 49.0 mmol) and 175 ml of dry tetrahydrofuran.

This solution was treated dropwise at 0° with n-butyl- lithium (1.6 M in hexane, 27 ml, 43.2 mmol) and the resulting light yellow solution was stirred at 0° for an additional 30 min. The base was cooled to -78° and a solution of J73 (5.00 g, 20.8 mmol) in 75 ml of dry tetrahydrofuran was added dropwise over 1 hr. After being stirred at -78° for an additional 6.25 hr, the reaction mixture was transferred via cannula to a flame-dried,

1 l three-necked flask containing anhydrous cupric chloride (6.75 g, 50.2 mmol, dried at 100°/house vacuum then 80°/l mm) in 450 ml of dry dimethylformamide and

100 ml of dry tetrahydrofuran at -78°. The initially light green solution quickly became dark brown as the bisenolate was added. After completion of the addition, the cooling bath was removed and the reaction mixture was stirred for 1 hr prior to being transferred via cannula to an ice-cold, stirred solution of water (1 &},

10% hydrochloric acid (50 ml), and saturated ammonium chloride solution (300 ml). This mixture was extracted with methylene chloride (1 x 300 ml, 3 x 150 ml) and the 141 combined organic extracts were washed with water (3 x 1 A) before drying. Concentration followed by chromatography on silica gel (elution with 10% ethyl acetate in hexane) afforded 2.62 g (53%) of 78. Further purification could be affected by recrystallization from hexane and sublimation (110°/0.3 mm)/ mp 214-215°; 1693 and

1301 cm"1; 1H NMR (5, CDC13 ) 6.05(111/ 4H), 3.59(rti/ 4H) / 2.60

(m/ 2H)/ and 2.33(s, 4H); l3C NMR (ppm, CDC13) 210.22,

133.21, 69.30, 67.12, 61.29, and 39.02; m/e calcd 238.0994, obs 238.1000.

Anal. Calcd for C, 80.64; H, 5.92.

Found: C, 80.63; H, 5.97.

From run to run, differing amounts of 7J5 and 19_ were isolated. Aldol product _75 was purified by preparative thin layer chromatography on silica gel (elution with 50% ether in hexane, R^ = 0.45), recrystallization from hexane, and sublimation at 60° and 0.2 mm, mp 163-164°; 2980, IU q a 142

1675, 1589, 1433, and 1331 cm- 1 ; NMR (fi, CDC13) 5.88

(m, 5H), 3.22(m, 2H), 3.12(m, 2H), 2.95(m, 2H), and 1.92

(d, J = 1.5 Hz, 3H); 13C NMR (ppm, CDC13) 205.36, 172.11,

137.42, 131.96, 130.60, 73.13, 70.40, 64.24, 61.38,

60.17, and 16.54? m/e calcd 222.1045, obs 222.1048.

Anal. Calcd for C^gH^O: C, 86.45; H, 6.35.

Pound: C, 86.36; H, 6.45

Compound ^79 was isolated by chromatography on silica gel (elution with ether/dichloromethane/hexane (1:1:6)), followed by preparative thin layer chromatography on silica gel (elution with 50% ether in dichloromethane,

Rf = 0.7. Final purification was affected by recrystal­ lization from carbon tetrachloride and sublimation at CHCl 100° and 0.2 mm, mp 118.5-119.5°; v 3 2970, 1693, 1357, I T l a X and 1285 cm"1? XH NMR (8, CDC13) 6.30(m, 2H), 6.01(m, 2H),

3.99 (s, 2H), 3.47 (m, 2H), 3.30(m, 2H), 2.65 (m, 2H), and

2.12(s, 3H); 13C NMR (ppm, CDC13) 207.49, 200.63, 135.18,

131.05, 81.12, 73.13, 66.15, 64.24, 59.96, 47.09, and

27.55; m/e calcd 274.0769, obs 274.0767.

Anal. Calcd for ci6^15<'1°2: (3/ ®^.94? H, 5.50. Found: C, 69.78; H, 5.48. 143

3, 3a/ 3b, 4, 6a/ 7a-Hexahydro-1, 3, 4, 7-[2]penten[l]ylC5]ylidyne-

7H-cyclopenta[a]pentalene-9,12-dione (76). Diene dione 7j3

(1.24 g, 5.20 mmol) was dissolved

in 40 ml of dry dioxane and selenium

dioxide (1.72 g, 15.5 mmol) plus

potassium dihydrogen phosphate

(2.56 g, 18.8 mmol) were added under a nitrogen atmosphere. The mixture was heated at the reflux temperature with stirring for 1.5 hr and allowed to come to room tenqperature prior to filtration and concentration. The residue was taken up in 100 ml of dry ether and an insoluble yellow solid was separated by filtration. The filtrate was again concentrated and the residue was taken up in a minimal amount of dichloro- methane. Red selenium was removed by filtration and the filtrate was extracted with saturated sodium bicarbonate solution (40 ml). The aqueous phase was extracted with additional dichloromethane (25 ml) and the combined organic layers were washed with water (40 ml) and dried.

Concentration afforded 0.9 2 g (75%) of 76. An analytical sample was obtained by repeated recrystallization from hexane and sublimation at 100° and 0.2 mm,. mp 222-223.5°; 144

2990, 2965, 1663, 1585, 1340, 1300, and 1286 cm” 1? XTlcW* XH NMR (6, CDC13) 6.61(s, 2H), 5.94(m, 4H), 3.66(m, 4H), and 2.67(m, 2H)? l3C NMR (ppm, CDC13) 200.14, 143.58,

132.72, 66.75, 66.21, and 60.81; m/e calcd 236.0837, obs 236.0842.

Anal. Calcd for ci6Hi2°2: C' 5.12. Found: C, 81.24; H, 5.10. la, 2, 2a,2b, 3, 5a, 6b, 6c-0ctahydro-lH-l,2,3,5b-ethanylylidyne- cyclobuta[l, 2, 3-cd]pentaleno [l, 2, 3-gh]pentalene-6, 8 (6aH) - dione (77). a solution of J ji (390 mg, 1.65 mmol) in 450

ml of dry benzene was deoxygenated

with nitrogen for 0.5 hr and then

irradiated through Pyrex under a

nitrogen atmosphere for 1.5 hr with

a bank of 3500 A lamps in a

Rayonette reactor. Two runs (780 mg total) were combined and the solvent was evaporated under reduced pressure to give 679 mg (87%) of 2 2 as a ligfrk golden oil which was used immediately without further purification? 1H NMR (6,

CDC13) 5.88(m, 2H), and 3.62-2.35(br m, 10H); 13C NMR (ppm,

CDC13) 209.00, 128.47, 76.56, 70.15, 63.27, 52.46, 48.73,

48.52, and 38.20; m/e calcd 236.0837, obs 236.0842. 145

la/2,2a,2b,3/5a,5b/6a,6b/6c-Decahydro-l/2, 3-ethanylylidene-

cyclobuta[cd]pentaleno[l,2,3-qh]pentalene-6,8 (1H)-dione

(84). To a magnetically stirred

solution of JjL obtained from 180 mg

(0.762 mmol) of 76 in glacial acetic

acid (10 ml) was added 750 mg of 70 activated zinc dust. The

reaction mixture was stirred at

room temperature for 6 hr and then poured into ice water

(75 ml) and dichloromethane (10 ml). The mixture was

filtered and the precipitate washed well with dichloro­ methane. The layers were separated and the aqueous phase was extracted with dichloromethane (2 x 10 ml) . The

combined organic layers were washed with water (2 x 50 ml),

saturated sodium bicarbonate solution (50 ml), and water

(50 ml) prior to drying. Evaporation left 140 mg of crude product which was purified by preparative thin layer

chromatography on silica gel (elution with 20% ether in dichloromethane, = 0.4) to yield 67 mg (37% from 7 6 ) of

84. An analytical sample was obtained by recrystallization from 95% ethanol followed by sublimation at 160° and 0.2 mm, mp > 360°; 1719 cm”1; h i NMR (6, CDC1.) 5.83(s, IH a X o 2H), 4.00-3.20(br m, 8H), and 3.20-2.74(m, 4H); l3C NMR 146

(ppm, CDC13) 219.71(d), 131.08(d), 61.38(d), 60.35(d),

60.17(d), 59.14(d), 58.11(d), 44.79(d), and 44.57(d), m/e calcd 238.0994, obs 238.1000.

Anal. Calcd for C 2 .6 ^ 1 4 ° 2 : C/ 8 8 *64; H ' 5.92. Found: C, 80.49? H, 5.97.

2, 2a, 3, 4, 4a, 4b, 5, 7a, 7b, 7c-Decahydro-3, 4, 5-[l]propanyl[3]- ylidene-lH-dicyclopenta[a,cd]pentalene-l,9-dione (85).

Unpurified photoproduct 1T_ obtained

from 4.16 g (17.6 mmol) of 76 was

divided into three portions

consisting of 1.56, 1.57, and 1.03 g.

Each of these samples was dissolved in 30 ml of glacial acetic acid and 5 g of activated zinc dust was added. The reaction mixtures were stirred i magnetically at the reflux temperature for 14 hr, allowed to cool to room temperature, and poured into 300 ml of ice water prior to filtration. The precipitate was washed with dichloromethane and the filtrate was extracted with methylene chloride (3 x 100 ml) . The combined organic layers were washed with water (100 ml), saturated sodium bicarbonate solution (100 ml), and water (100 ml) before drying. Concentration gave a gold-colored solid which was purified by chromatography on silica gel. Elution with 147

10% ethyl acetate in hexane returned 0.28 g of _78 while

2.35 g (60% based on unrecovered material) of &5 was

obtained upon elution with 30-50% ethyl acetate in hexane.

The analytical sample was obtained by sublimation at 160°

and 0.4 mm, followed by recrystallization from hexane

and another sublimation (130°, 0.2 mm); mp 216-218°;

1718 cm"1; % NMR (6, CDC1,) 5.70 (s, 2H), 3.61{m, 6H), IudA w 3.01(m, 4H) and 2.19 (m, 4H); 13C NMR (ppm, CDC13) 219.93,

131.57, 60.38, 59.17, 58.71, 57.65, 56.07, 43.82, and

42.54; m /e calcd 240.1150, obs 240.1153.

Anal. Calcd for 7 9 *9 7 * h, 6.71.

Found: C, 79.97; H, 6.69.

2,2a,3, 4,4a,4b, 5, 7a,7b,7c-Decahydro-3,4,5-[l]propanyl[3]- ylidene-lH-dicyclopentaCa, cd]pentalene-endo,endo-1,9-diol

(91) . Lithium aluminum hydride

(4.01 g, 106 mmol) was suspended in

25 ml of dry tetrahydrofuran con­

tained in a 300 ml three-necked

flask equipped with a nitrogen

inlet, magnetic stirring bar, condenser, and addition funnel. A solution of 85 (1.08 g,

4.50 mmol) in 40 ml of dry tetrahydrofuran was introduced dropwise over 1 hr and the reaction mixture was heated 148

at the reflux temperature for 13 hr. Excess hydride was quenched by dropwise addition of a saturated aqueous sodium sulfate solution (40 ml) at 0°. The mixture was filtered and the white salts were washed extensively with dichloromethane. The filtrate was washed with water

(2 x 50 ml) and saturated sodium bicarbonate solution

(50 ml) before drying. Concentration under reduced pressure afforded 0.81 g of 91. Dissolution of the lithium salts in 200 ml of 10% hydrochloric acid and extraction with methylene chloride gave an additional 0.21 g of 91

(total yield 93%). Recrystallization from ethyl acetate produced a white crystalline solid, mp 240-243° (corr);

3310, 2929, 2918, 2908, 1068, and 1050 cm"1; m/e max calcd 244.1463, obs 244.1468.

Anal. Calcd for C, 78.65; H, 8.25.

Found: C, 78.31; H, 8.35.

2,2a,3,4,4a,4b,5,7a,7b,7c-Decahydro-3,4,5-[l]propanyl[3]-

ylidene-lH-dicyclopenta[a, cd]pentalene-endo,endo-1,9-diol j

Di-p-toluenesulfonate ( 92) . Diol ^91 (780 mg, 3.19 mmol) j

i was dissolved in 25 ml of dry | ! i pyridine under nitrogen, p- j

i toluenesulfonyl chloride (2.43 g, j

12.76 mmol) was added in one i „149

portion, and the reaction mixture was stirred at room temperature for 48 hr before being poured into 250 ml of ice water and extracted with dichloromethane (3 x 100 ml).

The combined organic layers were washed with water (100 ml), cold 10% hydrochloric acid (1 x 100 ml, 1 x 150 ml), water (100 ml), saturated sodium bicarbonate solution

(100 ml), and water (100 ml). Drying and concentration gave 2.15 g of solid which was recrystallized from ethyl acetate to afford 1.28 g (73%) of 92, mp 169.5-170.5°

(corr); VKBr 2936, 1600, 1356, 1189, 1175, 1096, 1001, max 976, 957, 946, 927, 903, 896, 862, 840, 829, 810, 666,

573, and 557 cm-1? 1H NMR (6, CDC13) 7.73(d, J = 8.5 Hz,

4H), 7.23(d, J = 8.5 Hz, 4H), 5.80(s, 2H), 5.08-4.50(m,

2H), 3.53-1.33(br m, 14H), and 2.35 (s, 6H) ? m/e observed base peak at 208.

Anal. Calcd for C 3 qH 3 2 0 6 S 2 : C' 65.19? H, 5.84.

Pound: C, 65.23? H, 5.84.

2a, 4, 4a, 4b,5,7a, 7b, 7c-Octahydro~3,4,5-[l]propen[ l]yl[3]- ylidene-3H-dicyclopenta[a, cdlpentalene. C^g-Hexaquinacene

( 24) . Ditosylate 92^ (583 mg, 1.06

mmol) was dissolved in 20 ml of

dry dichloromethane under a nitrogen

atmosphere and 5 g of activated 150

neutral alumina (heated at 400°C for 24 "hr and allowed to

cool under vacuum) was added. The mixture was stirred

for 1 hr prior to removal of the dichloromethane on a

rotary evaporator. The coated alumina was further dried

under vacuum and then suspended under nitrogen in 2 0 ml

of dry toluene. The mixture was heated at the reflux

temperature with magnetic stirring for 2 days. After

being cooled to room temperature, the alumina was separated

by gravity filtration and washed with methylene chloride/

ether (1:1). Concentration of the filtrate afforded 215 mg

of white solid which was chromatographed on neutral

alumina. Elution with hexane gave C^g-hexaquinacene (24) which was recrystallized from ethyl acetate to yield 90.8 mg (41%) of white needles, mp 231.5-234° (bath preheated

to 226°); vKBr 3030, 2936, 2900, 1353, 743, and 720 cm"1? max ^cyclohexane l g 2 ^ (c _ 1 9 ,9 7 9 ),- ^ NMR (6 , CDC1-) 5.36 max o

(s, 6 H) and 3.76-3.20(m, 10H); 13C NMR (ppm, CDC13) 131.57,

60.1?, 54.91, and 53.06? m/e calcd 208.1251, obs 208.1254.

Anal. Calcd for C 1 6 H16: C, 92.26? H, 7.74.

Found: C, 92.21? H, 7.72. 151

Dodecahydro-3/ 4, 5- [ 1 ]propanyl [ 3 ]ylidene-lH-dicyclopenta~

[a,cd]pentalene. Cig-Hexaquinane (9_4) . C^g-Hexaquinancene

24 (59.4 mg, 285.4 mmol) was

dissolved in 10 ml of ethyl acetate

(distilled from potassium carbonate)

with gentle heating and the solution

was degassed with nitrogen for 15

min. After addition of 10% palladium on carbon (20.7 mg), the mixture was hydrogenated in a

Parr apparatus for 15 hr at 50 psig. The catalyst was separated by filtration through a Celite pad and the filtrate was concentrated under reduced pressure to give

53.8 mg (88%) of 94/ mp 177-180° (corr, from acetone) 7 vKBr 2988/ 2985/ 2936/ and 2858 cm"1; XH NMR (6, CDCl.) max 6 3.77-2.17(br m, 10H) and 2.07-1.13(br m , 12H); 13C NMR

(ppm, CDC13) 65.98(d)/ 54.47(d)/ 4 9 . 0 8 (d) / and, 30.25(t); gi/e calcd 214.1721/ obs 214.1725.

Anal. Calcd for C^gH 2 2 : c/ 89.65; H/ 10.35.

Founds C/ 89.32; H, 10.35. 152 Photolysis of C^g-Hexaquinacene (24). Synthesis of 111.

A magnetically stirred, degassed

solution of 24 (104 mg, 0.50 mmol)

in 450 ml of purified pentane

(stirred sequentially over concen­

trated sulfuric acid and aqueous potassium permanganate solution followed by drying over anhydrous calcium chloride and distillation) was irradiated through Vycor with a medium-pressure Hanovia light source

(450-W) for 33.5 hr under nitrogen. Concentration in vacuo afforded an oily solid which was purified by preparative thin layer chromatography on silica gel (elution with hexane) to afford 15.3 mg of an unidentified compound plus solvent residue (R^ = 0.52), 15.6 mg of 111 (R^=0.53,

22% based on 24 consumed), and 33.5 mg of recovered 2 4

(R^ = 0.35). Photoproduct 111 was further purified by column chromatography on 10% silver nitrate-silica gel.

Hexane elution removed solvent residue and elution with

20% ether in hexane gave 111 as a waxy solid. Sublimation at 40° and 1.2 mm onto a finger cooled by methanol at -78° circulated by a varistaltic pump furnished pure 111, mp

153-156°,• VCHCl3 2950 cm-1; NMR (6, CCl,) 5.63(s, 2H) max 4- and 3.57-2.13(br m, 14H) l3C NMR (ppm, CDC13) 133.32,

63.31, 62.38, 59.33, 56.70, 56.07, 49.37, 43.45, and

42.87; m/e calcd 208.1252, obs 208.1256. 153

Anal. Calcd for c^ 5 Hig: c/ 92.26; H, 7.74.

Found: C, 92.27; H, 7.81.

Hydroboration of C^g-Hexaquinacene (24). Synthesis of 115 and 116. Disiamylborane was prepared by dropwise addition

""•OH

of 2-methyl-2-butene (2.44 ml, 23.04 mmol, freshly distilled from lithium aluminum hydride) to borane-THF

(11.52 ml, 11.52 mmol, 1 M in tetrahydrofuran) at -78°.

This solution was stirred at 0° for 2 hr prior to slow addition to 24 (300 mg, 1.44 mmol) in 10 ml of dry tetra- hydrofuran at 0°. The resulting clear solution was stirred for 0.5 hr at 0° and then at room temperature for

21 hr before quenching the unreacted borane reagent with

5 ml of water. The reaction mixture was cooled to 0° and 10% aqueous sodium hydroxide solution (5 ml) was added followed by addition of 30% hydrogen peroxide at room temperature (periodic cooling in an ice bath was necessary because of heat evolution during the additon). The reaction mixture was stirred for 3 hr at room temperature, 154 then solid sodium chloride was added and the phases separated. The aqueous layer was extracted with tetrahydrofuran (3 x 25 ml) and the combined organic layers washed with brine, dried, and concentrated to afford 486 mg of white foam. Purification by preparative thin layer chromatography on silica gel (elution with tetrahydrofuran) gave 265.2 mg (70%) of a mixture of trio Is 115 and 316 , NMR (6, d5~pyridine) 5.73 (br s, 3H),

5.07-4.23(br m, 3H), and 4.17-1.57(br m, 13H); 13C NMR

(ppm, d(.-pyridine) 77.75, 77.02, 63.45, 63.38,

58.86, 58.21, 58.02, 55.13, 47.72, 46.92, 46.78, 40.98, and 40.22.

Synthesis and Separation of Tribenzoates 120 and 121.

limn 0CO1

A mixture of triols 115 and 116 (265.2 mg, 1.01 mmol) was dissolved in 20 ml of dry pyridine under nitrogen, treated dropwise with benzoyl chloride (710 ^,1, 6.10 mmol), and the reaction solution was stirred at room temperature for 155

12.5 hr. The solution was then poured into 125 ml of ice water and extracted with dichloromethane (3 x 150 ml) .

The combined organic layers were washed with water (50 ml),

10% hydrochloric acid (100 ml), water (50 ml), saturated sodium bicarbonate solution (50 ml), and water (50 ml), dried, and concentrated to afford 913 mg of an oil.

Purification by preparative thin layer chromatography on silica gel (elution twice with 15% ether in dichloro­ methane) yielded a fast-moving unknown compound, 442 mg of unsymmetrical tribenzoate 121(R^ =0.85) as a semi- solid, VKBr 2952, 1714, 1280, 1115, and 713 cm"1; XH NMR H132C (6, CDC13) 8.20-7.73(m, 6H), 7.60-7.17(m, 9H), 5.80-5.40

(m, 3H) and 3.93-1.67(br m, 16H); l3C NMR (ppm, CDC13)

166.23, 165.84, 132.73, 130.69 (2 lines), 129.62, 128.31

82.58, 81.76, 79.96, 63.31, 62.97, 62.48, 55.34, 54.96,

54.28, 53.84, 47.63, 47.14, 46.22, 37.77, 37.24, and

35.97; m/e calcd for M+-C Hq0 469.2015, obs 469.2024; 7 3 and 103 mg of symmetrical tribenzoate 120 (r ^ = 0.78, 94% total yield of 120 and 121) which was recrystallized from ethyl acetate, mp 218-220°, vJPj 2953, 1714, 1278, 1118, IIlQX 1113, and 714 cm"1; 1H NMR (6, CDC13) 8.17-7.90(m, 6H),

7.58-7.20(m, 9H), 5.63-5.40(m, 3H), 3.87-2.67 (m, 9H), and

2.33-2.00(m, 7H); l3C NMR (ppm, CDC13) 165.94, 132.78, 156

131.36/ 129.58, 128.31, 83.02, 63.60, 55,00, 46.27, and

35.68; m/e calcd for M ^ - C ^ O 469.2015, obs 469.2024.

Anal. Calcd for C^H^O^: C, 77.33; H, 5.96. Found: C, 77.46; H, 6.10.

Dodecahydro-3, 4, 5-[1Ipropanyl[3]ylidene-lH-dicyclopenta-

[a,cd]pentalene-l, 7,9-triol (115). A solution of 321 (442

mg, 0.77 mmol) in 60 ml of methanol

was treated with 30 ml of potassium

hydroxide solution (214 mmol, 40%

weight by volume in 5:3 methanol

and water) and the mixture was heated at reflux with stirring under a nitrogen atmosphere for 13 hr. It was then poured into 100 ml of ice water, extracted with tetrahydrofuran (4 x 75 ml), and the combined organic layers were washed with brine (75 ml), dried, and concentrated to afford 248 mg of crude 115.

Purification by preparative thin layer chromatography on silica gel (elution with tetrahydrofuran) gave 182 mg

(90%) of 115 (R^ = 0.5) which was recrystallized from ethyl acetate to yield a powdery white solid, mp 223-225°, vmax 3379' 2940, and 1025 cm"1; 1H NMR (6, d 5-pyridine)

5.62(br s, 3H), 5.00-4.27(m, 3H), and 3.82-1.77(br m,

16H); 13C NMR (ppm, de-pyridine) 77.73, 77.00, 63.40, 5 157

58.84/ 58.16/ 57.97/ 55.10, 47.72, 46.75, 40.98, and

40.25; m/e calcd for 244.1463, obs 244.1468.

Anal. Calcd for C16H22°3: G/ 73*25'* H ' 8 -45* Found: C, 73.11; H, 8.44.

Dodecahydro-3,4,5-[llpropanyl[3]ylidene-lH-dicyclopenta-

[a,cd]pentalene-l, 6,10-triol (116) . A suspension of 120

(103.4 mg, 0.18 mmol) in 60 ml of

methanol was treated with 10 ml of

potassium hydroxide solution (71

mmol, 40% weight by volume in 5:3

methanol and water) and the resulting solution was heated at reflux with stirring under a nitrogen atmosphere for 12 hr. The reaction mixture was concentrated and the residue was separated between tetrahydrofuran and water. The aqueous phase was extracted with tetrahydrofuran (4 x 50 ml), and the combined organic layers were washed with brine (50 ml), dried, and concentrated to afford 142 mg of crude 116, mp 215° dec; NMR (6, d^-pyridine) 5.97-5.33(m, 3H),

5.00-4.40(m, 3H), - and 3.77-1.63(br m, 16H); 13C KMR

(ppm, d_-pyridine) 77.73, 63.36, 58.79, 55.15, 46.95, and b 40.30; m/e calcd for 244.1463, obs 244.1468. APPENDIX A

Selected Proton Magnetic Resonance Spectra

158 i\ j I

c o c h 3 c o c h 3

0 . 6 159 160

tO

VO

r 161

CVJ

o

LTV

o

o 162

♦O

O H

O OJ

o

o

o lOi

o VO 163

to

0 01

o K\

o -St-

o ir\

o VO 164

«o

O i—I

o al

o K>i

O -=t

o* ITS

o# \o

o t— 165

to

o ■ rH

J°- — CVJ

o *

o -4-

o ir\ 166

o OJ

o

O LTV

o ■£>

o t- o — VD 168

O CVJ

o

o

o ir\

o 'vO

o t— 169 APPENDIX B

Carbon Magnetic Resonance Spectra

of Triols 115 and 116

170 HQ, .. OH HO,

Y" REFERENCES

1. F.W. Comer, F. McCapra, I.H. Qureshi, and A.I. Scott, Tetrahedro n, 23, 4761 (1967).

2. a. T. Kunimoto, M. Hori, and H. Umezawa, Biochem. Biophys. Acta, 298, 513 (1973). b. S. Takahashi, H. Naganawa, H. Iinuma, T. Takita, K. Maeda, and H. Umezawa, Tetrahedron Lett., 1955 (1971).

3. S. Nozoe, J. Furukawa, U. Sankawa, and S. Shibata, Tetrahedron Lett., 195 (1976).

4. P.E. Eaton, C. Giordano, G. Schloemer, and U. Vogel, J. Orq. Chem., 41, 2238 (1976).

5. M. Kaneda, R. Takahashi, Y. Iitaka, and S. Shibata, Tetrahedron Lett., 4609 (1972).

6. J. Oehldrich, J.M. Cook, and U. Weiss, Tetrahedron Lett., 4549 (1976).

7. R.B. Woodward, T. Fukunaga, and R.C. Kelly, J. Am. Chem. Soc., 86, 3164 (19 64).

8. I.T. Jacobson, Acta Chem. Scand., 21, 2235 (1967).

9. a. A. de Meijere, D. Kaufmann, and 0. Schallner, Angew. Chem., 83, 404 (1971); Anaew. Chem. Int. Ed. Enal., 1£/~?17 (1971). b. L.-U. Meyer and A. de Meijere, Chem. Ber., 110, 2545 (1977). c. A. de Meijere, Tetrahedron Lett., 1845 (1974).

10. C. Mercier, P. Soucy, W. Rosen, and P. Deslongchamps, Synth. Commun., 3, 161 (1973).

11. M.J. Wyvratt and L.A. Paquette, Tetrahedron Lett., 2433 (1974).

12. L.A. Paquette, I. Itoh, and W.B. Farnham, J . Am. Chem. Soc. 97, 7280 (1975).

172 173

13. a. L.A. Paquette and M.J. Wyvratt, J. Am. Chem. Soc., 96, 4671 (1974). to. D. McNeil, B.R. Vogt, J.J. Sudol, S. Theodoropoulos, and E. Hedaya, J. Am. Chem. Soc.. 96, 4673 (1974) .

14. L.A. Paquette, M.J. Wyvratt, H.C. Berk, and R.E. Moerck, J. Am. Chem. Soc., 100, 5845 (1978).

15. T. Fukunaga and R.A. Clement, J. Orq. Chem., 42 270 (1977).

16. a. K.C. Rice, B.E. Sharpless, U. Weiss, and R.J. Highet, Tetrahedron Lett., 3763 (1975). to. K.C. Rice, U. Weiss, T. Okiyama, R.J. Highet, T. Lee, and J.V. Silverton, ibid., 3767 (1975).

17. R. Srinivasan, J. Am. Chem. Soc., 94, 8117 (1972).

18. R. Mitschka, J.M. Cook, and U. Weiss, J. Am. Chem. Soc., 100, 3973 (1978).

19. P.E. Eaton, R.H. Mueller, G.R. Carlson, D.A. Cullison, G.F. Cooper, T.-C. Chou, and E.-P. Krebs, J. Am. Chem. Soc., 99, 2751 (1977).

2 0 . L.A. Paquette, M.J. Wyvratt, 0. Schallner, D.F. Schneider, W.J. Begley, and R.M. Blankenship, J. Am. Chem. Soc., 98, 6744 (1976).

21. R. Hoffmann, Acc. Chon. Res., 4, 1 (1971). 22. A.D. Baker, Acc. Chem. Res., 3, 17 (1970).

23. J.J. Bloomfield and R.E. Moser, J. Am. Chem. Soc., 90, 5625 (1968).

24. R. Fink, D. Van der Helm, and S.C. Neely, Acta Crvstalloar., Sect. B, 31, 1299 (1975).

25. S.C. Neely, R. Firik, D. Van der Helm, and J.J. Bloomfield, J. Am. Chem. Soc., 93, 4903 (1971).

26. D. Dougherty, J.J. Bloomfield, G.R. Newkome, J.F. Arnett, and S.P. McGlynn, J. Phys. Chem, 80, 2212 (1976). 174

27. T. Koopmans, Physica (Utrecht), 1, 104 (1934).

28. a. J.L. Meeks, H.J. Maria, P. Brint, and S.P. McGlynn, Chem. Rev., 75, 603 (1975). b. J.L. Meeks and S.P. M c S ’ly n n , J. Am. Chem. Soc., 97, 5079 (1975). c. 3T l . Meeks, J.F. Arnett, D.B. Larson, and S.P. McGlynn, J. Am. Chem. Soc., 97, 3095 (1975).

29. R. Gleiter, E. Heilbronner, L.A. Paquette, G.L. Thompson, and R.E. Wingard, Jr., Tetrahedron, 29, 565 (1973).

30. a. N.J. Leonard and P.M. Mader, J. Am. Chem. Soc., 72, 5388 (1950). b. J.F. Arnett, G. Newkome, W.L. Mattice, and S.P. McGlynn, J. Am. Chem. Soc., 96, 4385 (1974).

31. J.J. Bloomfield, Tetrahedron Lett., 587, 591 (1968).

32. G.E. Gream and S. Worthley, Tetrahedron Lett., 3319 (1968).

33. a. M. Van Dyke and N.D. Pritchard, J. Orq. Chem., 32, 3204 (1967). b. J.D. Albright and L. Goldman, J. Am. Chem. Soc., 87, 4214 (1965).

34. R. Bartetzko, R. Gleiter, J.L. Muthard, and L.A. i Paquette, J. Am. Chem. Soc.. 100, 5589 • (1978).

35. a. R. Gleiter, R. Bartetzko, P. Hofmann, and H.D. Scharf, Anaew. Chem., 89, 414 (1977)? Anaew. Chem., Int. Ed. Engl., 16, 400 (1977). b. R. Gleiter, P. Schang, and G. Seitz, Chem. Phys. Lett., in press. c. P. Schang, R. Gleiter, and A. Ricker, Ber. Bunsenqes. Phys. Chem., in press. d. D.O. Cowan, R. Gleiter, J.A. Hashmall, E. Heilbronner, and V. Hornung, Anqew. Chem., 83, 405 (1971) Anaew Chem., Int. Ed. Engl., 10, 401 (1971).

36. P. Bischof, J.A. Hashmall, E. Heilbronner, and V. Hornung, Helv. Chim. Acta, 52, 1745 (1969). 175

37. R. Gleiter, Anqew. Chem., Int. Ed. Engl.. 13, 696 (1974).

38. H.-D. Martin and R. Schwesinger, Chem. Ber., 107 3143 (1974).

39. R.C. Bingham, M.J.S. Dewar, and D.H. Lo, J. Am. Chem. Soc., 97, 1285 (1975).

40. H.H. Jaffe, Acc. Chem. Res.# 2, 136 (1969) and references cited therein.

41. S.H. Liggero, Z. Majerski, P.v.R. Schleyer, A.p. Wolf, C.S. Redvanly, H. Wynberg, J.A. Boerma, and J. Strating, J. Lab. Compds., 7, 3 (1971).

42. Ae. de Groot, J.A. Boerma, J. de Balk, and H. Wynberg, J. Orq. Chem., 33, 4025 (1968).

43. a. J. Meinwald and J.K. Crandall, J. Am. Chem. Soc., 88, 1292 (1966). b. J. Meinwald, J.C. Shelton, G.L. Buchanan, and A. Courtin, J. Orq. Chem., 33, 99 (1968).

44. J.M. Muchowski, Tetrahedron Lett., 1773 (196 6).

45. X. Creary, F. Hudock, M. Keller, J.F. Kerwin, Jr., and J.P. Dinnocenzo, J. Orq. Chem., 42, 409 (1977).

46. A.S. Kende, Orq. Reactions. 11, 261 (1960).

47. L.A. Paquette, R.H. Meisinger, R.E. Wingard, Jr., J. Am. Chem. Soc., 95, 2230 (1973).

48. a. E. Weitz and A. Scheffer, Chem. Ber., 54, 2327 (1921). b. E. Kyburz, B. Riniker, H.R. Schenk, H. Heusser, and O. Jeger, Helv. Chim. Acta, 36, 1891 (1953).

49. P.D.G. Dean, J. Chem. Soc.. 6655 (1965).

50. P. Muller and B. Siegfried, Helv. Chim. Acta, 57, 987 (1974). 176

51. a. T. Ho and G.A. Olah, Anaew. Chem., Int. Ed. Engl., 15, 774 (1976). b. M.E. Jung and M.A. , J. Am. Chem. Soc./ 99/ 968 (1977)? J. Ora. Chem., 42, 3761 (1977).

52. N.J. Hales, personal communication.

53. H. Stetter, V. Lohr, and A. Simos, Liebigs Ann. Cbem., 1977, 999.

54. L. Rand, W. Wagner, P. Warner, and L.R. Kovac, J. Org. Cbem., 27, 1034 (1962).

55. M.J. Wyvratt, Ph.D. Thesis, The Ohio State University, 1976.

56. J.L. Baas, A. Davies-Fidder, and H.O. Huisman, Tetrahedron, 22, 285 (1966).

57. C.G. Chin, H.W. Cuts, and S. Masamune, Chem. Commun., 880 (1966).

58. E. Piers and P.N. Worster, J. Am. Chem. Soc., 94, 2895 (1972), and references cited therein.

59. J. Grimshaw and R.J. Haslett, Chem. Commun. 174 (197 4).

60. L.A. Paquette, M.J. Wyvratt, O. Schallner, J.L. Muthard, W.J. Begley, R.M. Blankenship, and D. Balogh, submitted for publication.

61. E.L. Eliel, "Stereochemistry of Carbon Compounds," McGraw-Hill Book Co.: New York, 1962, pp. 206-207.

62. O. Schallner, private communication.

63. C.F. Lane, Synthesis, 135 (1975).

64. L.A. Paquette, W.B. Farnham, and.S.V. Ley, J. Am. Chem. Soc., 97, 7273 (1975). 177

65. E.E. Smissman, H.N. Alkaysi, and M.W. Creese, J. Orq. Chem., 40, 1640 (1975).

66. G. Kraiss, M. Povarny, P. Scheiber, and K. Nador, Tetrahedron Lett., 2359 (1973).

67. C.F. Lochow and R.G. Miller, J. Am. Chem. S o c . ,, 98, 1281 (1976).

68. W.S. Johnson, Anaew. Chem., Int. Ed. Engl., 15, 9 (1976).

69. K.W. Bowers, R.W. Giese, J. Grimshaw, H.O. House, N.H. Kolodny, K. Kronberger, and D.K. Roe, J. Am. Chem. Soc., 92, 2783 (1970).

70. K. Tsuda, E. Ohki, and S. Nozoe, J. Orq. Chem., 28, 783 (1963).

71. E. Wenkert and J.E. Yoder, J. Orq. Chem., 35, 2986 (1970).

72. P.A. Grieco and M. Miyashita, J. Orq. Chem., 39, 120 (1974).

73. A.E. Greene, J. Muller, and G. Ourisson, J. Orq. Chem., 39_, 186 (1974) .

74. Y. Ito, T. Konoike, T. Harada, and T. Saegusa, J. Am. Chem. Soc., 99, 1487 (1977).

75. B. Berkoz, L. Cuellar, R. Grezemkovsky, N.V. Avila, J.A,.. Edwards, and A.D. Cross, Tetrahedron, 24, 2851 (1968).

76. a. J.J. Pappas, W.P. Keaveney, E. Gancher, and M. Berger, Tetrahedron Lett., 4273 (1966). b. P.L. Stotter and J.B. Eppner, ibid., 2417 (1973).

77. Ruthenium dioxide dihydrate was treated with Clorox bleach in dichloromethane to form the yellow ruthenium tetroxide.

78. a. D.M. Piatak, H.B. Bhat, and E. Caspi, J. Orq. Chem., 34, 112 (1969). b. U.A. Spitzer and D.G. Lee, ibid., 39, 2468 (1974). c. S. Wolfe, S.K. Hasan, and J.R. Campbell, Chem. Commun.. 1420 (1970)o 178

79. This nomenclature is an adaptation of that first proposed by T. Jacobson, Ph.D. Thesis, University of Lund, Sweden, 1973.

80. a. L.A. Paquette, T.G. Wallis, T. Kempe, G.G. Christoph, J.P. Springer, and J. Clardy, J. Am. Chem. Soc., 99, 6946 (1977). b. K.G. Untch and D.J.Ttfartin, ibid., 87, 3518 (1965).

81. a. G.H. Posner and C.E. Whitten, Tetrahedron Lett. 4647 (1970). b. G.H. Posner, C.E. Whitten, and P.E. McFarland, J. Am. Chem. Soc., 9 4 , 5106 (1972) . c. T. Fox, J. Froborg, G. Magnusson, and S. Thoren, J. Orq. Chem., 41, 3518 (1976).

82. a. E.M. Kosower, W.J. Cole, G.-S. Wu, D.E. Cardy, and G. Meisters, J. Orq. Chem., 28, 630 (1963). b. J.K. Kochi, J. Am. Chem. Soc.. 777“ 5274 (1955).

83. a. R.C. Cookson, E. Crundwell, R.R. Hill, and J. Hudec, J. Chem. Soc., 3032 (1964). b. R.C. Cookson, E. Crundwell, and J. Hudec, Chem. and Ind.. 1003 (1958).

84. R.E. Ireland, D.C. Muchmore, and U. Hengartner, J. Am. Chem. Soc., 94, 5098 (1972).

85. P.E. E&ton, R.A. Hudson, and C. Giordano, J. Chem. Soc.. Chem. Commun., 978 (1974).

8 6 . L.A. Paquette, W.J. Begley, D. Balogh, M.J. Wyvratt, and D. Bremner, submitted for publication.

87. a. W.B. Motherwell, J. Chem. Soc., Chem. Commun., 935 (1973). b. P. Hodge and M.N. Khan, J. Chem. Soc., Perkin Trans. 1, 809 (1975).

8 8 . C. Mercier, P. Soucy, W. Rosen, and P. Deslongchamps, Synth. Commun., 3, 161 (1973).

89. G.H. Posner, G.M. Gurria, and K.A. Babiak, J. Orq. Chem., 42, 3173 (1977). 179

90. a. L.A. Paquette, Anqew. Chan., 90, 114 (1978); Anqew. Chem., Int. Ed. Engl., 17, 106 (1978). b. P.M. Warner in "Topics in Nonbenzenoid Aromatic Chemistry," Vol. II, Hirokawa Publishing Co., 1977.

91. R.C. Kelly, Ph.D. Thesis, Harvard University, 1965.

92. L.W. Pickett, M. Muntz, and E.M. McPherson, J. Am. Chem. Soc.. 73, 4862 (1951).

93. G.G. Christoph, J.L. Muthard, L.A. Paquette, M.C. Bohm, and R. Gleiter, J. Am. Chem. Soc., 100, 7782 (1978) .

94. E.D. Stevens, J.D. Kramer, and L.A. Paquette, J. Orq. Chem., 41, 2266 (1976).

95. P. Bischof, R. Gleiter, and E. Heilbronner, Helv. Chim. Acta. 5 3, 1425 (1970).

96. J.C. Bunzli, D.C. Frost, and L. Weiler, Tetrahedron Lett., 1159 (1973).

97. H. Bock, private communication.

98. P. Engel, private communication.

99. G. Baddeley and E.K. Baylis, J. Chan. Soc., 4933 (1965).

100. W.K. Anderson and T. Veysoglu, J. Orq. Chem., 38, 2267 (1973).

101. L.M. Jackman and B.C. Lange, Tetrahedron, 33, 2737 (1977).

102. R.G. Salomon and J.K. Kochi, J. Am. Chem. Soc., 95, 1889 (1973).

103. K. Jonas, P. Heimbach, and G. Wilke, Anqew. Chem., Int. Ed. Engl., 7, 949 (1968). 180

104. L.A. Paquette, J.M. Photis, and R.P. Micheli, J. Am. Chem. Soc., 99, 7905 (1977).

105. T. Fukunaga, private communication.

106. K.J. Klatounde, Acc. Chem. Res., 8, 393 (1975).

107. K.J. Klabunde, private communication.

108. L.T. Scott and G.K. Agopian, J. Am. Chen. Soc., 96, 4325 (1974).

109. D. Bosse and A. de Meijere, Anqew. Chem., Int. Ed. Engl., 13, 663 (1974).

110. H.C. Brown and E. Negishi, J. Am. Chem. Soc., 89, 5478 (1967).

111. H.C. Brown, "Organic Syntheses via Boranes," John Wiley and Sons, New York, N.Y., 1975.

112. J. Kramer, private communication.

113. H.C. Brown and G. Zweifel, J. Am. Chem. Soc., 88, 1433 (1966).