Organic Letters
Letter
synthetic applications.12 There is one report of the SO2F2-
mediated conversion of benzoic acid and cyclohexane-
carboxylic acid to the corresponding acyl fluorides.13
Unfortunately, the reported reaction times were too long
(>12 h) to be practical in a one-pot reaction. Our goal,
therefore, was to investigate the activation mechanism of
carboxylate 3 to determine if acyl fluorides 10 are viable
intermediates and, if so, how to accelerate their formation.
Investigations into the reaction mechanism started with the
sulfuryl fluoride-mediated conversion of 3-phenylpropionic
acid (1a; R = CH2CH2Ph) to the corresponding acyl fluoride
(10a, Scheme 2). Bubbling SO2F2, generated from 1,1′-
sulfonyldiimidazole (SDI), KF, and TFA,14 through a solution
of 1a and diisopropylethylamine (DIPEA) in acetonitrile led to
the quantitative formation of acyl fluoride 10a in approx-
imately 4 h. The only other significant IR stretch in the
carbonyl region was identified as anhydride 9a, which is
transient in this reaction, building to a maximum of ∼15%
conversion before being consumed to yield acyl fluoride 10a.
Fluorosulfate 4a and sulfate 8a never accumulated as
intermediates, which is consistent with the reports of their
high reactivity.10 Similar experiments with aryl acids gave
comparable results, although the reaction times were longer.15
We also examined the NAS reaction conditions described by
the Qin group and found that the amidation reaction likely
proceeds primarily through direct trapping of the acyl
fluorosulfate (4a).16
6 h compared to 12 h without activation.19 No rate
acceleration was observed when tetrabutylammonium
hexafluorophosphate was added, thus eliminating the ammo-
nium cation as the potential species responsible for the
observed activation.
We further optimized the halide-accelerated acyl fluoride
formation20,21 and then applied these conditions to a range of
carboxylic acids (Scheme 3). The reaction worked well for
Scheme 3. Accelerated Formation of Acyl Fluorides from
a
Carboxylic Acids Using SO2F2 and TBAB
We next explored methods to accelerate the formation of the
acyl fluoride (10a) from 3-phenylpropionic acid (1a). Little is
known about accelerating sulfuryl fluoride-mediated processe-
s.2a The only method that has been reported is the DBU-
mediated activation of fluorosulfate intermediates,17 which has
been proposed to proceed through a hydrogen-bonding
interaction between protonated DBU and the fluorine in the
S−F bond.17a Given the lack of direct precedence, we broadly
screened nucleophilic activators and monitored for an increase
in reaction rate.18 While most of the amine additives and
tetrabutylammonium iodide were not effective, a significant
rate acceleration was observed for tetrabutylammonium
chloride (TBAC) and tetrabutylammonium bromide (TBAB)
(Figure 1). This rate acceleration was also observed for slower
reacting aryl acids, such as p-methoxybenzoic acid, which
afforded the corresponding acyl fluoride in >95% conversion in
a
Isolated yields of acyl fluorides are reported using 1.0 mmol of
carboxylic acid 1. Yields of unstable acyl fluorides that were
derivatized to the N-hydroxysuccinimide (HOSu) ester are reported
in brackets. 19F NMR yields using 0.3 mmol of 1 are reported in
parentheses. Acyl fluoride yield was low due to volatility or instability
b
alkyl carboxylic acids 1a and 1b, affording acyl fluorides 10a
and 10b in 66% and 76% yields, respectively. Increasing the
steric bulk α to the acid to adamantyl acid 1c led to only a
slight decrease in the yield of acyl fluoride 10c after 1 h (61%
yield). The reaction also worked well for cinnamic acid,
affording acyl fluoride 10d in 75% yield after 2 h. Benzoic acid,
p-methylbenzoic acid, and p-chlorobenzoic acid were effective
substrates, affording 10e, 10f, and 10g in 30% (92% NMR
yield), 57%, and 54% yields, respectively. Electron-rich p-
methoxybenzoic acid was successful, although the reaction
required 3 h to reach 80% yield of 10h. Similarly, electron-poor
p-nitrobenzoic acid was also less reactive, requiring 6 h to
achieve 69% yield of 10i. The reaction was tolerant of meta-
and ortho-substituted acids, with 10j, 10k, and 10l all formed
in moderate to good yields. Heterocyclic substituents, such as
pyridines (1m, 1n), furan (1o), and thiophene (1p), were also
tolerated, although they all required longer reaction times to
achieve good to excellent yields of 10m−p.
Figure 1. In situ ReactIR data following the CO carbonyl stretch of
10a (1846 cm−1) for three reactions: TBAC as the additive (×, in
The identification of a stable intermediate (10) in these
NAS reactions, and the ability to accelerate its formation with
halide additives, enables the exciting possibility of utilizing a
□
○
yellow), TBAB as the additive ( , in red), and no additive ( , in
green). The reactions were sampled at 2 min intervals, and the
reaction profiles were normalized to the greatest peak intensity.
B
Org. Lett. XXXX, XXX, XXX−XXX