2300 J. Org. Chem., Vol. 40,No. 16,1975
Koster and West
Synthesis and Reactions of a Tetraquinocyclobutane Sandra K. Koster and Robert West* Department of Chemistry, University of Wisconsin, Madison, Wisconsin 53706 Received February 24,1975 Thermal dimerization of the cumulene 6 yields the bright purple tetraquinocyclobutane 3, which is reduced by benzopinacol to the diaryldiquinocyclobutene 4. Reduction of 3 electrolytically or by the dianion of 4 gives a stable anion radical whose electron spin resonance spectrum indicates that the unpaired electron is fully delocalized over the five-ring system. Direct oxidation of 4 proceeds through a neutral monoradical intermediate 9. The ESR spectrum of 9 indicates odd-electron delocalization over three aromatic rings.
The properties of triquinocyclopropanes 1 and their reduction products, diarylquinocyclopropenes (2), have been reported ear1ier.l This paper describes the synthesis and chemistry of four-membered ring analogs to these compounds: the tetraquinocyclobutane 3 [tetrakis(3,B-di-tertbutyl-4-oxo-2,5-cyclohexadien-1-ylidene)cyclobutane] and the diaryldiquinocyclobutene 4. 0
0
-
average twist angle of 36". The central four-membered ring is also distorted from planarity. As would be expected from the high symmetry of the molecule, the infrared spectrum of 3 is quite simple, showing only ten medium and intense bands, and the lH NMR spectrum shows only two singlets in the expected 9:l ratio. Although it is stable indefinitely as a solid and in hydrocarbon solutions, 3 reacts with nucleophiles. Solution in methanol results in the addition of 1mol of methanol to 3. Similarly, if 3 is dissolved in wet N,N-dimethylformamide a water adduct is isolated. Based on lH NMR evidence the most likely structures for these adducts are 6a, b. The ex-
redn
oxidn
R
R
R
1
\
R
R
OH
2
redn
c
oxidn
0 6a, R = CH:,
b, R - H pected 1:l:l:l ratio of tert-butyl resonances is observed in the 'H NMR of the methanol adduct in deuteriobenzene, 3 4 while the water adduct shows a 2:l:l pattern. Compound 3 is obtained by thermal dimerization of the Although triquinocyclopropanes are reduced to diaryldiquinoethylene 5.2 When heated in cyclooctane for 3-4 hr, quinocyclopropenes by hydroquinone,' 3 does not react at all with hydroquinone. Other conventional reduction techniques (e.g., Zn-HC1,Zn-HOAc) fail because of the reactivity of 3 and 4 with nucleophiles. However, when 3 is refluxed with a slight excess of benzopinacol in cyclooctane the desired diaryldiquinocyclobutene 4 is formed as a bright orange solid. The 'H NMR of 4 (C&) shows three 5 types of tert- butyl protons in a 2:l:l ratio (Figure 2) consistent with the assigned structure. The lH NMR spectrum 5 disappears completely and work-up of the reaction mixalso rules out the possibility of reduction diagonally across ture yields 25% of 3 as sparkling purple crystal^.^ The tethe ring to give a bicyclobutane structure, since this would traquinocyclobutane structure is related to those of 4-radibe expected to give two resonances of equal intensity in the alenes, many of which can similarly by synthesized by dimtert- butyl region. Although quite stable in the solid state, 4 erization of c ~ m u l e n e s . ~ ? ~ is oxidized back to 3 by atmospheric oxygen if left in soluIn organic solvents 3 gives brilliant purple solutions tion for a few hours. This diaryldiquinocyclobutenealso exwhich show complex electronic spectra in the visible and hibits the sensitivity to nucleophiles observed for 3. A gradultraviolet region (Figure 1).The spectrum of 3 is rather ual change in the electronic spectrum occurs when 4 is dissimilar to that of the triquinocyclopropanes 1,' but the visisolved in methanol, the final spectrum showing A,, 460 ble absorptions occur a t higher energy for 3. nm (t 3.87 X lo4). No attempt was made to determine the Molecular models indicate that a planar conformation is exact structure of the methanol adductb). possible for 1 but not for 3, and X-ray crystal structure deThe diarylquinocyclopropene analogous to 4, compound terminations on both compounds confirm this ded~ction.~,' 2 (R = tert-butyl), has its electronic absorption at 413 nm Compound 1 (R = tert-butyl) is almost planar with an avin benzene but shows a blue shift to 406 in acetonitrile. erage angle of twist of the six-membered rings of 5 . 5 O . In This suggests that the ground state of 2 is more polar than contrast 3 exists in a propeller-like conformation with an
opc=ceo
J. Org. Chem., Vol. 40,No. 16, 1975 2301
Synthesis and Reactions of a Tetraquinocyclobutane
I
I
=
.EL
i
id-.-I :00
I
400 500 WAVELFNGTI I, nm
602
I ~
Figure 1. Electronic spectrum of tetraquinocyclobutane 3,O.l mM in cyclohexane.
the excited state and is evidence for the importance of the resonance contribution of the cyclopropenium cation in the ground state, as pictured below. The absence of a similar 0-
2
$kH
HO
4
O
w
-
solvent shift in the spectrum of 4 is evidence that the cyclobutenyl cation or cyclobutadienyl dication does not contribute significantly to the ground state of 49 or a n p a y no more than to the excited state. The diaryldiquinocyclobutene 7 was reported8 while our work with diaryldiquinocyclobutene was in progress* Compound 7 was synthesized by oxidative dimerization of
@CwC*OH 7
,
8
LT
I
8.49
I
,
8.70 ai2
S W 108Hz
L
Figure 2. Proton magnetic resonance Spectrum of 4 in CsD6.
acetylene 8 and appears to have properties similar to those of 4. Compound 7 also reacts with alcohols to form “asymmetric adducts in which alcohol molecules seem to have reacted at one of the double bonds of the molecule”.8 Deprotonation of 4 to Its Dianion. Spectrophotometric titrations with various bases were attempted in order to convert 4 to its dianion. In acetonitrile with aqueous sodium hydroxide as base, the isosbestic point was lost after about 1 equiv had been added, indicating that reaction was occurring rather than simple deprotonation. Titration with potassium tert-butoxide was similarly unsuccessful. However, when the hindered bicyclic amine, 1,5-diazabicyclo[4.3.0]non-5-ene (DBN), was used as the base in dry tetrahydrofuran, simple deprotonation was observed to give the turquoise-colored dianion (Figure 3). All intermediate curves pass through two isosbestic points at 522 and 300 nm, implying that the monoanion is not stable with respect to the dianion under the conditions used. To complete the deprotonation an amount of DBN in excess of the theoretical 2 equiv was required, showing qualitatively that 4 is not a very strong acid. The Anion Radical of 3. Electrolytic reduction of 3 to its anion radical was carried out in tetrahydrofuran (THF). The reduced solution gave an ESR spectrum shown in Figure 4. The nine-line pattern indicates splitting by eight equivalent quinonoid protons, showing that the odd electron is completely delocalized over the four six-membered rings. Delocalization was similarly observed for the analogous triquinocyclopropane anion radicalg as well as for the anion radicals of the related diquinoethylene and diquinocyclopropanone.10The proton hyperfine splitting constant, 0.43 G, is accounted for by molecular orbital calculations discussed in the accompanying paper.1° The anion radical of 3 is stable for several d a w and can be exDosed at least briefly to air without decomposition. If degassing is carried out after such the original nine-line pattern is observed. To confirm that the species observed by electrolytic reduction of 3 was the anion radical, a solution of 3 plus the Same volume of an equimolar solution of the dianion of (prepared by DBN deprotonation) were mixed. The resulting solution showed exactly the same nine-line ESR Dattern. The electronic spectrum showed no absorptions d u e to 3 or the dianion of 4 but gave a new maximum at 615 nm which can be assigned to the anion radical. The Monoradical of 4. While the anion radical of 3 and the dianion of 4 should be redox intermediates in the tetraquinocyclobutane system under basic conditions, the expected intermediate under neutral conditions would be the monoradical 9 (the protonated anion radical). Two methods were employed to generate 9 in the ESR cavity, both
2302
J.Org. Chem., Vol. 40, No. 16, 1975
1 '
I
300
400
Koster and West
1
1
6CO
500 WAVELENGTH, nm
Figure 3. Spectrophotometric titration of 4 with DBN in THF solvent. Curve 1 shows the spectrum of pure 4. Curve 8 shows the spectrum of 4 in the presence of excess DBN sufficient to convert 4 almost completely to the dianion.
Figure 5. ESR spectrum of monoradical 9 formed by comproportionation of 3 and 4, at 100 and 180° in melted naphthalene.
and 4 undergo reversible comproportionation to 9 as shown below. The weakness of the signal shows that the equilibrium lies very far toward 3 and 4 at moderate temperatures.
-
Figure 4. ESR spectrum of the anion radical of 3, showing splitting by eight equivalent protons.
involving partial oxidation of 4. When 4 was treated with PbOn in xylene, no signal was observed below 90°, when a weak multiplet appeared. The signal intensity increased as the temperature was raised to 135O, and decreased to the previously observed level when the temperature was lowered. Radical 9 was also generated from 4 in melted naph-
+ HO
OH
3
4
'0 9
thalene using an equivalent amount of 3 as oxidant. The same pattern and reversible temperature effect was observed up to 180° (Figure 5). These results indicate that 3
9
The ESR spectrum of 9 is a seven-line pattern, with the two outermost lines partially obscured by noise (Figure 5). This indicates hyperfine splitting by six equivalent pro-
J. Org. Chem., Vol. 40, No. 16,1975 2303
Synthesis and Reactions of a Tetraquinocyclobutane
tons, i.e., delocalization over three of the benzenoid rings. Resonance structures can be drawn (see below) consistent
OH 9 with delocalization of the odd electron over three (and not four) rings, but only two of the rings are equivalent. Evidently accidental equivalence accounts for the equal splitting by six protons and the odd-electron density is too small in the fourth ring t o cause observable splitting. Scheme I summarizes the intermediates in the 3 + 4 redox systems and the relationships between them.
Scheme I
-
+ e l -
-le
+le
t 2H'
-H+
ESR electrolysis experiments were carried out using a 300-V battery source with the current kept at