Facile Redox Interconversion of 6,11-Diphenyldibenzo[b,f][1,4]diazocine and 2-(2-Aminophenyl)-1,3-diphenylisoindole: Reversible SET Ring Contraction and Expansion Processes

Article abstract of DOI:10.1002/ejoc.201403101

In our continuing attempts to convert tub-shaped dibenzo[1,4]diazocines or dibenzo[1,5]diazocines into necessarily planar Hückel aromatic ten-π-electron dianions or dihydro derivatives of the central diazocine ring, we have added requisite electrons by Na or Li metal in THF. Subsequent hydrolysis yielded no evidence for the formation of such Hückel aromatic products but in each case a profound rearrangement of the tricyclic diazocine had instead occurred. In the present study we have attempted to form the unknown aromatic 6,11-diphenyldibenzo[b,f][5,12]-dihydro[1,4]diazocine at 25°C by such a straightforward addition of two electrons to 6,11-diphenyldibenzo[b,f][1,4]diazocine. We were encouraged by the prior reduction of the unsubstituted [1,4]diazocine to 1,4-dihydro-[1,4]diazocine, which by X-ray and ¹H NMR evidence displays aromatic-like properties. However, this diphenyldibenzo[1,4]-diazocine upon reduction underwent instead an unusual, serendipitous rearrangement to yield quantitatively 2-(2-aminophenyl)-1,3-diphenylisoindole. Then in a purposive search for other reductants capable of reductively rearranging this [1,4]diazocine to its corresponding isoindole, we discovered three other reductants, namely o-diaminobenzene, titanium(II) salts, and concentrated aqueous hydriodic acid with visible light. Conversely, again in a serendipitous observation, it was found that O₂ in CHCl₃ with visible light could readily convert the isoindole in an oxidative rearrangement back into the [1,4]diazocine. A purposive method for achieving this oxidative rearrangement was then found to be treatment with DDQ. General mechanistic pathways are proposed via SET intermediates for both redox interconversions.

Full text of DOI:10.1002/ejoc.201403101

FULL PAPER
DOI: 10.1002/ejoc.201403101
Facile Redox Interconversion of 6,11-Diphenyldibenzo[b,f][1,4]diazocine and
-(2-Aminophenyl)-1,3-diphenylisoindole: Reversible SET Ring Contraction and
Expansion Processes[
2
‡]
[a]
[a][‡‡]
Lisheng Zhu,[a][‡‡‡] and Arnold L. Rheingold[b]
John J. Eisch,* and Wei Liu,
Keywords: Nitrogen heterocycles / Cyclic rearrangement / Electron transfer / SET oxidation / SET reduction
In our continuing attempts to convert tub-shaped dibenzo-
1,4]diazocines or dibenzo[1,5]diazocines into necessarily
this diphenyldibenzo[1,4]-diazocine upon reduction under-
went instead an unusual, serendipitous rearrangement to
yield quantitatively 2-(2-aminophenyl)-1,3-diphenyliso-
indole. Then in a purposive search for other reductants cap-
able of reductively rearranging this [1,4]diazocine to its cor-
responding isoindole, we discovered three other reductants,
namely o-diaminobenzene, titanium(II) salts, and concen-
trated aqueous hydriodic acid with visible light. Conversely,
[
planar Hückel aromatic ten-π-electron dianions or dihydro
derivatives of the central diazocine ring, we have added
requisite electrons by Na or Li metal in THF. Subsequent hy-
drolysis yielded no evidence for the formation of such Hückel
aromatic products but in each case a profound rearrange-
ment of the tricyclic diazocine had instead occurred. In the
present study we have attempted to form the unknown aro-
matic 6,11-diphenyldibenzo[b,f][5,12]-dihydro[1,4]diazocine
at 25 °C by such a straightforward addition of two electrons
again in a serendipitous observation, it was found that O
CHCl with visible light could readily convert the isoindole
in an oxidative rearrangement back into the [1,4]diazocine.
2
in
3
to 6,11-diphenyldibenzo[b,f][1,4]diazocine. We were encour- A purposive method for achieving this oxidative rearrange-
aged by the prior reduction of the unsubstituted [1,4]di-
azocine to 1,4-dihydro-[1,4]diazocine, which by X-ray and H
NMR evidence displays aromatic-like properties. However,
ment was then found to be treatment with DDQ. General
mechanistic pathways are proposed via SET intermediates
for both redox interconversions.
1
1
Introduction
IR and H NMR spectral features of 3, the overall product,
upon recrystallization from 95% ethanol, yielded only the
mechanically separable crystals of the Z-isomer (5a) and E-
isomer (5b) of N-(2-amino-1,2-diphenylethenyl)carbazole,
as was demonstrated by individual X-ray diffraction analy-
Background: Single-Electron Transfer Processes with
Dibenzodiazocines
Our interest in nitrogen analogs of cyclooctatetraene
arose from our attempts to generate the Hückel-aromatic
ten-π-electron, necessarily planar dianion 2 from the known
tub-shaped 6,7-diphenyldibenzo[e,g][1,4]diazocine (1) by
the addition of two equivalents of sodium or lithium metal
in THF suspension and its subsequent hydrolysis
[2,3]
sis.
(Scheme 1). Although the isolated product was found to
have the proper molecular mass and some of the expected
1]
[
[
‡] Cyclic Conjugated Imines, 4. Part 3: Ref.[
‡‡]Deceased, March 3, 2014, in a tragic traffic accident (see
Acknowledgments).
[
[
‡‡‡]Postdoctoral Research Associate with Professor John Eisch
a] Department of Chemistry, State University of New York at
Binghamton,
P. O. Box 6000, Binghamton, NY 13902-6000, USA
E-mail: jjeisch@binghamton.edu
http://www.binghamton.edu/chemistry
[
b] Department of Chemistry and Biochemistry, University of
California,
San Diego, La Jolla, California 92093-0332, USA
E-mail: arheingold@uscd.edu
http://www-chem.ucsd.edu/faculty/profiles/rheingold_arnold_
l.html
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201403101.
Scheme 1.
Eur. J. Org. Chem. 2014, 7489–7498
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7489

J. J. Eisch, W. Liu, L. Zhu, A. L. Rheingold
FULL PAPER
Clearly, the profound rearrangement of diazocine 1 upon because of the remaining ring strain in forming a planar
[
4b]
reduction with sodium or lithium must have involved trans- diazocine ring (about 23 kcal/mol. ) (Again each N center
annular bonding in the transition state between the in structures in 6–9 of Scheme 2 has an electron pair except
asterisked N and C sites in 1 (4 in Scheme 1). Each N center 7b that has been omitted for clarity.)
in structures 1–5 has an electron pair, which has been
In our continuing search for an isomer of 1 or 6 that
Because of the tub shape of 1, would form a stable and planar Hückel 10-π-electron di-
attaining the necessary coplanarity of 1 would require more anion and dihydro derivative (11), we have now selected the
omitted for clarity.[
4a]
[
4b]
activation energy
than the partial σ-bond stabilization known 6,11-diphenyldibenzo[b,f][1,4]diazocine (10). By ex-
amining the known properties of the unsubstituted dihydro-
anticipated between N* and C* in transition state 4.
In order to avoid the destabilizing ortho-H repulsions in diazocines, we note that the X-ray structure of 1,4-dihydro-
(bracketed by arrows), we then examined the electron- [1,4]diazocine (12, the enclosed substructure in 11) reveals
2
transfer reduction of an isomer of 1, namely, 6,12-diphenyl- a planar and aromatic connectivity of its C–C and C–N
dibenzo[]b,f[1,5]diazocine (6) (Scheme 2).[ Although the bonds of almost the same length. Also, the H NMR spec-
validity of structure 6 was assured both by its logical syn- trum exhibits diamagnetic ring current effects similar to
thesis via the bimolecular dehydration of 2-aminobenzo- those of heteroaromatic rings like furan or thiophene
4b]
1
[5]
[6]
phenone and by its limited spectroscopic data, its 3D (Scheme 3).
stereochemistry was not. The subsequently obtained X-ray
diffraction and C NMR spectroscopic data unequivocably
established the 3D-structure of 6 as that rendered in
Scheme 2, with coplanarity of the transannular C=N and
1
3
5
11
N=C groups and with the cross-ring N and N atom sepa-
6
12
rations of 3.294(3) Å and the C and C atom separation
of 3.000(2) Å. The cross-ring proximity of these atoms
raises the probability of transannular bonding in radical-
anionic or dianionic intermediates (i.e., 7a to 8) in SET pro-
cesses (Scheme 2). In fact, treatment of 6 with sodium or
lithium metal in THF must have led only to dianionic salt
Scheme 3.
Accordingly, we have now attempted to generate the
presently unknown 6,11-dibenzodiphenyl[b,f][5,12]-dihydro-
1,4]diazocine (11) by the straightforward addition of two
electrons from Na or Li metal in THF to the known 6,11-
diphenyldibenzo[b,f][1,4]diazocine (10) and hydrolysis of
the resulting dianion in the hope of preparing 11.
[
8
and ultimately none of the putative Hückel ten-π-electron
dianion 7a, because subsequent hydrolysis produced only 9.
Results
The Attempted Generation of 6,11-Diphenyldibenzo[b,f]-
5,12-dihydro-[1,4]diazocine (11) by the Sodium or Lithium
Metal Reduction of 10 and Subsequent Hydrolysis
The key probative experiment in the present investigation
as summarized in Scheme 3 was put to the test in the fol-
lowing manner. The requisite starting material (10) was pre-
pared according to our modification of the original pro-
[
7]
cedure of Olliéro and Solladié, which now uses the acid-
catalyzed condensation of 14 and 15 in toluene, instead of
benzene with p-toluenesulfonic acid (16) to generate cyclic
[
8]
[7,8]
dianil 10 (m.p. 197–198 °C; ref.
Scheme 4).
When 10 (one equiv.) was allowed to react with finely
184–189 °C)
(
divided pieces of freshly cut Na or Li metal (ten equiv.) in
THF at room temp. for 10 h for Na or for 24 h with Li,
followed by mechanical removal of metal and then hydroly-
sis, workup in either case provided a quantitative yield of
13 with no remaining 10 or any other product such as the
Scheme 2.
expected Hückel 10-π-electron system 11 (Scheme 5). The
most likely initial intermediates formed from such alkali-
Thus, despite the absence of the ortho-H group repul- metal additions to 10 would be radical-anion 10a and di-
[
9]
sions, present in the transition from 1 to 2 but absent in the anion 10b and either of these would undergo transannular
path from 6 to 7b, the Hückel dianion 7a is not formed rearrangement to the isoindole skeleton in 13 (see below).
7490
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 7489–7498

Reversible SET Ring Contraction and Expansion Processes
In this work, the sample of 10 isolated from 14 and 15
at 110 °C in toluene, was recrystallized from absolute eth-
[
7]
anol as yellow crystals, m.p. 197–198 °C (ref. 184–189 °C).
Its IR spectrum (CDCl ) displays one C=N stretch at
3
–1
1
1
625 cm ; the H NMR spectrum displayed 18 aromatic
13
13
protons; and the C NMR spectrum 11 C singlets, in-
cluding a C=N peak at δ = 169.7 ppm. The 3D X-ray
diffraction data for 10 is given in Figure 1 as a thermal
ellipsoid diagram. Especially noteworthy are the nonplanar
tub-like structure, the coplanarity of the two transannular
1
2
7
C=N groups and the cross-ring N and N and the C and
14
C
atom separations of 2.901(2) and 2.876(2) Å, respec-
tively. As with the nonplanar structure of 6, the cross-ring
proximity of these atoms enhances the likely significance
of transannular electronic delocalization/bonding in radical
anionic intermediate 10a (or depicted in Scheme 13, as reso-
Scheme 4. Method of Olliéro and Solladié.[7]
nance structures 28 and 29, where E = Na).
n
Scheme 5.
The sample of 13, obtained from 10 and Na in THF
(
Scheme 5), was recrystallized from absolute ethanol as pale Figure 1. Thermal ellipsoid (30%) diagrams for 6,11-diphenyldi-
[7]
benzo[b,f][1,4]diazocine (10). Selected bond lengths (atom separa-
tions) [Å] and bond angles of 6,11-diphenyldibenzo[b,f][1,4]di-
azocine (10) are shown in the following: N(1)–C(14) 1.281(2), N(1)–
C(26) 1.419(2), N(2)–C(7) 1.279(2), N(2)–C(21) 1.419(2), C(5)–C(6)
1.391(2), C(5)–C(7) 1.499(2), C(6)–C(14) 1.498(2), C(14)–C(15)
yellow crystals, m.p. 212–213 °C (ref. 201–202 °C). Even
when mixtures of 10 and Na metal were allowed to react in
different ratios or for shorter reaction times, 13 was found
as the only product.
1
2
3
.491(2), C(16)–C(17) 1.377(2), C(21)–C(26) 1.399(2), N(1)–N(2)
.901(2), C(7)-C(14) 2.876(2), C(7)–N(1) 3.122(2), C(14)–N(2)
.196(2),
C(14)–N(1)–C(26)
C(5)–C(6)–C(14)
N(2)–C(7)–C(5)
121.40(11),
118.77(11),
123.50(12),
C(7)–N(2)–C(21)
N(2)–C(7)–C(8)
N(1)–C(14)–C(15)
Confirmation of the Structures of Diazocine 10 and
Isoindole 13 as Originally Proposed by Olliéro and
120.09(11),
119.30(11),
118.15(12), C(16)–C(15)–C(14) 119.99(12).
[
7]
Solladié
The requisite diazocine 10 required for the attempted
synthesis of the planar Hückel aromatic system 11 and the
unexpected, serendipitous product 13 obtained by treating
The isoindole 13 first reported by Olliéro and Solladié as
1
0 with Na or Li in THF have both been reported in a brief the product obtained from a mixture of 14 and 15 at
[7]
[7]
paper by Olliéro and Solladié. Both had previously been 200 °C again was characterized in a manner similar to
prepared (Scheme 4)[ and partially characterized by ele- diazocine 10 but without X-ray data. Since the striking re-
8]
1
mental analyses, molecular mass estimate and IR and H ductive rearrangement of 10 into 13 by Na or Li has been
NMR spectroscopic analysis for which no actual measured observed for the first time in this study, it was essential to
values are given. Moreover, the melting points of their 10 obtain X-ray crystallographic and other spectral confirm-
and 13 are 10 °C lower than we have observed in our work. ation of its structure. Its IR spectrum (CDCl ) exhibits the
3
1
3
–1
1
However, the 3D X-ray structure and the C NMR spec- N–H stretches at 3478 and 3383 cm ; the H NMR spec-
troscopic data were lacking.
trum displayed 18 aromatic protons, as well as a broad two-
Eur. J. Org. Chem. 2014, 7489–7498
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7491

J. J. Eisch, W. Liu, L. Zhu, A. L. Rheingold
FULL PAPER
proton doublet for the NH group at δ = 3.47 ppm and
2
–
1
with IR bands at 3478 and 3383 cm . Finally, the thermal
ellipsoid diagram for the X-ray diffraction data for 13 is
exhibited in Figure 2.
Scheme 6.
presence of 13 and the absence of 10. Column chromato-
graphic separation provided 80% of 13, 10% of 14 and 15
combined and about 10% of phenazine (17) (excluding
other unknown amines) (Scheme 6).
These findings support the viability of path b and the
role of 15 as the reductant of 10 to 13 (Scheme 7). The
Figure 2. Thermal ellipsoid (30%) diagrams for 2-(2-aminophenyl)-
1
,3-diphenylisoindole (13). Selected bond lengths (atom separa-
tions) [Å] and bond angles of 2-(2-aminophenyl)-1,3-diphenylisoin-
dole (13) are shown in the following: N(1)–C(5) 1.387(2), N(1)–
C(19) 1.440(2), N(2)–C(24) 1.354(2), C(2)–C(3) 1.383(2), C(4)–C(5) condensation of putative diimino ortho-quinone 19 with 15
1
1
3
.471(2), C(11)–C(12) 1.437(2), C(13)–C(14) 1.364(2), C(14)–C(15)
.426(2), C(19)–C(24) 1.400(2), N(1)–N(2) 2.806(2), N(2)–C(5)
.182(2), N(2)–C(6) 3.657(2), C(6)–N(1)–C(5) 111.02(12), C(6)–
would lead to the phenazine (17) identified among the
products (Scheme 7).
[11]
N(1)–C(19) 125.53(12), C(3)–C(4)–C(5) 118.74(13), N(1)–C(5)–
C(12) 106.93(13), C(13)–C(12)–C(11) 119.38(13), C(14)–C(13)–
C(12) 118.95(15), C(24)–C(19)–N(1) 118.39(13).
Scheme 7.
Possible Reaction Pathways Leading from o-Dibenzoyl-
benzene (14) and o-Diaminobenzene (15) to Diazocine 10 or
to Isoindole 13
Additional Reductants of Diazocine 10 Causing
Rearrangement to Isoindole 13
Although Olliéro and Solladié prepared both 10 and 13,
as depicted in Scheme 5, and correctly identified their atom
The direct conversion of 10 into 13 by treatment of Na
connectivity, they offered no proof of the pathway(s) by or Li in THF can be counted as the first authentic and
[
12]
which these structures would form. In fact, they left the serendipitous
example of this reductive rearrangement
origin of 13 from 14 and 15 or from 10 and 15 as an open (Scheme 4). The conversion of 10 into 13 with o-diamino-
question to be answered by further investigation.[
10]
benzene and catalysis by 16 at 200 °C was carried out by us
Now the products 10 and 13 could arise from 14 and 15 after the method shown in Scheme 4 and thus is a second
by way of at least two distinct paths (Scheme 6): 1) path a, but purposive method for achieving this novel conversion.
through adduct 18 (possibly reversibly formed) with further
In order to gain further mechanistic insight into the
reductive dehydroxylation with reagent 15 to yield 13, pathway leading from diazocine 10 in its unusual rearrange-
accompanied with loss of phenazine (17); or 2) path b, ment to the 2-(2-aminophenyl)isoindole (13), we searched
formation of diazocine 10 with further reductive rearrange- for other reductants that would induce the same reductive
ment by 15 to 13 with loss of 17.
rearrangement. The third reductant employed were
To test the viability of path b, a dry mixture of 10 and titanium(II) salts, such as oligomers of TiCl2 (20a) and
1
5 in a 1:3 molar equiv. ratio was heated at 200 °C with 16. Ti(O-iPr) (20b) units, readily prepared at 25 °C in THF by
2
1
Hydrolytic workup, followed by both TLC and H NMR the alkylative reduction of the corresponding TiE salts (E
spectral analyses of the crude reaction mixture showed the = Cl or O-iPr) with n-butyllithium.
4
[
13–18]
Although the
7492
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 7489–7498

Reversible SET Ring Contraction and Expansion Processes
degree of association of TiE units could not be measured
A fourth alternative reductant system capable of con-
2
directly by their colligative properties, ESR measurements verting 10 to 13 proved to be aqueous hydriodic acid. Thus
of Ti(O-iPr) in THF showed that 20b is paramagnetic and trial reactions of a standard sample of 10 dissolved in 55%
2
exhibits a triplet signal indicative of two unpaired electrons. aqueous HI at various selected temperatures under labora-
The magnitude of the coupling constants of the electrons tory fluorescent light led to the following yields of 13
was in best accord with titanium centers located 1,3 to each (Table 1, Scheme 10).
other in an open chain[ (20b, Scheme 8).
18]
Table 1. Reactions of diazocine 10 to isoindole 13 in 55% aqueous
HI in Lab Light.[
a]
Products[b]
percentages)
Run 1
60 °C
Run 2
60 °C
(dark)
Run 3
25 °C
Run 4
0 °C
(
(
light)
Diazocine 10
Isoindole 13
15
59
38
43
18
17
70
13
88
12
–
o-Dibenzoylbenzene 14 26
[
a] See the Experimental Section for the details of standard run
1
2
size, hydrolytic workup, discharge of the I generated and the H
Scheme 8.
NMR analysis for the proportion of components 10, 13 and 14. [b]
The duration of reaction runs 1 and 2 was 2 h; that of reaction
runs 3 and 4, 6 h.
Since in parallel ESR studies, TiCl in THF proved to be
diamagnetic,
2
[
18]
we have made the likely assumption that
TiCl is also a trimer but having a cyclic structure (20a). In
2
reaction with unsaturated substrates like carbonyl or imino
derivatives, 20a and 20b act as if they deliver carbenoid-like
TiE units to the organic substrate in a singlet or triplet
2
form. In fact, the reaction of three equivalents of 10 with
two equivalents of trimeric 20b in refluxing THF leads
upon hydrolysis to a quantitative yield of 13 (Scheme 9). A
reasonable proposed pathway would be the addition of the
Scheme 10.
Ti(O-iPr) biradical to 10 to yield 21 followed by its subse-
2
quent rearrangement to 22. The superiority of 20b to pro-
vide the triplet monomer, Ti(O-iPr) (ȆȆ), to generate key
2
It is seen that the acidic hydrolysis of 10 into 14 is in
more serious competition with the redox reaction at 60 °C
than at 25 °C and that the redox reaction is photo-pro-
moted. The redox reaction can be readily ascribed to the
SET reaction of the iodide with protonated 10 to form radi-
cal 23, similar to 21 in Scheme 9, which can isomerize to 13
(Scheme 10). As is evident, by lowering the reaction tem-
perature the SET reaction can be fostered over hydrolysis
and could be used as a preparative method for 13 if a
chromatographic separation of the products is per-
intermediate 21 is clearly evident by employing the diamag-
netic reductant 20a in its place. At 25 °C for 48 h a 3:2 ratio
of 10 and 20a in THF gave after hydrolysis 62% of 13. A
similar mixture held at reflux for 48 h before hydrolysis
yielded only 6% of 13. This indicates that the correspond-
ing TiCl2 adduct 21 is thermally unstable and tends to
dissociate to 10.[17]
[
19]
formed.
Accidental Discovery of the Oxidative Rearrangement of
Isoindole 13 to Diazocine 10
An equally impressive oxidative rearrangement is the
reverse reaction of 13 into 10, occurring spontaneously and
completely at 25 °C in the presence of atmospheric oxygen
and visible light. This serendipitous reaction was revealed
by allowing a pure sample of 13, dissolved in CDCl in a
3
loosely stoppered NMR tube to stand in the lab for about
1
a week after recording its H NMR spectrum. Upon re-
1
cording the H NMR spectrum again after such storage,
the solution, now darker yellow, exhibited the proton spec-
Scheme 9.
trum of essentially pure 10.
Eur. J. Org. Chem. 2014, 7489–7498
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7493

J. J. Eisch, W. Liu, L. Zhu, A. L. Rheingold
FULL PAPER
A similar outcome was obtained with larger samples of though the kinetic equation for the rate of reductive re-
1
3 heated in CHCl under reflux in a silica gel suspension arrangement varies greatly for each reagent (En in
3
and with exposure to lab light. The recovered solid from Scheme 13), it appears reasonable to propose that such in-
this reaction consisted of 96% of 10 and 4% of 13. termediates would resemble 28, where E is respectively Na,
Encouraged by this accidental finding, we then under- E = H, E = Ti(OiPr) and E = H, with a free electron
1
2
3
2
4
7
took the purposive experiment of achieving such an oxidat- on C . Since the trans-situated imino (C=N) groups in 10
ive rearrangement of 13 to 10 by DDQ (24). Heating 1.0 are proximate to each other, the free-radical C in 28 could
·
molar equivalent of 13 with 2.2 molar equivalent of DDQ be delocalized transannularly, as shown in resonance struc-
at reflux in THF, followed by usual workup and column ture 29. Fragmentation of radical 28 at the C–N bond (at
chromatography led to 83% of pure 10 (Scheme 11). The arrow) is fostered in the transition state by the generation
quinhydrone complex (25) of DDQ with its hydroquinone of the aromatic delocalization of the incipient isoindolic
was the principal by-product of this reaction.
system 30. In the final step intermediate 30 could acquire
·
another E radical {Na, H or [ȆȆ Ti(OiPr) ]} to form 13
n
2
or its N,N-dimetallic derivative [Na or Ti(OiPr) ]. We offer
2
2
such an intermediate 28 as a working hypothesis for future
mechanistic studies aimed at the experimentally supported
[
20]
pathway followed by each reductant.
Scheme 11.
Finally in an attempt to apply this unusual reaction to
the synthesis of unknown diazocine derivative 27, the
known carbazole 26 was treated, first with O in lab light
2
and then with DDQ without success (Scheme 12). The
possible structural reasons for the failure of 26 to undergo
oxidative rearrangement to 27 will be considered in the next Scheme 13.
section.
The aforementioned Scheme 13 for reductive rearrange-
ment of 10 to 13 can be simply modified to account for the
general oxidative rearrangement of 13 to 10 by O or DDQ,
2
[
21]
as is shown in Scheme 14.
Scheme 14 combines two or
more steps involved in forming crucial intermediates in the
rearrangement. Formation of the intermediate 31a or 31b,
resonance representations of this cationic radical, is pro-
posed to result from the stepwise removal of two electrons
Scheme 12.
from 13 by the O biradicals with an intervening cyclization
2
of 31 to 33. Alternative pathways for this outcome could
also be involved. The conformation of resonance structure
Discussion
32 having the phenyl group alpha to the indolic N rotated
Our failed attempt to convert diazocine 10 has led to the out of the plane is thought to be necessary for the close
serendipitous discovery of the first authentic synthesis of 13 approach of the C-centered and the N-centered groups
[
10]
from 10. This achievement then motivated our search for leading to transition state 32. The necessity for close ap-
and discovery of other reductants causing this transforma- proach of these reactive centers is deduced from our failure
tion: o-DAB; TiE ; and HI. Thus the primary goal of this to cyclize the corresponding indole derivative 26 to the ex-
2
report has been the preparation of 13 from 10 efficiently.
pected cyclic dianil 27 (Scheme 12). The unrotatable benzo
That four such chemically diverse reductants can groups fused alpha, beta to the carbazolic nitrogen poses a
accomplish this conversion with facility raises the possibi- great steric barrier to the approach of the ortho amino
lity of similar crucial intermediates in each reduction. Al- group to either alpha cation derived from 26 (* asterisks).[
22]
7494
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 7489–7498

Reversible SET Ring Contraction and Expansion Processes
Scheme 14.
free argon. All solvents employed with organometallic compounds
were dried and distilled from a sodium metal/benzophenone ketyl
Conclusions
(
1) The attempt to generate the potential, planar Hückel mixture prior to use.[23]
ten π-electron 6,11-diphenyldibenzo[b,f]-5,12-dihydro[1,4]-
diazocine (11) by the addition of two equiv. of Na or Li
metal in THF to 6,11-diphenyldibenzo[b,f][1,4]diazocine
Routinely, the organometallic reaction mixtures were hydrolyzed
with deoxygenated water. Ether was added to the hydrolysate, the
organic layer was separated and the organic layer then dried with
(10), followed by hydrolysis, proved to be unsuccessful. In-
1
anhydrous Na
2
SO
4
. The volatile solvent was removed and H and
1
3
stead, this procedure led quantitatively to the first direct
C NMR spectra and TLC were recorded on the crude organic
conversion of 10 into 2-(2-aminopheny)-1,3-diphenyliso- products. Where preparative yields were to be corroborated,
indole (13).
2) This novel conversion of 10 into 13 by such reductive
rearrangement has then been found to occur with three
column chromatographic separation of products on silica gel with
a hexane/ethyl acetate eluent was carried out.
(
Analytical Methods: The IR spectra were recorded with a Perkin–
additional reductants: a) heating 10 and excess o-diamino- Elmer instrument, model 457, and samples were measured either
1
benezene (15) at 200 °C in the presence of p-toluenesulfonic as mineral oil mulls or as KBr films. The NMR spectra ( H and
1
3
acid (16) as catalyst; b) heating 10 with two equiv. of
titanium(II) isopropoxide in THF; and c) stirring 10 in a
solution of 55% aqueous hydriodic acid.
C) were recorded with a Bruker spectrometer, model Avance III
00, and tetramethylsilane Me Si was used as the internal standard.
The chemical shifts reported are expressed on the scale in parts per
million (ppm) from the Me Si reference signal. Melting points were
6
4
4
(3) All four reduction procedures convert 10 into 13 in
determined on a Thomas–Hoover Unimelt capillary melting point
apparatus and are uncorrected.
yields ranging from 70 to 100%. However, pure 13 can be
more readily isolated from the first and third procedures
using Na or Li metal or titanium(II) isopropoxide, respec-
tively.
Authentic 1H and 13C NMR spectra for comparison with those
of all known compounds listed in Supplemental Information are
accessible from AIST: Integrated Spectral Database System of Or-
ganic Compounds.[
(4) The oxidative rearrangement of 13 to 10 was another
24]
accidental discovery made in the course of this study. When
CDCl was allowed to stand in laboratory light in a loosely Preparation of Starting Materials and Products
3
capped NMR tube, it reverted to 10. Such adventitious air
6
,12-Diphenyldibenzo[b,f][1,4]-diazocine (10): The requisite o-di-
oxidation of 13 could be purposively achieved by heating
3 with two equiv. of DDQ.
benzoylbenzene (14) was prepared in strict adherence to a pub-
lished procedure[ from phthaloyl chloride and phenylmagnesium
25]
1
(5) Careful comparison of the reaction conditions pre- bromide in ethyl ether, in 27% yield and recrystallized from 95%
vailing for the facile reductive rearrangement of 10 into 13 ethanol, m.p. 146–147 °C.
and the equally smooth oxidative rearrangement 13 into 10
The procedure for preparing 10 has been modified from that em-
has led us to propose a similar crucial intermediate 28 and
chains of SET processes between the two redox products,
[7]
ployed by Olliéro and Solladié, namely by condensing 14 and o-
diaminobenzene (15) in a 1:3 molar ratio in benzene and heating
whose actual individual steps remain to be corroborated by at reflux for 10 days with p-toluenesulfonic acid monohydrate (16)
future kinetic and physiochemical studies.
in a ratio of 14:16 = 1:0.25. Instead, we have conducted the reaction
in refluxing toluene with the condenser attached to a Dean–Stark
trap. With this procedure the expected amount of water was col-
lected after 12 h. The resulting solution was subjected to rotary
evaporation. The dark residue was purified by column chromatog-
raphy on silica gel employing a 3:1 by volume hexane/ethyl acetate
eluent. From 5.73 g of 14 and 6.50 g of 15, 6.32 g of 10 (100%) was
Experimental Section
General Reaction and Hydrolytic Workup Procedures: All reactions
were carried out under a positive pressure of anhydrous, oxygen-
Eur. J. Org. Chem. 2014, 7489–7498
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7495

J. J. Eisch, W. Liu, L. Zhu, A. L. Rheingold
FULL PAPER
obtained, yellow crystals, m.p. 197–198 °C (ref.[7] 184–189 °C). H
1
(1H NMR and TLC identical with those of authentic 13), m.p. 212–
NMR (CDCl
3
): δ = 7.82 (d, 4 H), 7.43 (t, 2 H), 7.40 (q, 2 H), 7.37 213 °C.
13
(t, 4 H), 7.16 (q, 2 H), 7.00 (m, 4 H) ppm. C NMR (CDCl
3
): δ
When a reaction of the same scale was conducted with lithium
metal at 25 °C but for 24 h, again a quantitative yield was isolated.
=
169.73, 141.91, 137.8, 136.43, 130.99, 128.97, 128.74, 128.25,
127.11, 124.05, 121.18 ppm.
Second Method: Heating a 1:3 Molar Mixture of 10 with o-Diamino-
benzene (15) with Catalytic p-Toluenesulfonic Acid (16): See
Scheme 6, 10Ǟ13. Heating of an intimate mixture of 500 mg
2
-(2-Aminophenyl)-1,3-diphenylisoindole (13): According to a pub-
[
7]
lished procedure, an intimate mixture of 5.73 g (20 mmol) of 14,
.16 g (20 mmol) of 15 and 1.0 g (5.3 mmol) of p-toluenesulfonic
2
(
1.4 mmol) of 10, 500 mg (4.6 mmol) of 15 and 50 mg of 16 in an
acid monohydrate (16) was heated in an oil bath at 200 °CϮ5 °C
for 2 h, whereupon it became a viscous melt. Although the authors
claim they obtained a quantitative yield, the operative steps and
criteria are not given. In our procedure the reaction mixture was
stirred with a water/ethyl ether slurry and the organic layer was
oil bath at 200 °CϮ5 °C for 60 min. gave a viscous melt. After
cooling the reaction mixture with a water/ethyl ether slurry and the
ether lay was worked up and underwent column chromatography,
as in the preparation of authentic 13 (Cf. supra). The yield of pure
1
1
3 was 75% and 10% of phenazine was isolated ( H NMR and
separated and dried with anhydrous Na
solution and evaporation of the solvent left a dark residue of crude
3 and other amino components. Column chromatography on sil-
ica gel with a 3:1 v/v hexane/ethyl acetate eluent yielded 4.18 g
2 4
SO . Filtration of the dried
TLC). About 15% of unidentified amines were recovered.
Third Method: Heating 10 in THF with Titanium dichloride (20a) or
with Titanium Diisopropoxide (20b): See Scheme 9. The titanium(II)
reagents, 20a or 20b, can be readily prepared, just before use for
reduction of 10, by suspending 1.00 mmol of TiCl or Ti(O-iPr)
4 4
in 10 mL of THF under argon at –75 °C with stirring. Then 2.2
1
1
(85%) of 13, 10% of phenazine (17) (by H NMR and TLC) and
ca. 5% of other amines: 13 from 95% ethanol, yellow crystals, m.p.
[
7]
1
2
12–213 °C (ref. 201–202 °C). H NMR (CDCl
H), 7.30 (d, 3 H), 7.25 (m, 4 H), 7.17 (t, 2 H), 7.08 (t, 1 H), 7.04
m, 2 H), 6.96 (d, 1 H), 6.62 (t, 1 H), 6.58 (d, 1 H), 3.47 (b, 2 H)
3
): δ = 773 (q, 2
molar equivalents of n-butyllithium in hexane of known molarity
was transferred to the brown suspension of the respective TiE . As
4
(
13
the reaction mixture was allowed to come to 25 °C over 90 min,
the reaction suspension turned pitch black, signaling the respective
ppm. C NMR (CDCl
1
1
3
): δ = 143.17, 131.78, 130.16, 129.59,
29.54, 128.13, 126.45, 124.47, 124.41, 123.12, 122.5, 119.88,
18.29, 115.97 ppm.
[11,12,14]
formation of TiCl
Reaction of 10 with Titanium(II) Dichloride (20a): When 10
180 mg, 0.5 mmol) in 10 mL of THF was mixed at 25 °C with the
2 2
or Ti(O-iPr) .
Preparation of 2(9H-Carbazo-yl)aniline (27): This known com-
pound 27 was prepared in two steps from carbazole in two straight-
forward procedures.
(
foregoing batch of 1.0 mmol of 20a in 15 mmol of THF and stirred
for 48 h, the usual hydrolytic workup yielded 62% of 13.
9-(2-Nitrophenyl)-9H-carbazole (28): 9H-Carbazole (0.24 g,
Moreover, when a 2:1 mixture of 20a and 10 in THF was allowed
to reflux for 24 h, only 6% of 13 was obtained upon workup. This
result indicates that titanium dichloride adduct of 10, formed to
the extent of 62% at 25 °C, to dissociate largely at 70 °C in boiling
THF.
1.44 mmol), 2-fluoronitrobenzene (0.21 g, 1.51 mmol) and cesium
carbonate (0.13 g, 2.14 mmol) were heated to reflux in 5 mL of
DMSO for 4 h. Then the reaction mixture was poured into ice
water to give a yellow precipitate. After workup 9-(2-nitrophenyl)-
9H-carbazole (0.32 g, 86% yield) was obtained by recrystallization
from an ethyl acetate/hexane mixture as beige needles, m.p. 171–
Reaction of 10 with Titanium Diisopropoxide (20b): When 10
[
26]
1
1
72 °C (ref. 156 °C). H NMR (CDCl
3
): δ = 8.18 (m, 3 H), 7.83 (180 mg, 0.5 mmol) in 10 mL of THF was mixed at 25 °C with the
(m, 1 H), 7.68 (m, 2 H), 7.42 (t, 2 H), 7.33 (t, 2 H), 7.14 (d, 2 H) foregoing batch of 1.0 mmol of 20b in 15 mL of THF and stirred
13
ppm. C NMR (CDCl
3
): δ = 147.39, 140.78, 134.22, 131.39,
for 48 h, the usual hydrolytic workup yielded 13 quantitatively.
When a 1:1 molar equivalent reaction was carried out at 25 °C, the
usual workup yielded 41% of 13 with 59% of 10 being recovered.
131.25, 129.13, 126.31, 125.92, 123.84, 120.66, 120.56, 109.06 ppm.
2-(9H-Carbazol-9-yl)aniline (27): The 9-(2-nitrophenyl)-9H-carb-
azole (0.65 g, 2.25 mmol) and stannous chloride (1.61 g,
.13 mmol) were mixed in 15 mL absolute ethanol and the mixture
heated at reflux for 8 h. Then 15 mL aqueous 1 n sodium hydroxide
Fourth Method: Reaction of 10 with 55% Aqueous Hydriodic Acid:
With reference to the data in Table 1 and the reactions depicted in
Scheme 10, the sample composition in each of the four test tubes
was 150 mg (0.420 mm) of 10 admixed with 1.24 mL (10 mo-
lar equiv.) of 55% aqueous hydriodic acid (containing no stabilizer,
from Sigma–Aldrich). All Pyrex tubes were equipped with glass
stoppers and the tube for Run 2 only was completely enveloped in
aluminum foil, so as to exclude all lab (fluorescent) light. The sam-
7
was added and the suspension was stirred at room temperature for
1
9
h. The suspension was filtered and after workup 2-(9H-carbazol-
-yl)aniline (27) was obtained by recrystallization from absolute
ethanol (0.44 g, 75% yield) as colorless crystals, m.p. 124–125 °C
[
26]
1
(
3
ref. 119–121 °C). H NMR (CDCl ): δ = 8.16 (d, 2 H), 7.41 (t,
2
(
H), 7.29 (m, 4 H), 7.18 (d, 2 H), 6.96 (d, 1 H), 6.92 (t, 2 H), 3.55 ples were placed in water baths set at 60Ϯ2 °C, 25Ϯ2 °C and 0 °C,
): δ = 143.86, 140.64, 129.63, respectively. After reaction each sample was quenched with 10%
b, 2 H) ppm. 13C NMR (CDCl
3
1
1
29.60, 126.01, 123.35, 122.40, 120.31, 119.87, 118.98, 116.63,
10.13 ppm.
aqueous sodium bisulfite to remove the I2 and HI. The organic
products were extracted into ethyl ether and the ether extract dried
2 4
with anhydrous Na SO . After solvent removal the product mixture
was analyzed by TLC and quantitative H NMR spectroscopy.
Reactions Effecting the Reduction of 6,11-diphenyldibenzo-
b,f][1,4]-diazocine (10) to 2-(2-aminophenyl)-1,3-diphenylisoindole
13).
1
[
(
Oxidation of 2-(2-Aminiophenyl)-1,3-diphenylisoindole (13) to 6,11-
Diphenyldibenzo[b,f][1,4]diazocine (10)
First Method: Treatment of 10 in THF with Sodium or Lithium
Metal and Subsequent Hydrolysis: See Scheme 4. A solution of
Serendipitous Route: Isoindole 13 (100 mg) was mixed with silica
gel (SiliaFlash , P60 40–63 μm, pH = 7.2, Silicycle) (5.0 g) and
chloroform (15 mL) (containing 0.75% ethanol as stabilizer,
Fisher) and then magnetically stirred in the open air at room tem-
perature for 48 h. After workup [1,4]diazocine 10 was obtained in
®
5
4
3
40 mg of 10 in 15 mL of THF under argon was treated with either
00 mg (17 mmol) of sodium metal pieces (2–3 mm in size) at 25–
0 °C with stirring for 10 h. Quenching with water (some as H Ȇ)
2
and usual workup gave a quantitative yield of 13, yellow crystals
7496
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 7489–7498

Reversible SET Ring Contraction and Expansion Processes
4% yield. When the reaction was carried out under the same con-
ditions but at reflux for 8 h, then a 96% yield of [1,4]diazocine 10
was formed.
9
carboxyl groups at the 3- and 10-positions has estimated the
minimum repulsion energy of such ortho H groups in a transi-
tion of 1 to planar 3 at 20 kcal/mol (N. L. Allinger, W. Szkryb-
alo, M. A. DaRooge, J. Org. Chem. 1963, 28, 3007–3009). In
addition, the ring strain energy of producing an all-planar
eight-membered ring is assessed at about 23 kcal/mol more (A.
Streitweiser Jr., Molecular Orbital Theory for Organic Chemists,
Wiley, New York, 1962, p. 283). b) Up to 81 kcal/mol, the mean
bond stabilization energy for the fully formed Csp –Csp σ-
bond between C6 and C12, in: resonance structures 6Ǟ7 in
Scheme 2, see: J. Waser, K. N. Trueblood, C. M. Knobler,
Chem. One, McGraw-Hill, New York, 1976.
Purposive Route: 2-(2-Aminophenyl)-1,3-diphenylisoindole (13)
(360 mg, 1 mmol) was heated at reflux for 10 h in 15 mL of THF
with DDQ (24) (0.50 g, 2.2 mmol). After workup 6,11-diphenyldi-
benzo[b,f][1,4]diazocine (10) was separated by column chromatog-
raphy (eluent: hexanes/ethyl acetate, 10:1) in 83% yield, along with
other unidentified products. The melting point and NMR criteria
for the separated product were used to confirm that 10 was indeed
generated.
3
3
[
5] R. W. Koch, R. E. Dessy, J. Org. Chem. 1982, 47, 4452–4459.
Attempted Oxidations of 2-(9H-Carbazo-9-yl)aniline (26): See
Scheme 12. 2-(9H-Carbazol-9-yl)aniline (26) (100 mg, 0.39 mmol)
was mixed with silica gel (5.0 g) and 15 mL of chloroform and then
heated at reflux in the open air for 7 h. After workup only the
starting material was detected.
[6] H.-J. Altenbach, H. Stegelmeier, M. Wilhelm, B. Voss, J. Lex,
E. Vogel, Angew. Chem. Int. Ed. Engl. 1979, 18, 962–964; An-
gew. Chem. 1979, 91, 1028.
[
[
7] D. Olliéro, G. Solladié, Synthesis 1975, 246–247.
8] The original method of ref.[ employed refluxing benzene as
the reaction medium and required 10 days, whereas our use of
toluene required less than half a day. Their workup utilized
column chromatography and gave a yield of 72% of 13. Our
workup gave almost 100% of 13 and 10% of 17. The detection
of 17 is first evidence that 15 did function as a reductant (see
Scheme 7).
7]
Compound 26 (100 mg, 0.39 mmol) and DDQ (24) (0.27 g,
1
mmol) were heated at reflux in 15 mL of THF for 10 h. After
workup the starting compound 26 was recovered unchanged.
CCDC-956942 (for 10) and -956941 (for 13) contain the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[
9] The structural akin phenanthridines, phenanthridine itself, 6-p-
phenylphenanthridine and 6-p-biphenylylphenanthridine, have
been shown by alkali-metal reduction in THF and ESR spec-
troscopy to produce monomeric radical-anions and dianions,
similar to 10a and 10b; see: J. J. Eisch, R. M. Thompson, J.
Org. Chem. 1962, 27, 4171–4179.
Supporting Information (see footnote on the first page of this arti-
cle): Copies of the 1H NMR and C NMR and fully displayed
DEPT and IR spectra.
13
[7]
[10] In ref. the authors conclude their brief discussion with the
statement: “The mechanism of the formation of 4 [i.e., 2-(2-
aminophenyl)-1,3-diphenylisoindole] is presently under investi-
gation.” A subsequent search of the authors’ publications since
Acknowledgments
1975 reveals no further information on such promised research.
Our continuing studies on cyclic diimines have been partly sup-
ported by a grant from Dr. John M. Birmingham, Boulder Scien-
tific Company, Mead, Colorado. The 1H NMR and C NMR
spectra reported here were recorded at the regional NMR Facility
at Binghamton University on the 600 MHz instrument, obtained
from the National Science Foundation (NSF) under grant number
CHE-0922815. The HRMS measurements were obtained from the
Molecular Mass Spectrometry Facility in the University of Cali-
fornia, San Diego, with the help of Dr. Yongxuan Su of the Facility
Staff.
Finally, we, his co-authors, mourn the loss of Dr. Wei Liu, a
promising young scientist in lithium battery research, who in Feb-
ruary 2014 was appointed Research Manager for Novel Organic
Battery Electrolyte Development at the Shandong Hairong Power
Supply Materials Co., Ltd., Shandong Province, China. This article
is largely based upon his doctoral dissertation, Novel Aspects of
Epimetalation in Organic Synthesis, completed in October 2013,
and supervised by Professor John Eisch at Binghamton University,
Binghamton, New York, USA. Dr. Wei Liu collaborated exten-
sively in the composition of this article by E-mail exchange after
his return to China.
Further, it is remarkable that these authors have not ever re-
ported any attempt to interconvert 10 and 13.
13
[11] The detection of phenazine in our work (Scheme 7) is the first
actual evidence that 15 functions as a reductant in these reac-
tions. In Scheme 7 a possible pathway, out of several, for the
formation of 17 is offered. A comprehensive summary of such
phenazine syntheses is given in: A. Albert, Heterocyclic Chem-
nd
istry, 2 edition, Oxford University Press, New York, 1968, p.
180–182.
[
12] The term “serendipitous” is used throughout this report in its
primary meaning of the finding of valuable things not sought
for. By contrast, we also employ the expressions, purposive re-
ductive rearrangement of 10 to 13 or purposive oxidative re-
arrangement 13 to 10 to denote the intentional promotion of
these transformations.
[
[
[
13] J. J. Eisch, X. Shi, J. Lasota, Z. Naturforsch. B 1995, 50, 342–
350.
14] J. J. Eisch, X. Shi, J. R. Alila, S. Thiele, Chem. Ber./Recueil
1997, 130, 1175–1187.
15] J. J. Eisch, J. N. Gitua, P. O. Otieno, X. Shi, J. Organomet.
Chem. 2001, 624, 229–238.
16] J. J. Eisch, J. N. Gitua, Organometallics 2003, 22, 24–26.
17] J. J. Eisch, J. N. Gitua, Organometallics 2003, 22, 4172–4174.
18] J. J. Eisch, J. N. Gitua, D. C. Doetschman, Eur. J. Org. Chem.
[
[
[
2
006, 1968–1975.
[
[
[
1] J. J. Eisch, K. Yu, A. L. Rheingold, Eur. J. Org. Chem. 2014,
18–832.
2] J. J. Eisch, R. N. Manchanayakage, A. L. Rheingold, Org. Lett.
[
19] Such a column chromatography separation has been found to
be straightforward. Further modifications of this procedure
would be to study the use of solutions of anhydrous HI gas in
8
2009, 11, 4060–4063.
3] The melting point given for the E isomer of N-(2-amino-1,2-
2 2
CH Cl , as the reagent for converting 10 into 13.
diphenylethenyl)carbazole on page 4061, column A, paragraph
[20] That such a general intermediate 28 plays a pervasive role in
alkali metal and carbanion reductions has been recently re-
viewed: J. J. Eisch, Res. Chem. Intermed. 1996, 22, 145–187.
[21] The likely electronic steps involved in Scheme 14, electron loss
(13Ǟ31), cyclization (31Ǟ32Ǟ33) and electron loss
(33Ǟ10) were suggested by an anonymous reviewer, as well as
[
2]
2, line 14 in ref. as 5b, is incorrect; the correct m.p. is 124–
125 °C.
[
4] a) The necessary planarity of 1 required for Hückel aromaticity
involves two energy barriers. First, previous research on the
barrier to inversion of the optical active derivative of 1 bearing
Eur. J. Org. Chem. 2014, 7489–7498
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7497

J. J. Eisch, W. Liu, L. Zhu, A. L. Rheingold
FULL PAPER
by our reading of research articles on aminium cation radicals
[23] General detailed procedures for purification of solvents and
conducting organometallic reactions under an inert atmo-
sphere are also available: J. J. Eisch, Organomet. Synth. 1981,
2, 7–25; J. J. Eisch, Organomet. Synth. 1981, 2, 94–96.
[24] National Institute of Advanced Industrial Science and Tech-
nology; website: http://riodb01.ibase.aist.go.jp/sdbs/cgi-bin/
cre_index.cgi.
[25] F. Jensen, J. Org. Chem. 1960, 25, 269.
[26] H. G. Dunlop, S. H. Tucker, J. Chem. Soc. 1939, 1945–1956.
Received: August 18, 2014
(see J. H. Horner, F. N. Martinez, O. M. Musa, M. Newcomb,
H. E. Sheridan, J. Am. Chem. Soc. 1995, 117, 11124–11133).
We are grateful for the reviewer’s recommendations in formu-
lating the SET pathways given in Scheme 14.
[
22] The foregoing steric repulsion argument for the failure of ring
closure of biradical 33 to form 27 is based upon kinetic control.
Examinations of the structure of product 27 reveals aspects of
thermodynamic control hindering its formation, such as an
anti-Hückel 16-π-electron ring, two nonbenzenoid rings and
two repelling unshared electron pairs on the N-centers.
Published Online: October 14, 2014
7498
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 7489–7498