High-Yield Lithiation of Azobenzenes by Tin-Lithium Exchange

Article abstract of DOI:10.1002/chem.201500003

The lithiation of halogenated azobenzenes by halogen-lithium exchange commonly leads to substantial degradation of the azo group to give hydrazine derivatives besides the desired aryl lithium species. Yields of quenching reactions with electrophiles are therefore low. This work shows that a transmetalation reaction of easily accessible stannylated azobenzenes with methyllithium leads to a near-quantitative lithiation of azobenzenes in para, meta, and ortho positions. To investigate the scope of the reaction, various lithiated azobenzenes were quenched with a variety of electrophiles. Furthermore, mechanistic ¹¹⁹Sn NMR spectroscopic studies on the formation of lithiated azobenzenes are presented. A tin ate complex of the azobenzene was detected and characterized at low temperature.

Full text of DOI:10.1002/chem.201500003

DOI: 10.1002/chem.201500003
Full Paper
&
Azo Compounds
High-Yield Lithiation of Azobenzenes by Tin–Lithium Exchange
Jan Strueben,[a] Matthias Lipfert,[a] Jan-Ole Springer,[a] Colin A. Gould,[a] Paul J. Gates,[b]
Frank D. Sçnnichsen,[a] and Anne Staubitz*[a]
Abstract: The lithiation of halogenated azobenzenes by hal-
ogen–lithium exchange commonly leads to substantial deg-
radation of the azo group to give hydrazine derivatives be-
sides the desired aryl lithium species. Yields of quenching re-
actions with electrophiles are therefore low. This work shows
that a transmetalation reaction of easily accessible stannylat-
ed azobenzenes with methyllithium leads to a near-quantita-
tive lithiation of azobenzenes in para, meta, and ortho posi-
tions. To investigate the scope of the reaction, various lithiat-
ed azobenzenes were quenched with a variety of electro-
philes. Furthermore, mechanistic 119Sn NMR spectroscopic
studies on the formation of lithiated azobenzenes are pre-
sented. A tin ate complex of the azobenzene was detected
and characterized at low temperature.
Introduction
cially selective, as long as the starting material does not con-
tain competing halide functional groups.[20] However, if the
molecule contains other electrophilic groups, these are often
attacked by the lithiating agent.[21]
Azobenzene derivatives are important compounds in many
fields of research, for example, dyes,[1,2] and photoswitchable
systems. On irradiation with ultraviolet light, they undergo re-
versible photoisomerization from their trans form to their cis
form, which results in substantial changes in their electronic
and geometric properties.[3,4] Furthermore, azobenzenes show
a high resistance to photobleaching[5] and thermal decomposi-
tion.[6] These properties have led to wide use of azobenzenes
in biochemical research,[7,8] for example, in photoswitchable
cell adhesion,[9,10] in medical research, for example, photo-
switchable contrast media for magnetic resonance imaging,[11]
and in materials and polymer science, for example, photores-
ponsive polymer materials.[12–18] To tailor the properties of azo-
benzenes exactly to their intended application, flexible and ef-
fective methodologies for the functionalization of azobenzenes
are essential.
For the lithiation of azobenzenes, reported protocols all
employ a halogen–lithium exchange reaction.[22–27] Typically,
yields are quite low (max. 53%[22] for the para position, 47%
for the ortho position, and 54%[28] for the meta position). The
reason for these poor yields is the low tolerance of the azo
group towards reductive conditions: in competition with halo-
gen–lithium exchange, nucleophilic attack by the alkyl lithium
species on the azo group leads to hydrazine analogues of the
starting material 3 and 4 and, after quenching with trimethyl-
silyl bromide (TMSBr), 6 and 7 (Scheme 1).[29–33]
For the functionalization of aromatic rings in general,
a common synthetic strategy is to convert them to organolithi-
um or organomagnesium species and subsequently quench
the reaction with an appropriate electrophile.[19] For these sys-
tems, metalation is usually achieved in situ by halogen–metal
exchange, direct insertion into a carbon–halogen bond, or by
deprotonation. Halogen–metal exchange reactions are espe-
[a] J. Strueben, M. Lipfert,+ J.-O. Springer,+ C. A. Gould,
Prof. Dr. F. D. Sçnnichsen, Prof. Dr. A. Staubitz
Otto-Diels-Institute for Organic Chemistry, University of Kiel
Otto-Hahn-Platz 3-4, 24098 Kiel (Germany)
E-mail: astaubitz@oc.uni-kiel.de
[b] Dr. P. J. Gates
School of Chemistry, University of Bristol
Cantock’s Close, Bristol BS71TS (UK)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500003.
Scheme 1. Lithiation of para-iodoazobenzene and subsequent quenching
with TMSBr.[22]
Chem. Eur. J. 2015, 21, 11165 – 11173
11165
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
For a successful selective lithiation of azobenzene, the lithia-
tion reaction on the aromatic ring must be kinetically preferred
over attack on the diazo group: the desired aryl lithium species
should be much less nucleophilic than an alkyl lithium reagent.
We reasoned that if a tin–lithium exchange reaction on a stan-
nylated azobenzene of type 9 were faster than the attack of
the lithiating reagent on the diazo group, then selective lithia-
tion should be possible. We recently demonstrated the stanny-
lation of azobenzenes by a Stille–Kelly cross-coupling reaction
of the corresponding iodinated azobenzenes 8, catalyzed by
[Pd(PPh3)4], with hexamethyldistannane as the nucleophilic
component (Scheme 2).[34]
Herein we demonstrate that a transmetalation from trime-
thyltin to lithium leads to rapid and effective lithiation of azo-
benzenes in the para, meta, and ortho positions without any
decomposition of the azo group. This protocol increases the
yield of the lithiated species substantially and allows easy
access to compounds that are difficult to obtain by other
means.
Scheme 3. Proposed lithiation of 4-methyl-4’-trimethyl-stannylazobenzene
(11 a) and subsequent quenching with methyl iodide.
Methyllithium was chosen as a lithiating agent because of
its high nucleophilicity. A further advantage is its high stability
in THF compared to butyllithium reagents,[44,45] which have
a short half-life in THF of 107 min at 208C.[44,46] The reaction
conditions were optimized with respect to solvent, tempera-
ture, and lithiation time (Table 1). To analyze the efficiency of
the lithiation, methyl iodide was used as an effective electro-
philic quenching reagent, and the conversion was determined
by quantitative GC analysis (for details, see the Supporting In-
formation). All other possible side products were expected to
be chemically and thermally stable and thus suitable for GC
and GC-MS analysis. To establish reaction conditions for a selec-
tive transmetalation from tin to lithium, a very low tempera-
ture of À1308C was initially chosen. 2-Methyltetrahydrofuran
(MeTHF) was selected as solvent because it has a lower freez-
ing point than THF (À136 vs. À1088C) while maintaining com-
parable properties of activating organolithium reagents
through deaggregation.[47,48]
Scheme 2. Stannylation of mono-iodinated azobenzenes. Compounds 8a–
c,[35] 9a,[35] 9b,c,[36] 10a,[37] and 11 a[34] have been reported previously; the
other stannylated azobenzenes are described in this work.
Results and Discussion
Reaction optimization
After 4 min of lithiation at À1308C, the reaction showed no
side products but a conversion of only 72% (Table 1, entry 1).
When the temperature was raised to À788C, the transmetala-
tion led to full conversion to the desired product without ob-
servable side reactions (Figure 1). The same was true for THF
(Table 1, entries 2, 3). To evaluate the stability of the lithiated
intermediate, the total lithiation time was increased to 30 min,
but no side products were detected (Table 1, entry 4). In the
temperature range from À100 to À438C in THF, full conversion
to the desired product could be observed. At À168C, however,
the reaction produced various side products. Although it was
not possible to isolate these side products individually in pure
form and to quantify them, they could be identified as hydra-
zine derivatives by GC-MS (Table 1, entry 5).[49]
The three most important factors that influence the selectivity
of the two types of competing reactions (i.e., attack of the or-
ganolithium reagent on the azo group vs. tin–lithium ex-
change) are first the reactivity of the functional group that
should be exchanged with lithium (halogen vs. trialkyl tin), sec-
ond the temperature, and third the reactivity of organolithium
complexes in the reaction solvent. We hypothesized that it
should be possible to find reaction conditions under which the
formation of a lithium stannate complex 12 and subsequent
rearrangement to the lithiated azobenzene 13 would be fast
compared to nucleophilic attack of the alkyl lithium reagents
on the azo group. As a large group 14 element, tin is able to
react with nucleophiles by expanding its coordination sphere
to give a pentacoordinate reactive ate complex. Such inter-
mediates have been described before,[38,43] and, in one in-
stance, have even been isolated.[43] Lithium stannates typically
have lower reactivity towards electrophiles than other organo-
lithium reagents.[40,46] The resulting lithiated azobenzene
should have a significantly reduced nucleophilicity compared
to the alkyl lithium species and would not be capable of at-
tacking the azo group (Scheme 3).
In contrast to MeTHF and THF as solvents, the reaction pro-
ceeded entirely differently in diethyl ether. Under the same re-
action conditions as for THF at À1008C and À788C, the yields
and the conversion were much lower (29 and 32% yield, re-
spectively; Table 1, entries 7 and 8). The gas chromatograms of
the reactions showed that the lithiation in diethyl ether pro-
duced various side products (Figure 1). To ensure correct iden-
tification of the main side products, the lithiation reaction in
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11166
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. GC optimization of the lithiation of 4-ethyl-4’-trimethylstannylazobenzene (11 a).
Entry
Solvent
MeTHF
T [8C]
R1
R2Li
MeLi
t [min]
Starting material [%]
Yield of 14a [%]
1
2
3
4
5
6
7
8
9
10
À130
À78
À78
À78
À16
À78
À100
À78
À78
À78
SnMe3
SnMe3
SnMe3
SnMe3
SnMe3
SnMe3
SnMe3
SnMe3
I
4
4
4
30
4
4
4
4
4
30
25
<1
<1
<1
<1
<1
<1
<1
22[a]
<1
72
>99
>99
98
63
43
29
32
54
27
MeTHF
THF
THF
THF
THF
diethyl ether
diethyl ether
THF
MeLi
MeLi
MeLi
MeLi
nBuLi[b]
MeLi
MeLi
MeLi
nBuLi
THF
I
[a] Isolated from the reaction mixture, no GC quantification (see the Supporting Information).[b] Reverse order of addition; see text for details.
available and less expensive. Because of the low stability of
this reagent in THF, the reaction was performed in reverse
order by adding a solution of 4-methyl-4’-trimethylstannylazo-
benzene to a solution of n-butyllithium in THF at À788C.[52]
This reaction led to full conversion of the starting material, but
unselective transmetalation of the trimethylstannyl group was
observed. GC-MS analysis indicated butylation of the azo
group, corresponding to the side reaction with methyllithium
in diethyl ether and the hydrazine synthesis as described by
Katritzky and co-workers[30] (Table 1, entry 6; Scheme 4).
To provide a direct comparison between the tin–lithium ex-
change reaction and halogen–lithium exchange reactions,
which are standard in the literature, the reaction was also per-
formed with 4-iodo-4’-methylazobenzene. On adding methyl-
Figure 1. Gas chromatogram of the reaction mixtures after the lithiation of
lithium or butyllithium to the iodinated azobenzene 10, dis-
solved in THF, the reaction was incomplete and showed vari-
ous side products (Table 1, entries 9, 10).
4-methyl-4’-trimethylstannyl azobenzene (11 a) in THF (Table 1, entry 3;
green) and diethyl ether (Table 2, entry 8; blue). Signal a) triisopropylben-
zene, b) 4-methylazobenzene, c) hydrazine 16, and d) product 14a.
diethyl ether was performed on a larger scale to isolate the
side products. As quenching reagent, methyl iodide was used.
Besides the intended product 14a, the main side product in
1
the crude H NMR spectrum could be identified as the hydra-
zine species 15. Subsequent column chromatography gave hy-
drolyzed[50] species 16 in a purity of 90% and a yield of 34% of
isolated product.[51]
The different reaction kinetics were also visible to the naked
eye: In THF and MeTHF, the reaction mixture turned black im-
mediately after adding the methyllithium solution in THF, indi-
cating formation of an aryl lithium species. In diethyl ether, on
the other hand, no significant color change was observed at
À100 or À788C. Only when the reaction mixture was allowed
to warm to temperatures higher than À208C did an intense
color change to black occur. Because the absence of byprod-
ucts in THF at temperatures as high as À438C was striking, the
tin–lithium exchange reaction was performed with a solution
of n-butyllithium in hexanes. n-Butyllithium is more commonly
Scheme 4. Lithiation reaction in diethyl ether, followed by quenching with
methyl iodide.
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11167
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
NMR studies
One showed a signal at d=0 ppm, which is consistent with
tetramethylstannane.[56] A further 119Sn NMR signal was visible
at d=À312 ppm. This signal was assigned to ate complex 12,
on the basis of a literature precedent.[41,57] At À1038C, the rela-
tive intensity of the signal for the ate complex remained con-
stant, whereas the signal due to the starting material contin-
ued to disappear, and tetramethylstannane formed at the
same rate (based on the percentage change of the integral of
each species). The signal corresponding to the tin ate complex
could be observed at temperatures of up to À888C, but when
the reaction mixture was warmed further to À708C, the signal
could no longer be observed. After warming up to À608C, the
starting material had entirely disappeared, and only the signal
for tetramethylstannane was visible (Figure 2 and Figure SI-5 in
the Supporting Information).
To support the mechanistic hypothesis for the reaction and to
investigate the significantly lower selectivity in diethyl ether, in
situ NMR experiments were performed. As the reaction prog-
ress in MeTHF at À1308C was slow, those conditions were
chosen to mix the reactants in a NMR tube and immediately
freeze the NMR tube with liquid nitrogen in the glovebox. In
this way, any reaction progress could be avoided before insert-
ing the NMR tube into the precooled NMR spectrometer (see
the Supporting Information for a detailed protocol). The reac-
tions were then monitored by 119Sn NMR spectroscopy because
the fast relaxation behavior of this nucleus allowed efficient
sampling of the experiments. A fast scanning method was
needed because of the low concentration of the ate complex
in the reaction mixture. Because of the different relaxation be-
haviors of symmetric and asymmetric tin species,[53] the param-
eters for the asymmetric, fast-relaxing species could be opti-
mized (prescan delay of 30 ms, relaxation delay of 1 ms). This
allowed fast recording of 3000–6000 scans with an overall re-
cording time of 132/264 s per spectrum. Therefore, integration
of the signals could not be correlated to the concentration of
the species observed in solution, and only qualitative state-
ments can be made.[54]
The transmetalation reactions were recorded in a similar
manner for the corresponding meta- and ortho-substituted
species. 4-Methyl-3’-trimethylstannylazobenzene (11 b) showed
a chemical shift of À27 ppm at À1038C. As for the para conge-
ner, two new species appeared at d=0[58] and À313 ppm, cor-
responding to tetramethylstannane and a lithium–tin ate com-
plex (12b). The transmetalation reaction in the meta position
was very similar to the reaction of 4-methyl-4’-trimethylazo-
benzene (11 a), but appeared to be faster. The reaction was al-
ready completed at À788C (Figure SI-6 in the Supporting Infor-
mation). In the reaction of 4-methyl-2’-trimethylstannylazoben-
zene (11 c), the starting material showed a chemical shift of
À43 ppm. Although the signal corresponding to tetramethyl-
stannane at d=0 ppm appeared, no signal for the tin ate com-
plex was detectable in the temperature range of À103 to
The spectra were initially recorded at À1038C[55] and then
the temperature was raised stepwise by ca. 108C while moni-
toring the progress of the reaction. First the reaction of para-
stannylated azobenzene 11 a in MeTHF was investigated. At
À1038C, the starting material 11 a showed a chemical shift of
À27 ppm. After 4 min, two new species appeared (Figure 2).
Figure 2. In MeTHF: 119Sn NMR spectrum for the reaction of para-stannylated azobenzene 11 a at À1038C and t=0 min (blue). Reaction progress and the for-
mation of the tin ate complex could be observed at À1038C (t=4 min; red). At À608C the reaction was finished and only the signal of tetramethylstannane
was visible (green). In diethyl ether: The starting material at À808C (yellow); at À158C the reaction started (gray) and at 278C the reaction was finished, and
tetramethylstannane and the side product were visible (purple). *=First spectrum of at this temperature; the average heating rate of the NMR experiments
was 0.28CminÀ1
.
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11168
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
À588C. At this temperature the reactions of the para and meta
isomers 11 a and 11 b were almost complete, whereas that at
the ortho position had only begun to take place. However, like
for the reactions in the para and meta positions, only tetrame-
thylstannane could be observed as the final tin-containing
product, which indicates selective tin–lithium exchange. In di-
ethyl ether, a completely different reaction progress was ob-
served. On treatment of 4-methyl-4’-trimethylstannylazoben-
zene (11 a) with methyllithium at À1038C, the starting material
could be observed at d=27 ppm. However, after the reaction
mixture had melted at À1038C, neither a signal for tetrame-
thylstannane nor a signal for the ate complex could be ob-
served. Only after warming to À448C did a signal at d=
0 ppm, assigned to tetramethylstannane, became visible. No
further changes were observed up to À158C. At this tempera-
ture, however, the spectrum of the reaction mixture showed
depend on successful lithiation: the electrophile used for
quenching is also an important factor. To show that our newly
developed methodology for the lithiation of azobenzenes is of
general practical use, a variety of quenching agents were used
for 4-methyl-4’-trimethylstannylazobenzene (11 a) and 4-
methyl-3’-trimethylstannylazobenzene (11 b; Table 2). Quench-
ing with trimethylsilyl chloride resulted in yields of 94% for
17a and 95% for 17b (Table 2, entry 2). This compares favora-
bly to previous protocols for silylation of azobenzenes via halo-
gen–lithium exchange: the yields reported for these reactions
average around 20%[65] to a maximum yield of 42%[12] with
a trimethylsilyl halogenide electrophile.
Lithiation reactions are also often used for the introduction
of functional groups such as alcohol, carbonyl, or amide. The
quenching reaction with acetone gave alcohol 18a with
a yield of 89% and 18b in a slightly lower yield of 83%
(Table 2, entry 3). The reaction of tolualdehyde gave 19a and
19b in excellent yields ranging from 89 to 95% (Table 2,
entry 4). Aldehyde functionalization by quenching with N,N-di-
methylformamide gave 20a in a yield of 89% and meta com-
pound 20b in a yield of 83%.(Table 2, entry 5). A further classi-
cal methodology of synthesizing ketones from organolithium
reagents is the use of Weinreb’s amide (21).[66] In reaction with
the lithiated azobenzenes, it provided ketones 22a and 22b in
yields of up to 98% (Table 2, entry 6).
a
third, heretofore-unobserved signal at d=À36 ppm
(Figure 2). At 278C, the starting materials had disappeared and
only tetramethylstannane and this new stannyl species could
be observed. The chemical shift of this signal indicated an aryl-
stannylated compound. The reaction in the NMR tube was
eventually quenched with methyl iodide. Analysis by GC-MS in-
dicated the same side products as described in Figure 1. On
the basis of this quenching experiment in combination with
the isolated byproducts (Scheme 4), the signal occurring at d=
36 ppm must be the N-lithium salt of the hydrazine derivatives,
which is methylated on quenching. The different reactivity of
the tin–lithium exchange in diethyl ether and THF might be ex-
plained by different complexation of the lithiation agents and
the different reactivity of organolithium compounds in THF/
MeTHF and diethyl ether, although the situation is complex.
Neither diethyl ether nor THF (or even hexamethylphosphora-
mide) is a sufficiently strong Lewis base to deaggregate the
(MeLi)4 complex that has been shown to form in THF.[59–62]
Whereas THF has a significantly stronger deaggregation effect
on organolithium compounds such as phenyl lithium than di-
ethyl ether,[63] it must be assumed that, in the case of tin–lithi-
um exchange, methyllithium reacts in an aggregated form.[64]
More detailed DFT studies on the nature of the reaction and
the observed position-dependent differences in the reaction
rates in diethyl ether and TH, are underway. This activation of
the organolithium reagents enables reactions at much lower
temperatures compared to diethyl ether. The reactions in
MeTHF were already complete at À738C, whereas those in di-
ethyl ether only started at À158C. However, at this higher tem-
perature, the greater reactivity of all reactants led to decreased
selectivity, because alkylation/metalation of the azo group can
compete with transmetalation of the trimethylstannyl group.
Amides and thioamides are typically synthesized from aryl
lithium species by using phenyl isocyanate or phenyl isothio-
cyanate. These compounds were also appropriate electrophiles
and gave yields of 77 and 76%, respectively, for the para-lithi-
ated species and 80 and 81%, respectively, for the meta-lithiat-
ed species (Table 2, entries 7 and 8). Whereas simple electro-
philes can only react once, other electrophiles are capable of
multiple reactions. Such electrophiles are very attractive for
the synthesis of dyes with multiple chromophoric groups. Vari-
ous methane derivatives with two azobenzene units on one
carbon atom have been reported in the literature.[67–71] Typical-
ly, these compounds are synthesized from diamine precursors
and by subsequent azo coupling.[67,69,71] With benzoyl chloride
as electrophile, two equivalents of the lithiated azobenzene
could be added to the electrophile, whereby 96% of 25a and
71% of 25b could be obtained (Table 3, entry 1). Products 26a
and 26b could be obtained by using an analogous ester that
gave the products in both the para and meta positions in 79%
yield (Table 3, entry 2).
Only two protocols for the synthesis of triply azobenzene
functionalized methane derivatives have been reported to
date. In one procedure, the product was obtained by a conden-
sation reaction of nitrosobenzene with tris(4-aminophenyl)me-
thane in a yield of 35%, followed by oxidation of the corre-
sponding (triphenylazo)triphenylmethanol, an analogue of
27a, in a yield of 30% (which corresponds to an overall yield
of 10%).[67] The second protocol entails condensation of an al-
dehyde-functionalized azobenzene with unsubstituted azoben-
zenes by electrophilic substitution in sulfuric acid under very
harsh conditions. No yield was reported.[68]
Synthetic scope
The reaction of 4-lithio-4’-methylstannylazobenzene (13a) with
methyl iodide, as used in the optimization reactions, gave an
excellent yield of 96% of isolated 14a. The reaction conditions
were transferred to 4-methyl-3’-trimethylstannylazobenzene
(11 b) and 14b could be isolated in a comparable yield of 95%
(Table 2, entry 1). However, high conversion does not only
When three equivalents of 4-methyl-4’-trimethylstannyl azo-
benzene were lithiated and quenched with one equivalent of
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11169
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. Range of electrophiles employed for the reaction with the lithiated azobenzenes and the resulting products.
Entry
1
E
Product
Yield [%]
96
Product
Yield [%]
95
2
3
94
89
95
83
4
95
89
5
6
89
98
83
71
7
8
77
80
76
81
diethyl carbonate, the main products were triply azobenzene
functionalized methanol derivative 27a in 52% yield and
doubly azobenzene functionalized ketone 28a in 35% yield
(Table 3, entry 3a). Quenching the same reaction with one
equivalent of oxalyl chloride gave triply azobenzene sub-
stituted methanol 27a in 54% yield and ketone 28a in 23%
yield (Table 3, entry 4). The meta-substituted azobenzene ana-
logues could be obtained in 49% yield for trisubstituted meth-
anol 27b and 38% for disubstituted ketone 28b (Table 3,
entry 3b).
trophiles, we suspect that this might be different for the ortho-
lithiated azobenzene. Steric hindrance by the azo group could
affect the formation of the analogous products. On the other
hand, the lithiated species might be particularly stable due to
additional complexation of the lithium atom by the lone pair
of the adjacent nitrogen atom. In the literature, ortho lithia-
tions typically give higher yields than a lithiation in meta or
para position due to this effect.[33,71–73] When the ortho-lithiated
azobenzene was quenched with methyl iodide, the yield of iso-
lated ortho-methylated azobenzene derivative 11 c was almost
quantitative (99%). Even the more sterically hindered trime-
thylsilyl electrophiles gave silylated product 17c in a good
Although the para- and ortho-lithiated azobenzenes did not
show a significant difference in their reactivities towards elec-
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11170
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Conclusion
Table 3. Products with multiple azobenzene moieties. For structures of 25–
28, see Scheme 5.
The transmetalation of stannylated azobenzenes with meth-
yllithium in THF led to rapid and selective exchange of the
trimethylstannyl group without any attack on the azo
group. NMR spectroscopy showed that the reaction pro-
ceeded via a lithium–tin ate complex prior to rearrange-
ment to the lithiated species. Analysis by GC and NMR spec-
troscopy showed that the reaction was rapid and selective
in THF and MeTHF, but unselective in diethyl ether, in which
hydrazine derivatives were formed due to attack of the or-
ganolithium reagent on the diazo group. Furthermore, opti-
mized reaction conditions for performing the lithiation reac-
tion in THF at À788C could be applied for a high-yield gen-
eral method for lithiating azobenzenes in ortho, meta, and
para positions. The lithiated azobenzenes were quenched
with a wide variety of electrophiles. Products containing
one azobenzene unit could be obtained in yields ranging
from 71 to 98%. Quenching the lithiated azobenzenes with
electrophiles with functional groups of multiple reactivity al-
lowed the synthesis of doubly and even triply azobenzene
functionalized compounds in good yields.
Entry
1
E
Products
25a
Yield [%]
96
71
25b
26a
26b
79
79
2
27a
28a
27b
28b
27a
28a
52
35
49
38
54
23
3a
3b
4
Scheme 5. Products with multiple azobenzene moieties. For yields, see Table 3.
yield of 82% (Scheme 6). For further ortho products, see the
Experimental Section
Supporting Information.
Representative lithiation procedure for 4,4’-dimethylazobenzene
(14a): A solution of methyllithium (0.99 equiv, 440 mL, 700 mmol,
1.58m in diethyl ether) diluted in THF (1.58 mL) was added to a
solution of 4-methyl-4’-(trimethylstannyl)azobenzene (1.00 equiv,
Scheme 6. ortho-Functionalized azobenzenes.
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11171
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[26] T. Kozlecki, L. Syper, K. A. Wilk, Synthesis 1997, 681.
[27] M. Unno, K. Kakiage, M. Yamamura, T. Kogure, T. Kyomen, M. Hanaya,
Appl. Organomet. Chem. 2010, 24, 247.
250 mg, 700 mmol) in THF (10 mL) at À788C over 4 min. After
2 min, a solution of methyl iodide (1.10 equiv, 50.0 mL, 800 mmol) in
THF (0.32 mL) was added and the reaction mixture was warmed to
208C. After 6 h of stirring the solvent was evaporated and the resi-
due purified by column chromatography with CH2Cl2 (Rf =0.7) as
eluent. The solvent was evaporated and yellow crystalline needles
(140 mg, 666 mmol, 96%) were obtained. M.p. 1458C; 1H NMR
(500 MHz, CDCl3): d=7.82 (d, 3J=8.4 Hz, 4H, H-2), 7.32 (d, 3J=
8.4 Hz, 4H, H-3), 2.44 ppm (s, 6H, H-5); 13C NMR (125 MHz, CDCl3):
d=150.8 (C-1), 141.2 (C-4), 129.7 (C-2), 122.7 (C-3), 21.5 ppm (C-5);
[28]
M. D. Segarra-Maset, P. W. N. M. van Leeuwen, Z. Freixa, Eur. J. Inorg.
Chem. 2010, 2075.
[29] Y. Zhang, T. Qiang, M. Luo, Org. Biomol. Chem. 2011, 9, 4977.
[30] A. R. Katritzky, J. Wu, S. V. Verin, Synthesis 1995, 651.
[31] E. Ciganek, Organic Reactions,$Wiley, Hoboken, 2004.
[32] T. T. T. Nguyen, A. Boussonni›re, E. Banaszak, A.-S. Castanet, K. P. P.
Nguyen, J. Mortier, J. Org. Chem. 2014, 79, 2775.
[33] Compound 5 was also synthesized by our method in 93% yield (com-
pound SI-5 in the Supporting Information).
[34] J. Strueben, P. J. Gates, A. Staubitz, J. Org. Chem. 2014, 79, 1719.
[35] B.-C. Yu, Y. Shirai, J. M. Tour, Tetrahedron 2006, 62, 10303.
[36] J. Yamamoto, R. Yamawakiy, Y. Sumi, R. Okamoto, M. Saito, M. Kato, A.
Shibata, Nippon Kagaku Kaishi 1992, 1508.
~
IR (ATR): n=3023 (w), 2921 (w), 1601 (m), 1502 (m), 1306 (w), 1237
(w), 1149 (m), 1012 (m), 883 (m), 839 (m), 821 (s), 794 (m), 710 (m),
683 (s), 535 (m), 504 (s) 492 cmÀ1 (s); HRMS (EI sector): m/z calcd
for C14H14N2+: 210.1157; found: 210.1156 [M]+.
[37] D. Qiu, H. Meng, L. Jin, S. Wang, S. Tang, X. Wang, F. Mo, Y. Zhang, J.
Wang, Angew. Chem. Int. Ed. 2013, 52, 11581; Angew. Chem. 2013, 125,
11795. In this report, the compound was synthesized by a Sandmeyer-
type reaction. In this work, it was synthesized analogously to ref. [34].
[38] A. Maercker, H. Bodenstedt, L. Brandsma, Angew. Chem. 1992, 104,
1387.
[39] M. Saito, S. Imaizumi, T. Tajima, Eur. J. Inorg. Chem. 2010, 2153.
[40] N. H. Phillips, H. J. Reich, Pure Appl. Chem. 1987, 59, 1021.
[41] H. J. Reich, N. H. Phillips, J. Am. Chem. Soc. 1986, 108, 2102.
[42] H. J. Reich, J. P. Borst, J. Am. Chem. Soc. 1991, 113, 1835.
[43] M. Saito, S. Imaizumi, T. Tajima, K. Ishimura, S. Nagase, J. Am. Chem. Soc.
2007, 129, 10974.
Acknowledgements
This project was supported by the Special Research Area 677
“Function by Switching” of the Deutsche Forschungsgemein-
schaft (DFG), Project C10. The authors thank Hans J. Reich, Uni-
versity of Wisconsin, for helpful discussions.
Keywords: azo compounds · lithiation · metalation · synthetic
methods · tin
[44] T. L. Rathman, J. A. Schwindeman, Org. Process Res. Dev. 2014, 18, 1192.
[45] H. Gilman, B. J. Gaj, J. Org. Chem. 1957, 22, 1165.
[46] P. Stanetty, M. D. Mihovilovic, J. Org. Chem. 1997, 62, 1514.
[47] J. Clayden, Organolithiums: Selectivity for Synthesis, Pergamon, Amster-
dam, 2003.
[1] K. L. Hunger, Industrial Dyes: Chemistry, Properties, Applications, Wiley-
VCH, Weinheim, 2003.
[2] B. L. Feringa, W. R. Browne, Molecular Switches, 2nd ed., Wiley-VCH,
Weinheim, 2011.
[48] H. J. Reich, Chem. Rev. 2013, 113, 7130.
[49] NMR spectroscopic analysis of a fraction from an attempt to purify
these compounds by column chromatography lent further support to
their identification (Scheme 4 and Supporting Information).
[50] Protodestannylation of aromatic trimethyltin species is common: J. Lin-
shoeft, A. C. J. Heinrich, S. A. W. Segler, P. J. Gates, A. Staubitz, Org. Lett.
2012, 14, 5644.
[3] E. W.-G. Diau, J. Phys. Chem. A 2004, 108, 950.
[4] N. Siampiringue, G. Guyot, S. Monti, P. J. Bortolus, Photochemistry 1987,
37, 185.
[5] S. K. Yesodha, C. K. S. Pillai, N. Tsutsumi, Prog. Polym. Sci. 2004, 29, 45.
[6] C. Lagrasta, I. R. Bellobono, M. J. Bonardi, Photochem. Photobiol. 1997,
110, 201.
[51] The substance was contaminated with starting material 11 a and prod-
uct 14a (see the Supporting Information for details and spectra).
[52] Although at À788C, nBuLi is relatively stable in THF, for the reaction to
be conducted under the exact same conditions as for the reaction with
MeLi, nBuLi would need to be dissolved in THF and then transferred via
syringe or cannula. This would inevitably result in an uncontrolled and
irreproducible warming of the nBuLi/THF solution. In this context, note
that Table 1 merely represents the optimization process on a small
scale.
[53] R. R. Sharp, J. W. Tolan, J. Chem. Phys. 1976, 65, 522.
[54] For the experiments summarized in Table 1, the reactions were
quenched after 4 min, but in the case of the NMR experiments they
were not. The quenching reaction is an exothermic process, and at the
interface between the added droplet with the electrophile and the re-
action mixture containing the nucleophile, local heating will lead to the
observed reaction products. Moreover, diffusion at low temperatures is
so slow that mixing of the reaction components becomes important.
Whereas the reactions in Table 1 were all performed with stirring, this
was not possible in the NMR tube. Moreover the surface/volume ratio
of the NMR tube is much larger than of the reaction vessels used in the
experiments in Table 1, so that the cooling efficiency will be much
higher in the case of the NMR experiments. Therefore, the reactions in
the NMR tube were much slower than the stirred reactions described in
Table 1.
[7] M. Izquierdo-Serra, M. Gascón-Moya, J. J. Hirtz, S. Pittolo, K. E. Poskanzer,
ð. Ferrer, R. AlibØs, F. BusquØ, R. Yuste, J. Hernando, P. Gorostiza, J. Am.
Chem. Soc. 2014, 136, 8693.
[8] V. Chandrasekaran, T. K. Lindhorst, Chem. Commun. 2012, 48, 7519.
[9] J. Auernheimer, C. Dahmen, U. Hersel, A. Bausch, H. Kessler, J. Am.
Chem. Soc. 2005, 127, 16107.
[10] A. Goulet-Hanssens, K. Lai, W. Sun, T. E. Kennedy, C. J. Barrett, Biomacro-
molecules 2012, 13, 2958.
[11] S. Venkataramani, U. Jana, M. Dommaschk, F. D. Soennichsen, F. Tuczek,
R. Herges, Science 2011, 331, 445.
[12] G. S. Kumar, D. C. Neckers, Chem. Rev. 1989, 89, 1915.
[13] A. Natansohn, P. Rochon, Chem. Rev. 2002, 102, 4139.
[14] R. D. Mukhopadhyay, V. K. Praveen, A. Ajayaghosh, Materials Horizons
2014, 1, 572.
[15] Y. Deng, N. Li, Y. He, X. Wang, Macromolecules 2007, 40, 6669.
[16] Y. Yu, T. Maeda, J. I. Mamiya, T. Ikeda, Angew. Chem. Int. Ed. 2007, 46,
881; Angew. Chem. 2007, 119, 899.
[17] Y. Yu, M. Nakano, T. Ikeda, Nature 2003, 425, 145.
[18] Y. Zhao, T. Ikeda, Smart Light-Responsive Materials: Azobenzene-Contain-
ing Polymers and Liquid Crystals, Wiley-VCH, Weinheim, 2009.
[19] M. Schlosser, Organometallics in Synthesis - Third Manual; Wiley-VCH,
Weinheim, 2013.
[20] S. El Sheikh, H.-G. Schmalz, Curr. Opin. Drug Discov. Devel. 2004, 7, 882.
[21] T. Imamoto, T. Kusumoto, Y. Tawarayama, Y. Sugiura, T. Mita, Y. Hatanaka,
M. Yokoyama, J. Org. Chem. 1984, 49, 3904.
[55] At this low temperature, no temperature stability can be assumed. The
temperature varied by 7 K. This fact prevented quantitative kinetic stud-
ies, because of the sensitivity of the resonance of the 119Sn nucleus to
small temperature variations.
[22] F. A. Garlichs-Zschoche, K. H. Dçtz, Organometallics 2007, 26, 4535.
[23] D.-H. Qu, Q.-C. Wang, X. Ma, H. Tian, Chem. Eur. J. 2005, 11, 5929.
[24] A. K. Flatt, S. M. Dirk, J. C. Henderson, D. E. Shen, J. Su, M. A. Reed, J. M.
Tour, Tetrahedron 2003, 59, 8555.
ˇ
ˇ
ˇ
´
ˇ
[56] J. Holecek, K. Handlír˛, V. Cerny, M. Nµdvorník, A. Lycka, Polyhedron 1987,
[25] T. Soga, Y. Jimbo, K. Suzuki, D. Citterio, Anal. Chem. 2013, 85, 8973.
6, 1037.
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11172
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[57] H. J. Reich, B. Ö. Gudmundsson, D. P. Green, M. J. Bevan, I. L. Reich, Helv.
Chim. Acta 2002, 85, 3748.
[65] Average from SciFinder results for para-trimethylsilyl-substituted azo-
benzene derivatives at the time of writing.
[58] Because of the temperature variations and the SnMe4 product of this
reaction, no external references were added and the spectra were refer-
enced on SnMe4.
[59] H. J. Reich, J. P. Borst, R. R. Dykstra, P. D. Green, J. Am. Chem. Soc. 1993,
115, 8728.
[60] P. West, R. Waack, J. Am. Chem. Soc. 1967, 89, 4395.
[61] V. H. Gessner, C. Däschlein, C. Strohmann, Chem. Eur. J. 2009, 15, 3320.
[62] T. Stey, D. Stalke, Chemistry of Organolithium Compounds (Eds.: by Z.
Rappoport, I. Marek), 2004, 1, 47.
[63] H. J. Reich, D. P. Green, M. A. Medina, W. S. Goldenberg, B. Ö. Gud-
mundsson, R. R. Dykstra, N. H. Phillips, J. Am. Chem. Soc. 1998, 120,
7201.
[64] Aggregated complexes of organolithium reagents have been shown to
react, although this has not been demonstrated for MeLi itself: a) C.
Strohmann, V. H. Gessner, J. Am. Chem. Soc. 2008, 130, 11719; b) M. P.
Bernstein, D. B. Collum, J. Am. Chem. Soc. 1993, 115, 789; c) M. P. Bern-
stein, D. B. Collum, J. Am. Chem. Soc. 1993, 115, 8008; d) S. Liao, D. B.
Collum, J. Am. Chem. Soc. 2003, 125, 15114.
[66] M. Mentzel, H. M. R. Hoffmann, Prakt. Chem. 1997, 339, 517.
[67] D. Hellwinkel, H. Fritsch, Chem. Ber. 1990, 123, 2207.
[68] R. N. Sen, B. Sett, J. Am. Chem. Soc. 1924, 46, 111.
[69] S. Sengupta, S. K. Sadhukhan, Bull. Chem. Soc. Jpn. 2003, 76, 1223.
[70] S. Timofei, L. Kurunczi, T. Suzuki, W. M. F. Fabian, S. Mures¸an, Dyes Pigm.
1997, 34, 181.
[71] S. Miao, H. Li, Q. Xu, Y. Li, S. Ji, N. Li, L. Wang, J. Zheng, J. Lu, Adv. Mater.
2012, 24, 6210.
[72] G. W. Klumpp, Recl. Trav. Chim. Pay-B. 1986, 105, 1.
[73] D. I. Coppi, A. Salomone, F. M. Perna, V. Capriati, Angew. Chem. Int. Ed.
2012, 51, 7532; Angew. Chem. 2012, 124, 7650.
[74] V. Mallardo, R. Rizzi, F. C. Sassone, R. Mansueto, F. M. Perna, A. Salo-
mone, V. Capriati, Chem. Commun. 2014, 50, 8655.
Received: January 1, 2015
Published online on June 26, 2015
Chem. Eur. J. 2015, 21, 11165 – 11173
www.chemeurj.org
11173
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim