Radical decarboxylative fluorination of aryloxyacetic acids using N-fluorobenzenesulfonimide and a photosensitizer

Article abstract of DOI:10.1002/ejoc.201500038

Fluorinated methoxy arenes are emerging as important motifs in both agrochemicals and pharmaceuticals. A novel technique for the synthesis of monofluoromethoxy arenes through the direct fluorodecarboxylation of carboxylic acids was developed that uses photosensitizers and N-fluorobenzenesulfonimide (NFSI). Utilization of the oxidatively mild fluorine transfer agent NFSI enabled the synthesis of fluoromethyl ethers that were previously inaccessible with decarboxylative fluorinations performed with Selectfluor. Mechanistic studies are consistent with the photosensitizer effecting oxidation of the aryloxyacetic acid.

Full text of DOI:10.1002/ejoc.201500038

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
FULL PAPER
DOI: 10.1002/ejoc.201500038
Radical Decarboxylative Fluorination of Aryloxyacetic Acids Using
N-Fluorobenzenesulfonimide and a Photosensitizer
Joe C. T. Leung[a] and Glenn M. Sammis*[a]
Keywords: Radicals / Photochemistry / Photosensitizer / Fluorine
Fluorinated methoxy arenes are emerging as important mo-
tifs in both agrochemicals and pharmaceuticals. A novel
technique for the synthesis of monofluoromethoxy arenes
through the direct fluorodecarboxylation of carboxylic acids
was developed that uses photosensitizers and N-fluorobenz-
enesulfonimide (NFSI). Utilization of the oxidatively mild
fluorine transfer agent NFSI enabled the synthesis of
fluoromethyl ethers that were previously inaccessible with
decarboxylative fluorinations performed with Selectfluor.
Mechanistic studies are consistent with the photosensitizer
effecting oxidation of the aryloxyacetic acid.
Introduction
Fluorinated methoxy arenes are increasingly pervasive in
both the agrochemical and the pharmaceutical industries.[1]
The growing importance of these motifs in biologically
active molecules has led to a concomitant increase in the
development of new synthetic methodologies focused on
the preparation of fluoro ethers.[1–4] Mono-, di-, and tri-
fluoromethoxy-substituted arenes have traditionally been
prepared from either ionic[3] or carbene[4] reactive interme-
diates. Free-radical-based synthetic methods are attractive
alternatives, but there are only a few reports of their use in
the preparation of fluoromethoxy arenes.[5–9] A viable and
widely applicable radical-based approach to fluorinated
ethers would provide a complementary solution to this im-
portant chemical challenge.
The earliest reported example of a radical-based method
utilized xenon difluoride for the preparation of mono-
fluoromethoxy arenes.[5] Xenon difluoride is a versatile syn-
thetic reagent,[10] but its expense and high reduction poten-
tial[11] limit its utility. In 2012, we demonstrated that the
electrophilic fluorine sources Selectfluor[12] and N-fluoro-
benzenesulfonimide (NFSI)[13] are viable radical fluorine
Scheme 1. Synthesis of fluoromethoxy ethers by using radical de-
carboxylative fluorination; bpy = 2,2Ј-bipyridine.
We recently reported two different photochemical de-
transfer reagents that could be utilized in the thermal carboxylative fluorination methods for the preparation of
fluorination of α-phenoxy tert-butyl peresters [Scheme 1, fluoromethoxy arenes.[8,9] The first method[8] utilizes UV
Equation (1)].[6] Subsequently, the Li group demonstrated light and Selectfluor as both the oxidant and fluorine trans-
that Selectfluor, in combination with a silver catalyst, could fer agent to afford either monofluoromethoxy arenes or di-
also be employed for the synthesis of both alkyl fluorides fluoromethoxy arenes [Scheme 1, Equation (3)]. The second
as well as an example of a monofluoromethoxy arene method[9] utilizes visible light with a photoredox catalyst to
[Scheme 1, Equation (2)].[7]
effect the oxidation and decarboxylation [Scheme 1, Equa-
tion (4)]. Both of these reactions can be successfully em-
ployed if the arene is electron neutral or electron deficient.
However, several substrate classes, such as naphthyl and
electron-rich arenes, afford low yields of the desired fluorin-
ation product and instead provide products corresponding
to ionic ring fluorination.
[a] Department of Chemistry, University of British Columbia,
2036 Main Mall, Vancouver, BC V6T 1Z1, Canada
E-mail: gsammis@chem.ubc.ca
www.chem.ubc.ca
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201500038.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
J. C. T. Leung, G. M. Sammis
FULL PAPER
The most significant factor contributing to limited gener- acid 1a (Table 1), a substrate that is known to undergo
ality and scope for radical-based fluorination methodolo- photofluorodecarboxylation by using our previously devel-
gies is the ability of Selectfluor to directly fluorinate elec- oped Selectfluor-based methodologies.[8,9] All initial experi-
tron-rich aryl rings.[14] It has been postulated that the elec- ments were run in deuterated benzene to facilitate analysis
trophilic reactivity of the fluorinating agents correlates with by 1H NMR spectroscopy. As expected from previous stud-
their efficiency as an oxidant (i.e., higher reduction poten- ies that indicated a base facilitates the reaction,[8] the con-
tials).[15] The challenge is that most of the known sources trol experiment with just NFSI and 300 nm light (Table 1,
of atomic fluorine, such as molecular fluorine and xenon entry 1) led to fluorinated product 2a in only 7% yield, as
difluoride, have significantly higher reduction potentials [E determined by NMR spectroscopy.
= 2.87 V and 2.64 V (SHE), respectively][1a,11] than Se-
lectfluor [E = 0.57 V (SHE)][16] (Figure 1), and thus they
will likely have similar, or even more significant, substrate
Table 1. Effect of organic bases on the photodecarboxylative fluor-
ination of 1a.[a]
limitations.
Entry
Base (equiv.)
Yield[b] [%]
Figure 1. Comparison of fluorodecarboxylation reagents by in-
creasing reduction potential in reference to the standard hydrogen
electrode (SHE).
1
2
3
4
5
6
7
8
no base (0)
B1 (1.0)
B2 (1.0)
B3 (1.0)
B4 (1.0)
B5 (1.0)
B6 (1.0)
B7 (1.0)
B8 (1.0)
B8 (0.5)
B8 (0.25)
7
61[c]
46[c]
43[c]
0[d]
55
As a possible solution, we turned our attention to NFSI.
NFSI has been demonstrated to be a weaker electrophilic
source of atomic fluorine in aryl fluorination reactions than
Selectfluor and has been measured to be a weaker oxidant
[E = –1.00 V (SHE)][16] than any other known source of
atomic fluorine.[17] Unfortunately, NFSI was significantly
inferior to Selectfluor in our previously developed photo-
80
75[e]
81[e]
81[e]
65[f]
9
10
11
chemical decarboxylative fluorination methodologies.[8,9]
A
[a] Reaction conditions: base (as indicated), NFSI (4.0 equiv.), 1a
(1.0 equiv., 0.05 mmol) at 0.1 m in deuteriobenzene, irradiated at
300 nm for 2 h. [b] Determined by 1H NMR spectroscopy by using
new photodecarboxylative fluorination that can utilize a
mild oxidant, such as NFSI, has the potential to access a
wider substrate scope than any previous methodology that ethyl trifluoroacetate as an internal standard. [c] NFSI reacts with
base to form an ammonium fluoride. [d] Electrophilic aromatic
uses either thermal conditions or Selectfluor. Herein, we
fluorination of the base was observed. [e] Average value of three
trials. [f] Average value of two trials.
report a new approach to these photodecarboxylative
fluorination reactions by using a combination of NFSI and
a photosensitizer (Scheme 2).
Our optimization began by investigations into non-pyr-
idine amine bases B1–B4 that have solubility profiles similar
to that of NSFI (Table 1, entries 2–5). DBU (B1) and trialk-
ylamine bases B2 and B3 afforded fluoro ether 2a in moder-
ate yields, whereas B4 was incompatible with NFSI and
simply underwent electrophilic aromatic fluorination. The
moderate yields in entries 2–4 (Table 1) may be the result of
the direct fluorination of bases B1–B3 with NFSI.[15a] In-
deed, analysis of a mixture of NFSI and triethylamine
(Table 1, entry 3) by 19F NMR spectroscopy showed the
formation of a fluorotrialkylammonium species (δ ≈
63 ppm), which confirmed that fluorine transfer between
the amines had occurred. An experiment performed with
Scheme 2. This work: radical decarboxylative fluorination with
NFSI.
Results and Discussion
Reaction Development
Investigations into the development of a general photo- the use of an excess amount of triethylamine (B2) was per-
decarboxylative fluorination began with para-fluoroaryloxy formed to test if the newly formed fluorotrialkylammonium
2
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
Radical Decarboxylative Fluorination of Aryloxyacetic Acids
was a viable source of atomic fluorine. Analysis by
Unlike acetonitrile, acetone is a known triplet sensi-
19F NMR spectroscopy showed complete conversion of B2 tizer[18] and has been shown to catalyze the photodecarbox-
into fluorotriethylammonium, and fluorination product 2a ylation of alkyl carboxylates in the presence of single-elec-
was not observed after irradiation at 300 nm for 2 h. Thus, tron-transfer (SET) acceptors, such as phthalimides.[19] To
the fluorotrialkylammonium species is not a viable source determine if acetone was directly involved in the reaction,
of atomic fluorine and fluoro ether product 2a is formed we investigated the decarboxylative fluorination by using
exclusively from NFSI.
350 nm light. The emission range of the 350 nm light source
We next focused on less nucleophilic bases that were less is sufficiently narrow such that substrate 1a, NFSI,[20] and
likely to react with NFSI (i.e., B5–B8). Gratifyingly, hin- acetonitrile will not absorb, and thus, no reaction should
dered pyridine bases (Table 1, entries 7–9) afforded fluorin- occur. However, the emission range is well within the acet-
ation product 2a in good yields, and 2,6-lutidine (B6) per- one absorption profile.[21] As hypothesized, no reaction was
formed comparably to sterically hindered tert-butyl ana- observed upon running the decarboxylative fluorination in
logues B7 and B8. However, in solvents such as acetone, the acetonitrile, but the reaction in acetone provided 2a in 85%
B6–1a salt precipitated from solution. Thus, B8 was chosen yield. Visible-light sources failed to promote decarb-
as the base for further optimization. The amount of base oxylative fluorination, and no reaction was observed under
could be decreased to 0.5 equiv. without impacting the reac- thermal conditions, regardless of the solvent.
tion (Table 1, entry 10), likely because the fluorine transfer
Reaction time was the last reaction parameter we opti-
product from NFSI, bis(phenylsulfonyl)amide, can also mized. As shown in Figure 2, photodecarboxylative fluorin-
serve as a base in the reaction. A further decrease in the ation with NFSI was significantly faster in acetone than in
amount of base provided inconsistent results with yields acetonitrile. Within 10 min, the reaction in acetone was
ranging from 51 to 78%.
Ͼ90% complete with full conversion of aryloxyacetic acid
Photodecarboxylative fluorination with B8 and NFSI 1a after 1 h. For the reaction in acetonitrile to be Ͼ90%
shows broad solvent tolerance with high yields of fluorin- complete, 1 h of irradiation was necessary. Complete con-
ation in nonpolar solvents (Table 2, entries 1–4). With polar version of aryloxyacetic acid 1a required 2 to 3 h.
solvents, solubility problems were encountered; aqueous
acetone or acetonitrile mixed-solvent systems as well as
neat methanol (Table 2, entry 8) did not solubilize NFSI.
Fluorination failed to occur in neat DMSO, likely because
the solvent begins to absorb UV light at 330 nm. However,
trace amounts of DMSO could be added into acetone or
acetonitrile to improve substrate solubility if necessary. A
suitable balance of effective fluorination and reagent solu-
bility were found for acetone and acetonitrile (Table 2, en-
tries 6 and 7), both of which provided fluoro ether 2a in
high yields.
Table 2. Effect of solvent on the photodecarboxylative fluorination
of 1a.[a]
Figure 2. Effect of irradiation time on the photodecarboxylative
Entry
Solvent
Yield[b] [%]
fluorination. Reaction conditions: B8 (0.5 equiv.), NFSI
(4.0 equiv.), 1a (1.0 equiv., 0.05 mmol) at 0.1 m in deuterated sol-
vent (as indicated), irradiated at 300 nm for time as indicated. Con-
versions were determined by 1H NMR spectroscopy by using ethyl
trifluoroacetate as an internal standard.
1
2
3
4
5
6
7
8
benzene
toluene
CH2Cl2
CHCl3
DMSO
acetone
MeCN
MeOH
81
69
85
81
0[c]
80[d]
82[d]
0[e]
Substrate Scope
[a] Reaction conditions: B8 (0.5 equiv.), NFSI (4.0 equiv.), 1a
(1.0 equiv., 0.05 mmol) at 0.1 m in deuterated solvent, irradiated at
300 nm for 2 h. [b] Determined by 1H NMR spectroscopy by using
ethyl trifluoroacetate as an internal standard. [c] No photofluoro-
decarboxylation observed; 1a was recovered completely. [d] Average
value of four trials. [e] NFSI displayed poor solubility.
The first substrates investigated (i.e., 1a–f) were chosen
because they are all viable in the previously developed Se-
lectfluor-mediated methodologies[8,9] and, thus, provide a
benchmark for comparison (Table 3). Decarboxylative
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
J. C. T. Leung, G. M. Sammis
FULL PAPER
fluorination of 1a provided 2a in an average yield of 80%, 4-phenyl acid 1i required the addition of a small amount of
which was slightly lower than the related photodecarboxyl- DMSO for solubility, the decarboxylative fluorination af-
ative fluorination with Selectfluor (Table 3, entry 1). By forded 2i in 79% yield.
using the new NFSI conditions, decarboxylative fluorin-
ation of 1b, 1c, 1d, and 1f provided 2b, 2c, 2d, and 2f in
yields that were nearly identical to those obtained from the
Table 4. Substrate scope of the new photodecarboxylative fluorin-
ation.[a–c]
related decarboxylative fluorination by using our previous
methodologies (Table 3, entries 2–4, 6).[8,9] Further elec-
tronic deactivation of the aryloxy ring with two chlorine
substituents reduced the effectiveness of fluorination with
NFSI (Table 3, entry 5).
Table 3. Comparison of photodecarboxylative fluorination with
NFSI to fluorination with Selectfluor.[a]
[a] Reaction conditions: NFSI (4.0 equiv.), B7 (0.5 equiv.),
1
(1.0 equiv., 0.4 mmol) at 0.15 m in argon-sparged acetone irradiated
at 300 nm for 3 h. [b] Reaction conditions: NFSI (3.0 equiv.), B8
(0.5 equiv.), 1 (1.0 equiv., 0.4 mmol) at 0.15 m in argon-sparged
acetone irradiated at 350 nm for 3 h. [c] Yield of isolated product,
yields in parentheses represent yields determined by 1H NMR spec-
troscopy by using ethyl trifluoroacetate as an internal standard on
0.05 mmol scale in deuterioacetone and irradiated at 300 nm for
2 h.
Entry R1
R2
R3
Compound
Yield [%]
Selectfluor[d]
NFSI[c]
1
2
3
4
5
6
H
H
H
H
Cl
H
H
H
H
Br
H
F
1a
1b
1c
1d
1e
1f
(80)[e,f]
(82)[e,g]
68[h]
(94)
(84)
60
H
Br
H
Cl
H
73[h]
57 (72)
54[h] (70) 86
tBu
84
83
[a] Reaction conditions: NFSI (4.0 equiv.), 1 (1.0 equiv., 0.4 mmol)
at 0.15 m in argon-sparged acetone irradiated at 300 nm for 3 h.
[b] B8 (0.5 equiv.) was employed as the base unless otherwise indi-
cated. [c] Yield of isolated product, yields in parentheses represent
Encouraged by the successful use of phenyloxyacetic acid
substrates, fluorodecarboxylation was investigated with
naphthyloxyacetic acids, a substrate class that is incompat-
ible with both the original UV light/Selectfluor conditions[8]
and our photocatalytic conditions (Table 4).[9] For each of
these substrates (i.e., 1j–l), nonselective aryl fluorination
and oxidation of the naphthyloxy ring outcompeted the de-
sired decarboxylative fluorination reaction. By using our
newly developed methodology, 2j was isolated in near quan-
titative yield. Naphthyloxy substrates with an altered substi-
tution pattern (see compound 1k) or with an added elec-
tron-withdrawing group (see compound 1l) were also viable
1
yields determined by H NMR spectroscopy by using ethyl trifluo-
roacetate as an internal standard. NMR-scale reactions were per-
formed on 0.05 mmol scale in deuterioacetone and were irradiated
at 300 nm for 2 h. [d] Values taken from the literature are repro-
duced for ease of comparison. Yield of isolated product, yields in
parentheses represent yields determined by 1H NMR spectroscopy.
[e] Owing to the volatility of 2, isolation-scale reaction was not per-
formed. [f] Average value of four trials. [g] Average value of three
trials. [h] B7 (0.5 equiv.) was employed as the base.
With new, viable photodecarboxylative fluorination in the newly developed reaction and provided good yields
methodology in hand, we tested the principal hypothesis of of the isolated products. Increasing the electron density of
the study: whether the lower reduction potential of NFSI the aryloxy moiety further with a methoxy substituent (4-
would allow the decarboxylative fluorination of more elec- methoxyphenyloxyacetic acid) was still a limitation, and
tron-rich aryloxy acetic acids. All of the substrates shown nonspecific aryl fluorination outcompeted the desired de-
in Table 4[22] either provided low yields or failed completely carboxylative fluorination.
under our former conditions with UV light and Se-
We next investigated the decarboxylative fluorination by
lectfluor.[23] Whereas decarboxylative fluorination of sub- using a natural product analogue, apocynin derivative 3
strate 1g provided 2g in only 34% yield (as determined by (Scheme 3). The counterbalance of the electron-donating
NMR spectroscopy) with Selectfluor and UV light,[8] our and electron-withdrawing substituents on the aryl ring pro-
new NFSI conditions afforded 2g in 75% yield (as deter- vided an intriguing test for our newly developed photode-
mined by NMR spectroscopy), and ring fluorination was carboxylative fluorination methodology. Additionally, apo-
not observed. The decrease in the yield of the NMR prod- cynin derivative 3 was not a good substrate for our aqueous
uct was due to substrate volatility. Decarboxylative fluorin- Selectfluor conditions, as it only provided 4 in 23% yield
ation of substrate 1h proved to be unsuccessful with the (34% by NMR spectroscopy). However, the photosensi-
use of UV light and Selectfluor. With NFSI, this substrate tized reaction with NFSI in acetone was very successful,
proceeded with high conversion and in good yield. Whereas and 4 was isolated in 73% yield. Substrates without the
4
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
Radical Decarboxylative Fluorination of Aryloxyacetic Acids
counterbalanced electron-withdrawing acyl group, such Nonsensitized Photodecarboxylative Fluorinations
as para-methoxyarylacetic acid, afforded ring-fluorinated
products.
The proposed acetonitrile nonsensitized photoreaction
mechanism, presented in Figure 3, is analogous to the
mechanism for the previously developed decarboxylative
fluorination with Selectfluor. Excitation of the B-band tran-
sition (πǞπ*) of the phenoxy nucleus of 1 creates excited
state 7, which undergoes reversible single-electron trans-
fer[25] with NFSI. Following base-mediated deprotonation
and decarboxylation, radical 10 is fluorinated by the NFSI
radical anion through either direct radical abstraction of
fluorine or back electron transfer from the radical anion to
10, followed by ionic fluorination. Another possible mecha-
nism involves internal electron transfer (from the carboxyl-
Scheme 3. Photodecarboxylative fluorination of apocynin deriva-
tive 3.
ate to the benzenoid core) followed by radical decarboxyl-
ation of the resulting carboxy radical.
To further test our methodology, we examined estrone
derivative 5 (Scheme 4). Photodecarboxylative fluorination
of this steroid derivative presents a formidable challenge to
the Selectfluor-mediated system, as the tertiary benzylic po-
sition can be readily oxidized. Indeed, under conditions that
employed UV light and Selectfluor,[8] no fluorination prod-
uct was detected. However, photosensitized decarboxylative
fluorination by employing NFSI successfully provided fluo-
romethoxy steroid 6 in 49% yield.[24]
Figure 3. Proposed reaction mechanism for nonsensitized photo-
decarboxylative fluorination.
All of the photodecarboxylative fluorination optimiza-
tion experiments with NFSI support the mechanism indi-
cated in Figure 3. As established by the experiments listed
in Table 1 and Scheme 5 [Equation (1)], efficient decarb-
oxylative fluorination requires a base in the nonsensitized
reaction. The deprotonation step appears to facilitate
irreversible decarboxylation of 8, which supports the neces-
sity of a carboxylate for either the Strecker-type decarboxyl-
ation[26] mechanism or the internal single-electron-transfer
mechanism. Furthermore, no reaction is observed under
Scheme 4. Photodecarboxylative fluorination of estrone derivative
5.
Mechanistic Investigations
Future reaction developments for the photodecarboxyl-
ative fluorination with NFSI required a detailed under-
standing of the reaction mechanism, particularly as the sol-
vent selection appeared to influence the rate of reaction.
Photodecarboxylative fluorination with NFSI in aceto-
nitrile closely matched the overall reaction profile of the
decarboxylative fluorination with Selectfluor. If the new
conditions with NFSI were used in acetone, the reaction
rates significantly increased and the viable wavelengths
needed to promote the reaction broadened. To probe the
specifics of the reaction mechanism more deeply, we began
with a study of the nonsensitized photoreaction in aceto-
nitrile.
Scheme 5. Mechanistic investigations on the nonsensitized photo-
decarboxylative fluorination reaction in acetonitrile.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
J. C. T. Leung, G. M. Sammis
FULL PAPER
thermal conditions, which suggests that the reaction does ity of this result suggests that acetone acts as a single-elec-
not proceed through a Hunsdiecker-type reaction[27] with tron oxidant in the reaction.
acyl hypofluorite intermediates.
We next investigated the decarboxylative fluorination of
phenyl acetic acid (11), as it contains both aryl and carb-
oxylate chromophores. Using our standard reaction condi-
tions with NFSI in acetonitrile, no reactivity was observed
[Scheme 5, Equation (2)]. This further supports the as-
sertion that hypofluorites are not involved in the reaction
mechanism and that, additionally, the direct oxidation of
the carboxylate by NFSI does not occur.
Scheme 6. Photodecarboxylative fluorination of 1a by using a cata-
lytic amount of benzophenone.
Finally, to investigate the necessity of the aryloxy ring,
we irradiated α-alkoxy acid 13 [Scheme 5, Equation (3)]
with 300 nm light. As expected, the starting material did
not react under the standard conditions with NFSI. This
suggests that the α-oxygen atom is not critical but that the aryl
chromophore is required for reactivity.
If acetone acts as an oxidant, there are two functional
groups that may assist the oxidation: the aryl ring and the
oxygen atom of the ether. To determine which functionali-
ties were required, we examined the photodecarboxylative
fluorination of acetic acid derivatives 13, 17, and 11
(Scheme 7).[31] Photodecarboxylative fluorination of 13 pro-
vided complete decarboxylation, and several fluorinated
products were detected [Scheme 7, Equation (1)].[32] This
proved that aryl substitution on the substrate is not neces-
sary for the decarboxylation to occur. We next studied the
reactivity of 17 and phenylacetic acid (11) [Scheme 7, Equa-
tions (2) and (3)]. Substrate 17 possesses a carboxylic acid
with electronic properties similar to those of aryloxy acid
1, but the ether oxygen atom is more difficult to oxidize.
Phenylacetic acid has the carboxylic acid moiety but lacks
the ether oxygen atom. Both of these substrates were not
viable in the photodecarboxylative fluorination in acetone,
which suggests the ether oxygen atom is crucial for reacti-
vity.
Sensitized Photodecarboxylative Fluorinations
A photosensitizer, such as acetone, can alter the reaction
mechanism in two possible ways:[28] acetone can serve as a
mediator for energy transfer or it can act as a single-elec-
tron oxidant. If acetone serves as a mediator for energy
transfer, then the overall reaction mechanism would be sim-
ilar to that presented in Figure 3, with the exception that
acetone could facilitate the excitation of substrate 1 to ex-
cited species 7. Alternatively, photoexcited acetone can
serve as a single-electron oxidant (Figure 4). Photoexcited
ketone 15 can oxidize 1 in a reversible single-electron trans-
fer to lead to radical cation 8. Deprotonation by radical
anion 16, or from another base in solution,[29] leads to rapid
and irreversible decarboxylation to aryloxymethyl 9 and 10,
which can be fluorinated by NFSI to yield 2. The ketone
sensitizer can be regenerated by back electron transfer from
the bis(phenylsulfonyl)amidyl radical.
Figure 4. Proposed reaction mechanism for photosensitized photo-
decarboxylative fluorination through reversible electron transfer.
Scheme 7. Mechanistic investigations on the sensitized photode-
carboxylative fluorination in acetone.
To differentiate between these two mechanistic possibil-
ities, we explored adding a catalytic amount of the com-
monly employed photooxidant benzophenone[30] to the
On the basis of the mechanism depicted in Figure 4, radi-
nonsensitized reaction in acetonitrile. Addition of benzo- cal anion 16 could act as the base and facilitate decarboxyl-
phenone to the photodecarboxylative fluorination of sub- ation. Sensitized photodecarboxylative fluorinations, in
strate 1a in acetonitrile (Scheme 6) led to a similar accelera- both acetone and acetonitrile (with catalytic amounts of
tion in the reaction rate; after only 10 min, the benzophen- benzophenone), proceeded smoothly to give high conver-
one-sensitized reaction in acetonitrile provided results sions of 1a after only 10 min (Scheme 8), which thus sup-
analogous to those obtained for the reaction performed in ports the mechanism in which the photosensitizer promotes
acetone in terms of both conversion and yield. The similar- the photodecarboxylation as both an oxidant and a base.[33]
6
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:32
Pages: 9
Radical Decarboxylative Fluorination of Aryloxyacetic Acids
Acknowledgments
The authors thank Dr. P. Kennepohl for guidance during computa-
tional work. This work was supported by the University of British
Columbia (UBC) and a doctoral fellowship from the Natural Sci-
ences and Engineering Research Council of Canada (NSERC) to
J. C. T. L. We would also like to thank M. Rueda-Becerril and C.
Chatalova-Sazepin for assistance with preparing the manuscript.
[1]
a) T. Hiyama, Organofluorine Compounds: Chemistry and Ap-
plications, Springer, New York, 2000; b) F. R. Leroux, P.
Jeschke, M. Schlosser, Chem. Rev. 2005, 105, 827–856; c) P.
Jeschke, E. Baston, F. R. Leroux, Mini-Rev. Med. Chem. 2007,
7, 1027–1034; d) S. Purser, P. R. Moore, S. Swallow, V. Gou-
verneur, Chem. Soc. Rev. 2008, 37, 320–330; e) B. Manteau, S.
Pazenok, J. Vors, F. R. Leroux, J. Fluorine Chem. 2010, 131,
140–158.
Scheme 8. Base-free photosensitized decarboxylative fluorination.
Conclusions
We successfully developed a new photosensitized decarb-
oxylative fluorination reaction of aryloxyacetic acids by em-
ploying N-fluorobenzenesulfonimide (NFSI) that signifi-
cantly increases the substrate scope relative to that of
previously reported photodecarboxylative fluorinations.
Detailed mechanistic investigations into this new photo-
decarboxylative fluorination reaction uncovered a novel
fluorination mechanism that involves photoexcitation of a
sensitizer molecule, which facilitates single-electron transfer
from the benzenoid core of the aryloxyacetic acids, which
is followed by base-mediated decarboxylation and radical
fluorination.
Utilization of an oxidatively mild radical fluorine source,
NFSI, enabled the synthesis of fluoromethyl ethers that
contain more electron-rich aromatic components including
natural product derivatives. Most notably, photodecarbox-
ylative fluorination with NFSI enabled the class of naphthyl
fluoromethyl ethers to be synthesized, a class of compounds
that was previously inaccessible by using any of the photo-
chemical methods performed with Selectfluor. This consti-
tutes a significant advance in the synthetic utility of photo-
decarboxylative fluorination.
[2]
[3]
For reviews, see: a) F. R. Leroux, B. Manteau, J. Vors, S. Paz-
enok, Beilstein J. Org. Chem. 2008, 4, 13; b) J. Hu, W. Huang,
F. Wang, Chem. Commun. 2009, 7465–7478; c) see ref.[1]
For recent examples of ionic and cross-coupling methods, see:
a) G. K. S. Prakash, I. Ledneczki, S. Chacko, G. A. Olah, Org.
Lett. 2008, 10, 557–560; b) Y. Hagooly, O. Cohen, S. Rozen,
Tetrahedron Lett. 2009, 50, 392–394; c) M. Ochiai, A. Yoshi-
mura, K. Miyamoto, Tetrahedron Lett. 2009, 50, 4792–4795; d)
R. Koller, K. Stanek, D. Stolz, R. Aardoom, K. Niedermann,
A. Togni, Angew. Chem. Int. Ed. 2009, 48, 4332–4336; Angew.
Chem. 2009, 121, 4396–4400; e) O. Marrec, T. Billard, J. Vors,
S. Pazenok, B. R. Langlois, J. Fluorine Chem. 2010, 131, 200–
2007; f) M. Ochiai, A. Yoshimura, M. M. Hoque, T. Okubo,
M. Saito, K. Miyamoto, Org. Lett. 2011, 13, 5568–5571; g) C.
Huang, T. Liang, S. Harada, E. Lee, T. Ritter, J. Am. Chem.
Soc. 2011, 133, 13308–13310.
For recent examples of carbene-mediated methods see: a) K.
Fuchibe, Y. Koseki, T. Aono, H. Sasagawa, J. Ichikawa, J.
Fluorine Chem. 2012, 133, 52–60; b) K. Fuchibe, Y. Koseki, H.
Sasagawa, J. Ichikawa, Chem. Lett. 2011, 40, 1189–1191; c) J. B.
Sperry, K. Sutherland, Org. Process Res. Dev. 2011, 15, 721–
725; d) Y. Zafrani, G. Sod-Moriah, Y. Segall, Tetrahedron 2009,
65, 5278–5283; e) F. Wang, L. Zhang, J. Zheng, J. Hu, J. Fluor-
ine Chem. 2011, 132, 521–528; f) F. Wang, W. Huang, J. Hu,
Chin. J. Chem. 2011, 29, 2717–2721.
[4]
[5]
[6]
a) T. B. Patrick, K. K. Johri, D. H. White, J. Org. Chem. 1983,
48, 4158–4159; b) T. B. Patrick, K. K. Johri, D. H. White, W. S.
Bertrand, R. Mokhtar, M. R. Kilbourn, M. J. Welch, Can. J.
Chem. 1986, 64, 138–141.
M. Rueda-Becerril, C. Chatalova-Sazepin, J. C. T. Leung, T.
Okbinoglu, P. Kennepohl, J.-F. Paquin, G. M. Sammis, J. Am.
Chem. Soc. 2012, 134, 4026–4029.
Experimental Section
Photodecarboxylative Fluorination Optimization Studies: Solutions
containing aryloxyacetic acid 1 (1 equiv.), base (0.5–1.0 equiv.),
NFSI (1–4 equiv.), and ethyl trifluoroacetate (1.0 equiv.) in deuter-
ated solvent (0.1 m in 1) were partitioned to borosilicate NMR
tubes. One sample was set aside as the t = 0 sample, and the re-
maining sample(s) was(were) placed on a rotating carousel inside a
photochemical reactor (containing 16ϫ 8 W lamps) and exposed
to 300 nm light for 2 h. Analysis by NMR spectroscopy was per-
formed directly on the crude reaction mixtures.
[7]
[8]
F. Yin, Z. Wang, Z. Li, C. Li, J. Am. Chem. Soc. 2012, 134,
10401–10404.
J. C. T. Leung, C. Chatalova-Sazepin, J. G. West, M. Rueda-
Becerril, J.-F. Paquin, G. M. Sammis, Angew. Chem. Int. Ed.
2012, 51, 10804–10807; Angew. Chem. 2012, 124, 10962–10965.
M. Rueda-Becerril, O. Mahé, M. Drouin, M. B. Majewski,
J. G. West, M. O. Wolf, G. M. Sammis, J.-F. Paquin, J. Am.
Chem. Soc. 2014, 136, 2637–2641.
[9]
[10]
[11]
[12]
For a review on XeF2 see: M. A. Tius, Tetrahedron 1995, 51,
Isolation-Scale Photodecarboxylative Fluorination: The correspond-
ing aryloxyacetic acid 1 (1 equiv.), B8 or B7 (0.5 equiv.), and NFSI
(3–4 equiv.) were added to an argon-sparged solution of acetone
(0.15 m in 1) in an argon-filled borosilicate glass culture tube. The
reaction vessel was then placed on a rotating carousel inside a pho-
tochemical reactor (containing 16ϫ 8 W lamps) and exposed to
300 or 350 nm light for 3 h. Purification by flash column
chromatography (petroleum ether/diethyl ether) afforded fluoro-
methyl ether 2, 4, or 6.
6605–6634.
A. J. Bard, R. Parsons, J. Jordan, in: Standard potentials in
aqueous solution, vol. 6, CRC press, New York, 1985.
E. R. Banks, S. N. Mohialdin-Khaffa, G. S. Lal, I. Sharif,
R. G. Syvret, J. Chem. Soc., Chem. Commun. 1992, 595–596.
E. Differding, H. Ofner, Synlett 1991, 187–189.
For a review on Selectfluor, see: P. T. Nyffleler, S. G. Durón,
M. D. Burkart, S. P. Vincent, C.-H. Wong, Angew. Chem. Int.
Ed. 2004, 43, 192–212; Angew. Chem. 2004, 116, 194.
[13]
[14]
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:33
Pages: 9
J. C. T. Leung, G. M. Sammis
[15] a) G. S. Lal, G. P. Pez, R. G. Syvret, Chem. Rev. 1996, 96, 1737– [25] For reviews on photoinduced electron transfer (PET), see: a)
FULL PAPER
1755; b) T. Furuya, C. A. Kuttruff, T. Ritter, Curr. Opin. Drug
Discovery Dev. 2008, 11, 803–819.
[16] G. P. Girina, A. A. Fainzil’berg, L. G. Feoktistov, Russ. J. Elec-
trochem. 2000, 36, 162–163.
[17] There are multiple conflicting reports for the reduction poten-
tial of both Selectfluor and NFSI. We selected the most re-
cently reported values (see ref.[16]). See also: a) A. G. Gilicinski,
G. P. Pez, R. G. Syvret, G. S. Lal, J. Fluorine Chem. 1992, 59,
157–62; b) E. Differding, P. M. Bersier, Tetrahedron 1992, 48,
1595–1604.
J. Mattay, D. F. Eaton (Eds.), Photoinduced Electron Transfer,
Springer, New York, 1990, vol. 1; b) J. Mattay (Ed.), Photo-
induced Electron Transfer, Springer, New York, 1993, vol. 5.
[26] For a review on Strecker degradation, see: A. Schönberg, R.
Moubasher, Chem. Rev. 1952, 50, 261–277.
[27] For a general review on the Hundsdiecker reaction, see: R. G.
Johnson, R. K. Ingham, Chem. Rev. 1956, 56, 219–269.
[28] G. J. Kavarnos, N. J. Turro, Chem. Rev. 1986, 86, 401–449.
[29] The identity of the base is not clear as there are three possible
candidates: B8, 16, and bis(phenylsulfonyl)amide.
[18] N. J. Turro, Modern Molecular Photochemistry, University Sci-
ence Books, Suasalito, CA, 1991.
[30] Benzophone was specifically chosen, as it has been studied as
a triplet sensitizer for photodecarboxylative fluorination with
phenoxyacetic acid (1b). For early studies on benzophenone-
mediated photodecarboxylation of α-heteroatom acetic acids
see: R. S. Davidson, P. R. Steiner, J. Chem. Soc. Perkin Trans.
2 1972, 1357–1362.
[31] For a review, see: D. Budac, P. Wan, J. Photochem. Photobiol.
A: Chem. 1992, 67, 135–166.
[32] As product 14 is relatively unstable, we primarily focused on
the decarboxylation rather than the appearance of specific rad-
ical products.
[19] For a review, see: A. G. Griesbeck, Chimia 1998, 52, 272–283.
[20] NFSI absorbs at 270 nm with a shoulder at 277 nm. For an
absorbance spectrum, see the Supporting Information.
[21] The absorption maxima for acetone occurs at 270 nm and tails
to 330 nm. See: Tables of Physical & Chemical Constants, 16th
ed., 1995, chapter 3.8.7. Kaye & Laby Online, version 1.0, 2005:
www.kayelaby.npl.co.uk (retrieved April 15, 2013).
[22] With the exception of substrate 1g, which was originally re-
ported in a previous Selectfluor reactivity study (ref.[8]), all
compounds are new. See the Supporting Information for exper- [33] Whereas B7 and B8 are not mechanistically necessary as bases,
imental details.
in: sensitized photofluorodecarboxylations, the addition of
these pyridine bases leads to the formation of fewer fluorinated
byproducts with select substrates. See the Supporting Infor-
mation for additional details.
[23] Substrates 2g, 2h, and 2i could be synthesized by using our
photoredox catalysis method (ref.[9]) owing to the lower energy
light that was utilized.
[24] The remaining mass balance contained fluorinated products
from nonselective reactions.
Received: January 12, 2015
Published Online: ᭿
8
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O50038
/KAP1
Date: 05-02-15 12:47:33
Pages: 9
Radical Decarboxylative Fluorination of Aryloxyacetic Acids
Radical Decarboxylative Fluorination
J. C. T. Leung, G. M. Sammis* ......... 1–9
Photodecarboxylative fluorination method-
ology involving the use of photosensitizers
and N-fluorobenzenesulfonimide (NFSI)
provides facile access to monofluoro-
methoxy-substituted arenes through the di-
rect fluorodecarboxylation of carboxylic
acids. Utilization of NFSI enables the syn-
thesis of fluoromethyl ethers that were pre-
viously inaccessible with decarboxylative
fluorinations performed with Selectfluor.
Radical Decarboxylative Fluorination of
Aryloxyacetic Acids Using N-Fluoroben-
zenesulfonimide and a Photosensitizer
Keywords: Radicals
/
Photochemistry
/
Photosensitizer / Fluorine
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
9