Steric factors in the single-electron transfer carbolithiation and transannular reduction of 6,12-diphenyldibenzo[b,f][1,5]diazocine by organolithium reagents

Article abstract of DOI:10.1002/ejoc.201301181

The puckered tricyclic heterocycle 6,12-diphenyldibenzo-[b,f][1,5] diazocine, a novel probe for revealing single-electron transfer (SET) reactions, is here employed as a model substrate for investigating the carbolithiation of imines and the stereochemistry, the regiochemistry, and the electronic nature (polar or SET) of such R-Li additions. The chemical behavior of the following organolithium reagents (mainly in THF) toward the diazocine was studied: benzyl-, methyl-, phenyl-, n-butyl-, and phenylethynyllithiums, as well as the reactive allylmagnesium chloride and the typical tert-butylmagnesium chloride. The only reagent that failed to carbometalate or to reduce the diazocine transannularly was phenylethynyllithium. The tert-butyl Grignard reagent could not carbometalate but did reduce the diazocine with photocatalysis. The other five organometallics yielded both carbometalation of the diazocine and, upon hydrolysis, transannular SET reduction to 4b,9b-diphenyl-4b,5,9b,10- tetrahydroindolo[3,2-b]indole ( indoloindole ) in varying pro- portions. Three types of carbometalation products were observed in such hydrolyzed reactions: (1) exo-1,2-addition to a C=N bond by C₆H ₅CH₂Li, H₂C=CHCH₂MgCl, and CH ₃Li, as established by spectral and XRD data; (2) 1,4-addition to the ortho-phenyl carbon C_o=C-C=N linkage by phenyllithium with the subsequent elimination of LiH; and (3) 1,4-addition, as in type 2, by n-butyllithium but without the elimination of LiH. The foregoing modes of reactivity, the concomitant formation of the telltale SET reduction product, indoloindole , and the failure of PhC≡CLi to react at all with the diazocine are marshaled to support the conclusion that the organometallics reacting with the diazocine, by carbometalation and by transannular reduction, do so by SET processes and not by straightforward nucleophilic attack. The relative stabilities of sp³-, sp²-, and sp-hybridized carbon radicals possibly involved were assessed in concluding that PhC≡CLi is incapable of reacting through any SET pathway.

Full text of DOI:10.1002/ejoc.201301181

FULL PAPER
DOI: 10.1002/ejoc.201301181
Steric Factors in the Single-Electron Transfer Carbolithiation and
Transannular Reduction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine by
Organolithium Reagents[‡]
John J. Eisch,*[a] Kun Yu,[a][‡‡] and Arnold L. Rheingold[b]
Keywords: Electron transfer / Lithiation / Reduction / Regioselectivity / Steric effects
The puckered tricyclic heterocycle 6,12-diphenyldibenzo-
[b,f][1,5]diazocine, a novel probe for revealing single-elec-
tron transfer (SET) reactions, is here employed as a model
substrate for investigating the carbolithiation of imines and
the stereochemistry, the regiochemistry, and the electronic
nature (polar or SET) of such R–Li additions. The chemical
behavior of the following organolithium reagents (mainly in
THF) toward the diazocine was studied: benzyl-, methyl-,
phenyl-, n-butyl-, and phenylethynyllithiums, as well as the
reactive allylmagnesium chloride and the typical tert-butyl-
magnesium chloride. The only reagent that failed to carbo-
metalate or to reduce the diazocine transannularly was
phenylethynyllithium. The tert-butyl Grignard reagent could
not carbometalate but did reduce the diazocine with photo-
catalysis. The other five organometallics yielded both carbo-
metalation of the diazocine and, upon hydrolysis, transan-
portions. Three types of carbometalation products were ob-
served in such hydrolyzed reactions: (1) exo-1,2-addition to a
C=N bond by C6H5CH2Li, H2C=CHCH2MgCl, and CH3Li, as
established by spectral and XRD data; (2) 1,4-addition to the
ortho-phenyl carbon Co=C-C=N linkage by phenyllithium
with the subsequent elimination of LiH; and (3) 1,4-addition,
as in type 2, by n-butyllithium but without the elimination
of LiH. The foregoing modes of reactivity, the concomitant
formation of the telltale SET reduction product, “indolo-
indole”, and the failure of PhCϵCLi to react at all with the
diazocine are marshaled to support the conclusion that the
organometallics reacting with the diazocine, by carbometal-
ation and by transannular reduction, do so by SET processes
and not by straightforward nucleophilic attack. The relative
stabilities of sp3-, sp2-, and sp-hybridized carbon radicals
possibly involved were assessed in concluding that PhCϵCLi
nular SET reduction to 4b,9b-diphenyl-4b,5,9b,10-tetra- is incapable of reacting through any SET pathway.
hydroindolo[3,2-b]indole (“indoloindole”) in varying pro-
the electron-transfer reduction of the authentic, tub-shaped
Introduction
6,7-diphenyldibenzo[e,g][1,4]diazocine (1)[3] in the hope of
generating Hückel-aromatic, necessarily planar dianion 2.
However, after treatment with a sodium metal suspension
in THF and subsequent hydrolysis, the product was shown
not to be diamine 3 but a Z and E isomeric mixture of
enamines 4a and 4b, both of which were separated and their
individual structures determined by X-ray crystallogra-
phy.[3] Clearly, the profound rearrangement of diazocine 1
upon reduction with sodium (reduction with lithium
equally well) must have involved transannular bonding in
the transition state between the asterisked N and C sites in
1 (Scheme 1). Because of the tub shape of 1, attaining the
necessary coplanarity of 1 would require more activation
energy[4a] than the partial σ-bond stabilization anticipated
Background: Single-Electron Transfer Processes with
Dibenzodiazocines
Our interest in nitrogen analogs of cyclooctatetraene
originated from our reinvestigation of the preparation of
the purported 3,4,7,8-tetraphenyl-1,2,5,6-tetraazocine. We
determined that no such structure had been isolated but
that the putative substance is actually 2,4,5-triphenylimid-
azole.[2] In an extension of our studies, we then undertook
[‡] Cyclic Conjugated Imines, 3. Part 2: Ref.[1]
[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
between C6 and C12 of resonance structure
7 in
http://www2.binghamton.edu/index.php
Scheme 2.[4b]
[b] Department of Chemistry and Biochemistry, University of
California,
To avoid the destabilizing ortho H repulsions in 2, we
now have examined the electron-transfer reduction of an
isomer of 1, namely, 6,12-diphenyldibenzo[b,f][1,5]diazocine
(5) (Scheme 2).[5a] Although the validity of structure 5 was
assured both by its logical synthesis through the bimo-
lecular dehydration of 2-aminobenzophenone and by its
San Diego, La Jolla, California 92093-0332, USA
E-mail: arheingold@uscd.edu
[‡‡]Present address: Kun Yu, Ph.D. Innovative Drug Research
Centre, Hu Xi Campus, Chongqing University,
401331 Chongqing, People’s Republic of China
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201301181.
818
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832

SET Carbolithiation of Imines
Scheme 1.
Scheme 2.
spectroscopic data, its 3D stereochemistry was not.[5b] Our
electron-transfer investigations now envisaged two possible
and equally interesting outcomes: either the generation of
planar, aromatic dianion 6 of diazocine 5, lacking the steric
Scheme 3.
barriers inherent in 2 or 3, or the single-electron transfer discussed,[1] are that the reaction of diazocine 5 with di-
(SET) acceptance of sodium or lithium by 5 in THF to phenylmethyllithium (12) or with triphenylmethyllithium
form alternative dianion 7 with minimal disruption of the (13) in THF yielded quantitative amounts of 10b, with 92
[1,5]diazocine structure. Indeed, just such a reduction did or 90% of expected R group dimers 11, respectively. From
preferentially occur transannularly to yield 7 quantitatively these foregoing results, two important consequences can be
(Scheme 2).
drawn: first, by SET reduction, diazocine 5 can serve as an
In an intriguing and unexpected observation, we then electron-capture agent, because benzyllithium (8), diphenyl-
found that not only Na and Li in THF could transfer elec- methyllithium (12), and triphenylmethyllithium (13) in THF
trons to 5 to yield 7 but benzylic lithium reagents could do yielded quantitative amounts of 10b, with 100, 92, and 90%
so as well. Particularly illuminating were the reactions of of the expected R group dimers (i.e., 11) respectively; and
benzyllithium (8, Scheme 3).
second, diazocine 5 can function as a high-yielding oxidat-
Under the conditions of path a, benzylated adduct 9b ive coupling reagent for benzylic lithium compounds to
was exclusively formed, with the endo or exo stereochemis- form the R–R dimers (i.e., 11) from R–Li.
try of the benzyl group not yet defined pending XRD stud-
The composite behavior of diazocine 5 with benzylic rea-
ies. By contrast, under conditions of path b, the product gents 8, 12, and 13, each of them known to be monomeric
was quantitatively 10b. In the latter case, the reacting benzyl ion pairs in THF, is consistent with the operation of an
groups of 8 appear in the products as an amount of bi- SET mechanism[6] involving proposed caged biradical inter-
benzyl (11, for which R and R = PhCH2) equimolar to the mediate 14 in both the carbolithiation (14aǞ14b) and the
yield of 10b.[1] Further intriguing observations, previously transannular reduction (14Ǟ10a, Scheme 4).
Scheme 4.
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
819

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
Scope of the Present Research
and of donor solvent coordination. Of the R–Li and R–
MgCl reagents employed in THF or in toluene with
N,N,NЈЈ,NЈЈ-tetramethyl-1,2-ethylenediamine (TMEDA),
benzyllithium (8), allylmagnesium chloride (20), and tert-
butylmagnesium chloride (19) exist in solution as mono-
mers;[10] phenyllithium (17) and phenylethynyllithium (18)
are present as dimers in THF;[11] and methyllithium (15)
and n-butyllithium (16) are usually considered as tetrameric
in THF.[11,13]
In previous work, we observed that at low temperatures
of –78 °C a 1.0:1.1 ratio of formula equivalents of diazocine
5 and benzyllithium (8)[7] gave quantitative yield of 6-
benzyl-6,12-diphenyl-5,6-dihydrodibenzo[b,f][1,5]diazocine
(9b) upon hydrolysis. At reflux and extended time, 8, 12,
and 13 with a 1.0:2.2 equiv. of 5 and either 8, 12, or 13
gave only 10b upon hydrolysis. At low temperatures (about
–80 °C), neither 12 nor 13 gave any sign of carbolithiation.
Herein, we extend this study to the following typical po-
lar covalent lithium reagents: (a) formally sp3-hybridized at
the C–Li bond: methyllithium (15) and n-butyllithium (16),
(b) sp2-hybridized phenyllithium (17), and (c) sp-hybridized
phenylethynyllithium (18). So that the results of this study
might parallel the previous results, these reactions were also
conducted in THF, where possible, or at least in ethers as
the medium.[8a] Finally, we included two Grignard reagents,
tert-butylmagnesium chloride (19), a typical reagent, and
allylmagnesium chloride (20), a reagent of unusually high
reactivity toward carbonyl and imino linkages, adding to
such substrates in THF with a rapidity resembling that of
n-propyllithium (21) (exemplified with benzophenone anil
in THF at 25 °C, Scheme 5).[8b]
However, traditional molecular association measure-
ments have been based on colligative properties, depending
on the number of molecules concerned and not on their
nature. Over the last 30 years, however, new insight con-
cerning such lithium aggregates in solution has grown pro-
digiously.[12] Especially through exquisite XRD and 1H
7
NMR and Li NMR measurements made on THF solu-
tions of n-butyllithium, Keresztes and Willard[13] showed
that the oligomers [nBuLi]2·[THF]4, and [nBuLi]4·[THF]4
are in equilibrium with each other at 25 °C (NMR signals
of both complexes were detected by measurements at
–84 °C). From hexane, Barnett and co-workers[14] obtained
crystals of cubane-like tetramers of n-butyllithium-linked in
an infinite chain by TMEDA units, suitable for XRD analy-
sis. Finally, Qu and Collum[15] found that the 1,2-additions
of nBuLi to an optically active aldimine of pivaldehyde in
TMEDA/Et2O can involve four independent mechanisms
and both monomer- and dimer-based nBuLi pathways.
Such pathways display different rates of 1,2-addition and
varying stereoselectivities. Accordingly, we are thus aware
that in our carbolithiations involved with reactions of R–Li
in THF, we are dealing with more than one solvated oligo-
mer of R–Li, the varying reactivities of which may play an
important role in the overall reaction.
Scheme 5.
After we have analyzed the results of the carbometal-
ation of diazocine 5 by the selected R–Li and R–MgCl rea-
gents, we will consider what role the steric nature of these
individual organometallic solvated aggregates may play in
determining the stereochemistry, regiochemistry, and elec-
tronic nature of the carbometalation and transannular re-
duction of 5 (cf. Discussion).
In comparing allylmagnesium chloride (20) with organo-
lithium reagents in additions to diazocine 5, the Schlenk
equilibrium existing among all Grignard components in
ethers complicates interpretation of the carbomagnesi-
ation.[9] The complexation of the diazocine base with the
Lewis acidic MgCl2 tends to shift the equilibrium to the
right, favoring 22 as the carbomagnesiating agent remaining
in solution. Carbomagnesiation could thus arise from either
20 or 22, both existing as bis-THF complexes (Scheme 6).
Results
General Observations of the Reactions of Diazocine 5 with
the Selected Organometallic Reagents
In Table 1 are shown typical reaction runs of diazocine
5 with specific organolithium and organomagnesium rea-
gents in donor media as a function of selected ratio of rea-
gent equivalents, temperature, time, and illumination condi-
tions (Table 1, runs 1–6).[16] The only reagent giving absol-
utely no sign of reaction either in THF or in a (hexameth-
ylphosphoramide) HMPA/THF pair up to 80 °C was phen-
ylethynyllithium (18) (Table 1, run 6). Those lithium rea-
Scheme 6.
The Structure of Organolithiums in Solution
Besides such polar effects stemming from the hybridiza- gents that did react, either at –78 °C or up to 80 °C, were
tion of the carbon atom at the C–Li bond, possible steric methyllithium (15, Table 1, run 3), n-butyllithium (16
effects will arise from the molecular aggregation of the R– Table 1, run 4), phenyllithium (17, Table 1, run 5), and
Li unit of the reagent as a function of monomer structure benzyllithium (8,Table 1, run 1). The two Grignard reagents
820
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832

SET Carbolithiation of Imines
Table 1. Reactions of 6,12-diphenyldibenzo[b,f][1,5]diazocine (5) with the selected organometallic reagents under the given conditions
with R–Li and R–MgCl (cf. Scheme 1 and Table 2).
Run R–M
(medium)
Conditions
Products
(after hydrolysis)
Commentary[a]
T, t, 5/R–M[7]
1
PhCH2Li (8)
–78 °C (2 h)
25 °C (12 h)
ratio 1.0:1.1
98% 9c
2% 10b
run 1a: 1.0:2.2 ratio at 25 to 80 °C, 6 h
gave 100% of 10b and bibenzyl
(TMEDA)
(THF + C6H5CH3)
2
3
allylMgCl (20) (THF)
25 °C (48 h)
ratio 1.0:2.2
91% 9g
run 2a: 1.0:2.2 ratio at 80 °C, 48 h gave 25%
of 10b
MeLi (15) (EtO)2CH2 + THF
–78 °C (2 h)
25 °C (12 h)
ratio 1.0:2.2
29% 9d
71% 10b
steady CH4 evolution throughout (flammable gas,
condensable in liquid N2 trap)
run 3a: of 1.0:1.1 ratio left 55% of 5 unreacted
4
5
6
nBuLi (16) (THF)
–78 °C (6 h)
ratio 1.0:2.2
74% 9e
25% 9h
1% 10b
as 9h came from 9e, actual yield of 9e was at
least 99% (run 4). Run 4a: 1.0:2.2 ratio after
reflux in THF gave 30% of 10b
PhLi (17) (nBu2O + THF)
PhCϵCLi (18) (THF)
–78 °C (2 h)
25 °C (12 h)
ratio 1.0:1.1
84% 9f
16% 10b
during this reaction LiH was eliminated, as
revealed by H2 gas evolution upon hydrolysis
25 °C (12 h)
ratio 1.0:1.1 or
1.0:2.2
no reaction
run 6a: the same reaction ratios in HMPA/THF
heated at reflux for 2 h also gave no reaction
[a] Further comments on reaction run variants are given in the Results and Experimental Section under the reactions of the specific R–
M with diazocine 5.
examined were allyl reagent 20 in run 2 giving an excellent analytical and spectroscopic data that the structure of the
yield of 9g and tert-butyl reagent 19, which was completely organyl derivative is 9c, 6-benzyl-6,12-diphenyl-5,6-dihydro-
inert in the dark. However, with visible light it can reduce dibenzo[b,f][1,5]diazocine (Scheme 7).[1]
5 to 10b. All of these reactions gave the telltale side product,
indoloindole 10b, of the SET reduction of diazocine 5, ac-
cording to path b of Scheme 1. However, the proportion of
10b formed depended on the nature of R–Li and the par-
ticular reaction conditions. For example, 2% of 10b resulted
from a 1.0:1.1 ratio of diazocine 5 and benzyllithium (8) at
–78 °C, but a 1.0:2.2 ratio of 5/8 at 80 °C gave 100% of 10b
(Table 1, run 1a). From these observations there is evidence
that the SET reduction to 10b (after hydrolysis) plays at
Scheme 7.
least a partial role in all such carbolithiations (cf. see be-
low).
However, the stereochemistry of the 6-benzyl group at
C6 had remained undetermined. However, in this report we
can now present the single-crystal X-ray data demonstrat-
The further interesting finding in such reactions is that
the organyl derivatives of diazocine 5 isolated upon hydrol-
ing that the benzyl group in 9c is attached solely to the exo
ysis of these carbolithiations have specific stereochemical
face of 5 at the C6 position through an exo-syn-carbolithi-
1
ation (Figure 1). All other H NMR, 13C NMR, and IR
spectroscopic data (NH band at 3370 cm–1 and C=N
stretch at 1629 cm–1) and exact mass of 450.209 are in
agreement with this structure (cf. Supporting Information).
To determine whether 9c might undergo thermal inver-
sion of its tub-shaped configuration into the 6-endo-benzyl
isomer, 9c was heated at 200 °C under an atmosphere of
nitrogen without change. However, at 300 °C, it decom-
posed smoothly to 5 with the evolution of toluene.
and regiochemical features, which are valuable in specifying
mechanisms. The structures of the particular organyl deriv-
ative of 5 obtained from the specific R–Li or R–MgCl rea-
gent are tabulated in Table 2. Note that a given R–Li or R–
MgCl yields only the 1,2- or the 1,4-carbometalation prod-
uct and never both types of addition.
Particular Observations from the Reactions of Diazocine 5
with the Specific Organometallics Examined
Reaction with Allylmagnesium Chloride (20)
Reaction with Benzyllithium (8)
The reaction of a 1.0:2.2 molar ratio of diazocine 5 with
Consider the case of the reaction of diazocine 5 and 20 in THF was essentially complete after 2 h at 25 °C with
benzyllithium (8) at –78 °C in a 1.0:1.1 molar ratio (Table 1, the formation of 91% of 6-exo-allyl-6,12-diphenyl-5,6-di-
run 1). In a preceding publication, we already showed by hydro[b,f][1,5]diazocine (22). Upon proceeding for 48 h, 22
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
821

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
Table 2. Structures of carbometalation products obtained from the
interaction of diazocine 5 with various organolithium and organo-
magnesium reagents after subsequent hydrolysis.
Figure 1. Thermal ellipsoid (30%) diagrams for 6-exo-benzyl-6,12-
diphenyl-5,6-dihydrodibenzo[b,f][1,5]diazocine
(9c),
colorless
blocks, m.p. 209–211 °C. Selected bond lengths (atom separations)
[Å]: N1–C13 1.399(2), C12–C13 1.401(2), N1–C1 1.481(2), N2–C11
1.276(2), N2–C10 1.415(2), C1–C9 1.542(2), C1–C2 1.561(2), C2–
C3 1.512(2), C3–C4 1.393(2), C3–C8 1.400(2), C4–C5 1.390(2), C5–
C6 1.383(3), C6–C7 1.387(3), C7–C8 1.386(2), C9–C10 1.405(2),
C11–C12 1.486(3); and bond angles (°): C13–N1–C1 126.72(13),
C13–C12–C11 119.74(15), C11–N2–C10 119.01(15), N1–C1–C9
112.11(13), N1–C13–C12 120.94(16), N1–C1–C2 106.02(13), C9–
C1–C2 110.41(13), C3–C2–C1 115.61(13), C4–C3–C8 118.15(15),
C4–C3–C2 119.28(15), C8–C3–C2 122.52(15), C5–C4–C3
121.48(17), C6–C5–C4 119.45(17), C5–C6–C7 120.06(16), C8–C7–
C6 120.31(17), C7–C8–C3 120.52(17), C10–C9–C1 120.20(15), C9–
C10–N2 121.52(14), N2–C11–C12 123.45(16).
Scheme 8.
[a] Proportion of 10b depended on specific reaction conditions,
ranging from a detectable 1 to 100% (cf. commentary in Table 1).
mass for (9g)·H+ of 401.2012 with delta equal to 0.2 and a
began to reverse to regenerate 8% of 5 and 1% of 10b. The composition of C29H25N2. The IR spectrum has an N–H
conversion of the carbomagnesium salt (i.e., 22) of product stretch at 3360 cm–1 and a C=N stretch at 1625 cm–1; the
9g into dimagnesium salt 23, therefore, can be largely attrib- 1H NMR spectrum has 18 aromatic protons (1 H, 7.60–
uted to a thermal SET process (Scheme 8).
7.61 ppm; 3 H, 7.39–7.42 ppm; 3 H, 7.22–7.27 ppm; 4 H,
The structure of 9g was readily established by the XRD 7.06–7.08 ppm; 3 H, 6.95–6.98 ppm; 3 H, 6.67–6.82 ppm),
analysis of the solid, m.p. 223–225 °C (Figure 2) and cor- 5 protons from the allyl group (3 H, 5.18–5.44 ppm; 2 H,
roborated by the HRMS data and the IR, H NMR, and 3.08 ppm), and an N–H signal at δ = 4.00 ppm and the 13C
1
13C NMR spectroscopic data. The high-resolution mass NMR spectrum has the appropriate number of 13C signals
spectrum of protonated allyl adduct 9g yields a mass of of 25, including one N=C at δ = 169 ppm, N–CAr2 at δ =
401.2013, which compares favorably with the theoretical 65 ppm, and CH2=CHCH2CN at δ = 50.23 ppm.
822
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832

SET Carbolithiation of Imines
Scheme 9.
however, spectroscopic data permitted the structure of 9d
to be assigned as 6-exo-methyl-6,12-diphenyl-5,6-dihydrodi-
benzo[b,f][1,5]diazocine. First, the high-resolution mass
spectrum of protonated 9d yields a mass of 375.1856, which
compares favorably with the theoretical mass for 9d·H+ of
Figure 2. Thermal ellipsoid (30%) diagrams for 6-exo-allyl-6,12-di-
phenyl-5,6-dihydrodibenzo[b,f][1,5]diazocine (9g), colorless block,
m.p. 223–225 °C. Selected bond lengths (atom separations) [Å] and
bond angles [°]: N1–C26 1.286(2), N1–C12 1.420(2), C13–C18
1.400(2), N2–C13 1.412(2), N2–C25 1.496(2), C1–C6 1.388(2), C1–
C2 1.391(2), C2–C3 1.383(2), C18–C26 1.494(2), C3–C4 1.387(2), 375.1856 with delta equal to 0 and a composition of
C4–C5 1.387(2), C5–C6 1.394(2), C6–C25 1.543(2), C7–C12
C27H23N2. The IR spectrum has an N–H stretch at
1.408(2), C7–C25 1.540(2), C25–C27 1.556(2), C27–C28 1.498(2),
1
3321 cm–1 and a C=N stretch at 1616 cm–1; the H NMR
C28–C29 1.316(3); C26–N1–C12 118.27(13), C13–N2–C25
124.11(13), C6–C1–C2 120.42(15), C3–C2–C1 120.87(16), C2–C3–
C4 119.28(15), C13–C18–C26 120.11(14), C3–C4–C5 119.75(15),
spectrum has a CH3 singlet at δ = 1.92 ppm and the 13C
NMR spectrum has the appropriate number of 13C signals
C4–C5–C6 121.46(15), C1–C6–C5 118.21(14), C1–C6–C25
122.83(14), C5–C6–C25 118.90(13), C12–C7–C25 120.11(14), N2–
C25–C7 112.15(12), N2–C25–C6 110.87(12), C7–C25–C6
109.59(12), N2–C25–C27 106.37(12), C7–C25–C27 110.43(13), C6–
C25–C27 107.27(12), C7–C12–N1 122.08(14), N1–C26–C18
122.45(14), C18–C13–N2 120.99(14), C28–C27–C25 113.73(14),
C29–C28–C27 124.09(16).
at 23, including N=C at δ = 169 ppm, N–CAr2 at δ =
64 ppm, and CH3CN at δ = 36 ppm. Thus, 1,2-carbolithi-
ation of a C=N linkage took place, but only as the minor
product, compared with 10b.
Reaction with Methyllithium (15)
Diazocine 5 reacted relatively slowly with methyllithium
(15) in a 1.0:1.1 ratio [1,2-dimethoxyethane (DME)/THF at
–78 °C for 2 h and at 25 °C for 12 h] to consume 45% of 1
and produce organyl derivative 9d (17%) and indoloindole
10b (28%) in a 1.0:1.65 ratio. An identical reaction except
employing a 1.0:2.2 ratio of reactants (Table 1, run 6) led to
complete consumption of 5 and produced 29% of 9d and
71% of 10b (a ratio of 1.0:2.45). Throughout both reac-
tions, a continuous stream of a flammable gas was evolved.
As the gas was condensable in a liquid nitrogen gas, it is
most likely methane. The evolution of methane in this reac-
tion is further corroboration that methyl radicals are set
free in this reaction of CH3Li with 5. Those radicals that
escape coupling with radical 14 to form 14b (Scheme 4)
Figure 3. Thermal ellipsoid (30%) diagrams for 6-(o-biphenylyl)-
12-phenyldibenzo[b,f][1,5]diazocine (9f), colorless block, m.p. 254–
256 °C. Selected bond lengths (atom separations) [Å] and bond
angles [°]: N1–C14 1.284(2), N1–C13 1.417(2), N2–C7 1.275(2),
would readily be able to abstract a H· from THF. At these N2–C6 1.422(2), C1–C6 1.396(2), C1–C14 1.500(2), C21–C22
temperatures, abstraction of a H+ by carbanionic attack by
1.392(3), C21–C26 1.407(3), C22–C23 1.390(3), C23–C24 1.384(3),
C24–C25 1.385(3), C7–C21 1.494(2), C7–C8 1.495(3), C25–C26
CH3Li on THF would be too slow to produce methane.
1.400(3), C26–C27 1.490(3), C8–C13 1.405(3); C14–N1–C13
120.92(16), C7–N2–C6 121.69(16), C6–C1–C14 120.68(16), C22–
These results indicate that 10b is both the kinetically fa-
vored as well as the thermodynamically favored product
(Scheme 9).
C21–C26 119.66(17), C22–C21–C7 117.52(16), C1–C6–N2
122.88(16), C26–C21–C7 122.61(16), N2–C7–C21 117.01(16), C23–
C22–C21 121.40(18), N2–C7–C8 125.13(16), C21–C7–C8
117.85(16), C24–C23–C22 119.07(19), C23–C24–C25 120.13(18),
C24–C25–C26 121.58(18), C13–C8–C7 121.56(16), C25–C26–C21
118.10(18), C25–C26–C27 118.34(17), C21–C26–C27 123.54(16),
Resulting organyl product 9d, m.p. 179–181 °C, was sub-
jected to attempted crystal growth in nearly 30 solvents
without success. Only ultrafine needles, about 10–20 µm
wide and useless for XRD studies, resulted.[17] Fortunately, C8–C13–N1 121.48(17),N1–C14–C1 123.02(17).
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
823

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
Because of the failure to obtain crystals suitable for
Employing a 1.0:2.2 ratio of diazocine 5 and phenyllith-
XRD analysis, assignment of the 6-methyl group to an endo ium (17) in reactions in similar solvent pairs (THF/ethyl or
or exo configuration can be made only by analogy. Given n-butyl ether) and a similar temperature range (–75 to
that the stereochemistry of the entering benzyl and allyl 25 °C), however, gave indoloindole 10b as the major prod-
groups in the 1,2-carbometalations of diazocine 5 by benz- uct (70–80%) and the carbolithiated product as the minor
yllithium (8) and by allylmagnesium chloride (20) is shown (2–12%). These results are in keeping with kinetically fa-
to be exo by XRD data, namely, 9c (Figure 1) and 9g (Fig- vored product 26, formed at –78 °C, being reconverted into
ure 3), it is most likely that the methyl group in 9d has also 5 in the presence of an excess amount of phenyllithium (17)
the exo orientation.
at temperatures greater than –78 °C and thence undergoing
SET reduction into 10a (Scheme 12) (cf. Scheme 18 and ac-
companying text for a proposed manner).
Reaction with Phenyllithium (18)
Diazocine 5 was allowed to react with phenyllithium (18)
in THF or in THF/butyl ether mixtures in 1.0:1.1 or in
1.0:2.2 mixtures between –78 and 25 °C. In 1.0:1.1 mixtures,
the principal product was 6-(o-biphenylyl)-12-phenyldiben-
zo[b,f][1,5]diazocine (9f, 65–85%) and 10a as the minor
product (15–25%, Scheme 10).
Scheme 10.
Upon warming the reaction mixture to 25 °C, an off-
white precipitate formed. Hydrolysis of the mixture directly
or of the separated solid resulted in vigorous gas evolution,
and the collected gas was flammable and noncondensable
in a liquid nitrogen trap (thus, H2 gas).
Such a similar lithium hydride elimination has already
been observed in the 1,2-carbolithiation of pyridine by 18
by Ziegler and Zeiser (Scheme 11).[18] At low temperatures,
pyridine and 18 form adduct 24, but upon warming, LiH
was eliminated to yield 2-phenylpyridine (25, Scheme 11).
Scheme 12.
The relatively unstable phenyl radicals, released in the
SET conversion of 5 into 10a, have little opportunity to
couple to yield biphenyl (only about 6% detected) before
they abstract hydrogen atoms from the ethereal medium
(Scheme 12). A similar conversion of 5 or the 1,2-carbolithi-
ation product into 10a apparently takes place in the reac-
tions of methyllithium (15) with diazocine 5 (Table 1,
run 3).
Reaction with n-Butyllithium (16)
Diazocine 5 was allowed to react with n-butyllithium in
a 1.0:2.2 equiv. ratio[7] in THF at –78 °C for 6 h. The subse-
quent treatment with methanol at –78 °C, warming to
Scheme 11.
25 °C, and the usual hydrolytic workup with ether and
1
The structure of final product 9f was established by XRD water led to a dried ether extract, the H NMR spectro-
analysis and is shown in Figure 3. Thin-layer chromato- scopic analysis of which showed the presence of 74% of
graphic analysis of recrystallization mother liquors revealed butyl derivative 9e, 25% of butyl hydroperoxy derivative 9h,
no components other than 9f and 10b:
and 1% of indoloindole 10b (Table 1, run 4; Scheme 13).
That butyl hydroperoxide 9h was actually formed during
The high-resolution mass spectrum of protonated phenyl
adduct 9f yields a mass of 435.1858, which compares favor- the hydrolytic workup of run 4 was shown in the following
ably with the theoretical mass for (9f)·H+ of 435.1856 with way. A reaction run (Table 1, run 4a), identical with run 4,
delta equal to 0.5 and a composition of C32H23N2. The IR was allowed to stir at 25 °C for 12 h after hydrolysis. Col-
spectrum has two C=N stretches at 1633 and 1622 cm–1; umn chromatography of the organic residue on silica gel
the 1H NMR spectrum has 22 aromatic protons (2 H, 7.80– with an ethyl acetate/hexane eluent showed 91% of 9h had
7.81; 1 H, 7.56–7.57 ppm; 11 H, 7.21–7.44 ppm; 2 H, 7.15– formed with 7% of diazocine 5 and 2% of 10b remaining.
7.19 ppm; 5 H, 6.79–6.97 ppm; 1 H, 6.45–6.46 ppm) and
Pure samples of 9e, pale yellow with m.p. 150–151 °C,
the 13C NMR spectrum has the appropriate number of 13C and 9h, with m.p. 90–98 °C (dec.), were subjected to XRD
signals, 28, including two N=C at δ = 171 and 168 ppm. structural analysis and their structures and appropriate data
All these spectroscopic data are in accord with the X-ray are presented in Figures 4 and 5, respectively. Their struc-
structure of 9g, as given in the Supporting Information.
tures are completely corroborated by the IR, 1H NMR, and
824
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832

SET Carbolithiation of Imines
Figure 5. Thermal ellipsoid (30%) diagrams for 6-(2Ј-exo-butyl-5Ј-
endo-hydroperoxycyclohexa-3Ј,6Ј-dienyl)-12-phenyldibenzo[b,f]-
[1,5]diazocine (9h), m.p. 90–96 °C (dec.), recrystallized from X-ray
laboratory of A. L. Rheingold from dichloromethane right before
XRD structural analysis. Selected bond lengths (atom separations)
[Å] and bond angles [°]: O1–C8 1.4487(17), C12–C17 1.398(2), O1–
O2 1.4637(14), N1–C17 1.4244(17), N2–C11 1.2808(18), N2–C24
1.4155(18), C1–C2 1.520(2), C18–C19 1.4965(19), C2–C3
Scheme 13.
13C NMR spectroscopic data in the Supporting Infor-
mation. In addition, HRMS data confirm the exact mo-
lecular mass and formula of 9e as 417.2325 and C30H29N2. 1.5219(19), C3–C4 1.524(2), C19–C24 1.399(2), C4–C5 1.5539(19),
C5–C6 1.5012(19), C5–C10 1.5094(19), C6–C7 1.326(2), C7–C8
1.489(2), C8–C9 1.4975(19), C9–C10 1.337(2), C10–C11
1.4815(19), C11–C12 1.4967(19); C8–O1–O2 106.90(9), C18–N1–
C17 119.28(11), C11–N2–C24 119.35(12), C1–C2–C3 113.77(12),
C2–C3–C4 111.85(12), C12–C17–N1 120.81(12), C3–C4–C5
114.65(11), C6–C5–C10 111.70(11), N1–C18–C19 123.47(12), C6–
C5–C4 108.60(11), C10–C5–C4 112.41(11), C7–C6–C5 124.43(13),
C6–C7–C8 122.39(13), C24–C19–C18 121.79(12), O1–C8–C7
104.90(11), O1–C8–C9 110.84(11), C7–C8–C9 113.02(12), C10–
C9–C8 123.84(13), C9–C10–C11 120.44(12), C9–C10–C5
122.09(12), C11–C10–C5 117.34(12), C19–C24–N2 119.50(12),
N2–C11–C10 118.97(12), N2–C11–C12 122.52(12), C10–C11–C12
118.48(11), C17–C12–C11 119.59(12).
was attached solely to the ortho position of the 6-phenyl
group in diazocine 6 and a hydroperoxy group was located
on this 6-phenyl group with attachment trans to the n-butyl
group (Figure 5).
Figure 4. Thermal ellipsoid (30%) diagram for 6-(2Ј-exo-butyl-
Reaction with tert-Butylmagnesium Chloride (19)
cyclohexa-3Ј,6Ј-dienyl)-12-phenyldibenzo[b,f][1,5]diazocine
(9e),
In the same molar size as in the foregoing reactions of 5
with R–Li, the reaction of diazocine 5 with 19 in a 1.0:2.2
molar equivalent ratio in THF was heated to reflux for
12 h, while being exposed to ordinary fluorescent labora-
tory light. Usual hydrolytic workup showed the presence of
15% of 10b and 85% of unchanged 5.
If the foregoing reaction mixture and apparatus were
completely coated with aluminum foil, excluding all light, a
similar heating period and workup showed only unchanged
5. The former reaction was thus a photopromoted re-
duction.
pale yellow, m.p. 150–151 °C. Selected bond lengths (atom separa-
tions) [Å] and bond angles [°]: N1–C10 1.284(2), N1–C9 1.416(2),
N2–C7 1.285(2), N2–C15 1.414(2), C1–C2 1.520(3), C2–C3
1.519(3), C3–C4 1.525(3), C4–C5 1.569(3), C5–C21 1.498(3), C5–
C6 1.514(3), C6–C24 1.342(3), C6–C7 1.483(3), C7–C8 1.500(2),
C8–C25 1.394(3), C8–C9 1.398(3), C9–C28 1.398(3), C10–C11
1.482(2), C10–C16 1.499(2) C10–N1–C9 120.69(15), C7–N2–C15
121.09(15), C3–C2–C1 112.48(18), C2–C3–C4 112.42(17), C3–C4–
C5 115.37(17), C21–C5–C6 112.25(18), C21–C5–C4 109.24(17),
C6–C5–C4 112.46(16), C24–C6–C7 120.87(17), C24–C6–C5
121.55(18), C7–C6–C5 117.54(16), N2–C7–C6 117.25(16), N2–C7–
C8 122.90(16), C6–C7–C8 119.83(15), C25–C8–C9 118.97(16),
C25–C8–C7 120.21(16), C9–C8–C7 120.82(15), C28–C9–C8
119.57(16), C28–C9–N1 119.23(16), C8–C9–N1 120.83(16), N1–
C10–C11 117.79(15), N1–C10–C16 123.69(16), C11–C10–C16
118.52(14), C29–C11–C12 118.75(17), C29–C11–C10 121.30(16),
C12–C11–C10 119.90(16), C13–C12–C11 120.45(17).
Discussion
Electronic Mechanisms of the Carbolithiation and
Transannular Reduction of Diazocine 5 and the Possible
Intermediacy of SET Processes
The stereoselective formation of 9h from 9e and the reac-
tion pathway followed by 9h upon acid-promoted re-
arrangement are currently under investigation.[19]
The 3D structure of hydroperoxide 9h was elucidated by
As projected in Schemes 4 and 13 and accompanying
XRD measurements, which showed that the n-butyl group comments, the actual role of SET processes in the carbo-
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
825

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
lithiation of 5 can be inferred if radical pair 2f were to
dissociate into 2g and R· and 2g would then undergo SET
reduction into indoloindole dilithium 10a (Scheme 14).
Given that the hydrolysis product of 10a, namely, 10b, was
found in small or large proportions in all successful carbo-
lithiations of 5 (Tables 1 and 2), we conclude that SET pro-
cesses play a role in all such additions. Furthermore, in all
reaction runs under conditions favoring the formation of
thermodynamically controlled product 10b, namely, a
1.0:2.2 reagent ratio, higher temperature and longer reac-
tion time, the amount of 10b markedly increased. For exam-
ple, with n-butyllithium (Table 1, run 4) there was 1% of
10b at –78 °C but 30% of 10b after reflux in THF.
Scheme 16.
Now, in considering the probability of 18 behaving as an
SET reagent toward diazocine 5 (Scheme 16), we note that
this SET process generates phenylethynyl radical 29. In an
sp-hybridized orbital, radical 29 lies at a higher positive
value of ΔGo than in an sp2-hybridized phenyl radical or
the sp3-hybridized methyl or n-butyl radical. For the equa-
tion in Scheme 16 to proceed spontaneously to the right,
the overall Gibbs free energy ΔGo must be negative. We pro-
pose that in the electron-transfer process of organolithium
reagents with diazocine 5 lithium reagent 18 fails to react
because the ΔGo would be positive. Given that the relative
ΔGo values are unknown, we sought an estimated measure
of radical stability based upon the enthalpic C–H bond
dissociation energies (ΔHo in kcalmol–1), given in Table 3.
Arranged in order of the relative decreasing stability of rad-
icals the following sequence results: allyl (85) and benzyl
(85)Ͼ methyl (104) and phenyl (103)ϾϾ phenylethynyl (ca.
125).[24] Therefore, phenylethynyllithium (18) in any SET
process, be it carbolithiation or transannular reduction of
diazocine 5, would involve a much greater gain of Gibbs
free energy to have the clustered intermediates of 29, 30,
and THF solvation ending up having a final positive ΔGo
value (Table 3 and Scheme 16).
Scheme 14.
No other possible pathway than transannular electron
transfer can credibly be proposed for the reduction of 5 into
10a.
The further conclusion that SET processes play the only
role in such carbolithiations, exclusive of any role for nu-
cleophilic attack, can be supported by the following obser-
vations, summarized here and discussed hereafter: (1) Phen-
ylethynyllithium (18), unreactive toward diazocine 5, readily
adds to the carbonyl group of aldehydes,[20] benzophen-
one,[21] esters,[22] Weinreb amide 27,[15] and even the azo-
methine linkage of 28,[23] apparently in a clear nucleophilic
pathway (Scheme 15) (2) Some carbolithiations of 5 are
readily reversible with an excess amount of the allylic or
benzylic lithium reagent at higher temperatures. (3) The fi-
nal thermodynamically controlled product in such cases as
in point 2 is always dilithium salt 10a, and none of bis-
carbolithiated adduct 31 (Scheme 19).
Table 3. Carbon–hydrogen bond dissociation energies (ΔHo) of rad-
icals listed in decreasing value of ΔH and in increasing order of
radical stability.[24]
R·
ΔHo [kcalmol–1]
R–H
Ph-CϵC·
≈125
104
103
85
PhCϵC–H
H3C–H
C6H5–H
H2C=CHCH2–H
C6H5CH2–H
H3C·
·
C6H5
·
H2C=CHCH2
·
Ph-CH2
85
Scheme 15.
2. Reversibility of the Carbolithiation Adduct with
Benzyllithium (8)
1. Inertness of Phenylethynyllithium (18) toward Diazocine
5
A 1.0:1.1 molar ratio of 5/8 in THF at –78 °C upon
hydrolysis yielded 95% of 2c (Table 1, run 1). When
There is no evidence that lithium reagent 18 reacts with 1.2 equiv. more of 8 was then added to the unhydrolyzed
the foregoing carbonyl substrates other than in a nucleo- mixture resulting from run 1 and then heated to reflux in
philic carbolithiating process. The lack of any indication of THF, the hydrolyzed mixture yielded only 10b and a quanti-
an SET-reduction capability for 18 with carbonyl groups tative amount of bibenzyl (Table 1, run 1a). Such clear re-
can be attributed to the high positive Gibbs free energy ΔG versible carbolithiation can readily be rationalized by
involved in such a SET process as given in Scheme 16.
Scheme 17.
826
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832

SET Carbolithiation of Imines
be 31 and after hydrolysis 32, 6,12-exo,exo-dibenzyl-6,12-
diphenyl-5,6-dihydrodibenzo[b,f][1,5]diazocine, if the carbo-
lithiation were to have occurred by a straightforward,
stepwise nucleophilic attack. It is granted that the second
nucleophilic carbolithiation might be slower than the first
proposed attack, due to the steric crowding in the transition
state leading to putative adduct 31. In contrast with this
expectation, the second equivalent of benzyllithium in THF
already reacts at 25 °C to produce considerable proportions
of 10a (45%) from the monoadduct (55% 9c upon hydroly-
sis) [cf. run 1b in the Experimental Section (Scheme 19)].
Scheme 17.
Scheme 19.
By the principle of microscopic reversibility, if adduct 5d
were formed by the SET path of 5 + 8Ǟ 5aǞ5d, then
thermal reversal of 5d would proceed most readily by SET
path of 5dǞ5bǞ5aǞ5 + 8. As homolysis of a relatively
weak benzyl–C6 bond in 5d would be involved, a SET pro-
cess for 5dǞ5bǞ5a would be expected to be of lower en-
ergy than a carbanionic elimination of 5dǞ5 + 8. Likewise,
the thermal SET radical elimination of allyl radical from 22
leading to 23 should be facile for similar reasons
(Scheme 8).
Discussion of the Stereochemistry and Regiochemistry of
the Carbolithiation of Diazocine 5
Three of the carbometalations of diazocine 5, achieved
with allylmagnesium chloride (20), with benzyllithium (8),
and with methyllithium (15), occur in a 1,2-manner, as was
1
corroborated by HRMS and H NMR, 13C NMR, and IR
spectroscopic data. Furthermore, XRD analyses of the hy-
drolyzed allyl adduct (i.e., 9g, Figure 2) and the correspond-
ing benzyl adduct (i.e., 9c, Figure 1) established unequivoc-
ably that these 1,2-carbometalations occurred in a syn–exo
manner. Any crystals obtained from analogous methyl ad-
duct 9d proved unsuitable for XRD analysis. However, the
full array of spectroscopic data supports its most probable
structure as the 6-exo-methyl isomer as well. Carbolithi-
ation in this case must involve the approach of the mono-
meric (allyl, benzyl) or cluster (tetrameric methyl) reagent
to the exo face of the bucket-shaped eight-membered ring
in a stereoselective syn orientation.
Of the two allylmetalating agents, 20 and 22, present in
the allyl Grignard reagent (Scheme 6), 20 should be more
effective in allylmetalation because as the stronger Lewis
acid, it would complex preferentially with a diazocine nitro-
gen atom.
For reaction runs of diazocine 5 and lithium reagents R–
Li at temperatures of 25 °C and a reactant ratio of 1.0:2.2,[7]
it is not necessary to reckon with the reversibility of the
monoadduct, as with 5d for PhCH2–Li. Rather, intermedi-
ate complex 5e, with an R–Li unit on each N center, could
be involved as coordinated at both nitrogen atoms (i.e., 5f,
Scheme 18). In a concerted homolysis of the C–Li bonds
and the incipient σ-bond formation between C6 and C12 in
transition state 5f, the R radicals would be generated in a
·
solvent cage. With persistent radicals (R = PhCH2 ,
Ph2CH·, or Ph3C·) coupling to dimers could ensue ef-
·
·
ficiently. With short-lived radicals (R = CH3 and C6H5 )
H-atom abstraction from THF should occur preferentially.
In this regard, note the copious CH4 evolution in run 3 and
the low biphenyl yield (6%) in run 5 (Table 1).
The regiochemistry of such carbolithiations occurring in
a SET fashion could occur most readily in the aforemen-
tioned 1,2-manner and a conformationally conditioned 1,4-
fashion (Scheme 20). It should be borne in mind that the
R· could be monomeric, as may well be the case with
benzyl, as benzyllithium (8) is monomeric in THF. In con-
trast, methyllithium (10) is tetrameric, with the result that
its “R” group may be the tetrameric fragment [Me·-
(LiMe)3], left by the SET transfer of Li· to the diazocine.
Hence, in certain cases, transfer of a bulky “R” group to
radical 2g may determine whether 1,2- or 1,4-addition, or
Scheme 18.
3. The Absence of any 1,2-Bis-Carbolithiation Adduct (i.e.,
31) with an Excess Amount of Benzyllithium (8)
The final thermodynamic product of diazocine 5 with neither, occurs because of overwhelming steric factors. With
any excess amount of benzyllithium 8 would be expected to methyllithium, very little 1,2-addition occurs (29%) com-
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
827

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
pared with dominating SET reduction (71%; Table 1, hydrolyzed adduct that allylmagnesium chloride and benz-
run 3). Moreover, both phenyllithium (THF solvated dimer) yllithium added strictly in an exo,syn manner to the exo-
and n-butyllithium (THF solvated tetramer or dimer)[13] puckered face of the eight-membered ring of the diazocine.
gave no detectable 1,2-adduct but at –78 °C gave good By analogy, it can be inferred that methyllithium adds also
yields of the 1,4-adduct. The regioselectivity for 1,4-ad- in such a 1,2-fashion.
dition can be ascribed to the bulky, associated “R” radicals
involved. Both the solvated dimer or tetramer of nBuLi
present at –78 °C[13] would be too bulky for 1,2-carbolithi-
ation, but the solvated dimer should be more reactive than
(5) The 1,2- or 1,4-regioselectivity of carbolithiating the
lithiated Li–N1–C2 –C3=C4 intermediate appears to depend
·
upon the steric size of the R· or its associated cluster [R–
(RLi)n]·. Monomeric benzyllithium can easily deliver a
benzyl radical to C2, whereas the dimeric butyl cluster [Bu–
(BuLi)]· would find C4 sterically more accessible for adding
the butyl radical. A similar proposal is advanced for the
1,4-addition of the dimeric phenyllithium cluster.
the solvated tetramer in
a 1,4-SET carbolithiation
(Scheme 20). With the complex [2phenyllithium·4THF], a
similar transition state could be proposed (R = nBu, Ph)
(Scheme 20). Given that the 1,4-adduct of diazocine 5 and
phenyllithium (17) underwent loss of LiH upon warming
to 25 °C to produce 9f (Scheme 10), it was not possible to
establish the exo or endo stereochemistry of the 1,4-adduct.
Again, however, the n-butyllithium or phenyllithium dimer
should approach, for steric reasons, diazocine 5 for syn ad-
dition from the exo side.
(6) Coupling of the R radicals resulting from the SET
transannular reduction occurs in high yield only if such rad-
icals are relatively stable (termed persistent). Thus, with
benzyllithium in run 1 of Table 1, 100% of the expected
benzyl radicals was found as bibenzyl. By contrast, only
6% of the phenyl radicals expected from phenyllithium were
detected as biphenyl. Clearly, the reactive localized phenyl
radical has a much higher probability of abstracting a hy-
drogen atom from THF before encountering a phenyl radi-
cal or benzene molecule.
Experimental Section
Starting Reagents and Solvents: n-Butyllithium (16, 2.5 m in hex-
ane), phenyllithium (17, 1.8 m in dibutyl ether), methyllithium (15,
3.0 m in diethoxymethane), allylmagnesium chloride (20, 2.0 m in
THF), tert-butylmagnesium chloride (19, 1.0 m in THF) were used
as received from the Aldrich Chemical Company, with all transfers
being made under a dry, oxygen-free argon atmosphere with argon-
flushed gastight syringes. Likewise, benzyllithium (8), phenyllith-
ium (17) in diethyl ether, and phenylethynyllithium (18) reagents
were prepared, as described in the following paragraphs, transfer-
red, and allowed to react under an argon atmosphere.[25]
Scheme 20.
Conclusions
(1) A range of organolithium reagents in THF, contain-
ing sp2- and sp3-hybridized carbon atoms bonded to lith-
ium, namely, methyllithium, n-butyllithium, phenyllithium,
and benzyllithium, as well as allylmagnesium chloride, ef-
fect the carbometalation of diazocine 5 solely in either a
1,2- or a 1,4-regioselective manner.
(2) Accompanying such successful carbometalations as a
side reaction in greater or lesser proportion is the SET
transannular reduction of 5 to indoloindole 10b, a sign of
the possible role of electron transfer in the carbometal-
ations.
(3) The facility with which phenylethynyllithium acts as
a nucleophilic reagent toward a wide variety of carbonyl
substrates and even highly polar F3C–C=N linkages and
yet fails to add to diazocine 5 or to reduce 5 to 10b implies
that 18 cannot behave as an SET agent. In contrast, the
carbometalating agents in point (1) are, accordingly, in fact
behaving not as nucleophilic reagents but as SET agents in
both the carbometalation and the transannular reduction.
(4) In the successful 1,2-carbometalations of diazocine 5
Benzyllithium (8) was prepared as needed in two separate pro-
cedures: (1) in THF by the cleavage of benzyl methyl ether by lith-
ium metal pieces in THF at –10 °C and then analyzed by the Gil-
man double titration method[26] and (2) in toluene as the 1:1 com-
plex with TMEDA by treatment of TMEDA (1.0 equiv.) in toluene
with n-butyllithium in hexane (1.0 equiv.) and with subsequent
analysis by adduction of an aliquot with benzophenone.[25]
Phenyllithium (17) was prepared by treating a solution of bromo-
benzene (236 mg, 1.5 mmol, 1.3 equiv.) in diethyl ether (5 mL) with
n-butyllithium in hexane (0.46 mL, 1.15 mmol, 1.0 equiv.) at 0 °C
for 30 min to generate a colorless solution of 17. Phenyllithium was
analyzed by adduction of an aliquot with 9-fluorenone (207 mg,
1.15 mmol, 1.0 equiv.).Subsequent hydrolysis gave a yellow solu-
tion. Separating the organic layer after addition of ethyl ether, dry-
ing the organic layer, and removing volatiles gave an organic resi-
due, the 1H NMR spectrum of which was recorded. The relative
integrated intensities of 9-phenyl-9-fluorenol and unreacted 9-
fluorenone showed that 85% of 9-fluorenone had undergone the
it was demonstrated by XRD and spectral studies of the phenyl addition indicating an 85% yield of 17.
828 Eur. J. Org. Chem. 2014, 818–832
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

SET Carbolithiation of Imines
Phenylethynyllithium (18) was prepared by treating a solution of
phenylacetylene (102 mg, 1.0 mmol, 1.0 equiv.) in THF (5 mL) with
n-butyllithium in hexane (0.44 mL, 1.1 mmol, 1.1 equiv.) at –78 °C
for 30 min and then at room temperature for 30 min to give a color-
less solution of 18, which was analyzed by treating an aliquot with
D2O (100%), separating the organic layer after addition of ethyl
ether, drying the organic layer, and removing volatiles and re-
was run in an identical manner to that above and on the same
scale, except that the reaction molar ratio of 5/8 was 1.0:2.2. Al-
though the purpose of this reaction was to achieve the bis-benzyl-
lithiation of both C=N bonds of 5, the surprising result was that
the products consisted of a mixture of 55% (0.46 mmol) of mono-
benzyl adduct 9c, 45% (0.40 mmol) of indolo[3,2,b]indole 10b, and
0.38 mmol of bibenzyl (ratio of 10b:bibenzyl = 1.0:0.95). If the run
of such a reaction at a ratio of 1.0:2.2 was heated for 24 h at reflux
and then worked up after hydrolysis, starting diazocine 5 was con-
verted quantitatively into indolo[3,2,b]indole 10b (0.86 mmol)
(run 1a). In addition, bibenzyl (0.87 mmol) was found as the by-
product (ratio of 10b/bibenzyl = 1.0:1.01).
1
cording the H NMR spectrum of the organic residue. The relative
integrated intensities of the phenyl protons to the methylidyne pro-
tons, which had been 5.0 to 1.0 in the starting phenylacetylene,
were now 5.0 to 0, indicating 100% deuteriation at the CH group
and thus a 100% yield of 18.
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with the
Benzyllithium (8)–TMEDA Complex in Toluene (Reaction Molar
Ratio of 5/8 = 1.0:2.2, 1.1equiv. of TMEDA was added at –78 °C
and another 1.1 equiv. of TMEDA was added at r.t., run 1c): To a
yellow solution of diazocine 5 (309 mg, 0.86 mmol, 1 equiv.) in tol-
uene (15 mL) was added benzyllithium–TMEDA (0.95 mmol,
1.1 equiv.) at –78 °C, which generated a reddish black solution. Af-
ter stirring for 2 h at –78 °C, the mixture was then allowed to come
to room temperature over 12 h. Then, another aliquot of benz-
yllithium–TMEDA (0.95 mmol, 1.1 equiv.) was added to the reac-
tion mixture at room temperature, and the mixture remained a red-
dish black color. Finally, the reaction mixture was heated to reflux
for 12 h to give a black solution. The following hydrolysis and sub-
sequent usual workup with ethyl ether and water gave a slightly
yellow crude product. On the basis of the 1H NMR spectrum of
the crude product, 100% of diazocine 5 was converted into
All reactions were performed under a positive pressure of anhy-
drous, oxygen-free argon. All solvents employed with organometal-
lic compounds were dried and distilled from a sodium metal-benzo-
phenone ketyl mixture prior to use.[25]
Authentic 1H NMR and 13C NMR spectra for comparison with
those of all known compounds listed in the Supporting Infor-
mation are accessible from AIST: Integrated Spectral Database
System of Organic Compounds.[26]
Analytical Methods: The IR spectra were recorded with a Perkin–
Elmer instrument, model 457, and samples were measured either
as mineral oil mulls or as KBr films. The NMR spectra (1H and
13C) were recorded with a Bruker spectrometer, model Avance III
600, and tetramethylsilane was used as the internal standard. The
chemical shifts reported are expressed on the scale in parts per
million (ppm) from the Me4Si reference signal. Melting points were
determined with a Thomas–Hoover Unimelt capillary melting
point apparatus.
4b,9b-diphenyl-4b,5,9b,10-tetrahydroindolo[3,2-b]-indole
(10b,
0.86 mmol). In addition, bibenzyl (0.83 mmol) was found as the
byproduct (molar ratio of indolo[3,2-b]indole 10b and bibenzyl =
1.0:0.96).
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 was then dried
with anhydrous Na2SO4; the volatile solvent was removed and the
1H NMR and 13C NMR spectra and TLC were recorded on the
crude organic products. Where preparative yields were to be cor-
roborated, column chromatographic separation of products on sil-
ica gel with a hexane/ethyl acetate eluent was performed. Run num-
bers refer to the run entries in Table 1.
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Phen-
yllithium (17, Made by Mixing Bromobenzene with nBuLi in Et2O)
in THF (Reaction Molar Ratio of 5/17 = 1.0:2.2, –78 °C to r.t.,
run 5): To a white suspension of phenyllithium (0.97 mmol) in di-
ethyl ether (5.0 mL) at –78 °C was added a solution of diazocine 5
(318 mg, 0.88 mmol) in THF (10 mL), which generated a yellow
solution. After stirring for 2 h at –78 °C, the mixture was then al-
lowed to come to room temperature over 12 h to give a brownish
suspension. The usual workup was performed with ethyl ether and
water. Notably, a white precipitate was found in the reaction mix-
ture and evolution of a clearly flammable gas was observed during
hydrolysis, indicating the formation of LiH, and subsequent hydrol-
ysis then generated H2 gas. On the basis of the 1H NMR spectrum
of the crude product, the composition of the crude product was
10% of 4b,9b-diphenyl-4b,5,9b,10-tetrahydroindolo[3,2-b]indole
(10b) and 84% of 6-(o-biphenylyl)-12-phenyldibenzo[b,f][1,5]di-
azocine (9f). Recrystallization of the crude product from dichloro-
methane gave white crystals of 9f, m.p. 254–256 °C. The phenyl
group was added to the ortho position of the phenyl group on C6
as determined by single-crystal XRD analysis (Figure 3). The 1H
NMR, 13C NMR, and IR spectra are in complete accord with the
structural connectivity of structure of 9f. Finally, as 9f has not pre-
viously been reported, the high-resolution mass spectrum of pro-
tonated 9f was recorded, and its mass of 435.1858 compares favor-
ably with the theoretical mass for protonated 9f of 435.1856 with a
delta of 0.5 ppm and a composition of C32H22N2. The Supporting
Information has reproductions of the 1H NMR, 13C NMR, and
IR spectra, the HRMS recording, and the 3D structure of 9f (Fig-
ure 3, cf. also the Supporting Information).
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with the
Benzyllithium (8)–TMEDA Complex in Toluene (Reaction Molar
Ratio of 5/8 = 1.0:1.1, –78 °C to r.t., run 1): To a yellow suspension
of the benzyllithium–TMEDA complex (1.90 mmol) in toluene/
hexane (6.0 mL) at –78 °C was added a solution of diazocine 5
(620 mg, 1.70 mmol) in THF (5 mL), which generated a reddish
black solution. After stirring for 2 h at –78 °C, the mixture was
then allowed to come to room temperature over 12 h. The usual
workup was performed with ethyl ether and water. On the basis of
the 1H NMR spectrum of the crude product, the composition of
the crude product was 98% of 6-exo-benzyl-6,12-diphenyl-5,6-di-
hydrodibenzo[b,f][1,5]diazocine (9c) and 2% of indolo[3,2-b]indole
10b. Over 99% of diazocine 5 had reacted. Recrystallization of the
crude product from absolute ethanol gave yellow crystals of 9c,
m.p. 209–211 °C. The benzyl group was added exo to the C6 posi-
1
tion as determined by single-crystal XRD analysis. The H NMR,
13C NMR, and IR spectra are in complete accord with the struc-
tural connectivity of structure of 9c. The Supporting Information
1
has reproductions of the H NMR, 13C NMR, and IR spectra; the
HRMS recording; and the 3D structure of 9c (Figure 1).
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with the
Benzyllithium (8)–TMEDA Complex in Toluene (Reaction Molar
Ratio of 5/8 = 1.0:2.2, –78 °C to r.t., run 1b): The foregoing reaction
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Phen-
yllithium (17, Made by Mixing Bromobenzene with nBuLi in Et2O)
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
829

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
in THF (Reaction Molar Ratio of 5/17 = 1.0:2.2, –78 °C to r.t.,
run 5a): To a white suspension of phenyllithium (2.2 mmol) in di-
ethyl ether (10 mL) at –78 °C was added a solution of diazocine 5
and 5, respectively. About 1% of 10b was also present. The struc-
tures of 9e, 9h, and 10b were completely corroborated by the IR,
1H NMR, and 13C NMR spectroscopic data in the Supporting In-
(358 mg, 1.0 mmol) in THF (5 mL), which generated a yellow solu- formation.
tion. After stirring for 2 h at –78 °C, the mixture was then allowed
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Butyl-
to come to room temperature over 12 h to give a brownish solution.
The usual workup was performed with ethyl ether and water. On
the basis of the 1H NMR spectrum of the crude product, the com-
position of the crude product was 70% of 4b,9b-diphenyl-
4b,5,9b,10-tetrahydroindolo[3,2-b]indole (10b), 12% of 6-(o-bi-
phenylyl)-12-phenyldibenzo[b,f][1,5]diazocine (9f), 12% of remain-
ing diazocine 5, and 6% of biphenyl.
lithium (16) in THF (Reaction Molar Ratio of 5/16 = 1.0:2.2, –78 °C
to r.t., run 4b): To a yellow solution of diazocine 5 (950 mg,
2.66 mmol) in THF (40 mL) was added a solution of 2.5 m butyl-
lithium in hexane (2.34 mL, 5.82 mmol) at –78 °C, which generated
an orange solution. After stirring for 3 h at –78 °C, the mixture
was then allowed to come to room temperature over 12 h to give a
black solution. Hydrolysis and stirring of the resulting mixture at
25 °C for 12 h, followed by the usual workup yielded a bright yel-
low crude product. Column chromatographic separation of the
crude product on silica gel with a hexane/ethyl acetate eluent (4:1
volumetric ratio) was performed to give, as the major product, hy-
droperoxide 9h, m.p. 90–98 °C (dec.), in 91% yield. Product 9h was
crystallized from the separated portion in a mixture of hexane and
ethyl acetate (4:1 volumetric ratio). The residual crude product con-
tained 2% of indolo[3,2-b]indole (10b) and 8% of remaining di-
azocine 5.
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Meth-
yllithium (15) in THF (Reaction Molar Ratio of 5:15 = 1.0:1.1,
–78 °C to r.t., run 3a): To a yellow solution of diazocine 5 (358 mg,
1.0 mmol) in THF (15 mL) was added a solution of methyllithium
(10; 3.0 m in diethoxymethane, 0.36 mL, 1.1 mmol) at –78 °C,
which generated a bright yellow solution. After stirring for 2 h at
–78 °C, the mixture was then allowed to come to room temperature
over 12 h to give a black solution. Throughout the reaction, a flam-
mable gas, condensable by liquid nitrogen, evolved. Thus, the gas
was likely CH4 and not H2. The usual workup was performed with
ethyl ether and water. On the basis of the 1H NMR spectrum of
the crude product, the composition of the crude product was 28%
of 4b,9b-diphenyl-4b,5,9b,10-tetrahydroindolo[3,2-b]indole (10b),
17% of 6-exo-methyl-6,12-diphenyl-5,6-dihydrodibenzo[b,f][1,5]di-
azocine (9d), and 55% of remaining diazocine 5. Column chroma-
tographic separation of the crude product on silica gel with a hex-
ane/ethyl acetate eluent (8:1 volume ratio) was performed to give
9d in pure form. Recrystallization of 9d from dichloromethane gave
yellow crystals of 9d, m.p. 179–181 °C.
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Phenyl-
ethynyllithium (18) in THF (Reaction Molar Ratio of 5/18 = 1.0:1.1,
–78 °C to r.t., run 6): To a colorless solution of phenylethynyllith-
ium (1.1 mmol) in THF (10 mL) at –78 °C was added a solution of
diazocine 5 (358 mg, 1.0 mmol) in THF (5 mL). After stirring for
2 h at –78 °C, the mixture was then allowed to come to room tem-
perature over 12 h. The usual workup was performed with ethyl
ether and water. On the basis of the 1H NMR spectrum of the
crude product, no reaction took place.
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Meth-
yllithium (15) in THF (Reaction Molar Ratio of 5/16 = 1.0:2.2,
–78 °C to r.t., run 3): To a yellow solution of diazocine 5 (358 mg,
1.0 mmol) in THF (15 mL) was added a solution of methyllithium
(3.0 m in diethoxymethane, 0.72 mL, 2.2 mmol) at –78 °C, which
generated a bright yellow solution. After stirring for 2 h at –78 °C,
the mixture was then allowed to come to room temperature over
12 h to give a black solution. The usual workup was performed
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Phenyl-
ethynyllithium (18) in THF (Reaction Molar Ratio of 5/18 = 1.0:2.2,
–78 °C to r.t., run 6): To a colorless solution of phenylethynyl-
lithium (2.2 mmol) in THF (10 mL) at –78 °C was added a solution
of diazocine 5 (358 mg, 1.0 mmol) in THF (5 mL). After stirring
for 2 h at –78 °C, the mixture was then allowed to come to room
temperature over 12 h. The usual workup was performed with ethyl
ether and water. On the basis of the 1H NMR spectrum of the
crude product, no reaction took place.
1
with ethyl ether and water. On the basis of the H NMR spectrum
of the crude product, the composition of the crude product was
71% of 4b,9b-diphenyl-4b,5,9b,10-tetrahydroindolo[3,2-b]indole
(10b) and 29% of 6-exo-methyl-6,12-diphenyl-5,6dihydrodi-
benzo[b,f][1,5]diazocine (9d).
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Phenyl-
ethynyllithium (18) in THF (Reaction Molar Ratio of 5/18 = 1.0:1.1,
–78 °C to Reflux, with 1.1 equiv. of HMPA, run 6a): To a colorless
solution of phenylethynyllithium (1.1 mmol) in THF (10 mL) at
–78 °C was added HMPA (1.1 mmol) and then a solution of di-
azocine 5 (358 mg, 1.0 mmol) in THF (5 mL). After stirring for 2 h
at –78 °C, the mixture was heated to reflux at 80 °C for 2 h. The
usual workup was performed with ethyl ether and water. On the
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Butyl-
lithium (16) in THF (Reaction Molar Ratio of 5/16 = 1.0:2.2, –78 °C,
run 4): To a yellow solution of diazocine 5 (179 mg, 0.5 mmol) in
THF (15 mL) was added a solution of 2.5 m butyllithium in hexane
(0.44 mL, 1.1 mmol) at –78 °C, which generated an orange solu-
tion. After stirring for 6 h at –78 °C, the mixture was quenched by
methanol (5 mL), which generated a bright yellow solution. Then,
the reaction mixture was allowed to come to room temperature.
The usual workup was performed with ethyl ether and water. On
the basis of the 1H NMR spectrum of the crude product, the com-
position of the crude product was 74% of pale-yellow 6-
(2Ј-exo-butylcyclohexa-3Ј,6Ј-dienyl)-12-phenyldibenzo[b,f][1,5]di-
azocine (9e) and 26% of 6-(2Ј-exo-butyl-5Ј-hydroperoxycyclohexa-
3Ј,6Ј-dienyl)-12-phenyldibenzo[b,f][1,5]diazocine (9h). Products 9e
and 9h were separated by column chromatography on silica gel with
an ethyl acetate/hexane eluent (1:4 volumetric ratio) as pale-yellow
solids 9e, m.p. 150–151 °C, and 9h, m.p. 90–98 °C (dec.). Their
XRD structures and appropriate data are presented in Figures 4
1
basis of the H NMR spectrum of the crude product, no reaction
took place.
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Phenyl-
ethynyllithium (18) in THF (Reaction Molar Ratio of 5/18 = 1.0:2.2,
–78 °C to Reflux, with 2.2 equiv. of HMPA, run 6a): To a colorless
solution of phenylethynyllithium (2.2 mmol) in THF (10 mL) at
–78 °C was added HMPA (2.2 mmol) and then a solution of di-
azocine 5 (358 mg, 1.0 mmol) in THF (5 mL). After stirring for 2 h
at –78 °C, the mixture was heated to reflux at 80 °C for 2 h. The
usual workup was performed with ethyl ether and water. On the
1
basis of the H NMR spectrum of the crude product, no reaction
took place.
830
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832

SET Carbolithiation of Imines
planar 3 at 20 kcalmol–1. (N. L. Allinger, W. Szkrybalo, M. A.
DaRooge, J. Org. Chem. 1963, 28, 3007–3009). In addition, the
ring-strain energy of producing an all-planar eight-membered
ring was assessed at about 23 kcalmol–1 more. (A. Stre-
itweiser Jr., Molecular Orbital Theory for Organic Chemists,
Wiley, New York, 1962, p. 283); b) up to 81 kcalmol–1, the
Reaction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) with Allyl-
magnesium Chloride (20) in THF (Reaction Molar Ratio of 1/20 =
1.0:2.2, r.t., no light, run 2): To a yellow solution of diazocine 5
(150 mg, 0.42 mmol) in THF (15 mL) in a Schlenk flask was added
a solution of allylmagnesium chloride (2.0 m in THF, 0.46 mL,
0.92 mmol) at r.t., which generated an orange solution. After stir-
ring for 48 h at r.t., the usual workup was performed with ethyl
ether and water. On the basis of the 1H NMR spectrum of the
crude product, the composition of the crude product was 91% of
6-exo-allyl-6,12-diphenyl-5,6-dihydrodibenzo[b,f][1,5]diazocine
(9g), crystals, m.p. 223–225 °C; 1% of 10b from absolute ethanol;
and 8% of remaining diazocine 5. The XRD structure of 9g and
the appropriate data are presented in Figure 2. The Supporting In-
3
mean bond stabilization energy for the fully formed Csp –C
sp3
σ bond between C6 and C12 in resonance structures 6Ǟ7 in
Scheme 2: cf. J. Waser, K. N. Trueblood, C. M. Knobler, Chem.
One, McGraw-Hill, New York, 1976.
a) W. Metlesics, T. Resnick, G. Silverman, R. Tavares, L. H.
Sternbach, J. Med. Chem. 1966, 9, 633–634; b) R. W. Koch,
R. E. Dessy, J. Org. Chem. 1982, 47, 4452–4479.
a) For a review of SET processes occurring in organometallics,
ranging from Schlenk adducts through radical anions involved
in SET processes in addition reactions of RLi and RMgX rea-
gents to carbonyl and azomethine groups, see: J. J. Eisch, Res.
Chem. Intermed. 1996, 22, 145–187; b) for evidence of SET
involvement in addition reactions of allyl Grignard reagents to
azomethine groups in pyridines and quinolines, see: J. J. Eisch,
R. L. Harrell Jr., J. Organomet. Chem. 1970, 21, 21–27; J. J.
Eisch, D. R. Comfort, J. Organomet. Chem. 1972, 43, C17–
C19.
[5]
[6]
1
formation has reproductions of the H NMR, 13C NMR, and IR
spectra; the HRMS recording; and the 3D structure of 9g (Fig-
ure S2).
Reduction of 6,12-Diphenyldibenzo[b,f][1,5]diazocine (5) to 4b,9b-Di-
phenyl-4b,5,9b,10-tetrahydroindolo[3,2-b]indole (10b) with tert-But-
ylmagnesium Chloride (19) in Refluxing Tetrahydrofuran: The reac-
tion of diazocine 5 (1.0 equiv.) with 19 (2.2 equiv.) was performed
at reflux for 12 h, while the reaction apparatus was exposed to ordi-
nary fluorescent laboratory light. After hydrolytic workup, 15% of
10b was formed and 85% of 5 remained. If the foregoing reaction
mixture was completely covered by Al foil, excluding all light, the
same procedure left the reaction mixture with only 5.
[7]
[8]
Ratio of formula equivalents in this report are given as ratio
of molecular formula equivalents of 5 to empirical formula
equivalents of the specific R–Li or R–MgCl.
a) Commercially available tert-butyllithium in pentane could
not be included in this study of the reactions of 5 in ether
media in the temperature range of –78 °C to r.t. because it
readily decomposes such solvents. Diethyl ether is cleaved into
lithium ethoxide and ethylene with the evolution of isobutane.
Furthermore, any unreacted tert-butyllithium readily adds to
ethylene to generate neohexyllithium (P. D. Bartlett, S. Fried-
man, M. Stiles, J. Am. Chem. Soc. 1955, 77, 1771–1772); b)
H. Gilman, J. Eisch, J. Am. Chem. Soc. 1957, 79, 2150–2153.
Moreover, this lithium reagent is most hazardous to handle,
see footnote 12 in ref.[3] for the report of a fatal accident occur-
ring at the University of California at Los Angeles in 2008.
M. Kharasch, O. Reinmuth, Grignard Reactions of Nonmetallic
Substances, Prentice-Hall, New York, 1954, p. 104–109, passim.
W. Zarges, M. Marsch, K. Harms, G. Boche, Chem. Ber. 1989,
122, 2303–2309.
CCDC-928020 (for 9c), -928021 (for 9f), 948256 (for 9e), 948255
(for 9e), and -928022 (for 9g) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.u/data_request/cif.
Supporting Information (see footnote on the first page of this arti-
1
cle): Crystallographic data; copies of the fully displayed H NMR,
13C NMR, and DEPT spectra; copies of the IR spectra; copies of
the high-resolution mass spectra.
[9]
[10]
[11]
[12]
B. Qu, D. B. Collum, J. Org. Chem. 2006, 71, 7117–7119.
M. A. Beswick, D. S. Wright, Alkali Metals, in: Comprehensive
Organometallic Chemistry II (eds.: E. W. Abel, F. G. A. Stone,
G. Wilkinson, C. E. Housecroft), Pergamon Press, 1995, vol. 1,
chapter 1, p. 1–34.
See: I. Keresztes, P. G. Williard, J. Am. Chem. Soc. 2000, 122,
10228–10229 for the use of diffusion-ordered NMR spec-
troscopy to detect the equilibrium between such tetrasolvated
dimeric and tetrasolvated tetrameric nBuLi aggregates in THF
solution, as well as leading references to other NMR spectro-
scopic and cryoscopic studies achieving the same conclusions.
N. D. R. Barnett, R. E. Mulvey, W. Clegg, P. A. O’Neil, J. Am.
Chem. Soc. 1993, 115, 1573–1574.
B. Qu, D. B. Collum, J. Am. Chem. Soc. 2005, 10820–10821.
See commentary in Table 1and specific procedures in the Ex-
perimental Sections for additional variants of reaction runs.
These numerous attempts to obtain suitable crystals of 9d for
XRD analysis were performed in the laboratories of A. L. R.
in a special microscale apparatus in which any crystals formed
could be inspected under a microscope.
K. Ziegler, H. Zeiser, Justus Liebigs Ann. Chem. 1931, 485, 174.
J. J. Eisch, K. Yu, A. L. Rheingold, definitive research in pro-
gress.
Acknowledgments
Our continuing studies on cyclic diimines have been partly sup-
ported by a grant from Dr. John M. Birmingham of the Boulder
Scientific Company, Mead, Colorado. The 1H NMR and 13C NMR
spectra reported here were recorded at the regional NMR Facility
at Binghamton University with 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.
[13]
[14]
[15]
[16]
[17]
[1] J. J. Eisch, K. Yu, A. L. Rheingold, Eur. J. Org. Chem. 2012,
3165–3171.
[2] J. J. Eisch, T. Y. Chan, J. N. Gitua, Eur. J. Org. Chem. 2008,
[18]
[19]
392–397.
[3] J. J. Eisch, R. N. Manchanayakage, A. L. Rheingold, Org. Lett.
2009, 11, 4060–4063.
[4] a) The necessary planarity of 1 required for Hückel aromaticity
faces two energy barriers. Previous research on the barrier to
inversion of the optically active derivative of 1 bearing carboxyl
groups at the 3- and 10-positions estimated the minimum re-
pulsion energy of such ortho H groups in a transition of 1 to
[20]
[21]
O. F. Foote, D. W. Knight, A. C. L. Low, Y. F. Li, Tetrahedron
Lett. 2007, 8, 647–650.
M. N. Abualhasan, J. A. D. Good, K. Wittayanarakul, N. C.
Anthony, G. Berretta, O. Rath, F. Kozielski, O. b. Sutcliff, S. P.
Mackay, Eur. J. Med. Chem. 1963, 54, 483–498.
Eur. J. Org. Chem. 2014, 818–832
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
831

J. J. Eisch, K. Yu, A. L. Rheingold
FULL PAPER
[22] M. J. Chae, A. R. Jeon, T. Livinghouse, D. K. An, Chem. Com-
mun. 2011, 47, 3281–3283.
[23] T. F. Briggs, M. D. Winemiller, D. B. Collum, R. L. Parsons Jr.,
A. H. Davulcu, G. D. Harris, J. M. Fortunak, P. N. Confalone,
J. Am. Chem. Soc. 2004, 126, 5427–5435.
[24] T. H. Lowry, K. S. Richardson, Mechanism and Theory in Or-
ganic Chemistry, 3rd ed., Harper & Row, New York, 1987, p.
161 ff.
[25] General procedures for purification of solvents and conducting
organometallic reactions under an inert atmosphere are avail-
able: J. J. Eisch, Organomet. Synth. 1981, 2, 7–25, 94–96.
[26] National Institute of Advanced Industrial Science and Tech-
nology; website: http://riodb01.ibase.aist.go.jp/sdbs/cgi-bin/
cre_index.cgi.
Received: August 5, 2013
Published Online: November 26, 2013
832
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 818–832