<<

Aromatic and antiaromatic interactions in rigid polycyclic systems : an orbital symmetry model description

Citation for published version (APA): Schipper, P. (1977). Aromatic and antiaromatic interactions in rigid polycyclic systems : an orbital symmetry model description. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR8747

DOI: 10.6100/IR8747

Document status and date: Published: 01/01/1977

Document Version: Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website. • The final author version and the galley proof are versions of the publication after peer review. • The final published version features the final layout of the paper including the volume, issue and page numbers. Link to publication

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement: www.tue.nl/taverne

Take down policy If you believe that this document breaches copyright please contact us at: [email protected] providing details and we will investigate your claim.

Download date: 28. Sep. 2021

AROMATIC AND ANTIAROMATIC

INTERACTIONS IN RIGID POLYCYCLIC SYSTEMS

AN ORBITAL SYMMETRY MODEL DESCRIPTION

PROEFSCHRIFT

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE HOGESCHOOL EINDHOVEN, OP GEZAG VAN DE RECTOR MAGNIFICUS, PROF.DR. P. VAN DER LEEDEN, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN HET OPENBAAR TE VERDEDIGEN OP DINSDAG 14JUNI 1977 TE 16.00 UUR

DOOR

PIETER SCHIPPER

GEBOREN TE ROTTERDAM DIT PROEFSCHRIFT IS GOEDGEKEURD DOOR DE PROMOTORS PROF. DR. H.M. BUCK en PROF. DR. G.C.A. SCHUIT Aan mijn moeder

Ter nagedachtenis aan mijn vader

Aan Taos CONTENTS

Chapter 1 General introduction 7 References

Chapter 2 Generation of 7-norbornenyl cations and the 15 mechanism of their reaction with nucleophiles 2.1. Introduction 2.2. Preparation of norbornadienes substituted at the 7-position with elements of group VA 2.3. Generation of the 7-triphenylphosphonio- 7-norbornenyl dication: a nonclassical dication 2.4. Experimental References

Chapter 3 A Diels-Alder intermediate in the ionization 30 reaction of 9-chloro-9-methoxy-endo- lo- [4.2.1.o2•5]nona-3,7- 3.1. Introduction 3.2. Synthesis 3.3. Generation and properties of cations derived from 9-chloro-9-methoxy-endo-tri cyclo[4.2.1.o2 •5]nonanes 3.4. Reaction products 3.5. Kinetics of the reaction of pyridine with various 9-chloro-9-methoxy[4.2.1.o2 •5]- nonanes 3. 6. Discussion 3.7. Experimental References Chapter 4 Antiaromatic interaction in the 9-methoxy-9- 58 bicyolo[4.2.1]nona-2,4,7-trienyl cation 4.1. Introduction 4,2, Reaction products 4,3, Kinetics of the reaction of pyridine with 9-chloro-9-methoxybicyclo [4,2.1] nonanes 4.4. Generation and properties of cations derived from 9-chloro-9-methoxybicyclo­ [4.2.1]nonanes and -bicyclo[4,4.1]un ... de canes 4,5. Discussion 4.6. Experimental References

Chapter 5 Stereospecific reactions of the 9-phenyl­ 83 seleno-9-bicyclo [4. 2. 1] nona-2 ,4, 7-trienyl anion 5.1. Introduction 5.2. Preparation of the 9-phenylseleno-9-bi- cyclo[4.2.1]nona-2,4,7-trienyl anion 5.3. Stereospecific reactions 5,4, Discussion 5.5. Experimental References

Summary 97

Samenvatting 100

Curriculum vitae 103

Dankwoord 104 CHAPTER 1

General introduction

One of the most significant advances in the study of sol­ volytic mechanisms was the recognition by Winstein and Lucas 1 that which are not directly connected to the re­ action center can strongly affect the rate and the stereo­ chemical course of a reaction. If this phenomenon arises from an intramolecular nucleophilic attack of the at the reaction center, it is known as "neighbouring group participation"2 • This participation may result in an increase in reaction rate. In this case the neighbouring group provides "anchimeric assistance". Three major classes of neighbouring groups can be chosen on the basis of the kind of electrons available for participation: the nonbonding electrons as present on , sulphur, , phosphorus and the halo­ gens; electrons provide? by a rr system and a electrons of a saturated bond. An early example of a chemical consequence resulting from

R R

TsO H 3

!AcOH.

R R

TsO A cO H

7 the participation of the rr electrons of a is the enhanced rate of acetolysis of 3S-cholesteryl sulphonates (1), which proceeds 500 times faster than the acetolysis of satur ated 3B-cholestanyl compounds (2) 3 • The former process proceeds • with retention, whereas the latter one occurs with inversion of configuration4 • The intermediate 3 has been described as a homoallylic cations, the homologue of an allylic cation, which has an additional methylene group between the double bond and the cationic center. It was further demonstrated that a proper orientation of the leaving group for backside attack by the double bond is critical. The 3a-cholesteryl derivatives (1a) are less reactive than the 3S derivatives (1) and give, as the principal product on acetolysis, cholesta-3,5-diene6.

R R

H

OTs

When the double bond is situated in a symmetric position with respect to the reaction center, even larger rate enhance­ ments were observed 7 • The neighbouring group participation of the double bond of anti-7-norbornenyl tosylate (5) results in a rate acceleration of 10 11 with respect to the saturated analogue (4). The intermediate cation (6) has been described as a homoaromatic species, in which the overlap of p orbitals

4 5 gives rise to a set of molecular orbitals which are similar to the cyclopropenium cation. Three types of homocyclopropenium can be chosen on the basis of the number of interruptions of the a framelc. The mono-, bis- and trishomocyclopropenyl cations, respectively

8 exemplified by the cyclobutenyl cationlc,a (7), the norbornen­ yl cation (6)9 and the [3.1.0]hexenyl cation (8)w.

. A w. ~

7 6 8

The concept of has been extended to bi­ cycloaromaticity by Goldstein and Hoffmann in a series of papers in which they attempt to elucidate the nature of the interaction between ~ systems 1 1 • In their study the fundamen­ tal building block of an unsaturated compound is an intact conjugated polyene segment, which is designated in Figure 1 as an unbroken line, called a ribbon. Of the great variety of topologies which may be envisaged for the linkage of several ribbons, four have been selected. These topologies, the peri­ cyclic, the spirocyclic, the longicyclic and the laticyclic, are depicted in Figure 1 together with some examples. For the pericyclic topology Hlickel's 4n+2 ~-electron prerequisite for cyclic stabilization does apply. The other topologies are pre- dicted to be stabilized if there are more stabiliz than destabilizing interactions between the ribbons. An interaction between two ribbons is stabilized, when 4n+2 ~ electrons are involved. Thus, the 7-norbornenyl cation (6) as well as the 7-norbornadienyl cation (9), respectively examples of the pericyclic and longicyclic topology, are stabilized ions. The traditional tests for relative stabilities of inter­ mediates are most indirect. The rate of of a neu­ tral precursor is usually compar~d with that of an appropria­ tely hydrogenated derivative. However, the solvolysis of an unsaturated compound can give rise to rearranged products. In

9 Pericyclic topology (...... _____ ···~ .n C.,,__,; ~.. ~ 0 6 8 Spirocyclic topology Longicyclic topology c____.:.. D ..

9

Laticyclic topology n ... n ,. . 'l...g ...... ~... J Figure 1

this case it is difficult to decide which kind of activated complex is involved in the rate-determining step. Alternative­ ly, the character of the activated complex is correlated to the structure of the corresponding which may be observed as a long-lived species in superacid media. In this thesis the neighbouring group participation of rr and a electrons in polycyclic systems has been studied. Charge­ stabilizing groups were introduced at the reaction center in order to prevent rearrangement reactions. In this way it was possible to identify the intermediates and to ascertain their relative order of stability. Thus, for the first time, the study of antiaromatic interactions appeared to be accessible along kinetic and spectroscopic lines. Chapter 2 deals with the introduction of charge-stabiliz-

10 ing groups, derived from elements of group VA and VIA, at the 7~position of the norbornenyl system. Reaction of triphenyl­ phosphine with the homoaromatic 7-norbornadienyl cation afford­ ed 2,3-substituted products as a consequence of charge deloca· lization, whereas the related reaction with the classical 7- methoxy-7-norbornenyl cation produced the ?-substituted tri­ phenylphosphonium salt. The latter compound appeared to be a suitable precursor for the generation of the 7-triphenyl- , phosphonio-7-norbornenyl dication, a nonclassical dication. In Chapter 3 the study of ionization reactions of 9- chloro-9-methoxy-endo-tricyclo [4.2.1.o2•5]nona-3,7-diene to­ gether with some appropriate reference compounds is described, It was established that during the ionization of the diene anchimeric assistance is provided by an electron-delocalization process in which the two double bonds and one are involved. Under solvolytic conditions no rearrangement react­ ions took place, whereas in liquid so2 a rearranged structure was observed, which appeared to be in equilibrium with its neutral precursor. From this it was concluded that the struct­ ure of the intermediate depends on the position of the gegen­ ion. In Chapter 4 the stability and bicycloaromaticity of the 9-bicyclo [4.2.1] nona-2,4,7-trienyl cation is discussed. Rate constants for the reaction of anti-9-chloro-9-methoxybicyclo - [4.2.1]nona-2,4,7-triene and some appropriate reference com- pounds with pyridine were determined. The rate retardation of the triene with respect to the more hydrogenated derivatives indicates that the 9-bicyclo [4.2.1]nona-2,4,7-trienyl cation is destabilized by an antiaromatic interaction (4n electrons) between the cationic center and the butadiene moiety. This was confirmed by a study of the 9-methoxy-9-bicyclo[4.2.1] nona-2,4-dienyl cation under conditions of long life. PMR and CMR data indicate an asymmetric interaction between the 9- and the butadiene bridge. In Chapter 5 the generation of carbanions which are sta­ bilized due to homoconjugative interaction is described. An example is the 9-phenylseleno-9-bicyclo[4.2,1] nona-2,4,7-· 11 trionyl anion generated from the corresponding phenylseleno ketal. Electrophilic addition reactions at this anion proceed stereoselective, consequent on a stabilizing aromatic inter­ action (6w) of the anionic center (2w) with the butadiene bridge (4w). The highly preferential location of these 6w electrons controis the entrance of the electrophile.

12 References

1. S. Winstein and H.J. Lucas, J. Amer. Chern. Soc.,£.!_, 1576, 2845 (1939). 2. Reviews on neighbouring group participation: a. B.C. Capon, "Neighbouring Group Participation", Quart. Revs. , , 4 5 ( 1 9 64) . b. P.D. Bartlett, "Non-Classical Carbonium Ions", W.A. Benjamin, New York, 1965. c. S. Winstein, "Non-Classical Ions and Homoaromaticity", Quart. Revs., 141 (1969). d. P.R. Story and B.C. Clark, Jr., in "Carbonium Ions", Vol. 3, G.A. Olah and P. v. R. Schleyer, Ed., Wiley­ Interscience, New York, N.Y., 1972, Chapter 23. 3. S. Winstein and R. Adams, J. Amer. Chern. Soc., 0, 838 (1948). 4. C.W. Shoppee and G.H.R. Summers, J. Chern. Soc., 3361 (1952). 5. M. Simonnetta and S. Winstein, J. Amer. Chern. Soc., 22, 4183 (1955). 6. a. R.H. Davies, S. Meecham and C.W. Shoppee, J. Chern. Soc., 679 (1955). b. C.W. Shoppee and D.F. Williams, J. Chern. Soc., 686 (1955). c. J.H. Pierce, H.C. Richards, C.W. Shoppee, R.J. Stephen­ son and G.H.R. Summers, J. Chern. Soc., 694 (1955). 7. S. Winstein, M. Shatawsky, C. Norton and R.B. Woodward, J. Amer. Chern. Soc., 7, 4183 (1955). 8. a. T.J. Katz and E.H. Gold, J. Amer. Chern. Soc.,.§..§_, 1600 (1964). b. Mrs. A.E. van der Hout-Lodder, J.W. de Haan, L.J.M. van de Ven and H.M. Buck, Rec. Trav. Chim., 92, 1040 (1973). 13 9. M. Brookhart, A. Diaz and S. Winstein, J. Amer. Chem. Soc., ~. 3133, 3135 (1965). 10. S. Winstein, J. Sonnenberg and L. de Vries, J. Amer. Chem. Soc., .§J_, 6523 (1959). 11. M.J. Goldstein and R. Hoffmann, J. Amer. Chem. Soc., 93, 6193 (1971).

14 CHAPTER 2

Generation of 7-norbomenyl cations and the mechanism of their reaction with nucleophiles

2.1. Introduction

A vivid example of neighbouring group participation by nonallylic double bonds is found in the bicyclo [2.2.1] heptyl (norbornyl) series 1,2. The relative rates of acetolysis of 7-norbornyl tosylate (4) and the unsaturated analogues syn- 7-norbornenyl (10), anti-7-norbornenyl (5) and 7-norbornadienyl tosylates (11) are shown below: T~ T~ Th £;' f. j 4 10 5 11

4 1 0 11 1 014 k rel 10 The rate acceleration noted for 10 was attributed2 to cr parti- cipation of the bonding electrons, anti to the tosylate c1-c 6 group, which afforded the incipient allylic cation 12. The i11creased acceleration and the retention of configuration of 5 and 11 were ascribed to rr participation of the double

12 bond anti to the tosylate group 1. The intermediate cation 6 shows nucleophilic attack (AcOH) from the rearside, since in­ version of configuration at c7 is attended by an antiaromatic

15 activated complex (Chapter 5).

T~ AcOH

6

The structure of the intermediate cation (6) has been the subject of major controversy. When anti-7-norbornenyl tosylate (5) is solvolyzed under nonequilibrium conditions, tricyclic products (13) were obtained predominantly.

5 ~D This was interpreted by H.C. Brown 3 as an indication of two tr lie cations in rapid equilibrium (14), whereas S. Win­ stein~ described the cation in terms of a bishomocycloprope­ nium structure formed by the overlap of three carbon 2p orbi­ tals at c2, c3 and c7 , as visualized in 6a. The latter inter-

14 pretation was supported by the direct observation of 7-norbor­ nenyl cations in superacid solution. The PMR spectrum of 6 in FS0 3H-S02 indicates considerable positive charge at the c2 and 5 c3 positions • This could still be consistent with both inter­ pretations. However, the PMR spectrum of the 2-methyl-7-nor­ bornenyl cation 15 shows almost unchanged resonances of H3 and H7 and therefore is inconsistent with the tricyclic structure 16 in spite of the generation of a tertiary cation at c2. Thus these observations are in support of the nonclassical represen-

16 tation as implied in 156. Introduction at c7 of a charge-stabilizing , e.g. p-anisyl, appeared to cancel neighbouring group partici­ pation7. The 10 11 rate difference between 4 and S is leveled to a factor 3 in 18 and 19.

Moreover, the solvolysis of 19 is not stereospecific, but provides an epimeric mixture. If less stabilizing groups are introduced at the 7-position, the influence of the double bond on the reaction rate increases. From this it was conclud- ed that r. part ion is strongly related to the electron demand of the incipient carbonium ion7 In order to get more insight into the nature of non­ classical cations, the introduction of groups which are able to stabilize positive charge, such as derivatives of phos­ phorus and the heavier elements of group VA and VIA, has been examined. A similar method has proved to be successful in the oxidation of diphenylmethyltriphenylphosphonium methylide when the related radical and dipositive ion were formeds.

2.2. Prepoyation of norbornadienes substituted at the ?-position with elements of group VA9,10

A number of ?-substituted norbornadienes 20 f) could be prepared 9 by addition of nucleophiles derived from elements

17 of group VA and sulphur to a solution of the 7-norbornadienyl cation (9). The nucleophiles are mentioned in Table I. Cation 9 was generated according to the method of Story and Saunders11 by treatment of a solution of 7-norbornadienyl chloride in liquid so 2 with AgBF 4 at -70°. The reaction of the 7-norborna­ dienyl cation with the nucleophiles was observed by PMR. Re-

Cl~ -~·. + £I:? £1$ Ph3P~ 9 £lf

Table I PMR spectral dad' for 7-substituted norbornadienes in liquid so ( lJ values) 2

3 6 ~2

Compound X H H Other 2,3 7 + 20a S(CH ) 6,98 6.94 4,04 (m) 3.87 2. 87 (s, CH ) 3 2 3 (t,J=2.1) ( dt,J=O. 5, 1. 9) (t,J=1.6) + b S(C6H5)2 6,95 6,93 3.88 (m) 4.62 7.45 ( m, phenyl) (t,J=2.2) (dt,J:0.4, 1.9) (t,J=1.4) c k(C6 H5 ~ 7.10 6.60 4.20(m) 4.20 7. 65 (m, phenyl) (t.J~2.0) (dt,J:O.S, 1.9) (t,J=1.4)

d Sb(C6HS)3 7.08 7.00 4.40(m) 4.43 7. 72 (rn. phenyl) (t,J=2.0) (dt,J:O.S, 1.9) (t, J=l. 4) + e Bi(C6H5)3 6.80 6.40 3. 70 (m) 3.68 7.54 (rn, phenyl) (t,J=2.1) ( dt,J:O. s, 1. 9) (t,J=1.4) + NC D 6,82 6.60 4.40(m) 4.42 5 5 (t,J=2.2)

• J values refer to the observed line spacings in Hz and are not necessarily the true coupling constants.

18 markably, triphenylphosphine and pyridine afforded decomposi­ tion products only, as indicated by the appearance of broad upfield-signals, whereas the other nucleophiles produced 7- substi tuted norbornadiene salts. The PMR data of 20 (a.- are summarized in Table I. The assignments were made on the basis of a long-range coupling between the top and one of the vinylic proton pairs, which coupling only exists in "W" type configuratiorP Using the reaction conditions mentioned previously, all compounds were stable and two of them (20a,f) could be ob­ tained in crystalline form. The compounds 20a and ZOe appeared to be suitable for substitution at the ?-position with pyri­ dine (20f). However, a similar procedure with triphenylphos­ phine failed. This remarkable behaviour of triphenylphosphine towards the 7-norbornadienyl cation motivated the study of reactions with other tervalent phosphineslO. It appeared that reaction of phosphines with donating substituents like P(n-C H ) , P(OCH ) , P(C H ) and P(C H ) Cl with the cation 4 9 3 3 3 6 5 3 6 5 2 leads to decomposition of the ion. Phosphines with electron-

Table II PM:R spectral data for 7-substituted norbornadiene phosphonium compounds in liquid so 2 '' "''"~' x'b,_ ' 6~

X 4 H7 JPH(Hz)

+ PBr 7.10 7.00 4.40 4.38 42 3 + PCJ 7.00 6.96 4.42 4.28 35 3 + P(C H )Cl 6.90 6.50 4.20 4.10 15 6 5 2 + P(Cls>3 7.30 6.52 4.34 4.62 14 + P(OCH ) CCH 7.02 6.98 4.25 3.30 6 2 3 3

19 withdrawing substituents, as mentioned in Table II, give 7- substituted norbornadiene phosphonium salts. Only the bicyclic phosphine P(OCH 2) 3CCH 3 seems to show a dif­ ferent behaviour. In this case the decrease in basicity is due to introduction of molecular constraint on quaternization of the phosphorus atom13, It is known that reaction of the 7-norbornadienyl cation with nucleophiles under nonequilibrium conditions gives both 7- and 2,3-substituted products3 • 14. It appeared that the pro­ duct ratio varies with the nature of the nucleophiles. Strong nucleophiles react at the 2,3-positions, weak nucleophiles at the 7-position1 4. Therefore, the absence of 7-substituted products with the more nucleophilic phosphines may indicate substitution at the 2,3-positions. However, no evidence was found for the formation of tricyclic products in the previous reactions in liquid so 2. Perhaps their instability under the prevailing conditions leads to decomposition. For this reason the reaction with P(C6H5) 3 was carried out in dry in which the intermediate tricyclic product appeared to be more stable:

PPh3,AgBF4

CH30CH3

Table III NMR spectral data for compound 21 in acetone-d ( 8 values) 0

P·ositions 7

PMR 7.26 6.46 4.47 4.61 7.85

CMR 147.55 143.53 53.33 77.20 121.5, 129.8, 135.2

Compound 21 is characterized by its NMR spectral parameters as summarized in Table III. The IR spectrum shows a C=C stretch­ ing bond near 1570 em -1 characteristic of a norbornadienyl system. Compound 21 was isolated in a pure state. Spectral

20 parameters of compound 22 were obtained from mixtures of 21 and 22. In the IR spectrum of the mixture a C=C stretching was observed at 1615 cm- 1. This band, which is very strong, points to an asymmetrically substituted double bond. The PMR spectrum of 22 show~ a perturbed doublet (two ) at 6.33 ppm, a broad multiplet at 3.18 ppm (three protons) and a multiplet at 1.85 ppm (two protons). Extra aromatic signals were also ob­ served. The CMR spectrum shows, apart from the aromatic sig­ nals, two absorptions near 49.5 and 50.5 ppm, to be ascribed to vinylic carbon . Moreover, five signals were observed in the aliphatic region between 85 and 150 ppm. The overall spectral evidence suggests a structure as depicted above. Specific assignments in the CMR spectrum were not made with the exception of a signal near 85 ppm which is tentatively ascribed to c4 (J13C_ 31 p=100 Hz). Compound 22 slowly decom­ poses at room temperature. Dissolution of compounds 21 and 22 in liquid so2 also leads to decomposition of 22, whereas compound 21 is completely stable. Thus it would appear that 22 is the intermediate species in the reaction of triphenyl­ phosphine with the 7-norbornadienyl cation in liquid so2. It can be concluded that the reaction of the nucleophiles given in Table I and II, is thermodynamically controlled which re­ sults in ?-substituted products. In contrast, the reaction with triphenylphosphine is more kinetically controlled, lead­ ing to 2,3-substituted products in liquid so2 and both 7- and 2,3-substituted products in dry acetone.

2.3. GenePation of the 7-tPiphenylphosphonio-7-norbornenyl dication: a nonclassical diaation15

Stabilization of charge by double bond participation in 7-norbornenyl cations decreases with increasing electron­ donating ability of the 7-substituents (Section 2.1). Indeed, reaction of triphenylphosphine with 7-chloro-7-methoxynorbor­ nene (23b,c) in liquid so2 gives rise to ?-substituted pro­ ducts only, consequent on the absence of n-electron partici-

21 pation in the intermediate 7-methoxy-7-norbornenyl cation (25) (vide infra). Reaction of 7,7-dimethoxynorbornene (23a) with PC1 5 yields a mixture of e~imers of 7-chloro-7-methoxynorbornene, 23b (76%) and 23c (24%), which reacts rapidly with triphenylphosphine in liquid 60°) to yield a mixture of the epimeric phospho­ nium salts 23£ (28%) and 23g (72%). At -60°, the ratio of the epimers does not change, but at ca -14°, compound 23g isomerizes to 23f. The latter one could be obtained in crystalline form at room temperature. Roth 23f,g produce the dication 24 in FS0 3H-so 2 at -60°. Quenching of this dication with me­ thanol yields 23f only (Scheme I). The structure of compounds 23 was confirmed by their PMR and CMR spectra (Table V). The structure of cation 24 was assigned by comparing its PMR spec­ trum with that of the 7-norbornenyl cation (6) 14a and the 7-methoxy-7-norbornenyl cation (25) 14 f (Table IV). In the spec­ trum of 24 all resonances occur downfield with respect to the corresponding resonances of 6. This effect can be ascribed to the o- and o-electron withdrawing ability of the positive

Table IV PMR spectral data for7-norbornenylcat!ons In FS0 H-So ( 8 values) 3 2

+kA z;C)a 6 2

R H H H Other 2,3 1,4 Sn,6n

+ P(C H ) (24) 7.45 4.63 2.8-3.1 2.2-2.5 7.65 (phenyl) 6 5 3

H (6) 7.07 4.24 2.44 1. 87 3.24 (H ) 7

OCH (25) 6.58 3.70 2.2-2.4 1.6-1.8 4.64 (OCH ) 3 3 3.47

22 CH;t;H3

23a l"'' :£H3 CH;s +

23c 23b

1""'·'''so, Ph?t;H+ 3 CH;sh+ 3 +

23g 23f

24

Scheme I Formation of the 7-triphenylphosphonio-7-norborne­ nyl dication

23 1\) .j>. Table V PMR and CMR spectral datd' for compounds 23 ( o values)

d)3 6 2

Compound R Solvent Position 2,3 1,4 5,6 7 Other 1 R2

23a OCH OCH CDCJ PMR 5.96 2.74 (m) 1.90 (m,ex) 3.15 ( s) 3 3 3 ( t,J=2. 5) 1.05 (m,en) 3.11 (s)

CDC! CMR 134.0 44.9 23,7 119.6 52.2 ( 1) 3 49.5 (2)

23b OCH Cl CDC! PMR 5.95 2.93 (m) 1.85 (m,ex) 3.32 (s) 3 3 ( t,J=2. S) 1.05 (m,en)

CDCJ CMR 135.6 51.6 23.5 121.5 53.2 3

23c Cl CDC! PMR 5.92 2.93 (m) 1.85 (m,ex) 3.25 ( s) 3 ( t, }=2. S) LOS (m, en)

CMR 134.0 51.2 24.4 122.3 55.5 + 23f OCH so PMR 5.54 3.50 (m) 2.20 (m,ex) 3.01 7.65 3 P(C6HS)3 2 (t,J=2.5) 1.15 (m,en) (d,J=2) (br s, 15H)

so CMR 104.1 48.5 22.8 103.2 54.4 c 134.6 2 (d,JPC=10) ( d,JPC=10) (d,JPC=47) cP134.8 0 C 130.0 (d,JPC=13) C:12o.6 (d,JPC=175) + 23g OCH so PMR 6.15 3.77 (m) 1. 2-1.6 2.68 7.65 flC6H5)3 3 2 ( dt, J=6, 2. 5) (br m) (d,J=2) (br s, 1SH)

* J values are expressed in Hz. triphenylphosphonium group, which results in an increased delocalization of TI electrons to the ?-position. In contrast, the chemical shifts of cation 25 occur upfield with respect to comparable shifts of 6, which indicates absence of double bond participation. This is supported chemically by an ex­ change reaction of the epimers 23b and 23c in liquid so 2 at higher temperatures and by the initial formation of a mixture of products in the reaction of 23b,c with triphenylphosphine. First-order kinetics were observed for the isomerization of compound 2 to 23f in liquid so 2 at -14° (k=2.7 x 10- -l). In liquid so 2 with 10 vol.% FS03H compound 23f is stable, whereas 23g decomposes at a rate comparable to the rate of isomerization in liquid so 2. During the decomposition no 7- methoxycarbenium ion (25) was observed, although this cation was shown to be stable under these conditions by adding 23b,c. Thus, isomerization of 23g to 23f proceeds via the dication 24, which is not stable in liquid S0 2-FS03H (10:1). Double bond participation, which is increased by the PPh 3+ group, explains in this dication the irreversible formation of com­ pound 23f.

2.4. Experimental

PMR spectra were obtained on Varian Model T-60A and HA- 100 spectrometers, equipped with variable temperature probes. Chemical shifts are reported relative to TMS as internal standard. CMR spectra were obtained using a Varian Model HA-100 NMR spectrometer, equipped with FT accessory and variable temperature probe.

• 7-Norbornadienyl fluoroborate (9) A solution of 7-norbornadienyl chloride (1.0 g, 0.008 mol) in sulphur dioxide (10 ml) was added slowly during 15 min to a stirred suspension of silver fluoroborate (2.0 g, 0.01 mol) in sulphur dioxide (10 ml) at -70°. After the addition was

25 complete, the solution was stirred for an additional 15 min. The volume was reduced by application of a vacuum. After fil­ tration, the solution was transferred into an NMR sample tube.

• Reaction of 7-norbornadienyl fluoroborate with nucleophiles An equimolar amount of nucleophile was added to a solution of 7-norbornadienyl fluoroborate in sulphur dioxide in an NMR sample tube at 70°. After shaking vigorously, the PMR spectra were recorded at 20°. The PMR spectral data are· collected in Table I and II.

• 7-Norbornadienyltriphenylarsonium fluoroborate (20c) 7-Norbornadienyl fluoroborate was prepared in the usual way from 7-norbornadienyl chloride (2.0 g, 0.158 mol) and silver fluoroborate (3.11 g, 0.16 mol) in sulphur dioxide (40 ml). Triphenylarsine (4.9 g, 0.16 mol) was introduced into the solution and stirred for 30 min, whereupon removal of sulphur dioxide was started at reduced pressure. Concomitantly, me­ thylene chloride (50 ml) was added in small portions. The resulting mixture was filtered and poured into dry ethyl ether. After filtration the precipitate was recrystallized from a mixture of chloroform and carbon tetrachloride to give 20c (5.8 g), mp 220-224°. PMR: see Table I.

• 7-Chloro-7-methoxynorbornene To a stirred solution of 7,7-dimethoxynorbornene (13.53 g, 0.088 mol) in diethyl ether (15 ml) phosphorus pentachloride (18.72 g, 0.09 mol) was added in small portions at such a rate that the ether boiled gently. Sometimes heating and addition of some phosphorus oxychloride was required to initiate the reaction. Vacuum distillation provided pure 23b,c (6.89 g), bp 41-43° (2.0 mm). NMR: see Table V.

• 7-Methoxy-7-norbornenyltriphenylphosphonium chloride (23g) To a solution of 7-chloro-7-methoxynorbornene (1.0 g, 0.062

26 mol) in sulphur dioxide (10 ml) triphenylphosphine (1.65 g, 0.065 mol) was added. The solution was stirred for 30 min at -30°. Evaporation and recrystallization from chloroform and carbon tetrachloride yielded 23g (2.2 g), mp 182° (dec). NMR: see Table V.

27 References

1. a. S. Winstein, M. Shatavsky, C. Norton and R.B. Woodward, J. Amer. Chern. Soc.,]_]__, 4183 (1955). b. S. Winstein and M. Shatavsky, J. Amer. Chern. Soc., 592 (1956). c. S. Winstein and C. Ordronneau, J. Amer. Chern. Soc.,~. 2084 ( 1960). 2. S. Winstein and E.T. Stafford, J. Amer. Chern. Soc., 505 (1957).

3. H.C. Brown and H.M. Bell, J. Amer. Chern. Soc.,~. 2324 (1963). 4. a. S. Winstein, A.H. Lewin and K.L. Pande, J. Amer. Chern. Soc., ' 2324 (1963). b. S. Winstein, "Non-Classical Ions and Homoaromaticity", Quart. Revs., 141 (1969). 5. M. Brookhart, A. Diaz and S. Winstein, J. Amer. Chern. Soc., ~. 3135 (1966). 6. R.K. Lustgarten, M. Brookhart, S. Winstein, P.G. Gassman, D.S. Patton, H.G. Richey, Jr. and J.D. Nichols, Tetra­ hedron Lett., 1699 ( 1970). 7. a. P.G. Gassman, J. Zeller and J.T. Lumb, Chern. Comm., 69 (1968). b. P.G. Gassman and A.F. Fentiman, Jr., J. Amer. Chern. Soc., 2_!, 1545 (1969). c. P.G. Gassman and A.F. Fentiman, Jr., J. Amer. Chern. Soc., 92,2549 (1970). 8. P. Schipper and H.M. Buck, Phosphorus, 1, 97 (1971). 9. P. Schipper and H.M. Buck, Phosphorus, 1, 93 (1971). 10. P. and H.M. Buck, Phosphorus, 3, 133 (1973). 11. P.R. Story and M. Saunders, J. Amer. Chern. Soc.,!±, 4876 (1962). 28 12. E. I. Snyder and B. Franzus, J. Amer. Chem. Soc., ' 116 6 (1964). 13. L.J. Vandegriend, J.G. Verkade, J.F.M. Penn~ngs and H.M. Buck, J. Amer. Chem. Soc.,~ (1977), in press. 14. a. P.R. Story, J. Amer. Chem. Soc., 83, 3347 (1961). b. H. Tanida andY. Hata, J. Org. Chem., 30, 977 (1964). c. H. Tanida, T. Tsuji and T. Irie, J. Amer. Chem. Soc., ~. 864 (1965). d. A. Diaz, M. Brookhart and S. Winstein, J. Amer. Chem. Soc., ~. 3133 (1966). e. J.J. Tufariello, T.F. Milch and R.J. Lorence, Chem. Comm., 1202 ( 1967). f. R.K. Lustgarten, M. Brookhart and S. Winstein, Tetra­ hedron Lett., 141 (1971). 15. P. Schipper, W.A.M. Castenmiller, J.W. de Haan and H.M. Buck, J.C.S., Chem. Comm., 574 (1973).

29 CHAPTER 3

A Diels-Aider intermediate in the ionization reaction of 9-chloro-9-methoxy-endo-tricyclo [4.2.1.02,5] nona-3, 7-diene

3.1. Introduction

The study of strained polycyclic small ring compounds has revealed unusual solvolytic reactivities, numerous skeletal rearrangements and novel degenerate isomerizations 1 • In parti­ cular, when three- and four-membered carbocyclic rings are incorporated in norbornyl systems, enhanced reactivities were observed1c. The geometry of the system is crucial, as may be seen from the following examples: B:b B~ ~s

1.7 37

26 27

28

30 Only the with the cyclopropyl group in endo-anti posi­ tion shows an assisted ionization. The intermediate involved has been interpreted in terms of a nonclassical trishomoaro­ matic cation (26). However, the products of acetolysis were rearranged completely {27). Therefore, the alternative pair of classical ions (28) could not be excluded unequivocally 2 • Similar observations were made for the four-membered ring. However, the rate enhancements observed are smaller than for the cyclopropyl compounds, whereas the product mixtures are T?b

even more complex3,4. When solely rearranged products are observed, it is dif­ ficult to decide whether enhanced solvolytic rates arise from electronic factors or from factors, associated with possible low-energy routes prone to skeletal rearrangements. A charge stabilizing group, in particular methoxy, attached to the cationic center, has proved to be capable (Chapter 2) to suppress cationic rearrangements. In view of these observations ionization reactions of 7-chloro-7-methoxy derivatives of nor­ bornene with appropriately positioned four-membered carbocyclic rings (compounds 29-32) have been studied. The parent cations of the a-chloro ethers were generated as long-lived stable species by using polar solvents (e.g. liquid S02), Bronsted acids (e.g. H or HS0 F), Lewis acids (e.g. SbF , SbC1 or 2so 4 3 5 5 ) AlC1 3 or mixtures of these. Furthermore, the ionization reactions under short-life conditions were studied quantitati­ vely by monitoring the reaction of the a chloro ethers with pyridine in methylene chloride.

31 3. 2. Synthesis

The ketals 29a and 30a could be from ketone 33, which was readily available according to the method of Ant­ kowiak and Shechters.

0 II 3 CH'))H3 3 £b CH30H, BF3 hV CH ~H 33 34 29a j

l"'!''''

0 II 3 3 CH'))H ;}) CH30H,BF3 ~~ hv CH~H 35 36 30a j

Scheme II

Treatment of 33 with methanol in the presence of produced the triene ketal 34. Irradiation of 34 in ether through Quartz, using a high-pressure mercury lamp, yielded ketal 29a. Selective of 33 on NiB 26 produced compound 35, which was similarly converted to the monoene ketal 30a (Scheme II). Hydrogenation of compound 37, prepared according to the method of Anderson et aZ. 7 on Pd-C, yielded 38. Dechlorination, using metallic sodium in THF-t­ butyl alcohol 8 , gave tfie monoene ketal 31a. Similarly, the benzocyclobutene compound 32a was prepared from the Diels­ Alder adduct 399, obtained by the generation in situ of benzo­ cyclobutene in the presence of 5,5 dimethoxy-1,2,3,4-tetra­ chlorocyclopentadiene (Scheme III). 32 3 CH~OOCH3 CH~OOCHCl Cl Cl H2fPti-C Cl Najt-BuOH f. Cl Cl # Cl Cl 37 38

~~~'·©a Cl Cl

Scheme III

The ketals 29a-32a were converted to their corresponding a­ chloro ethers 29b,c-32b,c by PC1 510, The PMR spectra (Table XI) revealed the existence of two isomers by the occurrence R~~ R~~ R~~ ~ ~ ~

R1=Cl, R2=0CH 3 : b R ;0CH , R ;Cl: c 1 3 2 of two singlet methyl resonances. The ratio of epimers of the a-chloro. ethers is specific to each compound (Table IX). The epimers are in equilibrium as revealed by their interconversion in more polar solvents (liquid , ), similar to the so 2 cn 2c1 2 behaviour found for 7-chloro-7-methoxynorbornene (Chapter 2), and consequently are inseparable.

33 3.3. Generation and properties of cations derived from 9-chZoro-9-methoxy­ endo-tricycZo[4.2.1.02'5]nonanes

Compounds 29, the most unsaturated in the series, exhibit unusual reactivities with respect to 30 32. Dissolution of 29b,c in liquid so2 at -60° c resulted in the ionization of the C-Cl bond, as revealed by the NMR spectra (vide infra), whereas the cations derived from 30b,c-32b,c could be obtained in strong proton acids like FS03H only. Although addition of

Liq,S02

40

29QC 40 29b,c

9 8 7 6 5 4 3 ppm F 2 PMR spectra of 9-chloro-9-methoxytricyclo[4.2.1.o 2 •~- nona-3,7-diene in liquid so 2 at various temperatures

34 methanol to 29b,c in liquid so2 gives exclusively the starting ketal 29a, the structure of the intervening cation appeared to be rearranged completely to 40. Moreover, cation 40 is in equilibrium with its covalent precursors 29b,c, as indicated by the temperature dependent PMR spectrum (Figure 2), which shows an increasing amount of 29b,c at higher temperatures. The decrease in dipole moment of liquid so2 at increasing temperatures may account for this phenomenonll. When FS03H was added to the solution, 40 was observed only even at tem­ peratures up to 10° C. However, in this medium two isomers of 40 were present (40a,b, vide infra), which apparently inter­ change in pure liquid so , in which only one structure could 2 be observed. Structure 40 was also formed in aprotic media. Addition of AlC1 3 to 29b,c in CD 2cr 2 gave rise to 40, which was stable up to 20° C. The assignment of rearranged structure 40 is based on the NMR data, compared to those of the closely related structure 41. The latter compound was prepared by dis­ solution of its corresponding a-chloro ether in FS0 H-so . 3 2 The NMR spectral data are collected in Table VI. The PMR data of 40 in FS0 3H-S02 show two singlet methyl re­ sonances in a 1:4 ratio whose mean position (o 4.76) is 1.4 ppm downfield relative to the mean methoxyl position(o 3.38) in the covalent precursors 29b,c. Such shifts are typical for methoxy groups attached to cationic centers. The small separation (0.04 ppm) between the two methoxy peaks indicates existence of syn (40a) and anti (40b) forms, due to restricted rotation around the C-0 bond. This interpretation is confirmed by the CMR spectrum which shows two signals for each carbon in a ratio of 1:4. Furthermore, the PMR spectrum of both isomers of 40 shows comparable resonances at o 9.10, 7.57 and 8.87, 7.30 which are ~onsistent with methoxy stabilized allylic systems. The mutual coupling of 5 Hz indicates that the allylic system forms part of a five-membered ringl2 , Ion 41, which occurs similarly in two isomeric forms 41a and 41b in a ratio of 1:2, shows analogous spectral data (Table VI). In contrast with 40, the PMR data of ions derived from 30, 31

35 (,) (J) Table VI PMR and CMR spectral datl for ions 40a, b and 41a, bin FS0 H-so ( 8 values) 3 2

H3 C "- a b c d e g h 0 PMR 8. 87 7.30 6.00 3.9-4.2 4.2-4.5 4. 78 (s) b (dd,J=S, 2) (d,J=S) (br m) (br m) (br m)

d CMR 191.93 138.26 226.22 134.31 138.26 39.80 46.9 67.34 ~h f 40a 39.61 45.4

O...----CH3

PMR 9.10 7.57 6.00 3.9-4.2 4.2-4.5 4. 74 (s) (dd,J=5,2) (d,J=S) (br m) (br m) (br m) ~ CMR 196.33 134.28 225.44 135.29 136.86 40.61 46.4 67.34 40b 42.4 47.2

PMR 9.13 7.25 6.19 (m) 3.51(brm) 4.25 (br m) 2. 25 (br s) 4. 78 (s)

(dd,J=5,2) (d1 J=5) CMR 201.83 130.73 227.37 131.22 133.28 45,0 53.14 54.11 54.96 47.62 67.88 41a

PMR 8.89 6.93 ( dd, J=S, 2) ( d, J=5) 6.12 (m) 3.61 (br m) 4.05 (br m) 2.35(br s) 4.97 (s) 0~~ : ~ •+ .... CMR 196.37 133.95 228.83 129.83 134.86 45.62 52.72 55.26 55.81 47.32 67.88 41b •1 values are expressed in Hz. and 32 are consistent with the unrearranged structures 42, 43 and 44. The data are collected in Table VII, together with those of the 7-methoxy-7-norbornenyl cation (25}.

2 5 Table VII PMR spectral dat~ for 9-methoxy-9-tricyclo[4. 2. 1. o • ]nonyl carbenium ions in F,SO H-SO 3 2 ( 8 values)

H H H Other 1,6 2,5 3,4

3. 30 (m) 3. 30 (m) 6.50 (s) 2.11-1.56(m) 4.95 (s) 3.67 (m)

;~ 3.50 (m) 3.10 (m) 2.17 (m,ex) 6.83 4,80 (s) ~ 3.73 (m) 1.50 (m,e!!) (t,J=2.5) pcH3 +:

4.27 (m) 4.20 6.45 4. 90 (s) 7. 33 4,03 (m) (d,j=4.5) (t,J=2.5) (m, aromat)

H H H OCH 1,4 2,3 5,6 3

J;, 3.83(m) 2.23 (m,ex) 6.82 4.67 (s) 6 2 3. 53 (m) 1. 78 (m,e!!) (t,J=2,S) 25

• J values are expressed in Hz.

All spectra reveal the nonequivalence of the bridgehead positions, due to the restricted rotation around the C-0 bond 13 • The chemical shifts of the olefinic or cyclobutenyl protons of ions 42-44 do not indicate any charge delocalization onto these

37 positions. Apparently, the charge-stabiliz ability of the methoxy gr~up overwhelms the potential available " participa- tion of the double bond or the cr partie ion of the four- membered carbocyclic rings.

2 5 3.4. Re~otion products from 9-ohloro-9-methoxy-endo-trioyolo[4.2.1.o • ]- nonanes

Compounds 29b,c yield products in nucl lie substitu- tion reactions with an unrearranged structure. The reaction of 29b,c with methanol afforded the starting ketal 29a, whereas the reaction with pyridine in CH 2c1 2 at 0° C gave rise to a mixture of epimeric pyridinium salts 29d and 29e. At room temperature the isomer, with the pyridinium group syn to the

C 5H~H 3

29d j

C5H5N ~H, + Ll j CH~5 H 5

29e j cyclobutene moiety (29e), isomerizes to the anti epimer (29d). A similar reaction pattern was tound for the reference com­ pounds 30b,c-32b,c. In contrast with the previous behaviour of 29b,c, the pyridinium salt of rearranged structure 40 (40c) was observed when the reaction was performed in CH 2c1 2 with AlC1 3. Similar structures were obtained with other nucleophiles like (CH 3) under nonequilibrium conditions. These compounds were character- ized by their NMR data, as summarized in Table VIII.

38 Table VIII PMR and CMR spectral data• for tricyclo [ 4.3.0.0. 2, 5] nona-3,7- ( ova"' I ues)

X 3,4 7 8 9 1,6 2,5 OCH Other 3 + c H N PMR 6.16 (m) 5.10 5.53 3.6-3.9 3.3-3.6 3.77 7.98-9.03 6 5 6.50(m) ( d,J=2. 5) (t,J=2. 5) (br m) (br m) (aromat) + (CH ) S PMR 6.10(m) 4.95 4.60 3.3-3.7 3.82 2. 66(s, CH ) 3 2 3 6.35(m) (d,J=2.5) (t,J=2.5) (brm) CMR 140.31 153.14 90.92 63.30 34.47 46.00 58.30 28.35 (CH ) 3 137.70 44.09 44.09

•J values are expressed in Hz.

However, reaction with methanol afforded the starting ketal 29a again. This compound was also recovered after dissolution of 40c in methanol. The use of liquid so 2 instead of CH 2cl2- A1Cl3 gave similar results. The structure of the reaction

~H,

29b,c 40c

products of the reference compounds 30b,c-32b,c appeared to be independent of the reaction conditions used. In all cases

39 the unrearranged structures were observed.

3.5. Kinetics of the reaction of pyridine with various 9-chloro-9-methoxy- 2 ,:) [4.2.1.0 ~ ]nonanes

In order to study the reactivity of 29b,c under conditions of short life which proceed without skeletal rearrangements, the reaGtion of the a-chloro ethers of structures 29, 31, 32 and norbornene (23) with pyridine in CH 2c1 2 has been studied );H'

23b,c

quantitatively. The reactions were monitored by PMR spectro­ scopy. Under the solvolytic conditions used, all reactions afforded an epimeric mixture of pyridinium salts. The ratio's of epimers of the starting a-chloro ethers and those of the products were independent of time. They are collected in Table IX.

Table IX Composition of epimeric mixtures of starting a-chloro ethers (b, c) and their reaction products ( d, e) with pyridine ( %)

Compound b c d e

29 92 8 63 37

31 72 28 35 65

32 69 31 69 31

23 76 24 33 67

In all cases, the epimeric ratio's of the products differ from those of the starting a-chloro ethers. Thus, the reactions proceed with racemization. Furthermore, the reaction rates appeared to be proportional to the pyridine concentration, 40 which indicates second-order kinetics. The second-order rate constants, together with some activation parameters, are summarized in Table X.

Table X Rate data for the reaction of pyridine with various 9-chloro-9-methoxytricyclo[ 4. 2.1] nonanes

0 4 Compound Temp. C k x10 rel.k LlHf Llst 2 -1 -1 -1 I. mol sec ked mol eu

29b,c -11.0 9.23 20.0 108 45 11.4 -26.6

31b,c 20.0 19.3 8

32b,c 20.0 14.2 5.9 23b,c 20.0 2.4 35.6 7.3 12.0 -24.0

The large negative entropy of activation, observed for corn­ pounds 29 and 23 is in accord with a second-order process and similar to other "Menschutkin" reactions 14 • As can be seen from the data in Table X the reaction rate of 29b,c is enhanced with respect to the norbornyl system 23. The rates of endo­ fused cyclobutane and benzocyclobutene systems are enhanced to a lesser extent. Thus, the order of reactivities in nucleo­ philic displacement reactions is analogous to that observed under low-nucleophilic conditions.

3. 6. Dt~saussion

A. The stability of the 7-rnethoxy-7-tricyclo ~.3.o.o 2 • 5 ]­ nona-3,9-dienyl cation

The relative ease with which the C-CI bond of compounds 29b,c are ionized with respect to the reference compounds 30b,c-32b,c, may be attributed to the stability of the ulti­ mate ion 40, as revealed by its existence under relatively mild conditions (Section 3.3). The stability of 40, with respect to the unrearranged cationic structure 29+, evidently arises from the energy of the allylic system.

41 40

Probably, in 29+ no resonance stabilization is present in view of the results of the closely related ions 42, 43 and 44. The PMR data of the latter compounds did not reveal any charge delocalization (Section 3.3). However, the gain in resonance energy of 40 with respect to 29+ is reduced by the increased ring strain of the ring contracted structure . The difference in strain energy is probably in the order of 14 kcal/mol. This figure is the difference in standard free energy of bicyclo- ~.2.~ heptane and bicyclo[3.2.~ heptane systemsl5. The estim­ ated resonance energy of an allylic system is 60 kcal/moll6, Thus, compound 40 will be stabilized substantially with . + respect to t h e unrearrange d cat1on 29 . The existence of 40, under relatively mild conditions (liquid so 2), suggests an enhanced stabilization with respect to other comparable allylic systems. The related structure 41 does show quite different properties. Its preparation could only be achieved under strongly acidic conditions, such as

41

FS0 3H-S0 2. Comparison of the NMR data of 40 with those of 41 (Table VI) may indicate some additional charge stabilization of the allylic system in 40. The differences in the allylic resonances (65) in the PMR spectra of both isomers of 40 (1.53 ppm for both) are s ficantly smaller than the comparable values for both isomers of 41 (1.88 and 1.96 ppm). The CMR

42 values for these positions show a similar trend (Table VI). The charge may be delocalized by interaction with the cycle- double bond. This of interaction has been proposed by Olah in ion 46 to account for the fluorosulfonation of the c8-c 9 double bond of the protonated ketone 45,

HQ~·: FS03H H3C~······.. . . •+ : + ••••• •••• 45 ...... '..j~ whereas the isolated double bond in 46, with the less stabil izing methyl group at the allylic center, remained unaffect­ edl7. This kind of charge delocalization would be favoured in 40 relative to 41, in consequence of the shorter distance be­ tween the double bond and the allylic cation. However, the NMR parameters for these positions in both ions do not give any indication fa.£ this phenomenon. Moreover, it is highly unlikely that this kind of interaction would result in a net stabilization of the system, because the interaction of an allylic system with the monoene unit is antiaromatic (4TI electrons). Another possibility of charge delocalization in 40 may proceed via the Walsh orbitals 3 of the adjacent cyclo­ butane ring. This may be indicated by the downfield shift of the cyclobutane protons in 40 (o 3.9-4.5), with respect to the comparable positions in 41 (o 3.5-4.25).

40

B. The structure of the intermediate ions

The reaction of pyridine (N) with 29b,c (RX) turned out

43 to be a second-order process with racemization. This kinetic behaviour can be rationalized by assuming the reaction of pyridine with an intermediate ion to be the rate determining step:

k1 k2 RX R+x- RN+ + X k_, N

k k [NJ 1 2 k obsd = k_, + kz [N] when k_ »kz (N), kobsd=k k (N)/k_ , so that the reaction is 1 1 2 1 first order in pyridine and second order overalll8. However, the occurrence of an intermediate ion seems to be contradicted by the ionization of 29b,c under nonequilibrium conditions, which gave rise to the rearranged structure 40. Appa.rently, structure 29+ does not exist as a free . Thus, it seems reasonable to assume that the intermediate ion, under solvolytic conditions, is not free but "encumber ed"1 9 • This behaviour can be best understood on the basis of Winstein's ion pair scheme20, According to this mechanism the ionization of an alkyl halide A proceeds through a series of progressively more dissociated intermediates:

RX :;;;;=.===:=:: R+ X";;::====~ R+ II X-+====~ R+ + X- A B C D an "intimate" or "internal" ion pair B, a "solvent separated" or "external" ion C and a dissociated cation and gegen- ion D. Since the reaction of pyridine with compounds 29b,c in CH 2c1 2 proceeds with racemization, it can not proceed through stage B. The close association of the anion X in the intimate ion pair should effectively block nucleophilic attack from that side. Consequently, attack of nucleophiles at this stage would be expected to afford products of inverted configura­ tion. Because at stage D the rearranged structure 40 would arise, the reaction should proceed at stage C. The structural 44 dependence on the position of the gegenion X implicates that at those stages, in which R+ has its original structure, the charge at R+ has not completely developed, probqbly because there is some interaction between the cationic center of R+ and the gegenion x-21. Apparently, this observation applicates also to the external ion pair stage C, where racemization is possible. This conclusion is consistent with the equilibration of 29b,c and 40 in liquid so 2 and the formation of the starting

R'+ + X IV RX+==~R+ II X-+====~ R'+llx_/

II Ill'~ R'X v

40

II a

lila

.1GT G ___ _t!lla Tt 111 .1Gv v

LlGv-1 ------_j_

Reaction co-ordinate

Figure 3 Mechanism and corresponding free energy diagram for the ionization of 9-chloro-9-methoxytricyclo­ ~.2.1.o2•5]nona-3,7-diene

45 structure 29 after quenching ion 40 with methanol. When nu­ cleophiles, such as C and CH 30H, approach ion 40 in a thermo­ dynamically controlled reaction, the rearrangement to its initial structure 29 precedes its collapse with the nucleo­ phile. When stronger nucleophiles were used, such as pyri­ dine, ion 40 was trapped prior to rearrangement. Apparently, this reaction is kinetically controlled. The reaction pattern is visualized in Figure 3. Quenching of ion 40 at stage IV gives rise to the initial formation of ion pair III which can collapse to R'X (V). Alternatively, 40 in III may rearrange to structure 29+ re­ sulting in the formation of ion pair II, which subsequently collapses to RX. The free energy of activation ~Gtiiia of the former process, an association of ions, will probably be small­ er than that (6GTIII) of the latter one, a rearrangement re­ action (Figure 3). Therefore, in the first instance R'X (V) will be formed predominantly. However, when X is a weak nu­ cleophile, such as Cl and CH 30H in liquid so 2, R'X will dis­ sociate again. This process will lead to the ultimate formation of RX (I), which is thermodynamically most stable (6Gv-I=14.5 kcal/mol, see Section 3.5 A). When X is a stronger nucleo­ phile, such as pyridine, the dissociation of the initially formed compound R'X (V) does not take place and the kinetical­ ly controlled product (40c) is observed. Apparently, the interconversion of ion pairs II and III is induced by a balancing of the allylic resonance energy of 40 versus'its strain energy. At stage III the allylic reso­ nance is encumbered by the presence of the anion (vide supra). The release of ring strain after ring expansion to 29+ will now overbear the delocalization energy of the encumbered allylic resonance, which results in a conversion of III to II. The interconversion of II and III even at temperatures down to -80° C, suggests a lowered transition state. If one of the double bonds of 29 is saturated like in 30 and 31 or forms part of an aromatic system ~32), rearrangement reactions are not observed. Apparently, the two isolated double bonds are

46 + involved in the conversion of 29 to 40. These observations are in accord with several mechanisms which account for the observed path of rearrangement. One of these mechanisms is the concerted [3,3] sigmatropic shift ("Cope rearrangement") 2 2 with one transition state, observed in neutral molecules such as semibullvalene (47)23.

47

Alternatively, the reaction may proceed through Woodward and Katz's intermediate 24 which has been postponed to account for the rearrangement of dicyclopentadiene-1-keton 48. This re­ arrangement which proceeds through two transition states in-

0 ~~ A4~ ~48 49 l(:~ 0 volves fragmentation of the c,-c2 bond to yield intermediate 49 which may reclose to give either 48 or SO. The Cope re­ arrangement can be excluded, based on the observation of the kinetically controlled product 40c. This structure reveals that cation 40 is captured at the 9 position. At ion pair stage III (Figure 3), in which the nucleophile resides close to the

40 51

9 position, a Cope rearrangement would result in the formation of structure 51. The absence of the latter compound indicates

47 that this type of rearrangement does not occur. The involve­ ment of intermediate 52, in which the interaction of the allylic system with the butadiene unit results in a 6n elec­ tron homoaromatic system, accounts for all the observations. A rupture of the c1-c 2 bond in 4~ yields intermediate 52.

pcH CI-J: 3

crtio 1 2 ~ 52 40 II Ill Subsequently, the approach of a chloride ion, according to path a, produces ion pair II, whereas path b affords ion pair III. These processes depend clearly on the stability of the intermediate 52. When the double bond in the cyclobutene ring forms part of an aromatic system (32), no rearrangement re­ actions were observed. This may be due to a less stabiliz interaction of the resulting benzylic system with the butadiene moiety in the intermediate 53. Otherwise the rearrangement mechanism which has been encountered in 29, will lead to an untenable strained olefin (54), which furthermore has sacri­ ficed the initial aromatic ring.

53 54

The fragmentation of the c1-c 2 bond in compounds 29b,c to produce a stabilized intermediate 52, may provide anchi meric assistance to the ionization of the C-Cl bond, even under solvolytic conditions by which no rearranged structures were observed. The enhanced rate of the reaction between pyridine and compounds 29b,c with respect to the analogous

48 5 3 ~H, c5H5NC H~H CIJ)' f.. -<} j ' ', 29b,c 55 29d,e reaction of the reference compounds 31-, 32- and 23b,c, indeed suggests that the ionization of the C-Cl bond in 29b,c is as­ sisted by an electron delocalization process as depicted in 55. This kind of participation does not imply a stereoselective reaction. Therefore, the observed epimeric ~ixtures 29b,c and 29d,e are consistent with this mechanism.

49 Table XI o PMR spectral dat.l' for bi- and tricyclic compounds in co c1 ( 8 values) 2 2

8

H H OC:H Other R1 R2 Ht 6 H3,4 ' 2,5 7,8 3 a OC:H3 OC:H3 2.66 (m) 2.98 5.60 (s) 5.67 2.98 (s) (d,J=4) (t,J=2.3) 3.10 (s)

b Cl 3.10 (m) 3.16 5.86 (s) 5.93 3.43 (s) (d,J=4) (t,J=2. 3)

c OC:H3 Cl 3.10(m) 3.16 5.82 (s) 5.95 3.33 (s) (d,J=4) (t,J=2.3)

d c H N+ OC:H 4.13 (m) 3.46 5.93 (s) 6.00 3,23 (s) 8.30 (m,3H) 6 5 3 (d,J=4) (t,J=2. 3) 9.83 (d,J=7,2H)

e OC:H3 C H N+ 4. 37 (m) 2.83 5.80 (s) 6.17 3.27 (s) 8.30 (m, 3H) 6 5 (d,J=4) (t,J=2.3) 10.23 (d,J=7, 2H)

~0 a OC:H 2.08 (m) 3.00 (m) 6.16 (s) 1.60 (m,ex)3.20 (s) OC:H3 3 1.22 (m, en)

b Cl OC:H3 2.41 (m) 3;00 (m) 6.08 (s) 1.60 (m,ex)3.33 ~s) 1.38 (m, en) c Cl ~~ ~

a OC:H OC:H 2,73 (m) 2.73 (m)1.83 (m,ex) 6. 30 3.03 (s) 3 3 1.23 (m,en) (t, J=2. 3)

b Cl oc:H 3.00 (m) 2.83 (m) 2.00 (m,ex) 6.42 3.23 (s) 3 1. 33 (m, en) ( t, J=2. 3)

50 Table XI (continued)

Rl R2 H OCH Other H1,6 2,5 H3 •4 H7,8 3 c OCH Cl 3.00(m) 2.83 (m) 2.00 (m,ex) 6.42 3. 30 (s) 3 1.33 (m,en) (]=2. 3) d c H N+ OCH 4.06 (m) 3.17 (m) 2.17 (m,ex) 6.43 3.08 (s) 8.26 (m, 3H) 6 5 3 1.43 (m, en) (t,J=2.3) 9.73 (d,J=7,2H) + e OCH c H N 4. 30 (m) 2.83 (m) 2.46 (m,ex) 6.63 3.27 (s) 8.26 (m,3H) 3 6 5 2.43 (m,en) ( t, J=2. 3) 10.00 (d,J=7,2H)

a OCH OCH 3.15 (m) 3.73 5.56 3.03 (s) 7.02 (m,4H) 3 3 (d,J=4) (t,J=2.3) 3.15 (s) b Cl OCH 3. 33 (m) 3.81 5.64 3.43 (s) 7.12 (m,4H) 3 (d,J=4) (t,J=2.3) c OCH Cl 3.33 (m) 4,02 5.64 3.30 (s) 7.12 (m,4H) 3 (d,J=4) (t,J=2. 3) d c H N+ OCH 4,20 (m) 4.10 5.68 3.25 (s) 7.05 (m,4H) 6 5 3 ( d,J=4) (t,J=2.3) 8.45 (m,3H) 9.55 (d,J=7,2H) e OCH c H N+ 4.45 (m) 3.51 5.85 3,45 (s) 6.96 (m,4H) 3 6 5 ( d, J=4) (t,J=2.3) 8.45 (m,3H) 10.00 (d,J=7,2H)

5 :;s1 2 6 23

H H OCH Other R1 R2 H2 3 1,4 5,6 3 ' a OCH OCH 5,96 2.74(m) 1.90 (m, ex) 3.15 (s) 3 3 (t, J=2. S) 1.05 (m,en) 3.11 (s)

b Cl OCH 5.95 2.93 (m) 1. 85 (m, ex) 3.32 (s) 3 (t, J=2. 5) 1.05 (m,en)

51 Table XI (continued)

H H OCH Other Rl R2 H2 •3 1,4 5,6 3

c OCH Cl 5.92 2.93 (m) 1.85 (m,e.x) 3.25 (s) 3 ( t, J=2. 5) 1.05 (m,en)

d C H N+OCH 6.14 4.07 (m) 2.20 (m,e.x) 3.20 (s) 9.14-8,30 (m,3H) 6 5 3 (t,J=2. 5) 1.15 (m, eo) 10.00 (d,J=7,2H)

e OCH C H N+ 6.34 4.34 (m) 1. 2-1. s 3.26 (s) 9.14-8, 30 (m, 3H) 3 6 s (t1 J=2.S) (br m,ex-en) 10.37 (d, J=7, 2H)

*J values are expressed in Hz.

3.7. Experimental

• 9,9-Dimethoxybicyclo [4.2.1] nona-2,4,7-triene (34) Boron trifluoride etherate (0.5 ml, 48%) was added to a solu­ tion of trienone 33 (1.0 g, 7.5 mmol)S in methanol (25 ml) at 0°. The mixture was allowed to stand overnight in a refrige­ rator. After neutralization with saturated sodium bicarbonate at 0°, the solution was extracted with ether. The ether layer was washed.with saturated sodium chloride, dried (MgS0 4) and concentrated. The product was purified by chromatography on a silica gel column, eluting with ether-hexane mixtutes to give the dimethyl ketal 34 (0.95 g, 76%): &TMSCDC1 3 5.92 (m,4H), 5.21 (d,J=1.S Hz,2H), 3.12 (s). 3.07 (s) and 3.12 (br).

• 9,9 Dimethoxy-enda-tricyclo [4.2.1.o 2 • 5]nona-3,7-diene (29a) A solution of ketal 34 (1.0 g, 5.6 mmol) in ether was purged with nitrogen and irradiated at room temperature for 4 hr with an SP 500-W Philips high-pressure mercury lamp through Quartz. A positive nitrogen pressure was maintained above the solution during the photolysis. The solution was filtered and concen­ trated. The crude product was chromatographed on a column of

52 aluminum oxide, which was eluted with ether-hexane mixtures to give 29a (0.80 g, 80%): oTMS CDC1 3 5.87 (t,J=2.5,2H), 5.80 (s,2H), 3.10 (s), 2.98 (s), 2.98 (d,J=4), 2.66 (m,2H).

• Bicyclo [4.2. ·11 nona-2,4-dien-9-one (35) Dienone 35 was prepared by hydrogenation of trienone 33 using a nickel boride catalyst, which was obtained according to the procedure of Brown6. To a solution of nickel acetate (320 mg) in ethanol (20 ml) under a hydrogen atmosphere was introduced a solution of sodium borohydride (45 mg) in ethanol (20 ml). Hydrogenation was initiated by injecting trienone 33 (1.38 g). After the absorption of hydrogen (235 ml, 15 min) the reaction was stopped. The reaction mixture was poured into an aqueous solution of sodium bicarbonate (100 ml) and extracted with ether. The etheral solution was wa~hed with water, dried (MgS0 4) and concentrated. Analysis of the crude product with GLPC and PMR indicated 92% conversion to dienone 35. The mixture was chromatographed on a column of aluminum oxide with ether-hexane mixtures. A liquid product was isolated: oTMS CDC1 3 5.74 (m,4H), 2.67 (m,2H) and 2.20 (m,4H).

• 9, 9-Dimethoxybicyclo [4. 2. 1] nona-2, 4-diene (36) The ketalization of 35 was accomplished similar to the proce­ dure described for 33. The product was about 95% pure according to GLPC and PMR. The material was not purified furtheron, because of its easy conversion to ketone 35. NMR of 36: oTMS CDC1 3 5.93 (m,4H), 3.34 (s,6H), 2.75 (br s,2H) and 2.12 (m,4H).

• 9,9-Dimethoxy-endo-tricyclo[4.2.1.o2• 5] non-3-ene (30a) Compound 30a was prepared as described for 29a from 36 (1.0 g). Work-up afforded 30a (0.82 g):oTMS CDC1 3 6.16 (s,2H), 3.20 (s,6H), 3.00 (m,2H), 2.08 (m,2!-I) and 1.41 (br m,4H).

• 1,6,7,8-Tetrachloro-9,9-dimethoxy-endo-tricyclo[4.2.1.02 •5]­ non-7-ene (38) Diene 37 (3.84 g)7 was dissolved in ethanol (40 ml) and hydro­ genated at room temperature in the presence of 10% palladium 53 on charcoal (0.29 g). The solution was filtered and concen­ trated. The product, mp 75-76°, was isolated by chromatography on silioagel with carbon tetrachloride as eluent: aTMS CDC1 3 3.50 (s,6H), 3.11 (m,2H) and 1.83 (m,4H).

• 9,9-Dimethoxy-endo-[4.2.1.o2 • 5]non-7-ene (31a) Compound 38 (3.5 g, 11 mmol) was dissolved in a mixture of t-butyl alcohol (10 g) and freshly distilled tetrahydrofuran (60 ml) under nitrogen8 • Finely cut sodium (7 g) was added with s~irring. The mixture was heated to initiate the reaction and refluxed gently with stirring for 8 hr. The reaction mixture was cooled to 0° and methanol (SO mi) was added to destroy the excess of sodium and poured into water (200 ml). The aqueous solution was extracted with ether and the combined extracts were washed with water, dried (MgS0 4) and concen­ trated. Chromatography on silica gel with mixtures of carbon tetrachloride and chloroform afforded 31a (1.3 g): oTMS CDC1 3 6.30 (t,J=2.5,2H), 3.03 (s,6H), 2.73 (m,4H), 1.83 (m,2H) and 1.23 (m,2H).

• 1,4,4a,8b-Tetrahydro-9,9-dimethoxy-1,4-methanobiphenylene (32a) This compound was prepared from 39 (4.0 g, 10.9 mmol)9 in an identical way as described for 31a. Work-up and distillation yielded 32a (2.24 g), bp 110/0.01, mp 77-78°: oTMS CDC1 3 7.02 (m,4H), 5.56(t,J=2.5,2H), 3.73 (d,J=4,2H), 3.15 (m,2H), 3.15 (s,3H) and 3.03 (s,3H).

• 9-Chloro-9-methoxy-endo-tricyclo [4.2.1.o2 • 5]nona-3,7-diene (29b,c) To a stirred solution of 29a (5.2 g, 0.029 mol) in S ml of diethyl ether was added phosphorus pentachloride (6 g, 0.030 mol) in small portions, at such a rate that the ether boiled gently. Sometimes heating and addition of some phosphorus oxychloride was required to initiate the reaction. Vacuum distillation provided 29b,c (2.8 g), bp 57-59° (0.01 mm): oTMS CDC1 3 5.95, 5.93 (t,J=2.5,2H), 5.86, 5.82 (s,2H), 3.43, 54 3.33 (s,2H), 3.16 (d,2H) and 3.10 (m,2H).

• 9-Chloro-9-methoxy-endo-tricyclo[4.2.1.o2 •5]non-3-ene (30b,c) Bp 63-65° (0.01): cTMS CDC1 3 6.08 (s,2H), 3.33 (s,3H), 2.41 (m,2H), 3.00 (m,2H), 1.50 (m,4H).

• 9-Chloro-9-methoxy-endo-tricyclo [4. 2. 1. 02 ' 5] non-7-ene (31b,c) Bp 66-68° (0.01): cTMS CDC1 3 6.42 (t,J=2.5,2H), 3.30, 3.23 (s,3H), 3.00 (m,2H), 2.83 (m,2H), 2.00 (m,2H) and 1.33 (m,2H).

• 1,4,4a,8b-Tetrahydro-9-methoxy-9-chloro-1,4-methanobiphe nylene (32b,c) Bp aa 95-98° (0.002 mm): oTMS CDC1 3 5.64 (t,J=2.5,2H), 7.12 (m,4H), 3.81, 4.02 (d,J=4,2H), 3.43 (s), 3.30 (s). 3.33 (m).

• Kinetic measurements Equimolar amounts of pyridine and u-chloro ether dis­ solved in co 2c1 2 were mixed at -80° in an NMR sample tube (0.4-1.2 M solution). The runs were performed at temperatures as indicated in Table X. The progress of the reaction was followed by integrating the PMR methoxy signals of the a­ chloro ether and the ortho protons of the pyridinium substitu­ ent of the products at appropriate intervals. The rate con­ stants were determined graphically.

55 References

1. Recent reviews: a. R.E. Leone and P. v. R. Schleyer, Angew. Chern., Int. Ed. Engl., ~. 860 (1970). b. "Carbonium Ions", Vol. 1-4, G.A. Olah and P. v. R. Schleyer, Ed., Wiley Interscience, New York, N.Y., 1968, 1970, 1972, 1973. c. J. Haywood-Farmer, Chern. Revs., 1±, 315 (1974). 2. J.S. Haywood-Farmer and R.E. Pincock, J. Amer. Chern. Soc., ~. 3020 (1969). 3. R. Hoffmann and R.B. Davidson, J. Amer. Chern. Soc., 93, 5699 (1971). 4. a. M. Sakai, A. Diaz and S. Winstein, J. Amer. Chern. Soc., 2_l, 4452 (1970). b. M.A. Battiste and J.W. Nebzydoski, J. Amer. Chern. Soc., 2_l, 4450 (1970). 5. T.A. Antkowiak, D.C. Sanders, G.B. Trimitsis, J.B. Press and H. Shechter, J. Amer. Chern. Soc., 94, 5366 (1972). 6. C.A. Brown, Chern. Comm., 600 (1969). 7. C.M. Anderson, J.B. Bremner, I.W. McCay and R.N. Warrener, Tetrahedron Lett., 1255 (1968). 8. P.G. Gassman and P.G. Pape, J. Org. Chern.,~. 160 (1964). 9. A.J. Boulton and J.F.W. McOmie, J. Chern. Soc., 2549 (1965). 10. P. Schipper, P.B.J. Drie~sen, J.W. de Haan and H.M. Buck, J. Amer. Chern. Soc.,~. 4706 (1974). 11. N.N. Lichtin, "Ionization and Dissociation Equilibria in Solution in Liquid Sulfur Dioxide", in Progress in Physical , Vol. 1, S.G. Cohen, A. Streitwieser Jr., and R.W. Taft, Eds., Interscience, New York, 1963, pp 75- 108. 56 12. a. G.A. Olah, Gao Liang and Y.K. Mo, J. Amer. Chern. Soc., ~. 3544 (1972). b. G.A. Olah, Y. Halpern, Y.K. Mo and Gao Liang, J. Amer.

Chern. Soc.,~. 3554 (1972). 13. R.K. Lustgarten, M. Brookhart and S. Winstein, Tetrahedron Lett., 141 (1971). 14. M.H. Abraham, Chern. Commun., 1307 (1969). 15. a. R.B. Turner, P. Goebel, B.J. Mallon, W. von E. Doering, J.F. Colburn Jr. and M. Pomerantz, J. Amer. Chern. Soc., 2Q, 4315 (1968). b. S. Chang, D. McNally, S. Shary-Tehrany, M.J. Hickey and R.H. Boyd, J. Amer. Chern. Soc.,~. 3109 (1970). 16. W.G. Woods, R.A. Carboni and J.D. Roberts, J. Amer. Chern. Soc.,~. 5653 (1956). 17. G.A. Olah, G.K. Surya Prakash and Gao Liang, J. Org. Chern., .±]_, 2820 ( 1976). 18. R.A. Sneen, Ace. Chern. Res., 6, 46 (1973). 19. a. R.H. Boyd, R.W. Taft, Jr., A.P. Wolf and D.R. Christ­

man, J. Amer. Chern. Soc.,~. 4729 (1960). b. J.T. Keating and P.S. Skell in "Carbonium Ions", Vol. 2, G.A. Olah and P. v. R. Schleyer, Ed., Wiley Inter­ science, New York, N.Y., 1970, Chapter 15. 20. S. Winstein, B. Appel, R. Baker and A. Diaz, Chern. Soc. (London), Spec. Pub., _11, 109 (1965). 21. D.J. Raber, J.M. Harris and P. v. R. Schleyer in ''Ions and Ion Pairs in Organic Reactions", Vol. 2, M. Szwarc, Ed., Wiley-Interscience, New York, N.Y., 1974, Chapter 3. 22. H. Levy and A.C. Cope, J. Amer. Chern. Soc.,.§_£, 1684 (1944). 23. A.K. Cheng, F.A.L. Anet, J. Mioduski and J. Meinwald, J. Amer. Chern. Soc., ~. 2887 (1974). 24. R.B. Woodward and T.J. Katz, Tetrahedron,~. 70 (1959).

57 CHAPTER 4

Antiaromatic interaction in the 9-methoxy-9-bicyclo[4.2.~nona-2,4,7 -trienyl cation

4. 1. Introduction

I~ the general introduction (Chapter 1), the concepts of homoaromaticity and bicycloaromaticity were presented. The theory of homoaromaticity was developed by Winstein 1 , as a result of his interest in the long-range stabilization of ionic centers by remote carbon. In 1967, Goldstein2 used MO symmetry arguments to extend the concept of homoaromatic ions (Zn bridges) to bicyclo­ aromatic ions (3n bridges). In this study, a fully unsaturat­ ed bridged bicyclic ion was predicted to be stabilized, if the sum of n electrons, provided by the unique odd bridge and one even bridge, equals 4n+2. If the even bridges differ in length, the longer one dominates since its HOMO and LUMO are more proximate to the nonbonding level. Furthermore, a stabil­ ized ion was depicted as bicycloaromatic, if the total number of n electrons available from its three bridges, equals 4n. In this case, an isoelectronic bishomoconjugated reference compound would be destabilized. Thus, the 7-norbornadienyl cation (9) was predicted to be stabilized and bicycloaromatic, since its planar isoelectronic reference compound 56 is de­ stabilized according to Htickel's 4n+2 n-electron prerequisite

Q) + 9 56

58 for cyclic stabilization. On the contrary, the 9-bicyclo­ [4.2.1]nona-2,4,7-trienyl cation (57) was predicted to be destabilized, interaction of the cationic center with the bu­ tadiene bridge dominate~ and antibicycloaromatic since its pericyclic reference compound 58 is aromatic (6n electrons).

co+ 57 58

Subsequently3, a study on topology and provided a prerequisite for the stability of longicyclic ions, accord­ ing to which stabilization was achieved if the number of stabili interactions exceeds the number of those that are destabilizing. All interactions are supposed to be of equal importance. According to this concept the [4.2.1] cation 57 was predicted to be a stabilized species, since two of its interactions are stabilizing (monoene-butadiene and monoene­ cationic center), whereas only one is destabilizing (butadiene­ cationic center). In recent years, a number of cationic studies on the [4.2.1] system has been reported4 • However, attempts to gene­ rate the 9-bicyclo[4.2.i]nona-2,4,7-trieny1 cation (57) re­ sulted in deep-seated rearrangement and elimination. Thus, the solvolysis of syn-9-bicyclo[4.2.1]nona-2,4,7-trienyl tosylate (59, Z=H) afforded exo-dihydroindenyl acetate (exo-63) and indene (64) 4 • The 9-deuterated analogue 59 (Z=D) produced exo- 4 63 and 64 with effectively all of the deuterium at c2 b,c, This result eliminates the possibility of a simple 1,2 shift of 1 and 6 to the c9 cationic center. Interaction of the monoene and diene units, visualized by structures 60 and 61, has been suggested to account for the observed path of rearrangement. The introduction of electron-donating groups at the position of 59 (Z=Ph, p-An) did not affect the c9 path of rearrangement 4d. However, in these cases not only the eis,exo-dihydroindenyl derivatives (exo-63) were formed, but 59 . - (D-. . z - 59 b-l60 61 62

- (:Q'-z + (X}-z + QO-z exo-63 OAc endo-63 OAc 64

Z=H,Ph,p-An X=OTs,OPNB also the ais,endo-compounds. The singleness of the product formed, if Z=H, was interpreted as indicating the intermediacy of a bishomotropylium ion 62 (Z=H). This homoaromatic ion is known to yield only cis~exo-dihydroindenyl products with stereo­ electronic control 5 • In contrast, ion 62 (Z=Ph, p-An) is an open allylic ion which can be captured both exo and endo 4d. Supporting evidence for the bicycloaromaticity of the [4.2.1] cation (57) seemed to be provided by kinetic studies of the solvolysis of 9-bicyclo[4.2.1] nona-2,4,7-trienyl tosy­ late (59-0Ts, Z=H), which turned out to be much more reactive than the more saturated analogues 65 and 66 4c,d.

OTs

'I h ~ 59 65 66 4 k rel 2. 1 o 2

However, Kirmse and Voigt4e have dismissed n participation in the solvolysis of syn-59, based on the observation that the

60 rates of solvolysis of syn-59 and anti-59 agree within experi­ mental error. Therefore, the rate-determining formation of

tetracyclic tosylates 67 via an intramolecular ~iels-Alder reaction was proposed. Obviously, the rate enhancement for 59 !:Ts~ • syn-59

T~ 61 62

anti-59 anti-67

does not reflect the rate-d~termining formation of a bicyclo­ aromatic cation 60, but rather indicates the availability of a low-energy route of 59 for its conversion to aia-dihydro­ indenyl derivatives. In this case, the rate-determining step is controlled by the transition state of the skeletal re­ arrangement, which occurs earlier on the reaction coordinate than the formation of a [4.2.1] cation. Therefore, the nature of this species remains obscure. The methoxy group, attached to the cationic center, has proved to be successful in suppressing skeletal rearrangements (Chapters 2 and 3). Hence, ionization reactions of 9-chloro- 9-methoxy-bicyclo [4.2.1) nona-2,4,7-triene (68) have been studied, together with the more hydrogenated analogues 69 and 70.

;6H''\ ;t)H' :EH' ....-::

68 69 70

61 4.2. Reaction products of 9-ehloro-9-methoxybieyeZo[4.2.1]nonanes

The a-chloro ethers 68, 69 and 70 were prepared from their corresponding dimethyl ketals according to the method described previously. In contrast to the a-chloro ethers mentioned in Chapters 2 and 3 which consisted of a mixture of epimers, only one epimer of 68, 69 and 70 was observed. Stereoselective re­ actions of the 9-bicyclo[4.2.1]nona-2,4,7-trienyl system have been encountered in the reduction of trienone 33 with agents such as phenyllithium and sodium borohydride, which afforded the syn-epimers 71 and 72 only4a.

f.. '\ ? ;s~ fb ~ 71 33 N~ '\ 26~ 72

The of 71 was assigned on the basis of kinetic steric control, in which phenyllithium attacks the of 33 from its monoene side rather than across the buta­ diene bridge. Molecular models revealed that in trienone 33 the c7 and c8 double bond is nearly coplanar with the c1 , and bridge. This bridge is about 60° out of plane of c6 c9 the strained planar butadiene moiety. Thus, the entrance from the monoene side is relatively more favoureds. On a similar basis the stereochemistry of the structures 68, 69 and 70 was assigned. Supporting evidence was provided by the highly stereoselective reactions of the a-chloro ethers (vide infra). Reaction of compounds 68, 69 and 70 with nucleophiles

62 proceeds without skeletal rearrangements. Reduction of 5g with lithium aluminum hydride produced the protic compound 73. The

;t)H'"\ .....-.: ,I 73 + Ph;ESH 3 PPh3 "\ Cl- LiAIH4 ~ .....-.: 68

75 Scheme IV stereochemistry of 73 can be assigned unambiguously on the basis of its PMR spectrum. If H9 is syn-disposed with respect to the monoene bridge, its resonance signal exhibits a triplet (J=6 Hz) due to coupling with the bridgehead protons, while this coupling is absent in the anti-epimer4a • Reaction of 68 with triphenylphosphine in so2 provided the phosphonium salt 74. Its configuration was established on the basis of the coup- ling of phosphorus with the bridgehead protons (JPH 14 Hz) 1 '6 and furthermore the shielding of the monoene protons and the deshielding of the butadiene protons. Similar shielding effects of the triphenylphosphonium group have been encountered in the norbornene series (Table V), When phosphonium salt 74 was re­ duced with lithium aluminum hydride, compound 73 was recover­ ed. Thus, substitution proceeds with retention of configuration. A similar reaction pattern was observed for nucleophilic sub­ stitution with pyridine and subsequent reduction with lithium

63 aluminum hydride (Scheme IV). This stereoselect was es- tablished also for the more hydrogenated analogues 69 and 70. (Table XIII). Apparently, steric and electronic factors co­ operate in these series. Compound 68 appeared to be thermolabile. While no re­ arranged products were observed under solvolytic conditions, the indene derivatives 76 and 77 were produced quantitatively on heat . This epimeric mixture is similar to that found 4 d for 9-aryl 9-b lo [4.2.1]nona-2,4,7-trienyl derivatives, consequent on the intermediacy of an open allylic dihydro­ indenyl cation 62-0CH 3 (Section 4.1). CQ-ocH,

76 Cl + 68 CQ-ocH,

77 Cl

4.3. Kinetics of the reaction of pyridine with 9-chZoro-9-methoxybicyoZo­ [4. 2. 1] non'anes

The progress of the reactions of compounds 68, 69 and 70 with pyridine in methylene chloride was followed by integrating the PMR methoxy resonances of the a-chloro ethers and those of the reaction products. Correct second-order plots were obtained in the range of 5-50% conversion. The rate constants are listed in Table XII. Remarkably, the unsaturated compounds are con­ siderably less reactive than the more hydrogenated analogues. The presence of the butadiene moiety in 69 results in a rate -4 ', retardation of 7.6x10 as compared with the fully saturated analogue 70, whereas the fully unsaturated compound 68 is the

64 Table XII Rate constants for the reaction of pyridine with various 9-ch!oro-9-methoxy bicyclo [ 4. 2. 1] nonanes in

-1 -1 -1 l.mol sec kcal mol eu

12.0 3.36 12.9 -28.6 31.0 14.98 -5 -7 -80.0(ext) 2. 5x10 4.5x10 68

-30.6 14.54 10.2 -28.8 -44.2 3.90 -2 ?6H' -4 -80.0(ext) 4. 2x10 7.6xl0 69

-80.0 55.2 7.2 -30.6 ;£)' -91.6 17.0 70

-7 least reactive in the series (k68;k7 4.5x10 ). The large negative values for the entropy of activation, which are con­ sistent with a second-order process (Section 3.5) of 68, 69 and 70, agree within experimental error. Thus, the differences in rate are reflected in the differences in enthalpy of acti­ vation, which is manifested in an increase in AH+ of 2.7 kcal/ mol for 68 relative to 69 and 3.0 kcal/mol for 69 as compared with 70.

4.4. Generation and properties of oations derived from 9-ohloro-9-methoxy­ bioyolo[4.2.1]nonanes and -bioyalo[4.4.1]undeaanes

Attempts to generate the 9-methoxy-9-bicyclo [4. 2. 'I) nona- 2,4,7-trienyl cation (78) resulted exclusively in rearrange­ ment and elimination. Treatment of a solution of 68 in liquid

65 Table XIII PMR spectral dad' for bicyclo[4.2.1]n:onanes in CDC1 ( 8 values) 3

R H OCH Others 6 2-5 3

OCH 3.15 (m) 5.95 (m) 5.20 3.12 ( s) 3 ( d,J=1. 5) 3.22 (s) 5 4 Cl 3.55 (m) 5.97 (m) 5.40 3.42 (s) ;6"' ( d,J=1.5) 8 2 + CHN 4.15 (m) 6.20 (m) 5.48 0 .12 (s) 9.60 ( d,J=7 H ) 6 5 1 0 ( d,J=1. 5) 8.67 (m, HP) 8.30(m,Hml

OCH 2.60 (m) 5.72 (m) 2.05 (br m) 3.20 (s) 3 Cl 3.10 (m) 5.75 (m) 2.15 (m) 3. 52 (s) 1:)"' C H N+ 3.60 (m) 5.93 (m) 2.22 (m) 3.12 (s) 9.87 (d J=7,H ) 6 5 1 0 8. 78 (m, ~)

8.32(m,~)

OCH 2.42 (m) 1.62 (m) 1.62 (m) 3.17 (s) 3 3.30 (s) 1:)"' Cl 2.83 (m) 1.62 (m) 1.62 (m) 3.60 (s) c H N+ 2.92 (m) 1.71 (m) 1.71 (m) 3.08 (s) 9.67 (d,J=7,H ) 6 5 0 8.46 (br m,H ) m,p at -78° with FS0 H produced 1 chloroindene (81) quantita­ so 2 3 tively. Presumably, the path of rearrangement included an ini-

OCH3

+ H+ . -CH30H2 ;t)H'~ .. . ~ h c5 68 78 79 00 Cl- OQ 80 81 Cl tial 1,2 shift in cation 78, which produced the 9-methoxy­ indenyl cation 79. Subsequent elimination of methanol (80) and capturing of chloride afforded 81. The generation of the 9-methoxy-9-bicyclo [4.2.1]nona-2,4- dienyl cation (82) was more successful. Its preparation was accomplished by dissolution of a-chloro ether 69 in a mixture of FS0 3H and liquid so 2 at -60° C. The structure of 82 was elucidated on the basis of its PMR and CMR spectral data compared to those of the protonated ketone 83 and ketone 35. From the NMR data and a computer simulation of the H1-H 6 six­ spin system the coupling constants and the chemical shifts of 82 and 83 were obtained. These parameters are collected in Figure 4 together with the PMR spectrum of 82. All carbon and proton resonances of 82 occur at different field. The asymmetry on either side of the ion may be due to the asymmetric dispositioning of the methoxy group in conse­ quence of restricted rotation around the C-0 bond7 • However, this interpretation does not account for all the spectral data. The c2 resonance (o 117.0) occurs upfield with respect to ihe signal for the comparable position c5 (8 123.5), where­ as the resonance for the adjacent carbon c3 occurs downfield (o 133.6) as compared with that of c4 (6 124.2). The differ­ ence in the CMR resonances of the c2-c3 double bond (Ao 16.6) 67 1,2 2,3 3,4 4,5 5,6 Hz 10.2 10.1 7.2 10.5 7.5 82

7 6 5 4 3 2ppm

H H

H 5.92 J =5Hz 1,2 2,3 3,4 1, 2 Hz 8.8 10.5 7.0 35 83 Figure 4 PMR spectrum of 82 and the PMR and CMR spectral data (6 values) for 82, 83 in FS03H-S02 and 35 in so 2 at -60° indicates considerable polarization of this bond. The compara­ ble value for the c4-c5 bond (66=0.7) is insignificant. A further indication is obtained from the H -H six-spin system, 1 6 which shows nonequivalent couplings for J , (10.2 Hz) and 1 2 J 5 6 (7.5 Hz). The value of 10.2 Hz for the coupling between ' 68 the bridgehead proton H1 and the adjacent olefinic proton Hz is comparable with the coupling between the olefinic protons Hz and H (10.1 Hz). This indicates that the c -H bond is 3 1 1 coplanar with the plane of c1, c2 and c3. On the contrary, the value of 7.5 Hz for the coupling between the bridgehead proton H6 and the adjacent olefinic proton H5 points to a dihedral 0 angle of about ZS 8. Apparently, the carbon skeleton is dis­ torted to an asymmetric structure. The steric requirement of an asymmetric dispositioned methyl group is insufficient to account for this effect as is revealed by the PMR spectrum of the 7-methoxy-7-norbornenyl cation (Z5). This spectrum shows

25 absorptions for the bridgehead protons at different field, due to the shielding of the asymmetric positioned methyl group. However, no asymmetry in the H1-H 4 four-spin system was observed, indicated by the olefinic resonances H2, 3, which exhibit a triplet 2.5 Hz) due to equivalent couplings on either side of the ion (Table VII) 7 • Thus the methyl group did not alter the dihedral angle between the bridgehead and ole­ finic protons. Additional support that this observation also applies to ion 82, is provided by the NMR data of the proton­ ated ketone 83 (Figure 4). The spectra reveal equivalent ab­ sorptions for both sides of the ion. This requires a fast ex­ change of the proton on oxygen as shown below. Proton inver-

69 sion can occur via an intra- or intermolecular pathway. The first mechanism involves rotation around the C-0 bond, where­ as the latter one arises from proton exchange with the solvent. Apparently, the latter mechanism is operative here, since rotation around the C-0 bond would also be expected in the methoxy stabilized ion 82, in which it was not observed. The CMR and PMR spectra of 83 show the equivalent ab­ sorptions amidst the corresponding nonequivalent signals for 82. Mpreover, the observed coupling constant for J 1, 2 and J 5, 6 of 83 (8.8 Hz) is the mean value of the dissimilar coup­ lings of J and J in 82 (10.2 and 7.5 Hz, respectively). 1 2 5 6 Based on these' observations, ' an asymmetric distortion of the carbon skeleton of 82 due to the steric requirement of an asymmetric positioned methyl group can be excluded. All the available data are therefore consistent with an interaction of the cationic center with one of the double bonds of the butadiene system as implied in 82 (Figure 4). This type of interaction accounts for the induced polarization in one half of the butadiene system. Furthermore, the increased coupling of one of the bridgehead protons (H 1) with the adjacent pro­ ton of the polarized double bond (H 2) indicates a bending of the , , bridge into the direction of this bond9, Ap­ c1 c9 c6 parently, in the protonated ketone 83 this interaction occurs alternately on either side of the ion due to proton exchange on oxygen. As a consequence the time-averaged NMR spectrum of

83a 35 83b

structures 83a and 83b was observed. This spectrum fits well with the PMR spectrum of ion 82, in which the chemical shifts and coupling constants for comparable positio~s are averaged. This indicates that the methyl group has little influence on

70 these parameters (vide supra). Interaction of a cationic center with a double bond should give rise to charge delocalization between these posi­ tions. Indeed, the induced polarization in one half of the butadiene system is consistent with this observation. More­ over, the downfield shift of c9 in 82 relative to the carbonyl carbon in 35 (66=30.8) indicates less positive charge at the cationic center. The corresponding value for acetone and pro­ tonated acetone, in which no additional charge delocalization is present, comes to 45 ppmW. In order to get more insight into the nature of the interaction of a cationic centre with a butadiene system, the 11-methoxy-11-bicyclo [4. 4. 1] undeca- 2, 4, 8-trienyl cation ( 84) was studied. This system is more flexible with respect to structure 82, consequent on the presence of the c7-c 10 four­ carbon bridge. Therefore, in the case of an asymmetric inter­ action one would expect an enhanced skeletal distortion. This was confirmed by its NMR spectral data (Figure 5). The polari­ zation in one half of the butadiene moiety is increased with respect to 82, as indicated by the differences in the CMR resonances for C2 and c3 (11 o= 21 . 6). Moreover, larger differences in the constants for the coup­ lings between the bridgehead protons and the adjacent olefinic protons on either side of ion 84 were observed (J =8.6 Hz 1 > 2 and J 5 , 6=3.5 Hz). Finally, the downfield shift of the c11 carbon in 84, relative to the carbonyl resonance in ketone 86 (116=5.4)' indicates considerable electron delocaliza- tion to the cationic center. The downfield shift of 5.4 ppm on introduction of positive charge at c11 is even smaller than the comparable value for protonated allylic ketones (116=19.3) 11 •

0 OH . 6t3 6 209.8 229.1

71 . 1,2 2,3 3,4 4,5 5,6 4,6 Hz 8.6 9.0 5.8 9.5 3.5 1.4 84

7 6 5 4 3 ppm

0 QH )1208.4

H H

Hs.ee

J =6Hz 1,2 2,3 1, 2 3,4 Hz 6.0 9.8 6.0 86 85 Figure 5 PMR spectrum of 84 and the PMR and CMR spectral data (o values) for 84, 85 in FS0 H-S0 and 86 3 2 in so 2 at -60°

The NMR spectra of protonated ketone 85 (Figure 5) show, equivalent resonances for both sides of the ion similar to the protonated ketone 83, with the averaged dissimilar values for comparable positions of ion 84 (vide supPa). Thus, ions 82 and 84 can be described as homoallylic or bishomocyclopropenyl-

72 ic cations (see Discussion). Quench experiments with 82, 83, 84 and 85 using a solu­ tion of sodium methoxide in methanol at -75°, afforded the corresponding ketals and ketones.

4.5. Diseussion

The reaction of pyridine with compounds 68, 69 and 70 proceeds with second-order kinetics and retention of configu­ ration. This remarkable kinetic behaviour was encountered in the reaction of pyridine with the a chloro ethers of tri­ cyclo[4.2.1.o2•5]nonanes, which proceeds with racemization (Chapter 3). It can be rationalized similarly in terms of a rate-determining reaction of pyridine with an external ion­ pair. The previous observations regarding the ionic nature of external ion pairs also applies for 78a. There has to be still CsH')JH' C5H5N ~ Cl- ~ ;E)"' kJ [J;c,] k2 f. .4 &

68 78a 75 some interaction between cation 78 and the chloride anion at the external ion pair stage, since ionization of 68 under non­ equilibrium conditions afforded the rearranged 1-chloroindene (81), exclusively. Kim and Leffekl 2 concluded from a study of the reaction between pyridine and triphenylmethyl perchlorates that the enthalpy of activation of a Menschutkin reaction results mainly from the bond breaking process and concomitant solva­ tion changes, whereas the carbon-nitrogen bond forming pro­ cess is entropy controlled. Inspection of Table XII reveals that the entropies of activation for the a-chloro ethers 68, 69 and 70 agree within experimental error. In view of the previous observations, this indicates that the reaction of pyridine with these three compounds at the external ion pair stage is very similar. In contrast, the enthalpies of activa- 73 tion differ significantly. The ~Ht for the most unsaturated compound 68 is 2.7 kcal/mol beyond that of the diene 69, which surpasses that of the fully hydrogenated derivative 70 with 3 kcal/mol. Obviously, these values reflect the relative ease with which the C-Cl bonds are ionized. The differences are too large to be ascribed to an inductive effect of the double bonds. They rather suggest that the intermediates derived from 68 and 69 are destabilized. According to Htickel's 4n+2 n-electron prerequisite for cyclic stabilization, the in­ teraction of a cationic center with a butadiene moiety is destabilized. This is manifested in an increase in ~Hf of 3 kcal/mol in going from 70 to 69. The increase in ~Ht of 2.7 kcal/mol on going from the diene 69 to the triene 68 may be attributed to an enhanced inter­ action of the cationic center with the butadiene moiety in 68, due to the fact that the one carbon bridge is more con­ strained in the direction of the four carbon bridge. This re­ sult is in contrast with Goldstein's theory on topology and aromaticity, which predicted cation 78 to be more stable than 82 consequent on the presence of two stabilizing interactions as implied in 78b, in addition to the destabilizing 4rr-electron interaction which is present in both ions.

qcH3 OCH3 ...... h .. JJ. 78b 82b

Obviously, longicyclic stabilization as implied in 78b is a minor contributor to the overall stability of the trienyl cation. The homoantiaromatic interaction appears to be the dominant factor. Thus, the symmetry imposed rules for longi­ cyclic stabilization3 (vide supra) are not prevailing in this system. The results are in better agreement with Goldstein's original definition, according to which the stability of a

74 bicyclic ion is dominated by the interaction of the odd bridge with the longer even bridge 2 • In addition to the kinetic results, the presence of an interaction between the cationic center and the butadiene bridge is indicated by the NMR spectra of ions 82 and 84 under long-life conditions. The PMR and CMR data point to an asym­ metric structure for ions 82 and 84, due to interaction of the cationic center with one half of the butadiene system. The asymmetry in 84 is increased, with respect to 82, consequent on the larger flexibility of the former compound, which allows an increased skeletal distortion. Obviously, the asymmetric interaction is more favourable than the symmetric one, which is conceivable on the basis of MO symmetry arguments. The nonbonding (NBMO) will interact predominant­ ly with the HOMO (t2) and LUMO ($3) of butadiene, since these orbitals are energetically most proximate. The higher energy

0.60n0.60 0.3~0.37 -0.3n7-0.37 -0.6n00.60

0.37 0.37 o.6o I \..o.6o 0.60 0.60 0.37 -0.37 IJ'1 IJ'2 IJ'3 IJ'4 Figure 6 Hucke! MO's for butadiene

LUMO is symmetric with respect to the plane bisecting the bridgehead-bridgehead axis, whereas the lower energy HOMO is asymmetric with respect to this plane. Therefore, a symmetric interaction is only allowed with the LUMO, which will be de­ stabilizing. In contrast, an asymmetric interaction is allow-

Figure 7 Interaction of the NBMO with the HOMO of butadiene in ions 82 and 84

75 ed with the lower energy HOMO which will be stabilizing (Fi­ gure 7). This interaction may be of the homoallylic type with one overlap (C 2-cationic center). Alternatively, two overlaps may be involved (C 2- and c3-cationic center), which give rise to a bishomocyclopropenyl cation. Probably, the nature of ions 82 and 84 lies amidst both types. The interaction of c3 is smaller than that of c2 consequent on the differences in distance between these interacting atoms and the cationic center. Moreover, this effect is increased as a result of the sma~ler corresponding coefficient of c3 as compared with c2 of the HOMO of butadiene ( 6). Thus the structure of ions 82 and 84 can be represented as depicted in Figure 7.

Apparently, the mode of ~ participation in the ionization reaction of 68 and 69 is controlled by orbital symmetry. As a consequence, the energetically unfavourable symmetric inter­ action, which is commonly designated as antiaromatic, does not occur in the solvolysis. However, one would expect that this mode of ~ participation of the butadiene system should provide anchimeric assistance in the ionization reactions of 68 and 69. The rate retardation observed for the latter com­ pounds with respect to the saturated analogue 70 suggests that the resonance stabilization of the ultimate ions 82 and 84 is overbalanced by an increase in strain energy, due to the concomittant skeletal distortion which occurs in order to attain a maximum orbital overlap at one side of the ion. Ob­ viously, this energetically unfavourable situation occurs to avoid the even more unfavourable symmetric interaction. This disadvantageous interaction also accounts for the absence of alternate interactions on either side of ions 82 and 84. On exchange of the asymmetric interaction between the two double bonds the ions should pass through a symmetric transition state. On the contrary, alternation is possible for proton­ ated ketones 83 and 85. In these cases proton exchange with the solvent gives rise to the intermediacy of the parent neutral ketones. At this point the experiments of Breslow with respect to the preparation of the cyclopentadienyl cation (88)13 76 may be of interest. There is no doubt about the structure of this ion since ESR experiments clearly showed that it has a ground-state triplet, which perfectly fit~ with Hund 1 s rule for degenerate half-filled molecular orbitals. With re­ gard to orbital symmetry it is clear that the formation of the cation via the cyclopentadienyl bromide (87) and SbF 5 is a nonconcerted process. The reaction proceeds in a concerted way only when the LUMO of the butadiene fragment interacts with the incipient cation. Such a process needs a very high activation energy; therefore it is to be expected that the electron shift of a double bond to the incipient cation occurs with a simultaneous conrotatory motion of the other olefinic bond. This dynamic process is illustrated in Figure 8.

&~ Figure 8 Conrotatory motion of the double bonds during ionization of cyclopentadienyl bromide

This means that the rr-electron distribution in the activated complex is stored in a homocyclopropenyl cation (or allylic cation) plus an isolated double bond. This situation is fully comparable with the electron delocalization in the ions 82 and 84 as was firmly established by CMR . In both sys­ tems the only interaction can occur via the HOMO of the buta­ diene moiety with removal of one double bond by a conrotatory mode of motion:

OCH 3

82 77 According to these results it seems reasonable to assume that the formation of the cyclopentadienyl cation is a multi-step process catalyzed by the Lewis acid:

M SbF5 ~-~ 't- S T 87 88

The CMR study of the stable ions 82 and 84 clearly de­ monstrates in which way isolated double bonds can exist in carbonium ions. Until now similar examples were only encoun­ tered in metal-coordination complexesl4 , A closely related example in this field (89) has been published recently by PaquettelS,

89

4.6. Experimental

• anti-9-Chloro-9-methoxybicyclo[4.2.1] nona-2,4,7-triene (68) To a stirred solution of 68 (5.2 g, 0.029 mol) in diethyl ether (5 ml) PC1 5 (6 g, 0.03 mol) was added in small portions at such a rate that the ether boiled gently. Because of its thermal lability, 68 was purified by high-vacuum evaporation of the solvent and volatile components (POC1 3 and CH 3Cl) for 1 hr at 15° and subsequent recrystallization from pentane at 0 . -78 : oTMS CDC1 3 5.97 (m,4H), 5.40 (d,J;1.5,2H), 3.55 (m,2H) and 3.40 (s,3H).

• anti-9-Chloro-9-methoxybicyclo [4.2.1] nona-2,4-diene (69) This compound was prepared from ketal 36 according to the

78 previous procedure: oTMS CDC1 3 5.82 (m,4H), 3.50 (s,3H), 3.10 (m,2H) and 2.15 (m,4H).

• anti-9-Chloro-9-methoxyb lo ~.2.1] nonane (70) The ketalization of bicyclo[4.2.1] nonan-9-one 4a was accomplish­ ed as previously described for 36 (Chapter 3). Treatment of the dimethyl ketal with PC1 5 and work-up as for 68 provided 70: 8TMS CDC1 3 3.58 (s,3H), 2.83 (m,2H) and 1.60 (m,12H).

11-Chloro-11-methoxybicyclo[4.4.1] undeca-2,4,8-triene This compound was prepared from bicyclo[4.4.1] undeca-2,4,8- trien -11-one~ to the previous procedure:8TMS CDC1 3 5.84 (m,6H), 3.58 (s,3H), 3.30 (m,2H) and 2.54 (m,4H).

• oia~endo- and cis,exo-2-Methoxydihydroindenyl chloride (76, 77) Vacuum distillation of 68 afforded a mixture of 76 (82%) and 77 (181): 8TMS CDC1 3 5.73 (m,4H), 4.69 (m,2H), 3.71 (s,3H) and 3.50 (m,2H).

• Reaction of 68 with lithium aluminum hydride To a stirred solution of 68 (1 g, 5.5 mmol) in dry ethyl ether (10 ml) lithium aluminum hydride (0.5 g, 13 mmol) was added in 15 min. The excess hydride was hydrolyzed with 3 N sodium hydroxide and the ether was washed with water and evaporated. The residue was distilled to give 73 (0.45 g), bp 34-36° (0.1 mm): oTMS CDC1 3 5.93 (m,4H), 5.13 (d,J=1.5,2H), 3.80 (t,J=6, 1H) and 3.10 (m,5H).

• anti 9-Methoxy-9-bicyclo[4.2.1]nona-2,4,7-trienyltriphenyl- phosphonium fluoroborate (74) To a solution of 68 (1 g, 5.5 mmol) in 10 ml of liquid so 2 tr~ phenylphosphine (1.5 g, 5.7 mmol) was added. After 15 min of stirring at -50°, trimethyloxonium fluoroborate (0.8 g, 5.5 mmol) was added. The reaction mixture was stirred for an ad­ ditional 15 min and poured into dry ethyl ether, whereupon 74 (2.2 g) precipitated as white crystals: mp 230 232° (dec).

79 8TMS S0 7.82 (m,15H), 6.23 (m,4H), 4.97 (d,J•1.5,2H), 4.13 2 (d,J=14,2H) and 2.87 (d,J=1.5,

• anti 9-Methoxy-9-bi lo[4.2.1]nona-2,4,7-trienylpyridi- nium fluoroborate ( This compound was prepared from 68 and pyridine as described in the previous experiment: 6TMS CDC1 3 9.83 (d,J=7,2H), 8.50 (m,3H), 6.21 (m,4H), 5.45 (d,J=1.5,2H), 4.15 (m,2H) and 3.12 (s,3H).

• Reaction of 74 with lithium aluminum hydride To a suspension of 74 (2 g, 3.6 mmol) in dry tetrahydrofuran at -78° lithium aluminum hydride (0.1 g, 2.6 mmol) was added. The mixture was stirred for 1 hr and was then allowed to warm to room . The excess hydride was hydrolyzed with 3 N sodium hydroxide and the aqueous layer was extracted with ether. The ether extracts were washed with water, dried and concentrated. Vacuum distillation provided pure 73 (0.42 g), bp 34-36° (0.01 mm).

• Kinetic measurements Equimolar amounts of pyridine and a-chloro ether dissolved in cn 2c1 2 were mixed at -80° for compounds 68 and 69 and at -140° for 70 in an NMR sample tube (0.4-1.0 M solution). The runs were performed at temperatures as indicated in Table XII. The progress of the reaction was followed by integrating the PMR methoxy signals of substrate and product at appropriate in­ tervals. The rate constants were determined graphically.

80 References and notes

1. S. Winstein, . Rev., Chern. Soc., 22 (2), 141 (1969). 2. M.J. Goldstein, J. Amer. Chern. Soc.,~. 6356, 6357 (1967). 3. M.J. Goldstein and R. Hoffmann, J. Amer. Chern. Soc.,~. 6193 (1971). 4. a. T.A. Antkowiak, D.C. Sanders, G.B. Trimitsis, J.B. Press and H. Shechter, J. Amer. Chern. Soc., , 5366 (1972). b. D.C. Sanders and H. Shechter, J. Amer. Chern. Soc., 2i• 6858 (1973). c. A.F. Diaz, J. Fulcher, M. Sakai and S. Winstein, J. Amer. Chern. Soc., , 1264 (1974). d. A. Diaz and J. Fulcher, J. Amer. Chern. Soc., , 7954 (1974). e. w. Kirmse and G. Voigt, J. Amer. Chern. Soc., 96, 7598 (1974). f. A. Diaz and J. Fulcher, J. Amer. Chern. Soc., 798 ' (1976). 5. D. Cook, A. Diaz, J.P. Dirlam, D.L. Harris, M. Sakai, S. Winstein, J. Barborak and P. Schleyer, Tetrahedron Lett., ..:!_!, 1405 ( 1971). 6. a. T.A. Antkowiak and H. Shechter, J. Amer. Chern. Soc., 94, 5361 (1972). b. T.S. Cantrell and H. Shechter, J. Amer. Chern. Soc., ~. 5868 ( 1967). 7. R.K. Lustgarten, M. Brookhart and S. Winstein, Tetrahedron Lett., 141 (1971). 8. J. Karplus, J. Chern. Phys., 30, 11 (1959). 9. The restricted rotation around the C-0 bond and the asym­ metric interaction could give rise to two isomers of ion 82. The occurrence of one isomer may be ascribed to steric 81 factors introduced by the asymmetric distortion of 82. 10. G.A. Olah and A.M. White, J. Amer. Chern. Soc., 91, 5801 (1969). 11. G.A. Olah, Y. Halpern, Y.K. Mo and G. Liang, J. Amer. Chern. Soc., 94,3554 (1972). 12. C.B. Kim and K.T. Leffek, Can. J. Chern., , 3408 (1975). 13. M. Saunders, R. Berger, A. Jaffe, J.M McBride, J. O'Neill; R. Breslow, J.M. Hoffman, Jr., C. Perchonock; E. Wasser­ man, R.S. Hutton, V.J. Kuck, J. Amer. Chern. Soc., 3017 (1973). 14. a. M.D. Rausch and G.N. Schrauzer, Chern. Ind. (London) (1959). b. T.A. Manuel and F.G.A. Stone, Proc. Chern. Soc., London, 90 (1959). c. M. Brookhart and E.R. Davis, J. Am. Chern. Soc., 92, 7622 (1970). d. M. Brookhart, E.R. Davis and D.L. Harris, J. Am. Chern. Soc., 94, 7853 (1972). 15. L.A. Paquette, S.V. Ley, S. Maiorana, D.F. Schneider, M.J. Broadhurst and R.A. Boggs, J. Amer. Chern. Soc., 7, 4658 (1975).

82 CHAPTER 5

Stereospecific reactions of the 9-phenylseleno-9-bicyclo[4.2.~ nona-2,4,7-trienyl anion

5.1. Introduction

The theoretical analysis of bicycloaromatic stabilization in ~-bridged ions by Goldstein and Hoffmann 1 have led to con­ siderable effort directed towards experimental tests of their theory. In agreement with the prediction, the bicyclo [3.2.2]­ nonatrienyl anion (90-) appeared to be stabilized, as revealed by its generation and direct observation, whereas the expected instability of the parent cation 90+ was manifested in a ske­ letal rearrangement to 91 on its attempted generationla ,2,

e e+ go• 91

Moreover, the observed relative stabilities of anions 90-, 92 and 93, in the bicyclo[3.3.2]nonane series, were in support of the theory. The enhanced stability of 92 relative to 93 was attributed to bishomoaromatic stabilization (6n electrons), whereas the additional stability of triene 90 with respect

92 93 to diene 92 was ascribed to longicyclic n-electron delocali

83 zation throughout all three unsaturated bridges 3 • Similarly, longicyclic stabilization was suggested to account for the smooth preparation of the bicyclo[3.3.2]decatrienyl dianion ) 4.

94

The experimental results of studies of the 9-bicyclo­ [4.2.1) nona-2,4,7-trienyl system are less convincing. Both the 9-bicyclo [4.2.1] nona-2,4,7-trienyl cation (57+) and anion (57-) are stabilized according to Goldstein's prerequisite b+ 57 for longicyclic stabilization 1 • However, reactions that might have led to [4. 2. 1] cations provided only rearranged products. In addition, the results presented in Chapter 4 concerning a kinetic study on the methoxy-stabilized [4.2.1] cation which does not rearrange under solvolytic conditions, revealed its destabilizing nature. Until now, reports concerning the experimental verifica­ tion of the expected stability of the [4. 2. 1] anion are lack­ ing completely. In view of these observations the 9-bicyclo­ ~.2.1] nona-2,4,7-trienyl anion has been studied.

5.2. Preparation of the 9-phenylseleno-9-bicycZo[4.2.1]nona-2,4,?-trienyl anion

Attempts to prepare anion 57 by the classical ether cleavage methods, using the reaction between sodium-potassium alloy and 9-methoxybicyclo[4.2.1]nona-2,4,7-triene (73)

84 proved to be unsuccessful. Shaking of 73 in tetrahydrofuran at -90° with sodium-potass.ium alloy produced a reddish-brown solution. After quenching with methanol, the dimer 95 was recovered only. Compound 95 was identified by NMR and mass spectral analysis. ~·1 HH ~2• 73 95

Apparently, the butadiene moiety was reduced preferentially with respect to the carbon-oxygen bond. Therefore, the intro­ duction of hetero-atoms at the c9 position, which could be cleaved easier from carbon, was essential. The phenylseleno ketal of ketone 33 appeared to be an auspicious precursor. Recently it has been found6 that seleno ketals (96) are cleaved readily by n-butyllithium in tetrahydrofuran at -78°. This facile cleavage was attributed to the weak carbon-sele­ nium bond, which is comparable to carbon-bromine,and the ulti­ mate formation of a selenium-stabilized carbanion (97). The

R1"-. / SePh 1 R "' /H c n.Buli c R( "'SePh R2/ "-. SePh 96 97 98 acid quenching of anion 97 afforded the protic compound 98 6 • This reaction pattern could be applied to the bicycle [4.2.1]- 0 II PhS?f)SePh PhSeH,H+ l ~ + .cb """' O)SePh H 35 99 100

85 nona-2,4,7-trienyl system. Treatment of ketone 35 with phenyl­ selenol in the presence of HCl produced the seleno ketal 99 together with a small amount (6~) of 3-phenylselenoindene (100). Reaction of ketal 99 with n-butyllithium in tetrahydro­ at 78° gave a-seleno carbanion 101, as indicated by the formation of the protic epimers syn- and anti 9-phenylseleno­ bicyclo [4.2.1]nona-2,4,7-triene (102) on quenching with water. The stereochemistry of the products is readily assignable on the basis of their PMR spectra which exhibited a triplet 6 Hz) for the H9 proton of the syn-epimer, whereas a singlet was observed for the anti epimer. Subsequently, both epimers Ph;rsh PhSe n-BuLI H20 Jjh Ph?6 + h ~ '- ~ 99 101 syn-102 anti-1o2 of 102 were used as precursors in an attempt to generate the unsubstituted [4. 2.1] anion 57-. However, compounds 102 did not react with n-butyllithium in tetrahydrofuran at tempera­ tures up to -20°. At higher temperatures uncontrolled reactions took place leading to a complex mixture of products.

5.3. Stereospecific reactions of the 9-phenylseleno-9-bieycZo[4.2.1]­ nona-2.4.?-trienyl anion

The quench reaction of anion 101 with water, which pro­ vided the protic epimers syn- and anti-102, indicated a priori no stereoselectivity. However, stereospecific reactions were observed on the introduction of deuterium. Thus, quenching the

H'C:--o-0. PhSe~ iiJJ iiJJ 101 103a 104a

86 anion solution with n2o produced 103a and 104a. This result reveals that the epimers of 102 arise from two different me­ chanisms which both have a highly specific natu~e. Apparently, 103a is the result of an attack of D2o at the phenyl ring with subsequent transfer of hydrogen to c9 , syn with respect to the monoene unit. Alternatively, 104a is formed by a direct attack of n2o at c9 from the butadiene side of the anion. The use of other electrophiles gave similar results. In each case only two products of the types 103 and 104 were observed (Scheme V). PhSe ;s-o-x, Ph)) h ~ 101 103 104

E X Products (ratio)

CH 103b (62) 104b ( 38) 3 PhC(O)H C(H)(OH)Ph 103c (60) 104c (40)

Ph CO C(OH)Ph 103d (68) 104d (32) 2 2 (CH ) NC(O)H C(O)H 103e (52) 104e (48) 3 2

Scheme V Reaction of anion 101 with electrophiles

The stereochemistry of the products with an a-hydrogen at c9 (103) was established on the basis of their PMR spectra (vide supra). The configuration of the methyl-substituted compound 104b is tentatively assigned on the basis of product formation similar to that of 9-deutero-substituted compound 104a and other products (vide ), The stereochemistry of 104c could be established by its spectral properties (Table XIV and XV). All carbon and proton resonances occur at different field, which indicates an asymmetric structure. The differences in CMR chemical shifts are most pronounced for the bridgehead positions (~oc 1 6=3.97) and the butadiene system (~6=1.32 and 1.19), whereas' the corresponding value for the monoene unit is relatively small (0.53 ppm). The asymmetry can be

87 ()) ())

Table XIV PMR spectral data" for 9,9-disubstituted bicyclo [4.2.1] nona-2,4, 7-trienes in CDCI ( 8 values) 3

8

Compound R R Others 1 2

99 PhSe PhSe 3.13 (d,J=6) 6.13 (m) 5.13 (d,J=1.5) 7.33 (m,lOH)

syn-102 H PhSe 3.30 (t,J=6) 6.10 (m) 5. 33 ( d,J=l. 5) 7.20 (m,SH) 3.77 (t,J=6,H ) 9

anti-102 PhSe H 3.22 (6,]=6} 5.92 (m) 5.27 (d,J=l. 5) 7.22 (m,5H) 3.47 (s, H ) 9 104b PhSe CH 2.87 (m) 6.07 (m) 5.13 (d,J=l. 5) 7.27 (m,5H) 1. 37 (s, methyl) 3

104c PhSe C(H)(OH)Ph 3.43 (d,J=7) 6.07 (br m) 4.53 (d,J=l. 5) 7.13 (m, 10H) 3.53 (s,OH) 2.97 (d,J=6) 4.33 (s, methine)

104d PhSe C(OH)Ph 3.93 (m) 5.93 (m) 4.23 (d,J=1.5) 7.22 (m,1SH) 3.40 (s,OH) 2 104e PhSe C(O)H 3.32 (m) 6.13 (m) 5. 21 ( d, }=1. 5) 7.22 (m,SH) 9. 07 ( s, aldehyde)

*J values are expressed in Hz. Table XV CMR spectral data for 9,9-disubstituted bicyc1o[4.2.1}nona-2,4, 7-trienes in CDC1 ( f:i values) 3

8

Compound Rl R2 c c c c c c c c c Others 1,6 2,5 3,4 7,8 9 0 m p

99 PhSe PhSe 56.02 134.32 128.01 122.23 57.52 131.28 129.60 134.10 129.60 132.11 129.60 138.42 129.60

syn-102 H PhSe 49.10 135.32 127.22 125.10 42.13 132.41 130.00 134.41 129.87

anti -102 PhSe H 51.33 136.31 125.37 122.54 41.95 131.90 130.17 134. 10 128.10

104b PhSe cH 55.63 135.69 127.35 121.93 46.76 129.03 129.56 137.85 129.56 29.30 (CH ) 3 3

104c PhSe C(H)(OH)Ph 47.11 134. 19 128.63 119.63 63.70 128.37 130.04 136.92 129.64 51.08 135. 51 127.44 120.16 142.30 127.88 128.63 128.63 76.09 [C(H)OH)

104d PhSe C(OH)Ph 51.21 135.34 127.22 117.56 68.99 128.94 126.75, 129.25, 129.95, 84.73 [qH)Ph ) 2 2 148.48 136.57

()) CD induced only by the chiral benzylic alcohol function which

Ph OH PhSe ';oH 9 6 5 4

8 obviously is unable to rotate. Therefore, the previous spec­ tral observations are consistent with a syn disposition of the benzylic alcohol function with respect to the butadiene bridge. When the hydrogen at c10 is replaced by a phenyl group (104d), the asymmetry is cancelled as indicated by the NMR data (Table XIV and XV). In order to obtain additional structural information of compounds 104 (b-e), lanthanide shift reagents were used. Ad­ dition of small amounts of Eu(fod) 3 to a solution of 104c in CDC1 3, caused marked nonequivalent downfield shifts for the bridgehead protons H1, 6 (Figure 9). This points to an asymmetric positioning of the alcohol func­ tion at one side of the skeleton. The downfield shift of the monoene protons (H , ) is somewhat larger than that of the bu­ 7 8 tadiene.protons (H _ ). This indicates that in spite of the 2 5 syn disposition of the benzylic alcohol group with respect to the butadiene bridge (vide supra), europium is located somewhat more close to the monoene side of the molecule. In­ spection of Dreiding models of 104c revealed that this mode of complexation can be realized and in addition is sterically most favourable. For compound 104d the europium induced shift is very small, which has to be attributed to the crowded posi­ tion of the alcohol function. Therefore, no extra structural information could be obtained. The application of Eu(fod) 3 was more successful for 104e, with has a more approachable aldehyde group. In this case the downfield shift for the butadiene protons is significantly larger (Figure 9) than for the monoene protons, which is consistent with the depicted structure.

90 0 \\10 PhS;rssC H 6 5 k '4 6 1 ..& 3 2 160 104e 160

120 120

80 80

40 40

40 80 120 40 80 120

Eu(fod)3 (mg) Eu(fod)3 (mg) Figure 9 A plot of the induced by shift rea­ gent, ~vi, vs amount of added shift reagent for protons of 104e and 104c

5. 4. Discussion

The preparation of the unsubstituted 9-bicyclo[4.2.1]­ nona-2,4,7-trienyl anion appeared to be hampered by uncontrol­ led reactions. The reaction of sodium-potassium alloy with 9-methoxybicyclo[4.2.1]nona-2,4,7-triene (73) provided the dimer 95 only. Apparently, the reaction involves reduction of the butadiene moiety followed by dimerization and subsequent deprotonation of the ether solvent. Similarly, the 9-phenyl­ selenobicyclo [4.2.1] nona-2,4,7-triene (102) appeared to be inappropriate as a precursor. The experimental test of anion­ ic stability is usually provided by the preparation of the anion. However, the inability to produce anion 57-using the

91 previous methods does not demonstrate necessarily its insta­ bility. Apparently, the electron affinity of the butadiene moiety surpasses that of the carbon-oxygen and carbon-selenium bond. The reverse order of reactivities was observed if the ~.2.1] anion was stabilized by a-substituted selenium. Thus, 9-phenylseleno-9-bicyclo [4.2.1] nona-2,4,7- 1 anion (1~1) was readily formed in tetrahydrofuran at -78°. On quenching anion 101 with a variety of electrophiles, two products were recovered invariably. One arises from electrophilic substitu-

103 104 tion at the phenyl ring (103), whereas the second one results from a direct attack at the c9 position of 101. The mechanism which accounts for structure 103, apparently involves attack of the electrophile at the para position of the phenyl ring to produce intermediate 105. Subsequent intramolecular hydride transfer gives stereoselective structure 103. The formation of structure 103 for more than 50% (Scheme V) indicates con-

x+

101 105 103 siderable charge delocalization between c9 and the phenyl ring via selenium. The s s of product 104 with the electro- phile syn with respect to the butadiene moiety has to be attributed to stereoelectronic control, since stereoselectivi­ ty arising from kinetic steric control should give rise to

92 approach of the agent from the monoene side of the ion exclu­ sively (Chapter 4). Apparently, the negative charge at c9 resides predominant at the butadiene side of the ion, con­ sequent on the involvement of a bishomoaromatic (6u electrons) interaction (101 . The stereospecific reactions of the non-

101a classical carbanion 101 here encountered contrast with those of nonclassical cations. For instance, the reaction of the 7-norbornenyl cation (6) with nucleophiles afforded products with the substituent exclusively in anti position with respect to the interacting double bond (Chapter 2). The crucial factor

6 may be the number of electrons which are involved in the transition state. If a nucleophile approaches the 7-norbornenyl cation from the monoene side, it contributes two electrons to the homoaromatic 2u-electron system. As a consequence a 4u­ electron antiaromatic transition state occurs. In contrast, an electrophilic attack at a bishomoaromatic anion does not affect the number of electrons. Therefore, in this case the ficity arises from kinetic electronic control, in which the electrophile approaches the ion from the side with the highest electron density.

5.5. ExpePimental

• Reaction of syn-9-methoxybicyclo [4. 2. 1l nona-2, 4, 7-triene (73) with sodium-potassium alloy

93 To sodium-potassium alloy (0.5 g) under tetrahydrofuran (10 ml) 73 (0.3 g) was added at -78° under nitrogen atmosphere. The mixture was stirred vigorously for 1 hr. A reddish-brown color developed on contact and intensified as the reaction proceed­ ed. Subsequently, methanol (1 ml) was injected and the result­ ing mixture was poured into water (10 ml). The aqueous layer was extracted with ether and the combined ether layers were washed with water. After drying (K 2co 3) and evaporation of the solyent, the residue was chromatographed on silicagel with chloroform as eluent. This gave a pure fraction of 93: oTMS CDC1 3 5.57 (d,J=1.5,2H), 5.27 (m,ZH), 4.00 (t,J=6,1H), 3.73 (s,3H), 2.7 (m,1H) and 2.2 (m,2H).

• 9,9-Bis-phenylseleno-bicyclo [4.2.1] nona-2,4,7-triene (99) A stream of was passed through a solution of ketone 35 (5 g, 0.038 mol) and phenylselenol (10 g, 0.068 mol) in dry diethyl ether (10 ml) at 0° during 15 min. The resulting mixture was poured into saturated sodium bicarbonate solution and extracted with chloroform. The organic layer was washed with water, 7% aqueous potassium hydroxide and water, dried (K 2co 3) and concentrated. Chromatography on silicagel with benzene-hexane (20/80) as eluent gave 99 (12.2 g, 78%): oTMS CDC1 3 7.33 (m,phenyl), 6.13 (m,4H), 5.13 (d,J=1.5,2H), 3.13 (m,2H). Furthermore, a fraction of 100(1.2 g, 6%) was obtained: oTMS CDC1 3 7.12 (m,phenyl), 6.63 (d,J=S.S,1H), 6.22 (dd,J=5.5 and 1.5) and 4.92 (br,s).

• anti-9-Phenylselenobicyclo [4.2.1] nona-2,4,7-triene(anti-102) • syn-9-Phenylselenobicyclo [4.2.1] nona-2,4,7-triene(syn-102) A solution of selenonketal 99 (1.0 g, 2.3 mmol) in dry tetra­ hydrofuran (10 ml) was treated with a 20% solution of n-butyl­ lithium in hexane (1.56 ml) at -78° under a nitrogen atmosphere. After stirring for 30 min water (1 ml) was added. The reaction mixture was stirred for an additional 30 min at -78° and then allowed to warm up to room temperature. The mixture was poured into water and extracted with chloroform. The organic layer

94 was washed with water, dried (K 2co 3) and concentrated. Chro- matography on sil 1 with be_nzene-hexane (20/80) as eluent gave a pure fraction of syn-102 (58%) and anti-102 (24%). NMR: see Table XIV and XV.

• Reaction of 101 with D2o provided: 103a (56%): oTMS CDC1 3 7.22 (m,4H), 6.11 (m,4H), 5.33 (d,J= 1.5,2H), 3.77 (t,J=6,1H) and 3.33 (t,J=6,2H). 104a (22%): oTMS CDC1 3 7.22 (m,5H), 5.91 (m,4H), 5.23 (d,J= 1.5,2H) and 3.20 (br,m,2H).

• Reaction of 101 with methyl iodide provided: 103b (52%): oTMS CDC1 3 7.07 (m,4H), 6.03 (m,4H), 5.30 (d,J= 1.5,2H), 3.71 (t,J=6,1H), 3.28 (t,J=6,2H) and 2.27 (s,3H). 104b (32%): NMR see Table XIV and XV.

• Reaction of 101 with benzaldehyde provided: 103c (48%): oTMS CDC1 3 7._12 (m,9H), 6.12 (s,1H), 5.91 (m,4H), 5.13 (d,J=1.5,2H), 3.52 (t,J=6,1H), 3.13 (t,J=6,2H) and 2.80 (s,OH). 104c (28%): NMR see Table XIV and XV.

• Reaction of 101 with benzophenone provided: 103d (44%): oTMS CDC1 3 7.17 (m,14H),S.9 (m,4H), 5.82 (s,OH), 5.19 (d,J=1.5,2H), 3.52 (t,J=6,2H) and 3.07 (t,J=6,2H). 104d (22%): NMR see Table XIV and XV.

• Reaction of 101with dimethylformamide provided: 103e (45%): oTMS CDC1 3 10.2 (s,aldehyde), 7.2 (m,4H), 6.03 (m,4H) 5.3 (d,J=1.5,2H), 3.77 (t,J=6,1H) and 3.30 (t,J=6,2H). 104e (44%): NMR see Table XIV

95 References

1. a. M.J. Goldstein, J. Amer. Chem. Soc.,~. 6356, 6357 (1967). b. M.J. Goldstein and R. Hoffmann, J. Amer. Chem. Soc., 3, 6193 (1971). 2. J.B. Grutzner and S. Winstein, J. Amer. Chem. Soc.,~. 2200 (1972). 3. a. M.V. Moncur and J.B. Grutzner, J. Amer. Chem. Soc., ~. 6449 (1973). b. M.J. Goldstein and S. Natowsky, J. Amer. Chem. Soc., 6451 (1973). 4. M.J. Goldstein, S. Tomoda and G. Whittaker, J. Amer. Chem. Soc., 96, 3676 (1974). 5. S. Winstein, M. Ogliaruso, M. Sakai and J.M. Nicholson, J. Amer. Chem. Soc., 89, 3656 (1967). 6. a. W. Dumont, P. Bayet and A. Krief, Angew. Chem. Internat. Ed., 804 (1974). b. D. Seebach and A.K. Beck, Angew. Chem. Internat. Ed., 806 (1974).

96 SUMMARY

This thesis deals with the neighbouring group participa­ tion of cr and " electrons in rigid polycyclic systems. The ionization reactions of these systems have revealed unusual solvolytic reactivities which are often accompanied by skelet­ al rearrangements. When solely rearranged products are ob­ served, it is difficult to decide whether enhanced solvolytic rates arise from electronic factors or from factors asso­ ciated with possible low-energy routes prone to skeletal re­ arrangements. Therefore, charge-stabilizing groups were intro­ duced at the reaction center in order to prevent rearrangement reactions. In this way it was possible to elucidate the nature of the intermediates. Even the study of antiaromatic inter­ actions appeared to be accessible along kinetic and spectro­ scopic lines. Reaction of the homoaromatic 7-norbornadienyl cation with nucleophiles derived from elements of group VA and VIA afford­ ed 2,3- and ?-substituted products dependent on their basicity. While the low-nucleophilic tris(pentafluorophenyl)phosphine afforded solely the ?-substituted phosphonium salt the ana­ logous reaction with the more nucleophilic triphenylphosphine gave rise to 2,3-substituted products. In contrast, when charge-stabilizing groups were introduced at the cationic center, only ?-substituted products were observed. Thus, re­ action of triphenylphosphine with 7-chloro-7-methoxynorbor­ nene produced the ?-substituted triphenylphosphonium salt. The latter compound appeared to be a suitable precursor for the generation of the 7-triphenylphosphonio-7-norbornenyl

97 dication, a nonclassical dication. Unusual reactivities were observed in 7-norbornenyl systems with an endo-fused cyclobutene ring. It was establish­ ed that during the ionization of 9-chloro-9-methoxy-endo­ tricyclo[4.2.1.o2•5]nona-3,7-diene anchimeric assistance was provided by an electron-delocalization process in which the two double bonds and one sigma bond are involved. The inter­ mediate is described as a "sandwich complex" of cyclopenta­ diene and. the cyclobutenyl cation in which an aromatic (6n electrons) interaction occurs. This intermediate may reclose to give the starting structure or the rearra~ged tricyclo­ [4.3.o.o2•5)nona-3,7-diene, dependent on the reaction condi­ tions. The 9-bicyclo[4.2.1]nona-2,4,7-trienyl cation has evoked interest because of its possible homoaromaticity and bicycloaromaticity. Support evidence for its bicycloaroma- ticity seemed to be provided by kinetic studies. However, the products of solvolysis were completely rearranged. In this case also, the introduction of a methoxy group at the cation­ ic center was successful in suppressing rearrangement react­ ions. The ionization reaction of 9-chloro-9-methoxybicyclo­ [4.2.1] nona-2,4,7-triene under solvolytic conditions proceeds without rearrangement. A kinetic study revealed that the triene is much less reactive than the more hydrogenated deri­ vatives. This rate retardation could be explained by assuming an antihomoaromatic interaction between the cationic center and the butadiene moiety. This was confirmed by a study of the 9-methoxy-9-bicyclo[4.2.1]nona-2,4-dienyl cation under conditions of long life. PMR and CMR data point to an asym­ metric interaction between the 9-carbon and the butadiene bridge. Obviously the mode of n participation is governed by the Woodward-Hoffmann rules according to which the symmetric interaction is "forbidden" and the asymmetric is "allowed". Finally, routes to the generation of the [4.2.1] anion have been developed. The 9-phenylseleno-9-bicyclo [4.2.1] nona- 2,4,7-trienyl anion could be generated from the correspond-

98 ing phenylseleno ketal. Electrophilic addition reactions at this anion proceed stereoselective, consequent on a stabiliz­ ing aromatic interaction·(6n) of the anionic center with the butadiene bridge. In this case the highly preferential lo­ cation of these 6n electrons gives rise to kinetic electronic control in which the electrophile approaches the ion from the side of the highest electron density.

99 SAMENVATTING

In dit proefschrift wordt de neighbouring group partiei­ pation beschreven van cr- en n-elektronen in rigide polycycli­ sche systemen. Ionisatiereakties van deze systemen vertoonden onverwachte solvolytische reaktiviteiten, die vaak gepaard gaan met skeletomleggingen. Wanneer slechts omleggingsproduk­ ten worden waargenomen is het moeilijk te bepalen of verhoog­ de reaktiesnelheden samenhangen met elektronische faktoren of het gevolg zijn van energetische energieroutes, welke be- schikbaar zijn door skeletomleggingen. Teneinde deze omleg­ gingsreakties te voorkomen, werd het reaktiecentrum gesub­ stitueerd met groepen die lading kunnen stabiliseren. Op deze manier was het mogelijk de struktuur van intermediairen op te helderen. De bestudering van antiaromatische interakties met behulp van kinetische en spektroskopische technieken bleek zelfs mogelijk te zijn. Reaktie van het homoaromatische 7-norbornadienyl kation met nukleofielen afgeleid van elementen van groep VA en VIA leidde afhankel k van hun basiditeit tot 2,3- en 7-gesub­ stitueerde produkten. Zo leverde het zwak nukleofiele tris­ (pentafluorofenyl)fosfine slechts 7-gesubstitueerde norbor­ nadienylfosfonium zouten op, terwijl uitsluitend 2,3-gesub­ stitueerde produkten werden verkregen met het sterker nukleo fiele trifenylfosfine. Daarentegen werJen slechts 7-gesub­ stitueerde produkten gevonden met het methoxy-gestabiliseerde 7-norbornenyl kation. De reaktie van trifenylfosfine met 7 chloor-7-methoxynorborneen gaf uitsluitend het 7-methoxy-7- norbornenyltrifenylfosfonium zout. De laatste verbinding bleek een geschikte precursor te zijn voor de bereiding van 100 het 7-trifenylfosfonio-7-norbornenyl dikation, een niet­ klassiek tweewaardig positief kation. Een onverwachte reaktiviteit werd waargeno~en in 7-nor- bornenyl systemen met een cyclobuteen in endo positie. Vastgesteld werd, dat tijdens de ionisatie van 9-chloor-9- methoxy-endo-tricyclo[4.2.1.02•5]nona-3,7-dieen anahimeria assistanae wordt verleend door twee dubbele banden en een sigma band. Het intermediaii wordt beschreven als een "sand­ wich complex" van cyclopentadieen en het cyclobutenyl kation, waartussen een aromatische (6w) interaktie optreedt. Dit intermediair kan, afhankelijk van de reaktiekondities, zowel de oorspronkelijke struktuur opleveren als het omgelegde tricycle [4. 3. 0. o2 •5] nona-3, 7-dieen. Het 9-bicyclo [4.2.1] nona-2,4,7-trienyl kation is in de belangstelling gekomen vanwege zijn mogelijk homoaromatische en bicycloaromatische karakter. Het bicycloaromatsiche ka­ rakter leek bevestigd te worden door bestudering van de reaktie.kinetiek. De sol volyse van het [4. 2. ·11 systeem leverde echter slechts omleggingsprodukten op. Ook in dit geval bleek een methoxygroep aan het reaktiecentrum omleggingsreakties te kunnen voorkomen. De ionisatiereaktie van 9-chloor-9-methoxy­ bicyclo [4.2.1! nona-2,4,7-trieen onder solvolytische kondities verloopt zonder skeletomleggingen. Een kinetische studie toonde aan, dat het trieen veel minder reaktief is dan de meer gehydrogeneerde derivaten. Deze verminderde reaktiesnelheid kon verklaard worden door aan te nemen, dat de interaktie tussen het kationcentrum en het butadieenfragment antiaroma­ tisch is. Dit werd bevestigd door een spektroskopisch onder­ zoek van het 9-methoxy-9-bicyclo[4.2.1] nona-2,4-dienyl kation onder langlevende kondities. PMR en CMR gegevens wijzen erop, dat het kationcentrum een interaktie vertoont met 66n van de dubbele banden van het butadieenfragment. Klaarblijkelijk wordt het interaktiemodel gekontroleerd door orbitalsymmetrie: de interaktie met een dubbele band kan alleen geschieden via de hoogst bezette MO van butadieen (HOMO). Tens lotte werd het [4. 2. 1) anion bestudeerd. Het 9-fenyl-

101 seleno-9-bicyclo[4.2.1] nona-2,4,7-trienyl anion kon bereid worden uithetovereenkomstige fenylselenoketaal. Dit anion ondergaat elektrofiele additiereakties selektief aan de buta­ dieenzijde. De hogere elektronendichtheid aan deze kant, die het gevolg is van een bishomoaromatische interaktie (6~) van het anioncentrum met butadieen, verklaart de stereoselektivi­ teit.

102 CURRICULUM VITAE

Na het behalen van het diploma HBS-b aan de van Olden­ barnevelt HBS te Rotterdam werd in september 1962 begonnen met de studie aan de Rijksuniversiteit te Leiden. Het kandi­ daatsexamen in de scheikunde met bijvakken wiskunde, natuur­ kunde en mineralogie (F-richting) werd in maart 1966 afgelegd. De studie werd voortgezet aan de afdeling Theoretische Orga­ nische Chemie van prof. dr. L.J. Oosterhoff. Onder leiding van dr. H.M. Buck werd een ESR onderzoek verricht aan organo­ fosforverbindingen. In april 1969 werd het doktoraal examen afgelegd in de theoretische organische chemie met als bij­ vakken theoretische natuurkunde en wiskunde. Gedurende de periode april 1969 tot januari 1971 werd de funktie vervuld van doktoraal assistent bij de afdel Thea- retische Organische Chemie van de Rijksuniversiteit te Leiden. In januari 1971 volgde de aanstelling als wetenschappelijk medewerker bij de vakgroep voor Organische Chemie van de Technische Hogeschool Eindhoven.

103 DANKWOORD

Aan het onderzoek dat in dit proefschrift is beschreven hebben velen een bijdrage geleverd op het gebied van synthese en spektroskopie. Met velen oak heb ik bijzonder stimulerende diskussies mogen voeren. Voor deze hulp ben ik hen zeer er­ kentelijk. Verder wil ik diegenen bedanken die hebben bijgedragen tot de uiteindelijke vormgeving van dit proefschrift.

104 STELLINGEN

1 De stereospecifieke syn additie van jodiumazide op de dub­ bele band van cyclobuteen in 9,10-dicarbomethoxytricyclo­ [4.2.2.02•5]deca-3,7-dieen is meer in overeenstemming met anchimeric assistance van de carbomethoxy groep dan met de dominerende rol van "twist strain" faktoren. G.Mehta en P.N. Pandey, J. Org. Chern., 40, 3631 (1975).

2 De voorkeurskonformatie van squaleen in een polair solvent kan niet zonder meer vergeleken worden met die in het kristalrooster. J. Ernst, W.S. Sheldrick en J.H. Fuhrhop, Angew. Chern., 88, 85 (1976).

3 Bij kwantumchemische berekeningen aan sigmatrope methyl­ shifts leiden voor de hand liggende modellen tot niet­ realistische beschrijvingen van de transition-state. N.D. Epiotis, R.L. Yates en F. Bernardi, J. Amer. Chern. Soc., 97, 4198 (1975).

4 De voorgestelde struktuur van het sulfinaatkomplex van 1,5-dimethyl-9-thiobarbaralaan-9,9-dioxide is niet in overeenstemming met het gepresenteerde PMR spektrum. L.A. Paquette, u. Jacobsson en s.v. Ley, J. Amer. Chern. Soc., 98, 152 (1976).

5 De gepresenteerde spektrale gegevens van 2 azobicyclo­ [3.2.1] octa-3,6-dieen en die van de meer verzadigde sys­ temen duiden meer op sterische en substituent effekten dan op hetero-homokonjugatie in eerstgenoemde verbinding. A.G. Anastassiou en H. Kasmai, J.C.S. Chern. Comm., 201 (1975). 6 Het mechanisme van de stereospecifieke omzetting van oxiranen in thiiranen met behulp van 3-methylbenzothiazol- 2-thion is uitstekend verklaarbaar wanneer het optreden van een intermediair episulfonium ion wordt verondersteld. v. Calo, L. Lopez, L. Marchese en G. Pesce, J.C.S. Chern. Comm., 621 (1975).

7 Omleggingsreakties van carboniumionen in sterk zure media zijn niet a priori vergelijkbaar met die onder solvolyti­ sche kondities.

T.S. Sorensen, Ace. Chern. Res.,~~ 257 (1976).

8 De door Olah gepresenteerde PMR en CMR spektra van 3- methyl-endo-tricyclo[5.2.1.02•6]deca-4,8-dienyl kation wijzen niet op een interaktie van het allyl systeem met de geisoleerde dubbele band. G.A. Olah, G.K. Surya Prakash en G. Liang, J. Org. Chern. , !!_, 2820 (1976) .

9 De bewering van Pagni dat de hydrideabstraktie van alkanen in sterk zure media in feite een elektronen-overdrachts mechanisme is, wordt onvoldoende door de experimenten ondersteund. R.M. Pagni, P.A. Bouis en P. Easley, Tetrahedron Lett., 2671 (1975) •

10 Wanneer radikalen afgeleid van fosforverbindingen geen fosforradikalen genoemd worden, verdient het aanbeveling te vermelden bij welke fosforsplitsing deze naamgeving niet meer van toepassing is. John Ernsley and Dennis Hall, "The Chemistry of Phos­ phorus", Harper and Row, London, 1976, Chapter 9.

11 De agressiviteit van de nederlandse automobilist is om­ gekeerd evenredig met de kapaciteit van het wegennet. 12 Het toepassen van een "vrije opvoedingsmethode" blijkt bij een aanzienlijk aantal kinderen tot stoornissen in het koncentratievermogen te leiden.

13 Indien het ethisch reveil betekent, dat men het justitieel apparaat kan inschakelen om feiten niet volgens de geest maar volgens de letter van de wet te vervolgen, vormt het een bedreiging voor de demokratie.

14 Een grondpolitiek die de eigen grond onder de voeten doet wegzinken, moet wel op zwakke fundamenten gebaseerd zijn.

P. Schipper Eindhoven, 14 juni 1977