Organic Letters
Letter
a
exceeding the boiling point of THF using back pressure
regulators was only briefly contemplated, as blockage of the
backpressure regulator by the suspension was expected.
Scheme 3. Scope of Functionalized Ferrocenyl Azides
We were now in the position to merge both processes into
one continuous flow setup (Scheme 2). Some fine adjustment
of flow rates and residence times was performed (see
a constant argon flow and a pumping system for solvent were
employed to form a stable segmented flow regime.
Iodoferrocene 1a was introduced through an injection valve,
while solutions of n-butyllithium in n-hexane and tosyl azide in
THF were fed to the system by syringe pumps and mixed at 0
°C for 19 s. Finally, the reaction mixture was merged with an
argon stream resulting in a segmented flow and warmed to 60
°C with a residence time of 15 min in a coiled tube reactor.
Insoluble lithium para-toluenesulfinate precipitated from the
reaction mixture resulting in a triphasic system of salt, liquid,
and argon. The reaction mixture was eventually collected in a
flask containing an aqueous sodium bicarbonate solution to
quench any remaining reactive intermediates and remove
inorganic byproducts. Using this flow setup, azidoferrocene
(2a) was prepared in a yield of 82%. To showcase the
practicality of our setup in the preparation of multigram
quantities of azidoferrocenes, we performed a run with 1a at a
productivity of 4 g/h and achieved similar yields of 2a.
We next applied our optimized flow conditions to a broad
set of functionalized ferrocenyl halides 1b−n (Scheme 3). The
preparation of substrates 1e−1g was achieved by selective
bromine−lithium exchange of 1,1′-dibromoferrocene (see
azide 2b was obtained in an excellent yield. For the iodo-
substituted compound 2c, the yield dropped significantly to
26% in batch as the iodine−lithium exchange was less selective.
However, in flow the selectivity could be enhanced providing
2c in 54% yield. Double azidation of 1,1′-dibromoferrocene
gave the valuable diazide 2d in an improved yield of 94%
compared to previously reported 59% in batch.9a Due to its
explosive nature, the preparation of this compound in flow
(75% yield) benefited significantly from the better safety
profile of our setup. Notably, the tert-butyl ester present in
substrate 1e was tolerated giving the corresponding azide 2e, a
precursor for ferrocene amino acids (vide infra). Camphor-
derived ferrocenyl terpene 2g was obtained in 66−69% yield.
The planar chiral substrate 1h was prepared by diastereose-
lective lithiation−iodination of enantioenriched Ugi’s amine.
Formation of the corresponding azide 2h proceeded with
retention of configuration. Heteroaromatic ferrocenyl azide 2f
could be obtained in good yield. As ethynyl-substituted
ferrocene derivatives are frequently used in molecular wires,4
we prepared an array of ethynyl-substituted ferrocenyl azides
(2i−2n). These compounds were obtained in good to
excellent yields in batch and flow. Both, electron-donating
(2l) and electron-withdrawing (2m) groups attached to the
aromatic moiety were tolerated. Ethynyl azide 2i was prepared
in 77−79% yield presenting the opportunity for sequential
click chemistry.11,24 For cyclopropyl-substituted azide 2j the
yield dropped slightly (56% in batch, 69% in flow). Generally,
we assume that yields of ferrocenyl azides 2 are slightly lower
in flow than in batch due to traces of water in the flow reactor
resulting in hydrolysis of the intermediary ferrocenyllithiums 3.
The reduction of ferrocenyl azides by palladium-catalyzed
hydrogenation is well documented7c,25 but not suitable for
ferrocenyl azides 2g and 2i−n containing olefin or alkyne
a
Starting from the corresponding ferrocenyl bromides [Br] or iodides
[I]. Yields of isolated products are given. Batch reaction conditions:
substrate (1 equiv), n-butyllithium (1.1 equiv), tosyl azide (1.2 equiv)
in THF (0.05−0.16 M) at −78 °C, 30 min, warmed to 25 °C over 16
h. Flow reaction conditions: substrate (0.2 M in THF; v = 2 mL/
min), n-butyllithium (1.2 M in n-hexane; v = 0.4 mL/min), tosyl azide
(0.75 M in THF; v = 0.8 mL/min), argon (v = 0.2 mL/min), 0 to 60
°C, 15 min residence time. Bromine−lithium exchange was
performed for 1 h. n-Butyllithium (2.2 equiv) and tosyl azide (2.3
b
c
moieties. Treatment of ferrocenyl halides with ammonia in the
presence of a copper/iron catalyst26 would not allow the
synthesis of halogenated ferrocenyl amines 5b and 5c. Thus,
we decided to reduce ferrocenyl azides 2a−2n to the
corresponding amines 5a−5j (Scheme 4) under Staudinger
conditions. We found that treatment of ferrocenyl azides with
triphenylphosphine (PPh3) in a mixture of THF and water at
65 °C provided the corresponding amines 5a−5j in
consistently high yields. Among the products obtained,
bromoferrocenyl amine 5b is particularly valuable. A previous
synthesis required five steps starting from 1,1′-dibromoferro-
cene and resulted in 31% yield.8e We now prepared 5b in two
steps and 83% yield. The tert-butyl ester substituted ferrocenyl
amine 5d was obtained in 94% yield and is useful to access
C
Org. Lett. XXXX, XXX, XXX−XXX