.
Angewandte
Communications
Table 1: Reduction of olefins by diimide generated in situ from
to generate segments of the liquid phase and O2, a residence
time unit (RT) made of perfluoroalkoxy copolymer (PFA),
a heat exchanger (HE), and a static (BPR1) as well as an
adjustable back pressure regulator (BPR2).[14] Preliminary
experiments using ethanol as the solvent revealed that at
N2H4·H2O and O2 (Scheme 1).[a]
Entry Substrate
N2H4·H2O Conversion Selectivity Yield
[equiv]
[%][b]
[%][b]
[%][c]
a backpressure of 20 bar, a liquid flow rate of 0.4 mLminÀ1
,
1
2
4
>99
>99
99
and a gaseous stream of 2 mLminÀ1 are necessary to maintain
both a suitable segment pattern and a proper residence time
(10 min).
4
>99
>99
98
The reduction of allylbenzene to propylbenzene was
chosen as the model transformation for a detailed optimiza-
tion study of the continuous process. Initially, a solvent
screening was performed using a variety of alkyl alcohols and
5 equivalents of hydrazine monohydrate at different temper-
atures (Supporting Information, Table S1 and Figure S4). At
808C, a clear trend with respect to reaction rate (MeOH <
3
4
5
4
4
4
>99
>99
>99
>99
>99
>99
92
95
92
EtOH < iPrOH < nPrOH < n-butanol < n-pentanol)
was
observed. The most reasonable explanation for this phenom-
enon is related to the increasing solubility of oxygen gas in
this series of alcohols.[15] We therefore decided to utilize
nPrOH for all further optimization experiments, as the
conversions in this solvent were significantly higher than in
MeOH, EtOH, or iPrOH, and the boiling point of nPrOH,
compared to the higher homologues, is still low enough to
allow for a convenient work up.
6
5
>99
>99
88
7[d]
8[f]
4
5
>99[e]
>99[e]
>99[e]
>99[e]
87
97
9
4
5
>99
>99
>99
93
58
The allylbenzene model reaction was further tested at
higher temperatures, resulting in a conversion of 94% at
1008C when using 5 equivalents of N2H4·H2O (Table S2). To
drive the reaction to completion and simultaneously increase
the throughput, the substrate concentration was increased
from 0.1m to 0.5m. This intensified process allowed us to
reduce the amount of hydrazine hydrate to 4 equivalents
while still maintaining quantitative conversion with perfect
selectivity in only 10 min reaction time;[16,17] whereas lower
amounts resulted in incomplete reactions, even at elevated
temperatures. It has to be stressed that mixtures of oxygen
and organic solvents have a rather high potential for
explosions, and therefore the small volumes and channel
dimensions of a continuous flow (micro)reactor minimizes
possible flame propagation.[18] The installed heat exchanger
cools the reaction mixture to ambient temperatures before
depressurization. As anticipated, the use of synthetic air
instead of O2 resulted in dramatically reduced substrate
consumption (53%). Control experiments without hydrazine
(< 1% conversion) or nitrogen instead of oxygen (8%
conversion) were also carried out.
With the optimized conditions for allylbenzene in hand,
we next evaluated the continuous-flow transfer-hydrogena-
tion system for a variety of functionalized alkenes. The
suitability of the substrates for the reaction was pre-evaluated
through calculations at the M06-2X level[19] for the hydrogen
transfer from cis-diimide to the corresponding olefin. Nota-
bly, the computational results provided valuable guidance
prior to experimental assessment of the reactivity (Table 1;
see also Table S4). Most of the chosen substrates (1–5, 7, 9)
underwent total consumption without any need for reoptim-
ization from the conditions used for allylbenzene, whereas
some other olefins required a larger excess of hydrazine (6, 8,
10), or a combination of more hydrazine, longer reaction
10
66[g]
11[h]
12[i]
5
5
>99
>99
>99
>99
94
91
[a] Conditions (unless otherwise noted): alkene (0.5m, 0.5 or 1 mmol) in
nPrOH, 0.4 mLminÀ1 liquid flow rate, 2 mLminÀ1 gas flow rate, 20 bar
backpressure, at 1008C, with a 10 min residence time (10 mL coil).
[b] Given as percent of GC-FID peak area. [c] Yield of isolated product.
[d] Conc.=0.33m. [e] Determined by 1H NMR spectroscopy. [f] nPrOH/
H2O (1:1) as solvent. [g] Given as percent of HPLC peak area at 215 nm.
[h] Run at 1208C with a 20 min residence time (16 mL coil). [i] Run at
1208C with a 30 min residence time (26 mL coil). Cbz=benzyloxycar-
bonyl, TBDMS=tert-butyldimethylsilyl.
times, and slightly increased temperatures (11, 12). In terms of
selectivity, olefins containing nitro groups (1, 3) resulted in
the desired products in near quantitative yields.[20] Silyl ether
protecting groups (5) and halogen atoms (9) easily resisted
the applied conditions. Moreover, Cbz-protected amines (7;
Cbz = benzyloxycarbonyl), which are prone to deprotection
in transition-metal-catalyzed hydrogenations, were reduced
in almost quantitative yields.[21] Carbonyl functional groups
are somewhat critical when hydrazine is used as a reagent;
somewhat surprisingly however, the reduction of ethyl
cinnamate (11) afforded the desired product in almost
quantitative yield without any detectable amount of the
corresponding hydrazide.
A comparison study for the reduction of cinnamate 11
using a standard batch balloon method was carried out in n-
pentanol to mimic the experimental conditions (1208C) of the
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 10241 –10244