Full Paper
(entry 10), which is in agreement with the very low yield ob-
tained under nitrogen atmosphere (entry 11), proving
vessel was either irradiated through a UV glass filter (390 nm
cutoff) or a solution of benzophenone in ethanol (0.5m, path
length 14 mm, see the Supporting Information, page S5).
After optimization of the reaction conditions, the scope and
limitations of the procedure were investigated by using differ-
ent tertiary aliphatic amines (Table 3). Good to excellent yields
were observed in most cases, illustrating the efficiency of the
developed photocyanation method. Longer reaction times
were required for N,N-dimethylalkylamines (entries 4 and 6).
This behavior could be explained by the oxidation potentials
of tertiary amines, which decrease with increasing chain length
(e.g., Me3N, +0.82 V; Et3N, +0.79 V; nBu3N, +0.62 V; measured
vs SCE).[23] Remarkably, high regioselectivities were observed
when different a-protons were available for substitution (en-
tries 4–8, 10 and 11). In most cases, the reaction took place at
the less hindered carbon; however, electronic effects also seem
to play an important role.[23] When triethanolamine (1i) was
subjected to the photocyanation, the OH groups in the corre-
sponding a-aminonitrile were silylated (entry 9). To increase
the molecular complexity of the substrates, we decided to in-
vestigate the photocyanation of (À)-nicotine (1l), (+)-sparteine
(1m), orphenadrine (1n), atropine (1o), gramine (1p), and
strychnine (1q, entries 12–17). As anticipated, similar results
were encountered. However, no reaction was observed for qui-
nine (1r), DABCO (1,4-diazabicyclo[2.2.2]octane; 1s), and hex-
amethylenetetramine (1t, entries 18–20). A potential explana-
tion could be the violation of Bredt’s rule in the formation of
bridgehead iminium ions. On the other hand, DABCO is well
known to act as an excellent charge-transfer quencher, in par-
ticular for singlet oxygen.[24] The same behavior has been re-
ported for Cinchona alkaloids such as quinine (1r).[25] To ex-
plore the roles of quenchers in the reaction, inhibitor and
light–dark cycle experiments were performed to check the po-
tential involvement of singlet oxygen (1O2), radical chain reac-
tions, or light-independent processes in the photocyanation
(see the Supporting Information, pages S51–S53). However, no
clear mechanistic conclusions could be drawn from the results.
To demonstrate the application of photogenerated a-amino-
nitriles in the synthesis of new compounds and natural prod-
ucts, 2-(dipropylamino)butanenitrile (2b) and 5-cyanonicotine
(2l) were selected as exemplary candidates for further derivati-
zation. First, a one-pot C-alkylation/reductive decyanation was
attempted. The a-aminonitriles were deprotonated with lithi-
um diisopropylamide (LDA) at À788C and alkylated with 1-io-
doheptane. In situ reduction of the alkylation product with
NaCNBH3 afforded the heptyl-substituted compounds 3a and
3b in 77% and 44% yield, respectively (Scheme 1). Using an-
other well-known reactivity mode of a-aminonitriles, com-
pounds 2b and 2l were subjected to a Bruylants reaction with
n-pentylmagnesium bromide in THF at À208C, furnishing the
pentyl-substituted compounds 4a and 4b in 64% and 45%
yield, respectively (Scheme 1). These examples underline the
synthetic potential of photocatalytic postfunctionalizations of
amines and alkaloids.
a
oxygen to be essential for the photocyanation. Control experi-
ments in the absence of light (entry 13) or catalyst (entries 14
and 15) demonstrated that the combination of light and rose
bengal is required for an appreciable reaction rate although an
uncatalyzed background reaction takes place. Other cyanating
agents such as KCN also produced a-aminonitrile 2a (entry 12)
but the yield was significantly lower than with TMSCN
(entry 9). Reduction of the catalyst loading to 0.5 mol% still af-
forded a 90% yield after only 3 h (entry 16), indicating that
even lower catalyst loadings might still be effective.
Remarkably, loadings of 0.1 mol% (Table 2, entry 2) and
0.01 mol% (entry 3) still afforded good to excellent yields. With
0.001 mol% (entry 4) and 0.0001 mol% (entry 5) of rose
bengal, the reaction continued affording acceptable results
Table 2. Reduction of catalyst loading.[a]
[d]
Rose bengal
[mol%]
Average yield TON[c]
[%][b]
TOF [hÀ1
]
1
2
3
4
5
6
7
1
0.1
97.3Æ0.3
86.2Æ0.6
69.0Æ1.6
45.6Æ0.7
24.7Æ1.0
21.7Æ2.3
11.4Æ0.1
85.9Æ0.4
28.6Æ0.1
249Æ2
748Æ7
0.01
0.001
0.0001
0.00001
–
5760Æ17
1920Æ6
34267Æ80
133333Æ1100
11 422Æ27
44444Æ367
1030000Æ240000 343333Æ80000
–
–
[a] Reaction conditions: 210 mmol of tributylamine (1.0 equiv), TMSCN
(4.0 equiv), and 2.0 mL of MeCN were used. The reaction mixture was
stirred and irradiated under air bubbling with a 24 W fluorescent house-
hold bulb for 3 h at room temperature. Each reaction was run at least in
triplicate (see the Supporting Information, page S4). [b] Determined by
1H NMR spectroscopy by using CH2Br2 as an internal standard. [c] TON=
n(product)/n(rose bengal) after subtraction of background reaction
(Table 1, entry 7). [d] TOF=(n(product)/n(rose bengal))/3 h after subtrac-
tion of background reaction (Table 1, entry 7).
with yields between 25–45% after 3 h. A catalyst loading of
0.0001 mol% (10 ppm, entry 5) already corresponds to the
lowest value ever reported for an organic dye, which gave
a turnover number (TON) in excess of 130000. The limit of the
catalyst was reached at 0.00001 mol% or 0.1 ppm (entry 6,
TON>1000000). In this case, the yield dropped to 21.7%, but
was still twofold higher than the yield of the uncatalyzed back-
ground reaction. Full absorption at lmax of rose bengal
(562 nm) occurs in entries 1–3 as judged by UV/VIS spectrosco-
py and could explain the lower TON and turnover frequency
(TOF) values at these three catalyst loadings.
To prove that UV emissions from the CFL lamp play no
major role in the photocyanation, UV exclusion experiments
were performed at 0.1 mol% catalyst loading. Only a moderate
decrease in the yield could be observed when the reaction
Furthermore, the direct application of photogenerated a-
aminonitriles in the synthesis of (Æ)-crispine A as well as an
enantioselective approach to tetraponerines T7 and T8 could
Chem. Eur. J. 2016, 22, 5409 – 5415
5411
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim