Angewandte
Communications
Chemie
most of substrates are either simple unfunctionalized alkynes
or alkenes.
We used the fluorination of the alkyne 1a as our model
reaction to study the HF insertion protocol (Table 1). We used
a faster reaction, and we could even reduce the gold catalyst
loading to 0.5 mol%, although longer reaction time was
needed (entry 15). It should be noted this reaction is not
sensitive to moisture and oxygen as the reactions can be
conducted at ambient atmosphere without any precautions.
With the optimized reaction conditions in hand, we
explored the scope and functional-group tolerance of our
HF insertion protocol (Table 2). First we evaluated aromatic
terminal alkynes (3a–u). Phenyl acetylene and alkyl-group-
substituted phenyl acetylenes all gave excellent yields of a-
fluoroketones (3a–c). Electron-rich phenol and phenol ether
groups were also well tolerated (3d–g), and this kind of
functional-group tolerance is not possible when strong
oxidants like Selectfluor are used.[25] Electron-withdrawing
functional-groups (ketone, ester, and acetonitrile) on the aryl
group of the alkynes all gave excellent yields (3h–j). Nonbasic
amine groups (TsNH, TsMeN) are also well tolerated (3k,l).
Halogen-substituted (F, Cl, Br) phenyl acetylenes all gave
excellent yields of the products 3. In general, substitution
patterns (ortho, meta, para) and the presence of either
electron-donating or electron-withdrawing groups on the
aromatic ring of alkynes have little influence on the efficiency
of the reactions.
Table 1: Development of reaction conditions.[a]
Entry
[Au]
Activator
2
HF Reagent
Yield
[%][b]
1
2
3
4
5
6
7
8
PPh3AuCl
PPh3AuCl
PPh3AuCl
PPh3AuCl
AgOTf
AgNTf2
AgSbF6
NaBARF
AgOTs
AgNTf2
AgNTf2
AgNTf2
AgNTf2
AgNTf2
AgNTf2
AgNTf2
AgNTf2
AgNTf2
–
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2b
2c
2a
2a
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Pyridine-HF
Et3N-HF
trace
70%
trace
trace
trace
44
54
77
89
85
PPh3AuCl
(RO)3PAuClc
tBuXPhosAuCl
XPhos
We also tested alkynes containing fused aromatics and
heteroaromatics (naphthalene, thiophene, benzothiophene,
indole, benzofuran), and they all gave excellent results
(Table 2, 3q–u). It should be noted that these heteroaromatics
are highly electron rich and can be easily oxidized, but they all
tolerated our reaction conditions. An alkenyl-substituted
alkyne (enyne) also worked well, and the highly functional-
ized fluoroketone 3v was obtained in good yield.
9
JohnPhos-AuCl
SPhos-Au-Cl
Dppf(AuCl)2
JohnPhos-AuCl
JohnPhos-AuCl
JohnPhos-AuCl
JohnPhos-Au-NTf2
(0.5% loading)
10
11
12
13
14
15
trace
n.r.
n.r.
n.r.
Pyridine-HF
90[d]
We then moved our focus to reactions of aliphatic alkynes.
Simple unfunctionalized aliphatic alkynes worked very well
(Table 2, 3w–y). And various common functional groups
(allyl, ether, ester, halogen, tosylate, mesylate, amide) are
well tolerated (3z-ah), and in all cases excellent yields of the
products 3 were obtained. Finally, we explored the feasibility
of our method in synthesis of complex target molecules. We
were glad to find that a structurally complex and highly
biologically important glycoside, an estrone derivative, and
a cholesterol derivative were suitable substrates and excellent
product yields (3ai–ak) were obtained.
In general, our method did not work for internal alkynes
(sluggish reaction). Because of the lower reactivity of internal
alkynes, the generation of a gold carbene is very slow.[16a–e] For
alkynes with strong nucleophilic groups present (e.g. prop-
argyl alcohol), our method did not give a clean transforma-
tion, and it could be due to the nucleophilic groups competing
with the N-oxide in the addition to the alkyne.
The a-fluoroketones 3 are versatile building blocks in
synthesis (Scheme 2). It can be simply reduced by NaBH4 to
give the fluorohydrin 4 in quantitative yield.[26] 3 can also react
with aldehyde to give the aldol condensation product 5
(fluorinated a,b-unsaturated ketone).[27] Reaction of 3 with
a Michael acceptor gives the 1,4-addition product 6 via an
enamine intermediate.[28] And reductive amination of 3 gives
the b-fluoroamine 7 in good yield.[29]
[a] Reaction conditions: 1a (0.2 mmol), N-oxide 2 (0.28 mmol), pyridine-
HF (1.3 mmol HF, 6.5 equiv), ligand-AuCl (5 mol%), and activator
(5 mol%) were stirred in DCM (1 mL) at RT in a sealed polypropylene
tube for 1.5 h. [b] Determined by GC-MS analysis. [c] R=2,4-di-tBuC6H3.
[d] Reaction time 24 h. DCM=dichloromethane, Dppf=1,1’-bis(diphe-
nylphosphanyl)ferrocene, JohnPhos=2-dicyclohexylphosphinobiphenyl,
SPhos=2-dicyclohexylphosphino-2’,6’-dimethoxybiphenyl, XPhos=2-
dicyclohexylphosphino-2’,4’,6’-triisopropylbiphenyl, Tf=trifluorometha-
nesulfonyl.
8-methylquinoline N-oxide (2a) as our oxidant and pyridine-
HF as our fluorine source. First, screening of silver activators
indicated AgNTf2 was the best choice (entries 1–5). Then we
investigated the ligand effects (entries 6–11).[23] In general,
gold catalysts derived from Buchwald-type ligands gave good
results, and among them the JohnPhos-based gold catalyst
gave the best chemical yield (entry 9). It was interesting to see
that the bidentate-ligand-based (Dppf) gold catalyst only
gave trace product (entry 11). Other N-oxides such as
pyridine N-oxide (2b) and dichloropyridine N-oxide (2c)
could not mediate this reaction (entries 12 and 13). We also
tested another HF-based fluorination reagent Et3N-HF, but
no reaction took place (entry 14). The low reactivity of the
Et3N-HF system may be due to high basicity of Et3N.[8]
Previously, we found that excess amounts of silver activators
are harmful to gold-catalyzed reactions, and preformed
cationic gold catalysts usually have the best efficiency.[24]
Indeed, the preformed gold catalyst JohnPhos-Au-NTf2 gave
Our proposed mechanism is shown in Scheme 3. Our
preferred mechanism is the carbene pathway, the key step of
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!