Organic Letters
Letter
Table 1. Reaction Optimization
a
b
entry
catalyst
yield
c
1
2
3
4
[Ru(p-cymene)Cl2]2
PdCl2
Rh2(OAc)4
CuOTf
JohnPhos(AuCl)
B(C6F5)3
AgPF6
no reaction
traces
dec of 1a
29%
20%
15%
d
5
6
7
8
84%
86%
e
AgPF6
9
aq. HPF6
no reaction
a
Reaction conditions: In an oven-dried test tube, catalyst (3 mol %)
was put under vacuum and flushed with Argon for three times. Dry,
degassed DCM (0.5 mL) was added. Aniline 9a (0.4 mmol) and aryl/
aryl-diazo 1a (0.2 mmol) was dissolved in 0.5 mL of dry, degassed
DCM and added to the reaction mixture in one portion. The reaction
mixture was stirred at rt until the aryl/aryl diazo 1a was consumed.
b
c
Isolated reaction yield of 10a. Both staring materials were
d
e
recovered. NaBArF4 (5 mol %) was used as an additive. A 1:1
stoichiometry of aniline 9a and aryl/aryl-diazo 1a was used. dec =
decomposition
rationalized by an acid base reaction between the Brønsted
acid catalyst and aniline 9a that shuts down catalytic activity of
the Brønsted acid (Table 1, entry 9).
With the optimized conditions in hand, we next studied
different anilines 9 in the reaction with diphenyl diazomethane
1a (Scheme 1). Different electron-donating, electron-with-
drawing substituents and halogens were well tolerated in all
positions of the aromatic ring of 9, and the corresponding N−
H functionalization products 10a−o were isolated in high
yield. We observed slightly reduced reaction yields in the case
of ortho-substituted anilines, which can be reasoned by
increased sterical hindrance due to the ortho substituent
(10k−o). Moreover, 1-naphtylamine and benzo[d][1,3]dioxol-
5-amine proved compatible with the present reaction
conditions and the products were formed in 68% and 56%,
respectively (10p,q). Additionally, we investigated the syn-
thesis of 10a on a 1 mmol scale, and to our delight we were
able to isolate 10a in slightly reduced yield (78%).
Next, we investigated different aryl/aryl diazoalkanes 1 in
the reaction with aniline 9a. Electron-donating and chloro
substituents were well tolerated in all positions of the aromatic
ring (10r−w). Overall, we observed slightly reduced yields
compared to the unsubstituted diazoalkane 1a. When
introducing electron-withdrawing substituents instead of
electron-donating substituents, we observed a strong depend-
ence of the position of the substituent, while a 3-nitro
substituent was well tolerated, a 4-nitro substituent reduced
the yield significantly (10x−z). This observation could be
reasoned by the mesomeric effect of the nitro group leading to
an activation of diphenyl diazomethane 1x and subsequent
higher reaction yield compared to unsubstituted diphenyl
diazomethane 1a (97% vs 86%, respectively). Furthermore, 4-
nitro substituted diphenyl diazomethanes 1y,z deactivated the
diazoalkane and longer reaction times and lower yields were
observed (Scheme 1).
Figure 1. Application of diaryl carbenes in organic synthesis.
reaction of diphenyl diazomethane 1a with aniline 9a in the
presence of different metal catalysts or under photochemical
(SI), Table S1).
However, carbene transfer catalysts based on rhodium, gold,
copper, ruthenium or palladium did not provide the desired
product in reasonable yield (Table 1, entries 1−5). In the case
of JohnPhos(AuCl) and CuOTf, the N−H functionalization
product 10a was formed in low yield and the diazine was
formed as the major byproduct (Table 1, entries 4 and 5). In a
next step, we investigated the borane-catalyzed reaction, using
B(C6F5)3 as catalyst we isolated the N−H functionalization
product in 15% yield after reaction overnight (Table 1, entry
6). Unexpectedly, simple AgPF6 proved highly efficient and the
reaction was completed after only 10 min of reaction time at
ambient temperature and we were able to isolate 10a in 84%
yield (Table 1, entry 7). Intrigued by the promising results
using AgPF6 as catalyst, we further investigated different silver
salts, catalyst loadings, solvents and the stoichiometry of the
model reaction (for details, please see the SI, Table S1).13
However, no further improvement of the reaction yield was
obtained. A 1:1 stoichiometry of 1a and 9a in DCM as solvent
proved optimal to yield the product 10a in 86% yield (Table 1,
entry 8). To evaluate the participation of HPF6 as a hidden
Brønsted acid catalyst, we investigated HPF6 as a catalyst. Yet,
even after 24 h of reaction time, we did not observe any
reaction and both starting materials remained untouched. This
excludes the participation of HPF6 as catalyst, which can be
B
Org. Lett. XXXX, XXX, XXX−XXX