STRUCTURE AND REACTIVITY OF POLYCYCLICCROSS-CONJUGATED ir-ELECTRON SYSTEMS

KLAUS HAFNER

Institute of Organic Chemistry, Technische Hochschule Darmstadt.Germany

ABSTRACT

In contrast to the well-known monocyclic conjugated systems with (4n + 2)and 4nit-electrons, non-benzenoid polycyclic conjugated it-electron systemwhich should not obey Hückel's rule contain the element of cross-conjugation.In bicyclic and linear annelated polycyclic conjugated systems this cross-conjugation is not associated with branching of the it-electron system, but thiswill however occur in pericondensed tn- and polycyclic compounds. Theparticipation of the element of cross-conjugation in the it-electron systems ofsuch polycycles should affect their properties and result in characteristicdifferences in bonding character and reactivity compared with monocyclicconjugated compounds with the same number of it-electrons.

To obtain experimental support for these theoretical predictions, severalnon-benzenoid bicyclic as well as tn- and tetracyclic pericondensed it-electronsystems were synthesized and studied with respect to the magnetic criteriaof aromaticity and the connection between structure and reactivity. Thesuccessful preparation of most of the described hydrocarbons has centred on a

single basic and rational synthetic principle.

INTRODUCTIONIn the last few decades, the chemistry of aromatic compounds has occupied

the interest of chemists to an increasing extent. The synthetic accessibility ofnumerous new cyclically conjugated it-electron systems and their theoreticalunderstanding has given rise to vigorous development of an interesting fieldof organic chemistry. Owing to the advancement of quantum chemistry theconcept of the aromatic sextet1 of electrons has been deprived of its leadingrole and has been replaced by the postulate that planar monocyclic con-jugated systems with a closed shell configuration of(4n + 2)it-electrons shouldin general possess special electronic stability2. Besides the well knowncyclopentadienyl anion, the tropylium cation and the related tropone andtropolones which support the electron sextet3, the successful synthesesof cyclopropenylium cations4, of the equally stable negatively chargedlOir-electron systems—cyclononatetraenyl anion5 and cyclooctatetraenyldianion6—as well as of the annulenes7 and bridged annulenes8 have con-firmed these views. In studying these cyclically conjugated it-electronsystems the more or less arbitrary nature of a differentiation between aro-matic and non-aromatic compounds based solely on criteria of special types

153

KLAUS HAFNER

of reactivity and kinetic stability must not be overlooked. More recentlya thermodynamic criterion has been applied whereby cyclic conjugatedsystems are considered aromatic if cyclic delocalization of ic-electrons makesa notable negative contribution to their heat of formation. Furthermoresuch aromatic' molecules are capable of sustaining a strong induced diamag-netic ring current in the presence of an externally applied magnetic field.This property manifests itself in the n.m.r. spectra9 or magnetic susceptibili-ties'° of such molecules. As an extension of this definition of aromaticity,Breslow1' predicted the precisely opposite phenomenon, namely destabiliza-tion by it-electron delocalization. He termed this phenomenon 'antiaromati-city signifying that in such molecules resonance would lead to an increasein energy. In accord with this it has recently been possible to show that inseveral cyclic 4n n-electron systems a paramagnetic ring current is induced7" 2•

In the past five years several monocyclic conjugated systems with (4n + 2)and 4n n-electrons have been most successfully studied from this point ofview7'8" . But nonbenzenoid polycyclic conjugated n-electron systemswhich should not obey HUckel's rule are also well worth consideration. Incontrast to monocyclic molecules these systems contain the element of cross-conjugation. In bicyclic and linear annelated polycyclic conjugated systemsthis cross-conjugation is not associated with branching of the it-electronsystem, but this will, however, occur in pericondensed tn- and poly-cycliccompounds'4. The participation of the element of cross-conjugation in then-electron systems of such polycycles should affect their properties and resultin characteristic differences in bonding character and reactivity comparedwith monocyclic conjugated compounds with the same number of n-elec-trons. It was, therefore, of interest to study such molecules in respect of thethermodynamic and magnetic criteria of aromaticity, so learning moreabout the connection between structure and it-electron delocalization on theone hand and the ability of polycyclic systems to sustain an induced dia-magnetic or paramagnetic ring current on the other.

Prompted by the desire to answer these questions we have synthesizedseveral polycyclic nonbenzenoid conjugated compounds. Systems composedof 5- and 7-membered rings seemed to be most promising because thesecontain the cross-conjugated patterns of double bonds as found in the wellknown penta- and hepta-fulvenes3. Thus, azulene (I), for example, combinesthese two cross-conjugated systems, the result being a resonance stabilizedbicyclic iOn electron system, a formal combination of the two charged

(a) (I) (b)

\+RHR \

(III) (II)

154

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

sextets of the cyclopentadienyl anion (Ia) and the tropylium cation (Ib).In accord with such considerations electrophilic reagents react with azulenein a way reminiscent of their behaviour towards fulvenes and the cyclo-pentadienyl anion. In this way attack takes place at the 5-membered ringand is concomitant with the formation of a tropylium cation (11)15. Nucleo-philes on the other hand, as might be expected from their behaviour towardsheptafulvenes and tropylium cations, react with the positively polarized7-membered moiety of azulene with formation of a cyclopentadienyl anion

CH

{R2 lBxe —HX

/+ HX/c'cH

+ R2N—(CH=CH)—CHO(cH2

1120

çCH __C(CH*)

(CH -(c' \\ -R2NHCH),'H r!.4R2

NR2HII ci-!(c \ -R2HCH)—.-

//(CH

(CH('c-' \\CH).

(C*CH\CH)

(111)16. A lack of reactivity towards dienopbiles17 indicates a pronouncedperipheral ic-electron delocalization in azulene. The n.m.r. spectrum ofazulene (Figure 1) showing the proton signals in the region characteristicfor benzenoid compounds confirms the assumption and indicates stronginduced diamagnetic ring currents in the 5- and 7-membered rings1 8•

The successful preparation of numerous nonbenzenoid polycyclic con-

H5C6

+ H3CCH=CHCH(IV)

OCH_CH*

Bj—H20

(V)

H5C6CH3 H5C6CH3

(VI)

C6H5NH

CH3

(1)

155

KLAUS HAFNER

jugated ic-electron systems has centred on a single basic and rationalsynthetic principle, the substitution of mono- or poly-cyclic systems by anacid amide or its vinylogues via condensation or electrophilic substitutionto give an enamine or immonium salt, subsequent intramolecular cycliza-tion and final n-elimination.

For the syntheses of polycycles with 5-membered rings cyclopentadieneproved to be a suitable starting material. In the case of the synthesis ofazulene'9 the reaction scheme consists of base catalysed condensation ofcyclopentadiene (IV) with 5-(N-methylanilino)2,4-pentadienal (V) to givethe fulvene (VI) and subsequent intramolecular electrophilic substitutioninvolving and all-cis transition state with elimination of N-methylaniline. Thedriving force for the amine elimination is the gain in delocalization energyof the bicyclic lOit-electron system. That this reaction is so facile is under-standable in terms of Baker's 'rigid group principle'20. Several new bi- andpoly-cyclic systems are accessible by appliçation of similar reaction schemes.A few representative examples are discussed below.

PENTALENE AND HEPTALENEJust as azulene (I) can be considered as a cs-bridged and therefore planar

system related to cyclodecapentaene, so may pentalene (VII) and heptalene(VIII) be viewed respectively as planar cr-bridged cyclo-octatetraene orcyclododecahexaene. Unlike azulene (I) the two latter compounds lackresonance stabilization. They are 4n ic-electron systems and therefore, ifthe Hückel rule is applicable, non- or even anti-aromatic.

CO CO(VII) (VIII)

Dauben's2' elegant synthesis of heptalene (VIII) has shown it to be therm-ally unstable. The highly reactive bicyclic l2ir-electron system is obtainedonly at temperatures below —70°C. The n.m.r. spectrum of heptalene (VIII)(Figure 1) shows signals between 3.8 and 5.2 r indicating either a smallparamagnetic ring current or, more probably and, in accord with the resultsof quantum chemical calculations22'2 3, a lack of ic-electron delocalization.

As for heptalene (VIII), theory23 also predicts thermal instability and highreactivity for pentalene (VII). The few well known derivatives of pentalene,such as dibenzo[ae]pentalene (IX)24 or hexaphenylpentalene (X)25, do not

H5C6 C6H5

EZ1IjIII1I H5C6——C6H5(IX) H5C6 C6H5

(X)

provide the best information about the bicyclic 8ic-electron system. In thefirst case two of the double bonds are part of benzenoid systems whilst inthe second case the inductive and mesomeric effects of the phenyl groupscannot be neglected. The synthesis of pentalene as such has so far proved tobe elusive.

156

POLYCYCLIC CROSS-CONJUGATED ic-ELECTRON SYSTEMS

In view of the theoretical predictions it is not surprising that attemptedsynthesis of pentalene (VII) in a way analogous to that used for azulene (I),i.e. by ring closure of 6-(2'-dimethylaminovinyl-l')-fulvene (XI), was notsuccessful. Although amine elimination takes place even at low temperatureonly ill-defined tarry products could be isolated26.

7

Figure 1. N.m.r. spectra of azulene (I) and heptalene (VIII) (H. J. Dauben and D. J. Bertelli21) incarbon tetrachioride.

Attempted synthesis of simple stabilized pentalene derivatives provedto be more informative. Just as pentafulvene is stabilized by electron-donating groups at the exocyclic carbon atom27, so might pentalene—abicyclic fulvene—be stabilized by such groups in the 1- or 3-position.6-Dimethylaminofulvene (XII)28 is resonance stabilized by the contribution

3.8—4.5

cc4.7-5.2

TMS

1.6—3.2 ccTMS

I I I1 2 3 4 5 6 7 8 9 10

0(IV)

H2

+

.CHOHC

CH

CH3)2

-H20

JF(CH3)2—(CH3)2NH

N(CH3)2I-J(XI)

157

w(VII)

KLAUS HAFNER

of a dipolar 67t-electron canonical form. Similar stabilization was to beexpected for 1-aminopentalenes.

N(cH3)2

(XII)

N(CH3)

In an attempt to verify this supposition sodium cyclopentadienide wascondensed at —20°C with the salt (XIV) obtained by reaction of N, N,N',N'-tetramethyl-3-aminocrotonamide (XIII) with triethyloxonium fluoroborate.The resulting fulvene (XV.) rapidly lost dimethylamine even below 0°C anda yellow crystalline thermally unstable product was isclated in 20 per centyield. The constitution of this product was shown to be that of the tautomer

CH3CH,N(CH3)2

(CH3)2N

(XIII)

N(cH3)2

(XVI)CH3

R(CH

(XVII)

0÷{(c2H5)3orBF4e

—(C2H5)20

N(cH3)2

H3C N(CH3)2

CH3CH..,N(CH3)2I i( BF4e

(cH3),ri 02H5(XIV)

+ JNa —C2H,OH— —NaBF4

(CH3)2f H3

oc f\J (XV)

(XVII) of the desired pentalene derivative (XVI)29. This unexpected resultpoints once again to the thermodynamic instability of the bicyclic andpossibly antiaromatic 8ir-electron system. In a similar fashion intramolecularcyclization even at 20°C of the 6-methyl-6-(2'-methyl-2'-dimethylamino-vinyl-1')-fulvene (XIX), obtained by reaction of sodium cyclopentadienidewith the immonium salt (XVIII) leads to 25 per cent of 1,3-dimethyl-3-dimethylamino-2,3-dihydropentalene (XX) together with 15 per cent of thetautomer (XXII) of the desired 1,3-dimethylpentalene (XXI)30. These resultscontrast with the corresponding but successful preparation of azulenederivatives where the products are aromatic lOir-electron systems.

It was therefore apparent that one amino substituent does not confersufficient stability upon the bicyclic octatetraene system. We therefore

158

CH3 CH3 CH3 CH3

(CH3)2N=C ,-e CH -(CH3)2H N(CH3)2

BF4—NaBF4

(XXI) CH3

XIX)

sought a method for the further enhancement of the thermodynamic stabilityand turned our attention to a synthesis of a 1.,3-bis(amino)pentalene.Accordingly 3-(dimethylamino)- 1(2H)-pentalenone (XXV) was prepared bycondensation of sodium cyclopentadienide with the alkylation product(XXIII) from N,N,N',N'-tetramethyl-malonamide and triethyloxoniumfluoroborate and following cyclization of the aminofulvene derivative(XXIV) in boiling xylene. The bicyclic fulvene (XXV) proved to be a suitableprecursor for the 1,3-bis(dimethylamino)pentalene (XXIX)31

The conversion of derivatives of bicyclo[3,3,O]octanes to pentalenes hasfrequently been studied but without success. Of significance in this context

01/(CH3)2N— C\

CH2/(CH3)2N—C\

+ [(c2H5)30]WBF4e

-(C2H5)20

0C2H5

(CH)2N-c" BF4CH2

(CH3)2 N— C\\0 (XXIII)

+ Na —NaBF4—C2H5OH

(CH3)2

H2(xxv)

140°C

-(CH)2NH

159

F(CH3)2{CH2\:J /\\(CH3)2N 0

(XXIV)

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

Na +(XVIII)

—(CH3)2iH

120°C

H3H3C (CH3°2

(XX)

CH3

K' I )H2\Lm/

(XXII) 'H2

KLAUS I-IAFNER

e0 0

0 00(XXVI)

(CH) fi(CH)

e32(XXV) (XX VII)

is an experiment by Dauben32: namely the reaction of the bicyclo[3,3,0]-octadiendione (XXVI) with strong bases. The expected enolization of thediketone was not observed; an enol was not detected even in traces. Con-version of the twofold c,13-unsaturated carbonyl system to the bicycloocta-tetraene is not energetically favourable. As expected, 3-(dimethylamino)-1(2H)pentalenone (XXV) shows a different behaviour. The yellowishcompound can be converted with bases such as potassium tert-butoxide orGrignard reagents at 20°C in ether to a blue thermally not too stable material(XXVII), the ultra-violet spectrum of which resembles that of hexaphenyl-pentalene (X) and which regenerates the starting material with protic acids.The n.m.r. spectrum confirmed the assumption that the enolate of (XXV) ispresent. Unfortunately attempted isolation of the enolate met with no moresuccess than did its conversion to an enol ether31.

An access to simple, stable pentalenes was, however, opened when thepentalenone derivative (XXV) was reacted with dimethylamine in thepresence of perchioric acid. The stable crystalline symmetrical immonium

N(CH) (CH )2

H

2

+[(CH,)2NH2]CIO4 f\ H2 -H20 \L./ c104e0 N(CH3)2(XXVIII)

(XXV)-HC1O4//

//+HCiO4

N(CH3)2 (CH3)2 N(CFI3)2

CH3)2- CH3)2 -

CH32(XXIX)

160

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

salt (XXVIII) is formed. The deprotonation to 1,3-bis(dimethylamino)-pentalene (XXIX) occurs easily with strong bases in aprotic solvents, i.e.with isopropylmagnesium chloride in ether. The substituted pentalene isobtained as deep blue crystals, soluble with a blue colour in polar aproticsolvents, stable to heating to 120°C or for some time towards oxygen,sublimable at 115°C in high vacuum and melting with decomposition at163°C. In protic solvents the compound is quickly converted to violet andstrongly fluorescent materials of unknown constitution whilst with mineralacids in aprotic solvents the pentalene derivative (XXIX) reacts as an enamineregenerating the immonium salt (XXVIII)3'.

Figure 2. Ultra-violet spectra of 1,3-bis(dimethylamino)pentalene (XXIX) in methylene chlorideand of hexaphenylpentalene (X) in dioxane (E. LeGoff25); vertical rulings show the predicted

absorption maxima for (XXIX).

The ultra-violet spectrum (Figure 2) of the pentalene derivative (XXIX)resembles that of hexaphenylpentalene (X) and agrees satisfactorily withSCF-calculations33. The calculated values for ic-electron densities and bondorders are well consistent with enamine character and require a highelectron density at C-2, the centre of the trimethinecyanine group. The sameconclusion must be drawn from the n.m.r. spectrum (Figure 3) where, inaddition to the olefinic A2B-spectrum of the unsubstituted 5-memberedring, the doublet for the two protons at C-4 and C-6 and the triplet for theproton at C-S with a coupling constant of 3.4 Hz, one observes 12 methylprotons as a singlet at 6.92 and the C-2 proton as a high field singlet at 7.20 t,This latter value is characteristic for 3-protons of enamines. It follows that

161

4— V

200 300 1.00 500 600 700 800 nm

KLAUS HAFNER

the n.m.r. spectrum of this bicyclic compound offers no indication of anysizeable ring current.

It is apparent that the amino groups in the 1- and 3-positions of (XXIX)exert a strong influence on the bicyclic it-electron system which may beaccompanied by a significant stability increase. The reactivity of the moleculeis essentially that of the enamine function. This is reflected not only in the

7.20

IMS

Figure 3. N.m.r. spectrum of 1,3-bis(dimethylamino)pentalene (XXIX) in hexadeuteroacetone.

above mentioned reversible protonation but also in 1,2-cycloaddition withdimethyl acetylenedicarboxylate which takes place even below 0°C. Theexpected tricyclic adduct (XXX) is not isolatable due to its rapid valenceisomerization to the red azulene derivative (XXXI)3 '. Such a tendency toundergo cycloaddition to activated acetylenes has been similarly observed,

P(cH) N(CH3)23 2

KIII1COOCH,(CH3)2 (CH3)2N COOCH

(XXIX)(XXX)

CH3)2I,-'' COOCH3OOCH3

t(CH3)2

(XXXI)

162

N(CH3)2

N(CH3)2

6.92

/.31

0 1I I

2 3 4 5 6 7 8 9 10T

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

albeit at elevated temperatures, for hexaphenylpentalene (X)25. The inherentinstability of the pentalenoid it-electron system is once again demonstratedby the reaction of (XXIX) with active methylene groups. Thus followingattack at C-i elimination of dimethylamine is in the sense required to producecompound (XXXII) rather than to regenerate the energetically less favouredpentalene (XXXIII)34. It may be seen from the foregoing examples that theformation of the less stabilized bicyclic 8it-electron system is not favoured.This is surely due to resonance destabilization which is in turn obviouslymoderated or perhaps even cancelled by electron-donating substituents.

f4(CH3)

H2(CH3)2N' CH

A possibly promising route to pentalene itself or to simple alkyl or arylderivatives involves 6-(2'-dialkylamino-vinyl-l')-fulvenes (XXXIV). Theseare easily prepared by condensation of co-amino-acroleins or J3-dialkyl-aminovinyl ketones with cyclopentadiene and undergo intramolecularcyclization in boiling piperidine. This cyclization proceeding by Michaeladdition is followed by isornerization and the resonance stabilized 1-dialkyl-amino-2,3-dihydropentalenes (XXXV) are obtained in 50 to 70 per centyield. Reduction with lithium aluminium hydride or alternatively with alkylor aryl lithium and subsequent amine elimination yields the hithertounknown 1,2-dihydropentalene (R' = H) or its 3-alkyl or -aryl derivatives

(XXXIV)

2

(XXXV)

163

(CH3)2

N(CH3)2

X+ H2CY

(XXIX)

(XXXII) (XXXIII)

KLAUS HAFNER

NR2

(XXXV)

+L1R'

or LiAIH4

R R2

IH2Li H2

+ H20

tJ5H2(XXXVI)

H2

—R2NH

H1\ R2

(JIH2

(XXXVI) respectively30' 35• The attractive possibility of dehydrogenation ofthese hydrocarbons is currently under investigation.

CYCLOPENT[cd}AZULENEIt should be emphasized that isolatable compounds which may be for-

mulated as pentalenoid have been known for several years. Thus the.cyclopent[cd]azulene system (XXXIX) can be viewed as a pentalene deriva-tive with a pen-fused cycloheptatriene ring. Here the pentalene moiety isagain stabilized by an electron donor, the cycloheptatriene ring exhibitingits tendency towards the 6ir-electron cation. The preparation of this yellowl2it-electron hydrocarbon is related to the basic principle of azulene syn-thesis. The aldimmonium salt (XXXVIII), easily prepared by Vilsmeier

CH3

(XXX VII)

4-(CH3)2N-CHO+Poc13—HPO2CI2

CH3

(T/L_CH3

e HC\ CH3ci N(CH3)2(XXXVIII)

lB t_HCI

CH

HA H2t(CH3)2

CH3

(XXXIX)

-(CH3)2NH

164

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

formylation of 4,6,8-trimethylazulene (XXXVII), undergoes base-catalysedintramolecular cyclization involving one of the activated methyl groups atthe 7-membered ring thus giving 65 per cent of the tricyclic system (XXXIX)36.

In good agreement with theoretical predictions37 based on the SCF-method the x-ray analysis38 of the 2-phenyl derivative of (XXXIX) shows avirtually planar structure and nearly constant bond lengths in the 7-memberedring. The 5-membered rings show alternation of bond length in the range1.465 to 1.356 A (Figure 4). The n.m.r. spectrum of (XXXIX) (Figure 5) isquite compatible with these results, does not bear comparison with the

Figure 4. Experimental and calculated bond lengths of 2-phenyl-5,7-dimethyl-cyclopent[ccl]azulene.

isomeric acenaphthylene and is, therefore, clearly not that which would beexpected for an ethylene-bridged azulene. It shows resonances for the7-membered ring protons and for the methyl protons at the same positionsas observed for the comparable 4,6,8-trimethylazulene (XXXVII). However,the two AB-systems of the four 5-membered ring protons appear at higherfield than do the corresponding azulene protons. This indicates that the7-membered but not the 5-membered rings of (XXXIX) may be comparedwith the corresponding ring of azulene. Thus the larger ring of (XXXIX) iscapable of sustaining an induced diamagnetic ring current whilst in the twosmaller rings this current is much reduced, perhaps even to be replaced by aslight paramagnetic effect39. In this light the hydrocarbon seems to representa superposition of the aromatic azulene and the non- or even anti-aromaticpentalene.

This dualism of behaviour of the hydrocarbon (XXXIX) is mirrored byits chemical properties. Thus on the one hand attack by electrophules at the5-membered rings40 is reminiscent of azulenes whilst on the other handfacile 1,2-cycloaddition to one of the 5-membered ring double bonds is asexpected for a pentalene derivative and moreover regenerates the azulenoid

165

r9

Experimental*Calculated (SCF calculation with vartat ion)

CH3

CH3

2 3-3 0

CH3

CH3

I I I1 2 3 4 5 6

Figure 5. N.m.r. spectra of 5,7-dimethyl-cyclopent[cd]azulene (XXXIX) and 4,6$-trimethyl-azulene (XXX VII) in carbon tetrachioride.

KLAUS HAFNER

7.2—7.5

+ N2CFICOOC2H.5/Cu

-N2

CH

(lL\_CH3

(XXXIX)

CH3

(XL)

CH3

H2 CH3

(XLII) (XLI)

166

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

it-electron system. It is not surprising that the tricyclic hydrocarbon (XXXIX)reacts with ethoxycarbonylcarbene—generated by copper catalysed thermo-lysis of ethyl diazotate—to give the cyclopropane derivative (XL) bycycloaddition to the 1,2-bond41. The product (XLII) isolated is, however,a derivative of the hitherto rather inaccessible 2H-benz[cd]azulene42 whichwarrants interest as an isomer of phenalene. Compound (XLII) probablyarises by spontaneous valence isomerization of (XL) to the cross-conjugatedtricyclic compound (XLI) which then undergoes a hydrogen shift.

CH3

CH3A

CH3

(XXXIX)

CH3

CH3

CH3OOCCH300C

25°C +CH3OOC—CC----COOCH3

CH3

ç1)_CH3

jCOOCH3COOCH3

B

CH

CH3

CH3OOCCH300C

+

(XLIII) (XLIV)

140°C 140°C

CH3

ç_CH3

COOCH3

(XLV) (XLVI)

Furthermore the cyclopent[cd]azulene (XXXIX) combines with dimethylacetylenedicarboxylate at 25°C in the 1,2- or 3,4-position to regenerate theazulenoid it-electron system and to form the blue cyclobutene derivatives(XLIII) and (XLIV) in the ratio 4:1. The high bond order of the 1,2- and3,4-linkages in (XXXIX) and the polarization consistent with participationof the resonance structures (XXXIXA and B) permit the suggestion that thereaction of (XXXIX) with electron-deficient alkenes36 and alkynes are two-

167

KLAUS HAFNER

step cycloadditions proceeding via resonance stabilized dipolar inter-mediates. Thermal valence isomerization of the adducts in boiling xyleneleads to the green aceheptylenes (XLV) and (XLVI) in yields of about 90 percent43. In these tricyclic systems the unstable heptalene is fused in a pen-position with a cyclopentadiene ring which, in accord with its enhancedstability in the anionic form, acts as an electron acceptor and stabilizes theheptalene moiety.

ACEHEPTYLENE

Whilst several substituted aceheptylenes are accessible by a route36'44analogous to that used for the preparation of cyclopent[cd]azulenes(XXXIX), the same is not true for the parent hydrocarbon (LI). This wasprepared starting from 4-methylazulene (XLVII) which reacts with sodiumN-methylanilide to give the sodium azuleniate (XLVIII)15. Further treat-ment with the immonium salt (XLIX) yields the azulene derivative (L). Inanalogy to the synthesis of azulene (I) the dienamine (L) undergoes thermalintramolecular cyclization with loss of N-methylaniline to give 35 per centof the aceheptylene (LI)45.

The picture presented by x-ray analysis and n.m.r. studies of aceheptylene(LI) is similar to that of cyclopent[cd]azulene (XXXIX) being one of a super-position of the nonbenzenoid structures of azulene and of heptalene. Inthe ground state, therefore, the molecule should neither be represented as anonbenzenoid analogue of pleiadiene—i.e. as a 1,8-diene bridged azulene—nor as a 1,10-ethylene bridged heptalene. The x-ray analysis46 of the 3,5,8.10-tetramethyl derivative of (LI) (Figure 6) shows in accord with SCF-calcula-lions37 a symmetrical planar structure and bond alternance in the 7-mem-bered rings. Signifying a diamagnetic ring current in the 5-membered ring

H,C\_+

C6H, NaW L) /CH2 I

(XLVIII) CH2(XL VII)

H5C5N_ ,,CH3+ ,N—CH=CH—CH=NN

H,C C6H5

(XLIX) dO?

CH,NH. —NaC1O4

H,C6

H3C

H,c(N\ /CHcH—cH

(LI) (L)

168

e

POLYCYCLIC CROSS-CONJUGATED m-ELECTRON SYSTEMS

1.393

Figure 6. Experimental and calculated bond lengths of 3,5,8,1-tetramethy1aceheptylene.

the n.m.r. spectrum of (LI) (Figure 7) shows resonance for the two 5-mem-bered ring protons as a low field doublet. In contrast it is probable that the7-membered rings sustain a weak paramagnetic ring current for the signalsdue to the 7-membered ring protons appear at higher field than observedfor corresponding azulenic protons39. The n.m.r. spectrum of 3,5-dimethyl-aceheptylene (LII) (Figure 7) adds further support to this suggestion, themethyl protons showing signals above 8 t and thereby contrasting withmethyl proton resonance for analogous azulenes. One might justifiablyimagine aceheptylene as a superposition of two azulene and one heptaleneunit such that the diamagnetic ring current of the 5-membered ring mayweaken the intensity of the paramagnetic ring current of the heptalenemoiety. We may expect it to combine the chemical properties of azulene (I)and heptalene(VIII).

The aceheptylene system indeed reacts with electrophiles with substitu-tion in the 5-membered ring. However, it is surprising that substitution takesplace preferentially in the 4- and 6-positions, that is to say in the 7-membered

CH3

CH(LII)

+C(CN)2

169

— CH3(CN)

(CN)2

CH3

(LIII)

E x per me nt o

*C1 Ltd (SCF calculation)

KLAUS HAFNER

8.0- 8.3

TMS

Figure 7. N.m.r. spectra of aceheptylene (LI), 3,5-dimethylaceheptylene (LII) and 4,6,8-tn-methylazulene (XXXVII) in carbon tetrachioride.

ring44'47—49. This is probably a reflection of kinetic versus thermodynamiccontrol. On the other hand the hydrocarbon reacts with dienophiles in aDiels—Alder fashion involving one of the formal 7-membered ring dieneunits (LII —÷ LIII)36'44.

PENTALENO{6,6a,1,2-del]HEPTALENEWe have seen from the properties of cyclopent[cd]azulene (XXXIX) and

of aceheptylene (LI) that in pen-condensed tricyclic systems of two 5-andone 7-membered ring or of two 7- and one 5-membered ring we have com-bined both aromaticity and non- or even anti-aromaticity. It follows that byappropriate combination of these structural elements it should be possible

170

hO- 5.0

7.2

CH3

c?-cH3CH3

7.5

UJ1. 5 6 7 8 9 10

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

to construct either highly aromatic or alternatively rather antiaromaticmolecules. These, as will be seen from the following two examples, seem tobe realizable possibilities.

Addition of a second 7-membered ring in the pen-positions of the azuleneportion of cyclopent[cd]azulene (XXXIXor alternatively of a second

iTh

(XXXIX) /5-membered ring to aceheptylene (LI) gives pentaleno[6,6a,1,2-def]heptalene(LV). This nonbenzenoid isomer of pyrene is formally not only a combina-tion of two azulenes but also one of pentalene and heptalene. Quantumchemical calculations33'39'5° would favour the first of these two combina-tions as the more meaningful image; all four rings should sustain induceddiamagnetic ring current and the l6ir-electron system should have highelectronic stability51.

In order to add substance to this suggestion we determined to synthesizethe tetracyclic system (LV). To this end our basic synthetic principle couldonce again be applied. In this way either the aceheptylene derivative (LI V)36can be converted directly to the trimethyl derivative of (LV)47 or alterna-tively the easily accessible 4,6-dimethyl-1,8-cyclopentenoazulene (LVI)could be subjected to modified Vilsmeier reaction with 3-(N-methylanilino)-acrolein and phosphorous oxychioride to yield the aldimmonium salt(LVII) This on treatment with bases undergoes ring closure with theneighbouring activated methyl group. Spontaneous elimination of N-methyl-aniline then yields the hydrocarbon (LVIII) which is transformed in 65per

N(CH3)2

H3CçCH3cio?

H3C2/J(3 3 FI3C Cl-I3

(LIV)

171

(LV)

(LI)

1 2 3 4 5 6 7 8 9 10

FigureS. N.m.r. spectra of 5-methy1penta1eno[66a,1,2-def]hepta1ene (LIX) in deuterochioroformand of 46,8-trimethy1azu1ene (XXX VII) in carbon tetrachioride.

172

(LVI)

(LVII)

lB

H5C_+ N—CH==CH--CHO

H3C+POCI-HPO2CI2

—H2

CH3

HCI

H5C6NNH

H3C7

(LIX) (LVIII)

73

1.6 -3,5

TMS

7.2

-CH3CH3

2.8

2.2

TMS

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

cent yield to (LIX)52 on treatment with ch)oranil in benzene at 20°C. The easeof this dehydrogenation is noteworthy and indicative of the thermodynamicstability of the product.

The n.m.r. spectrum (Figure 8) of (LIX) is just as might be predicted ontheoretical grounds39. The chemical shifts of the multiplets for the nine ringprotons lie in the range 1.6 to 3.5 r thus pointing to a high diamagnetic ring

Figure 9. Experimental and calculated bond lengths of 5-methylpentaleno[6,6a,1,2-def]heptalene(LIX).

current in all four rings. Furthermore the dipole moment of 1.5 D is incom-patible with charge localization in the ground state. It is, however, surprisingthat the x-ray analysis53 reveals a not completely planar structure with aslight propeller-like distortion of the heptalene moiety and a correspondinglyslight deviation towards pyramidal geometry about the two central C-atoms.These distortions must be too small to influence the ir-electron system to anysignificant extent. Contrary to SCF-calculations33 which predict a uniformbond alternance throughout the molecule, the x-ray analysis shows azulene-like bond order in rings A and C (Figure 9) but relatively pronounced bondalternance for the remainder of the molecule. The hydrocarbon would seemto be composed therefore of one azulene portion which is conjugated with aa triene unit (LIXa). The chemical behaviour of the hydrocarbon (LIX), e.g.

173

1.3821.378

E xperimentat*C lid (SCF co1cua±ion with variafion)

KLAUS HAFNER

CR3

(LIXa)

towards catalytic hydrogenation, dienophiles or electrophiles, closelyresembles that ofazulene52.

AZULENO[8,8a,1,2-del]HEPTALENEWhilst pentaleno[6,6a,1,2-def]heptalene (LV) shows a degree of n-elec-

tronic stabilization, the same should not be expected for the 18n-electronazuleno[8,8a,1,2-def]heptalene (LXVIII). Here the azulene unit is combinedwith heptalene, three 7-membered rings sharing one 5-membered ring.Correspondingly and in spite of its belonging to the (4n + 2) it-electronseries this molecule should possess only small it-electronic stability. More orless intensive paramagnetic ring currents in the 7-membered rings shouldcounterbalance a diamagnetic ring current in the 5-membered ring andcause a relative upfield shift of the proton of the 5-membered ring39. So as toobtain experimental support for these theoretical predictions we syn-thesized this hydrocarbon.

\/(LXViii)

At first sight the synthesis of (LXVIII) by annelation of a third 7-memberedring to aceheptylene (LI) appeared to be a simple matter, but unfortunatelydid not in fact prove to be realizable owing to the exclusive substitution ofthis hydrocarbon by 3-(N-methylanilino)-propenal in the presence ofphosphorous oxychioride in the 7-membered rings. Thus, for example,3,5,8,10-tetramethylaceheptylene yields the propenals (LX) and (LXI)which are transformed in the presence of base to the benzo derivatives(LXII) and (LXIIfl47.

An appropriate aldimmonium salt (LXVII) could nonetheless be obtainedin a slightly roundabout fashion. The 7,10-dihydro derivative (LXV) of the3-methylaceheptylene (LXIV), which is prepared from the latter by partialreduction with lithium aluminium hydride, reacts with electrophiles as theazulene derivative which it is, exclusively in the 3-position referring to azulene.Thus the dihydro immonium salt (LXVI) could easily be obtained and thisin turn dehydrogenated by chioranil to regenerate the aceheptylene structure

174

(LI)

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

H5C6(1) + N—CH=CH—-CHO

+P0c13

(2) OHe/H20

H3C CH3

+

3 HC T CH3HCCH

(LXI) CHO20— 20

H3C I '7 \T 'i' \J H3C J''LCH3H3C CH3 \\ I /-_-, -(

(LXII) H3C I \\ll(LXIII) N—'in (LXVII). Base-catalysed ring closure with the activated methyl groupyields seven per cent of the desired hydrocarbon (LXVIII), a brownishyellow, thermally labile crystalline substance which is fluorescent in solu-tion47'54. The hydrocarbon (LXVIII), like the otlr previously discussedpolycyclic systems, forms an isolatable complex with trinitrobenzene but ismoreover distinguished by its remarkably pronounced basicity In this way(LXIX) is soluble with violet colour even in 2N sulphuric acid from which itcan be regenerated unchanged on dilution with water. The constitution ofthe conjugate acid (LXX) was inferred from its n.m.r. spectrum54.

The ultra-violet (Figure 10) and n.m.r. spectra (Figure 11) of hydrocarbon(LXVIII) and its methylated derivative (LXIX) accord with quantumchemical predictions39'55 and show most unusual features. Thus in contrastto the ultra-violet spectrum of the pentaleno[6,6a,1,2-def]heptalene deriva-tive (LIX) (Figure 10) which shows the longest absorption maximum at569 nm with further broad absorption extending up to approximately800 nm and unusual fine structure, the longest absorption of azuleno-[8,8a,1,2-def]heptalene (LX VIII) is found in the near infra-red. An absorptionmaximum at 1075 nm extends with shoulders of small extinction up to the

175

P.A.C.—28/3-—D

H

(LX)

KLAUS HAFNER

CH3(1) LJAJH4

(2) H20

(LXIV)(LXV)

(LXVII)

H5C6+ )—CH==CH_CH0H,C+ Pod3— HPO,C1,

CH5

_H5C6NRH

-H2

Hd1e\H3C H,

B (LXVI)

(LX VIII)

region of 1500 to 1700 nm, a most unusual longwave absorption for a hydro-carbon.

The n.m.r. spectrum reveals a similar surprise. The singlet for the protonof the 5-membered ring which at 5.5 r appears at much higher field than theprotons of the 5-membered ring of azulenes and tricyclic hydrocarbonsindicates a small or even nonexistent diamagnetic ring current in this partof the molecule. On the other hand the proton signals of the 7-memberedrings between 5.6 and 8.4 r—a region normally reserved for the protons of

CH3

(LXX)

176

(LXIX)

20 15 10 5x103cm1

200 300 /400 500 600 700 800 900 1000 1 200 1 /400 1 600 1 800A

2000nm

Fogire 10. Ultra-violet spectra of 5-methylpentaleno[6,6a,l,2-def]heptalene (LIX) and azuleno-[8,8a,I,2-def]heptalene (LXVIII) in n-hexane.

9.68

_JJJL-CH3

CH3

7.2

7.5 TMS

Figure 11. N.m.r. spectra of azuEeno[88a,1,2-def]heptalene (LXVIII) and 1 1-methyl-azuleno-[8,8a,1,2-def]heptalene (LXIX) in carbon disuiphide and of 4,6,8-trimethylazulene (XXXVII) in

carbon tetrachloride.

177

cepCH3

4.0

3.0

2.0 —

CH3

f I I

1.0 -

KLAUS HAFNER

saturated carbons—indicate strong paramagnetic ring currents in the three7-membered rings. An unusually high shielding of the ring protons and ofthe methyl group not observed so far in such systems is the result This isillustrated by the position of the signal for the methyl protons of (LXIX) at9.68 t, a region where proton signals of methyl groups at saturated carbonsare normally found. These significant peculiarities in the ultra-violet andn.m.r. spectra of the tetracyclic hydrocarbon (LXVIII) demonstrate itsextraordinary position in the series of the polycyclic conjugated non-benzenoid systems known so far.

An investigation of the reactivity of this compound together with a moredetailed quantum chemical and an x-ray analysis will hopefully yield furtherinteresting information about the interrelations between bond structureand reactivity in nonbenzenoid polycyclic ic-electron systems. Meanwhileit would be of much interest to investigate the properties of a pentacycliccompound related to (LXVIII) by inclusion of a further pen-fused 7-mem-beredring. Synthetic routes are being explored.

CONCLUSIONTo conclude, one can say that polycyclic nonbenzenoid ic-electron

systems differ, at times considerably, in their bonding structure and chemicalproperties from the well known monocyclic conjugated compounds with thesame number of it-electrons. The more cross-conjugated elements participatein the conjugation the more noticeable are these differences. The Hückelrule, reliable for monocyclic it-electron systems seems to be valid to someextent for bicyclic conjugated compounds and evidently also for tn- andtetra-cyclic linearly fused it-electron systems. It has no validity for pen-condensed tn- and tetra-cyclic compounds with cross-conjugated structuralelements. Just as the reactivity of azulene can be understood as a combinationof cyclopentadienyl anion and tropylium cation, the chemical properties,in part even the physical properties, of the discussed tn- and tetra-cyclicpen-condensed hydrocarbons may be interpreted qualitatively as a super-position of the bicyclic structures of pentalene, heptalene and azulene.

ACKNOWLEDGEMENTSIt is a pleasure to acknowledge gratefully the cooperation of my associates

F. Bauer, K. R. Bangert, R. Fleischer, W. Friebe, K. Fritz, E. Goedecke,G. Hafner-Schneider, D. Jung, R. Kaiser, H. J. Lindner, U. MUller-Westerhoff,V. Orfanos, W. Rieper and J. Schneider whose efforts, persistence and abilityare largely responsible for the results described in this paper.

Finally I would like to thank the 'Deutsche Forschungsgemeinschaft'and the 'Fonds der Chemischen Industrie' for generous support.

REFERENCES1 j• w•Armit and R. Robinson, J. Chern. Soc. 127, 1604 (1925).2 E. HUckel, Z. Phys. 70, 204 (1931); Grundzüge der Theorie ungesattigterund aromatischer

Verbindungen, Verlag Chemie: Berlin (1938).D. M. G. Lloyd. Carbocyclie Non-Benzenoid Aromatic Compounds, Elsevier: London (1966).

178

POLYCYCLIC CROSS-CONJUGATED it-ELECTRON SYSTEMS

' A. W. Krebs, Angew. Chem. 77, 10 (1965); Angew. Chem. internat. Edit. 4, 10(1965);R. Breslow, J. T. Groves and G. Ryan, J. Amer. Chem. Soc. 89, 5048 (1967).T. J. Katz and P. 1. Garratt. J. Amer. Chem. Soc. 85, 2852 (1963); 86, 5194 (1964);E. A. LaLancette and R. E. Benson, J. Amer. Chem. Soc. 85, 2853 (1963); 87, 1941 (1965).

6 T. 1. Katz, J. Amer. Chem. Soc. 82, 3784 (1960).F. Sondheimer, J. C. Calder, J. A. Elix, Y. Gaoni, P. J. Garratt, K. Grohmann, G. Di Maio,J. Mayer, M. V. Sargent and R. Wolovsky in Aromaticity, Spec. Pub!. No. 21, pp 75—107.Chemical Society: London (1967);G. W. Brown and F. Sondheimer. J. Amer. Chem. Soc. 91, 760(1969);J. Griffiths and F. Sondheimer, J. Amer. Chem. Soc. 91, 7518 (1969) and earlier papers.

8 E. Vogel in Aromaticity, Spec. Pubi. No. 21, pp 113—147, Chemical Society London (1967);E. Vogel, H. Haberland and J. Ick, Angew. Chem. 82, 514 (1970); Angew Chem. Internat.Edit., 9, 517(1970), and earlier papers;R. H. Mitchell and V. Boekelheide, J. Amer. Chem. Soc. 92, 3510(1970) and earlier papers.J. A. Elvidge and L. M. Jackman, J. Chem. Soc. 859(1961);L. M. Jackman, F. Sondheimer, Y. Amiel, D. A. Ben-Efraim, Y. Gaoni, R. Wolovsky andA. A. Bothner-By, J. Amer. Chem. Soc. 84,4307(1962);J. M. Gaidis and R. West, J. Chem. Phys. 46, 1218 (1967);F. Baer, H. Kuhn and W. Regel, Z. Naturforschung, 22a, 103 (1967).

10 H. J. Dauben Jr, J. D. Wilson and J. L. Laity, J. Amer. Chem. Soc. 91, 1991 (1969).' R. Breslow, Chem. & Eng. News, 43, No. 26. p90 (1965);M. J. S. Dewar, Advanc. Chem. Phys. 8, 121 (1965).' K. G. Untch and J. A. Pople, J. Amer. Chem. Soc. 88,4811(1966);H. P. Figeys, Chem. Commun. 495 (1967);G. Schröder and J, F. M. 0th. Tetrahedron Letters. 4082 (1966);J. F. M. 0th and J. M. Gilles, Tetrahedron Letters, 6259(1968);J. F. M. 0th, G. Anthoin and J. M. Gilles, Tetrahedron Letters, 6265 (1968).

13 R. reslow, J. Brown and J. J. Gajewski, J. Amer. Chem. Soc. 89,4383 (1967);R. Breslow and M. Douek, J. Amer. Cheni. Soc. 90,2698(1968);R. Breslow, Angew. Chem. 80, 573 (1968); Angew. Chem. Internat. Edit. 7, 565 (1968).

14 K. Hafner, Angew. Chem. '75, 1041 (1963); Angew. Chem. Internat. Edit. 3, 165 (1964); Z.Chemie, 8,74(1968).K. Hafner, A. Stephan and C. Bernhard, Liebigs Ann. Chem. 650,42(1961);K. Hafner, H. Peister and J. Schneider, Liebigs Ann. Chem. 650, 62 (1961).

16 K. Hafner and H. Weldes, Liebigs Ann. Chem. 606,90(1957);K. Hafner, C. Bernhard and R. MUller, Liebigs Ann. Chem. 650, 35 (1961);K. Hafner, H. Peister and H. Patzelt, Liebigs Ann. Chem. 650, 80 (1961).

17 w Treibs, Naturwissenschaften, 47, 156(1960);K. Hafner and K. L. Moritz, Liebigs Ann. Chein. 650,92(1961).

18 W. U. Schneider, H. J. Bernstein and J. A. Pople, J. Amer. Chem. Soc. 80,3497(1958);D. Meuche, B. B. Molloy, D. H. Reid and E. Heilbronner, Helv. Chim. Acta, 46,2483 (1963).

19 K. Hafner, Liebigs Ann. Chem. 606, 79 (1957).20 W. Baker,/. Chem. Soc. 200,201,209, 1114,1118(1951); 1443,1447,1452,2991,3163(1952);

Jndustr. Chim. Beige, 17, 633 (1952).21 H. J. Dauben Jr and D. J. Bertelli, J. Amer. Chem. Soc. 83, 4659 (1961).22 T. Nakajima and S. Katagiri, Molec. Phys. 7, 149 (1963);

0. Chalvet, R. Daudel and J. J. Kaufman, J. Phys. Chem. 68, 490 (1964).23 P. C. den Boer, D. H. W. den Boer, C. A. Coulson and T. H. Goodwin, Tetrahedron, 19, 2163

(1963);T. Nakajima, Y. Yaguchi, R. Kaeriyama andY. Nemoto, Bull. Chem. Soc. Japan, 37,272(1964);R. Zahradnik and J. Michi, Coil. Czech. Chem. Commun. 30, 3173 (1965);M. J. S. Dewar in Aromacity, Spec. Pub!. No.21, pp 177—215, Chemical Society: London (1967).

24 C. T. Blood and R. P. Linstead, J. Chem. Soc. 2255 and 2263 (1952);C. C. Chuen and S. W. Fenton, J. Org. Chem. 23, 1538 (1958).

25 E. LeGoff, J. Amer. Chem. Soc. 84, 3975 (1962).26 K. Hafner, K. H. VUpel and W. Bauer, unpublished results.27 K. Hafner, K. H. Häfner, C. KOnig, M. Kreuder, G. Ploll, G. Schulz, E. Sturm and K. H.

Vopel, Angew. Chein. 75, 35 (1963); Angew. Chem. Internat. Edit. 2, 123 (1963);K. Hafner, G. Schulz and K. Wagner, Liebigs Ann. Chem. 678, 39 (1964).

179

KLAUS HAFNER

28 K. Hafner, K. H. Vopel, G. PloD and C. König, Liebigs Ann. Chem. 661, 52 (1963).29 K. Hafner and F. Goedecke, unpublished results.30 K. Hafner and R. Kaiser, Angew. Chem. 82, 877 (1970); Angew. Chem. Jnternat. Edit. 9,

892(1970).' K. Hafner, K. F. Bangert and V. Orfanos, Angew. Chem. 79, 414 (1967); Angew. Chem.Internat. Edit. 6, 451 (1967).

32 H. J. Dauben, Jr., S. H. K. Jiang and V. R. Ben, Hua Hsüeh Hsüeh Pao, 23, 411 (1957); Chem.Abstr. 52, 16309h (1958).U. MilIler-Westerhoff; S. Dãhne; personal communications.K. Hafner and V. Orfanos, unpublished results.F. Sturm and K. Hafner, Angew. Chem. 76, 862 (1964); Angew. Chem. Internat. Edit. 3, 749(1964).

36 K. Hafner and J. Schneider, Liebigs Ann. Chem. 624, 37 (1959);K. Hafner and K. F. Bangert, Liebigs Ann. Chem. 650, 98 (1961).' N. K. DasGupta and M. A. Au, Theor. Chim. Acta, 4, 101 (1966);P. Hochmann, R. Zahradnik and V. Krasnièka, Coil. Czech. Chem. Commun. 33, 3478 (1968);U. Muller-Westerhoff, personal communication.

38 H. J. Lindner, J. Chem. Soc. (B), 907 (1970).D. Jung, Tetrahedron, 25, 129 (1969).

40 K. Hafner and it. Fleischer, unpublished results.41 K. Hafner and W. Rieper, Angew. Chem. 82, 218 (1970); Angew. Chem. Internat. Edit. 9, 248

(1970).42 K. Hafner and H. Schaum, Angew. Chem. 75, 90 (1963); Angew. C/tern. Internat. Edit. 2, 95

(1963);V. Boekeiheide and C. D. Smith, J. Amer. Chem. Soc. 88, 3950 (1966).' K. Hafner and it. Fleischer, Angew. Chem. 82, 217 (1970); Angew. Chem. Internat. Edit. 9, 247(1970).'' K. Hafner and G. Schneider, Liebigs Ann. C/tern. 672, 194 (1964).'1 K. Hafner, U. Hafner-Schneider and F. Bauer, unpublished results.

46 E.Carstensen-Oeser and 0. Habermehl, Angew. Chem. 80, 564(1968); Angew. Chem. Internat.Edit. 7, 543 (1968);R. Qasba, F. Brandl, W. Hoppe and R. Huber, Acta Crist. B25, 1198 (1969).' K. Hafrer, W. Friebe, 0. Hafner-Schneider and F. Bauer, unpublished results.

48 F. Haselbach, Tetrahedron Letters, 1543 (1970)."Electrophilic substitution of azulenes takes place in the 7-membered ring only when thepreferred 1- and 3-positions are blocked [K. Hafner and K. L. Moritz, Liebigs Ann. Chem.656, 40 (1962)].° R. Zahradnik, Angew. C/tern. 77, 1097 (1965); Angew. Chetn. Internat. Edit. 4, 1039 (1965)." For diocyclopenta [ef—kl]heptalene, isomeric with (LV), similar prediction proved to bejusti-fied (A. G. Anderson Jr, A. A. MacDonald and A. F. Montnna, J. Amer. Chem. Soc. 90, 2993(1968)).

52 K. Hafner, R. Fleischer and K. Fritz, Angew. Chem. 77, 42 (1965); Angew. C/tern. Internal.Edit. 4, 69 (1965).H. J. Lindner, Chem. Ber. 102, 2456 (1969).K. Hafner, 0. Hafner-Schneider and F. Bauer, Angew. Chem. 80, 801 (1968); Angew. C/tern.Internat. Edit. 7, 808 (1968)." R. Zahradnik and U. MUller—Westerhoff, personal communication.

180