Angewandte
Communications
Chemie
Table 1: Reaction optimization studies.
Scheme 1. Fluorinated side products generated in the palladium-cata-
lyzed trifluoromethylation reaction.
Entry
Ligand
Additive
2a [%]
3a [%]
2a/3a
1
2
3
4
5
6
7
8
9
10
11
12
13[b]
14[b]
L1
L2
L3
L4
L5
L6
L7
L8
L8
L8
L8
L6
L8
L8
–
–
–
–
–
–
–
–
23
18
11
19
27
37
23
37
68
73
47
57
74
7
21
18
13
19
20
19
20
21
2
1.0
10
5.0
0.4
4
1.1:1
1.0:1
0.8:1
1.0:1
1.4:1
1.9:1
1.2:1
1.8:1
34:1
73:1
4.7:1
11:1
>99:1
1.8:1
results in the absence of the trifluoromethyl silane additive,
the yield and regioselectivity obtained with L8 were consid-
erably higher when TESCF3 (30 mol%) was added (entries 6,
8, 10 and 12). Finally, performing the reaction in 2-MeTHF at
908C with L8 as the supporting ligand, along with a sub-
stoichiometric amount of TESCF3 afforded the desired
product 2a in 74% yield with excellent regioselectivity
(> 99:1) (entry 13). In addition to this significant improve-
ment in regioselectivity, the addition of the TESCF3 additive
also allowed the reaction to be conducted at a lower temper-
ature (entry 13 vs. entry 14).
TMSCF3
TESCF3
TIPSCF3
TESCF3
TESCF3
–
We subsequently examined the substrate scope using
these optimized reaction conditions, and found this protocol
to be applicable to the fluorination of a variety of 1,2-
disubstituted cyclic vinyl triflates (Table 2). In addition, the
fluorination of 1,2,2-trisubstituted vinyl triflates, for which
regioisomer formation is not an issue,[12d] was possible with L3
as the ligand and CsF as the fluoride source. TESCF3 was not
required for these processes. Interestingly, our new ligand
(L8) developed for vinyl triflate 1a, did not perform well for
1,2,2-trisubstituted vinyl triflates, which likely stems from its
sterically encumbered nature.
[a] Reactions were run at 0.1 mmol scale. Yields were determined by
19F NMR analysis of the crude reaction mixture using 1-fluoronaphtha-
lene as an internal standard. [b] 2-MeTHF, 908C.
1-Cyclohexenyl triflates with substituents at the 4- (2a and
2c), 3- (2e), or 2- (2b) position were all excellent substrates
(Table 2, n = 1). Benzofused (2 f) and oxygen-containing six-
membered cyclic triflates (2g) were compatible as well.
Moreover, the method could be used to access fluorinated
analogues of biologically active terpene and steroid deriva-
tives (2h, 2i, and 2j). In the case of 1j, isomerization of the
terminal double bond to the more thermodynamically stable
internal position occurs under these reaction conditions.
In general, the fluorination of 1-cyclopentenyl triflates
was more difficult, presumably due to the higher energy
eletrophiles,[12a–c] provided low yields of the desired vinyl
fluoride 2a and a substantial amount of the undesired
regioisomer 3a (close to 1:1 ratio, entry 1–4). Replacing the
6-methoxyl group of the ligand with a methyl substituent (L5
and L6) led to improved yields, indicating that ligand rigidity
may be important in this transformation (entry 5, 6).[13] After
further evaluation of ligands possessing a trimethylmethoxy-
substituted top ring (L7 and L8),[14] a novel biarylphosphine
ligand L8 was found to provide 2a in moderate combined
yield (58%) though still with poor regioselectivity (1.8:1;
entry 7, 8).
À
barrier for C F reductive elimination from the respective
palladium(II) complex (Table 1, n = 0).[16] Thus, vinyl triflate
1k without additional substitution on the double bond
provided the desired cyclopentenyl fluoride in low yield.
However, substrates possessing an additional substituent at
the 2-position reacted efficiently to provide the corresponding
cyclic vinyl fluorides in good to excellent yield (2l, 2m, and
2n).
During our investigation of the palladium-catalyzed
trifluoromethylation of vinyl sulfonates,[15b] we made the
serendipitous discovery that the corresponding vinyl fluoride
was formed as a side product with relatively high regioselec-
tivity (6.3:1, Scheme 1). We hypothesized that the presence of
Although 1-cycloheptenyl triflate 1o was fully consumed
under these conditions, the fluorinated product was not
obtained (Table 1, n = 2). GC/MS analysis of the crude
reaction mixture indicated the formation of the correspond-
ing alkyne or allene product, implying that the vinyl triflate
starting material decomposed through b-hydrogen elimina-
tion. Consistent with this hypothesis, seven-membered cyclic
vinyl triflates without b-hydrogen atoms were fluorinated in
good yields (2p, 2q, and 2r).
À
trifluoromethylsilanes, which were used as CF3 sources in
trifluoromethylation reactions, might be responsible for the
improved regioselectivity of the fluorination process. Indeed,
the use of TMSCF3, TESCF3 or TIPSCF3 as substoichiometric
additives (30 mol%) drastically improved the regioselectivity
(34:1, 73:1, and 4.7:1, respectively) (Table 1, entry 9–11).
Although catalyst based on L6 and L8 gave comparable
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!