A convenient photocatalytic fluorination of unactivated C-H bonds

Article abstract of DOI:10.1002/anie.201400420

Fluorination reactions are essential to modern medicinal chemistry, thus providing a means to block site-selective metabolic degradation of drugs and access radiotracers for positron emission tomography imaging. Despite current sophistication in fluorination reagents and processes, the fluorination of unactivated C-H bonds remains a significant challenge. Reported herein is a convenient and economic process for direct fluorination of unactivated C-H bonds that exploits the hydrogen abstracting ability of a decatungstate photocatalyst in combination with the mild fluorine atom transfer reagent N-fluorobenzenesulfonimide. This operationally straightforward reaction provides direct access to a wide range of fluorinated organic molecules, including structurally complex natural products, acyl fluorides, and fluorinated amino acid derivatives. The direct fluorination of unactivated C(sp³)-H bonds is catalyzed by the inexpensive photocatalyst tetrabutylammonium decatungstate (TBADT). This convenient reaction provides direct access to a wide range of fluorinated organic molecules, including structurally complex natural products, acyl fluorides, and fluorinated amino acids.

Full text of DOI:10.1002/anie.201400420

Angewandte
Chemie
DOI: 10.1002/anie.201400420
À
C H Activation
À
A Convenient Photocatalytic Fluorination of Unactivated C H
Bonds**
Shira D. Halperin, Hope Fan, Stanley Chang, Rainer E. Martin, and Robert Britton*
À
Abstract: Fluorination reactions are essential to modern
medicinal chemistry, thus providing a means to block site-
selective metabolic degradation of drugs and access radio-
tracers for positron emission tomography imaging. Despite
current sophistication in fluorination reagents and processes,
fluorination of C H bonds. In this context, success has been
realized in allylic[10] and benzylic[11] fluorination. However,
the identification of reagent systems that effect selective
3
À
fluorination of unactivated C(sp ) H bonds, or those not
adjacent to sp2-hybridized carbon atoms or other functional
groups, which facilitate the formation of radicals or anions,
remains a significant challenge.[12]
À
the fluorination of unactivated C H bonds remains a signifi-
cant challenge. Reported herein is a convenient and economic
À
process for direct fluorination of unactivated C H bonds that
Recently, Groves and co-workers disclosed a protocol for
À
exploits the hydrogen abstracting ability of a decatungstate
photocatalyst in combination with the mild fluorine atom
transfer reagent N-fluorobenzenesulfonimide. This operation-
ally straightforward reaction provides direct access to a wide
range of fluorinated organic molecules, including structurally
complex natural products, acyl fluorides, and fluorinated
amino acid derivatives.
the fluorination of unactivated C H bonds through a radical
process which relies on a manganese porphyrin catalyst
working in concert with an oxidant (iodosylbenzene), AgF,
and a catalytic amount of tetrabutylammonium fluoride
(TBAF).[12] The groups of Lectka[13] and Inoue[14] have also
demonstrated that N-oxyl radicals are capable of catalyzing
3
À
the fluorination of C(sp ) H bonds in unfunctionalized
cycloalkanes (e.g., cyclohexane, cyclooctane) when using
Selectfluor as the fluorine-transfer reagent. An earlier
report by Sandford, Chambers, and co-workers also describes
a catalyst-free electrophilic fluorination reaction of hydro-
carbons using Selectfluor in refluxing MeCN.[15] For the
reasons discussed above, we have been interested in devel-
T
he incorporation of a fluorine atom into an advanced drug
candidate is a common strategy in medicinal chemistry.[1,2]
Oftentimes, single hydrogen to fluorine substitution
a
improves druglike properties by blocking undesired metab-
olism at a specific site, increasing lipophilicity or binding
affinity, or altering drug absorption, distribution, or excre-
tion.[1] In addition to the medicinal advantages often pre-
sented by fluorination, valuable pharmacokinetic information
can be gleaned from non-invasive positron emission tomog-
raphy (PET) imaging of 18F-labeled drugs in vivo.[3] The
critical role fluorine plays in the drug discovery process is
highlighted by the fact that one-third of the so-called block-
buster drugs are fluorinated in at least one position.[4] Since
the fluorination of cortisone[5] signalled a new era of
fluoropharmaceuticals, further advances have relied on the
discovery of mild fluorination reagents which can be handled
safely and obviate the use of fluorine gas or its surrogates.[2,6]
While present sophistication in the field includes the addition
of electrophilic fluorine to alkenes,[6,7] and fluorination of aryl
triflates[8] and palladium aryl complexes,[9] there is much
interest in the development of reactions that effect the direct
oping a convenient protocol for the direct fluorination of
3
À
unactivated C(sp ) H bonds, and were particularly intrigued
by a process that would involve the generation and trapping
of alkyl radicals with fluorine-transfer reagents. In this regard,
Sammis and co-workers have calculated[16] the homolytic N F
À
bond dissociation energy (BDE) for several electrophilic
fluorinating reagents, including N-fluorobenzenesulfonimide
(NFSI; BDENF = 63 kcalmolÀ1), and demonstrated the utility
of NFSI in fluorine transfer to alkyl radicals generated
through decarboxylative processes.[17] Considering that [18F]-
NFSI can be readily prepared from the reaction of [18F]F2 and
NaN(SO2Ph)2,[18] the development of a NFSI-based fluorina-
3
À
tion of unactivated C(sp ) H bonds could also serve as
a platform technology for the production of new radiotracers
for PET imaging. Bearing this in mind, we envisioned
a reaction which combines the hydrogen abstraction ability
of a polyoxometalate with a fluorine-transfer reagent to
3
À
À
[*] S. D. Halperin, H. Fan, S. Chang, Prof. R. Britton
Department of Chemistry, Simon Fraser University
Burnaby, British Columbia (Canada)
directly transform unactivated C(sp ) H bonds into C F
bonds.
A variety of polyoxometalates have proven effective as
E-mail: rbritton@sfu.ca
photocatalysts in oxidative transformations,[19] and the deca-
Dr. R. E. Martin
4À
tungstate anion W10O32 has been studied extensively for
Medicinal Chemistry, Small Molecule Research, Pharma Research &
Early Development (pRED), F. Hoffmann-La Roche AG
Grenzacherstrasse 124, 4070 Basel (Switzerland)
these purposes.[20-22] The tetrabutylammonium salt of deca-
tungstate (TBADT; 3; Table 1) is a well-characterized
polyoxometalate catalyst which is particularly efficient at
abstracting hydrogen atoms from saturated hydrocarbons[21]
and has been utilized in carbon–carbon bond formation,[23]
[**] This work was supported by an NSERC Discovery Grant to R.B.,
a MSFHR Career Investigator Award to R.B., and NSERC Post-
graduate Scholarships for S.D.H., H.F., S.C.
oxidation of alcohols,[24] epimerization of unactivated C H
À
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201400420.
centers,[25] and carbonylation reactions.[26] From a practical
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 1: Demonstration and optimization of the TBADT-catalyzed fluo-
rination of bornyl acetate (1).
presumably derived from acetonitrile displacement of fluo-
ride, were also observed, various inorganic bases (e.g.,
Na2CO3, NaHCO3, K2CO3) were added in an effort to
consume the small amounts of coincident hydrofluoric acid
deemed deleterious to the desired process. Optimally, the
addition of a catalytic amount of base (< 20 mol%) substan-
tially increased conversion, and when 10 mol% of NaHCO3
was employed (entry 7), bornyl acetate was fluorinated in
a total yield of 56% (ca. 80% yield considering the unreacted
starting material).
While a full mechanistic understanding of this reaction
will require further experimentation, we suggest the mecha-
nism depicted in Scheme 1, a mechanism which is consistent
with both the photocatalytic reactivity of TBADT and
Entry
TBADT
[mol%]
Conc.
[m]
Additive[a]
Yield
[%][b]
1
2
3
4
5
6
7
2
1
0
2
2
2
2
0.2
0.2
0.2
1.0
2.0
2.0
2.0
none
none
none
none
none
Na2CO3
NaHCO3
23 (32)
9 (12)
0
31 (43)
33 (46)
37 (52)
40 (56)[c]
[a] Used 0.1 equiv. [b] Yield of isolated 2; combined yield of isolated
fluorinated products within parentheses. [c] An additional 30% of
starting material was recovered from this reaction.
Scheme 1. Proposed catalytic cycle for the TBADT/NFSI fluorination of
3
À
unactivated C(sp ) H bonds.
perspective, TBADT is readily prepared in one step through
a complex self-assembly process from inexpensive sodium
tungstate[21] and absorbs light at wavelengths that do not
interact with common organic reagents (e.g., l = 350–
400 nm). Photocatalysis with TBADT also involves mild
reaction conditions (e.g. room temperature, wet solvents)
and is functional-group tolerant. Herein we report the
discovery of an operationally simple and economic photo-
fluorine-transfer capacity of NFSI. Thus, photoexcitation of
the decatungstate anion (W10O324À) produces a short-lived
excited state[28] which rapidly decays to a long-lived reactive
intermediate (denoted W*)[28,29] capable of hydrogen atom
abstraction, thus producing W10O325ÀH+. Fluorine atom trans-
fer from NFSI to the coincident carbon-centered radical (or
a single-electron transfer from the radical species to NFSI to
catalytic system for the direct fluorination of unactivated
3
C(sp ) H bonds, and it exploits the hydrogen abstraction
generate a carbocation, followed by fluoride transfer[16]
)
À
ability of TBADT and fluorine-transfer capacity of NFSI.
Considering both the steric bulk and redox potential of
decatungstate catalysts,[27] this process may mimic in vivo
provides the fluorinated product. This proposal is further
supported by the fact that reaction of bornyl acetate (Table 1,
entry 7) in the presence of the radical scavenger TEMPO
provided no fluorinated products. Bearing in mind the size
and charge of the decatungstate catalyst, the selectivity
À
oxidative metabolism by selecting C H bonds prone to
À
metabolic degradation and replacing them with C F bonds,
À
a potentially transformative reaction for lead optimization in
medicinal chemistry.
observed in the C H fluorination of bornyl acetate is
À
consistent with the avoidance of sterically congested C H
As depicted in Table 1, we initially investigated combina-
tions of TBADT and NFSI in acetonitrile for the fluorination
bonds and destabilizing electrostatic interactions.
À
Having demonstrated the direct C H fluorination of
of bornyl acetate, which contains 11 unique and unactivated
bornyl acetate using the photocatalytic TBADT/NFSI system,
without further optimization we explored the scope of this
3
À
C(sp ) H bonds. Photoexcitation of the decatungstate cata-
lyst was achieved by irradiating the reaction mixture with two
15 Watt black light bulbs (centered at l = 365 nm). We were
delighted to find that by using 2 mol% of TBADTand a small
excess of NFSI, bornyl acetate was converted into the the
fluoroacetate 2 in modest yield (23%). Increasing the
reaction time (up to 60 h) had little effect on this result,
whereas reduction in catalyst loading significantly impacted
the yield of 2 (entry 2), and no product was observed in the
absence of TBADT or light. We next examined the effect of
the reaction concentration and found that fluorination
proceeded optimally at high concentrations of bornyl acetate
(e.g., 2.0m, entry 5). Finally, as acetamide side products,
reaction through the fluorination of several small organic
3
À
molecules, all of which possess multiple unactivated C(sp ) H
bonds (Table 2).[30] In general, the TBADT/NFSI fluorination
reaction is functional-group tolerant, and aliphatic ketones,
esters, and lactones were all selectively fluorinated. The
fluorination of camphor (entry 1) is notable as this substrate
was incompatible with the fluorination process reported by
Groves and co-workers,[12] and 6-fluoroisocamphanone (6),
presumably from the rearrangement of a cation generated
following hydrogen abstraction from C6 of camphor and SET
to NFSI,[16] was also produced in this reaction. In line with
observations summarized above, examination of the products
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Table 2: Scope of photocatalytic fluorination.
fluorinate at positions most remote from carbonyl or O car-
boxy functions. Additionally, no difluoro compounds were
detected in any of these reactions. These electronic-gearing
factors are also evidenced in the monofluorination of menthyl
acetate (9, entry 3), and occurs equally at the most sterically
accessible methine and the methylene most remote from the
acetoxy group. Further studies on saturated esters were
consistent with these results. Thus, fluorination of ethyl
isovalerate (18), methyl valerate (20), and methyl 4-methyl-
valerate (23; entries 6–8), occurred predominantly at the
methylene or methine position most remote from the
carboxylic ester functionality. However, further increasing
the chain length resulted in a decreased selectivity, and
fluorination of methyl hexanoate (25) provided an approx-
imate 2:1 mixture of the C5 and C4 fluorinated esters 26 and
27 (entry 9), respectively. Notably, the conversion in fluori-
nation reactions of linear esters was much improved by the
addition of water, however, the overall yields were compro-
mised by concomitant hydrolysis and formation of the
corresponding fluoroacids.
Entry Substrate
Major product(s)[a]
Yield
[%][b]
1
48
30
2
3
4
43
53
To compare the present catalyst system with that reported
by Groves,[12] the fluorination of sclareolide (28) was also
investigated. As depicted in Scheme 2, the TBADT/NFSI
system is highly efficient for fluorination of sclareolide and
both catalyst/reagent systems are similar in producing (2S)-2-
fluorosclareolide (29) as the major product. Considering the
relatively low BDE for acylhydrides (CO-H BDE ꢀ 89 kcal
5
6
7
50
75
70
8
9
49
51
[a] Yields are given within parentheses, and unless otherwise indicated,
were calculated from a 1H NMR spectrum recorded on the crude reaction
mixture by using an internal standard. [b] Yield based on unreacted
starting material; combined yield where two products are depicted.
[c] Yield of isolated product.
depicted in Table 2 highlights the selectivity for fluorination
at sterically accessible branched methines or methylenes
which are also remote from electron-withdrawing groups. The
inductive influence of electron-withdrawing groups on the
selectivity of fluorination is further exemplified in the
reactions of camphor (4, entry 1), the corresponding lactone
7 (entry 2), and bornyl acetate (1; Table 1), all of which
Scheme 2. Site-selective fluorination of sclareolide, benzaldehyde, and
amino acid derivatives. The yields for 35, 36, and 38 are calculated
from 1H NMR spectra recorded on crude reaction mixtures with the
addition of an internal standard.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
molÀ1),[31] we also explored the direct conversion of aldehydes
into acylfluorides, which are a relatively stable but underu-
tilized class of acyl halide. By employing the optimized
reaction conditions, benzaldehyde (31) was converted into
benzoylfluoride (32), and subsequently transformed directly
into N-benzylbenzamide (33) by reaction with benzylamine.
The facile conversion of an aldehyde directly into an acyl
fluoride is a rare example of this potentially useful reaction,
which was previously limited to strongly electrophilic fluori-
nating reagents (e.g., BrF3,[32] CsSO4F[33]).
[7] T. Barker, D. L. Boger, J. Am. Chem. Soc. 2012, 134, 13588 –
13591.
[8] D. A. Watson, M. Su, G. Teverovskiy, Y. Zhang, J. Garcia-
Fortanet, T. Kinzel, S. L. Buchwald, Science 2009, 325, 1661 –
1664.
[9] E. Lee, A. S. Kamlet, D. C. Powers, C. N. Neumann, G. B.
Boursalian, T. Furuya, D. C. Choi, J. M. Hooker, T. Ritter,
Science 2011, 334, 639 – 642.
[10] M.-G. Braun, A. G. Doyle, J. Am. Chem. Soc. 2013, 135, 12990 –
12993.
[11] J.-B. Xia, C. Zhu, C. Chen, J. Am. Chem. Soc. 2013, 135, 17494 –
17500.
Considering the value of fluorinated amino acids to drug
discovery[34] and their potential utility in PET imaging,[35] we
examined the fluorination of the methyl esters derived from
valine and isoleucine. While irradiation of the free base of
valine methyl ester with TBADT/NFSI provided none of the
desired product, we were delighted to find that after 17 h,
reaction of the corresponding hydrochloride salt 37 pro-
ceeded smoothly to afford 3-fluorovaline (38). In this case, the
reaction was necessarily carried out in a mixture of MeCN/
H2O (2:1) to ensure solubility of all reagents. Likewise,
overnight irradiation of a solution of the hydrochloride salt of
isoleucine methyl ester with TBADT/NFSI resulted in the
formation of a diastereomeric mixture of 4-fluoro isoleucine
derivatives 35 along with a small amount of the 3-fluoro
amino ester 36.
[12] W. Liu, X. Huang, M.-J. Cheng, R. J. Nielson, W. A. Goddard III,
J. T. Groves, Science 2012, 337, 1322 – 1325.
[13] S. Bloom, C. R. Pitts, D. Curtin Miller, N. Haselton, M. G. Holl,
E. Urheim, T. Lectka, Angew. Chem. 2012, 124, 10732 – 10735;
Angew. Chem. Int. Ed. 2012, 51, 10580 – 10583.
[14] Y. Amaoka, M. Nagatomo, M. Inoue, Org. Lett. 2013, 15, 2160 –
2163.
[15] R. D. Chambers, M. Parsons, G. Sandford, R. Bowyden, Chem.
Commun. 2000, 959 – 960.
[16] M. Rueda-Becerril, C. C. Sazepin, J. C. T. Leung, T. Okbinoglu,
P. Kennepohl, J.-F. Paquin, G. M. Sammis, J. Am. Chem. Soc.
2012, 134, 4026 – 4029.
[17] J. C. T. Leung, C. Chatalova-Sazepin, J. G. West, M. Rueda-
Becerril, J.-F. Paquin, G. M. Sammis, Angew. Chem. 2012, 124,
10962 – 10965; Angew. Chem. Int. Ed. 2012, 51, 10804 – 10807.
[18] H. Teare, E. G. Robins, E. Arstad, S. K. Luthra, V. Gouverneur,
Chem. Commun. 2007, 2330 – 2332.
In summary, exploiting the hydrogen abstraction capa-
bility of a readily prepared and inexpensive decatungstate
photocatalyst in combination with the fluorine-transfer agent
[19] C. L. Hill, J. Mol. Catal. A 2007, 262, 2 – 6.
[20] M. D. Tzirakis, I. N. Lykakis, M. Orfanopoulos, Chem. Soc. Rev.
2009, 38, 2609 – 2621.
[21] C. L. Hill, Synlett 1995, 127 – 132.
NFSI, we have developed a convenient method for the
[22] C. Tanielian, Coord. Chem. Rev. 1998, 178 – 180, 1165 – 1181.
[23] S. Protti, D. Ravelli, M. Fagnoni, A. Albini, Chem. Commun.
2009, 7351 – 7353.
[24] I. N. Lykakis, C. Tanielian, R. Seghrouchni, M. Orfanopoulos, J.
Mol. Catal. A 2007, 262, 176 – 184.
[25] L. A. Combs-Walker, C. L. Hill, J. Am. Chem. Soc. 1992, 114,
938 – 946.
[26] B. S. Jaynes, C. L. Hill, J. Am. Chem. Soc. 1995, 117, 4704 – 4705.
[27] R. F. Renneke, M. Pasquali, C. L. Hill, J. Am. Chem. Soc. 1990,
112, 6585 – 6594.
3
À
fluorination of unactivated C(sp ) H bonds. This new process
provides direct access to a variety of fluorinated organics,
including natural products, acyl fluorides, and fluorinated
amino acid derivatives, many of which would otherwise be
inaccessible or require multistep reaction sequences. Consid-
ering the steric and electronic bias of hydrogen atom
abstraction by the photoexcited decatungstate, an important
À
aspect of this reaction system is its ability to fluorinate C H
bonds prone to oxidative metabolism, a potentially enabling
capability for medicinal chemistry. Additionally, fluorination
of other amino acid derivatives or indeed small peptides with
[18F]NFSI may well provide new radiotracers for PET imaging
or improved procedures for preparing existing imaging
agents.
[28] D. C. Duncan, T. L. Netzel, C. L. Hill, Inorg. Chem. 1995, 34,
4640 – 4646.
[29] D. Dondi, M. Fagnoni, A. Albini, Chem. Eur. J. 2006, 12, 4153 –
4163.
[30] Typical experimental procedure (fluorination of (À)-bornyl
acetate): TBADT (0.02 equiv), NFSI (113 mg, 1.2 equiv), and
NaHCO3 (3 mg, 0.1 equiv) were added to a 2m solution of (À)-
bornyl acetate (59 mg, 0.3 mmol) in acetonitrile (0.15 mL). The
resulting suspension was sparged with N2 for 5 min. The reaction
vessel was then sealed, and the mixture was stirred while being
irradiated between two 15 Watt UVB-BLB lamps (centered at
l = 365 nm) at room temperature for 17 h. After work-up and
purification (see the Supporting Information), the major fluo-
rinated isomer 2 was obtained as a colorless oil (26 mg, 40%).
[31] D. Ravelli, M. Zema, M. Mella, M. Fagnoni, A. Albini, Org.
Biomol. Chem. 2010, 8, 4158 – 4164.
[32] M. Ochiai, A. Yoshimura, M. M. Hoque, T. Okubo, M. Saito, K.
Miyamoto, Org. Lett. 2011, 13, 5568 – 5571.
[33] S. Stavber, Z. Planinsek, M. Zupan, J. Org. Chem. 1992, 57,
5334 – 5337.
[34] X.-L. Qui, F.-L. Qing, Eur. J. Org. Chem. 2011, 3261 – 3278.
[35] P. Laverman, O. C. Boerman, F. H. M. Corstens, W. J. G. Oyen,
Eur. J. Nucl. Med. 2002, 29, 681 – 690.
Received: January 14, 2014
Published online: && &&, &&&&
À
Keywords: amino acids · C H activation · photochemistry ·
polyoxometalates · tungsten
.
[1] W. K. Hagmann, J. Med. Chem. 2008, 51, 4359 – 4369.
[2] K. L. Kirk, Org. Process Res. Dev. 2008, 12, 305 – 321.
[3] S. M. Ametamey, M. Honer, P. A. Schubiger, Chem. Rev. 2008,
108, 1501 – 1516.
[4] D. OꢀHagan, J. Fluorine Chem. 2010, 131, 1071 – 1081.
[5] J. Fried, E. F. Sabo, J. Am. Chem. Soc. 1954, 76, 1455 – 1456.
[6] T. Liang, C. N. Neumann, T. Ritter, Angew. Chem. 2013, 125,
8372 – 8423; Angew. Chem. Int. Ed. 2013, 52, 8214 – 8264.
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
À
C H Activation
S. D. Halperin, H. Fan, S. Chang,
R. E. Martin, R. Britton*
&&&& — &&&&
A Convenient Photocatalytic Fluorination
À
of Unactivated C H Bonds
The direct fluorination of unactivated
a wide range of fluorinated organic
3
À
C(sp ) H bonds is catalyzed by the in-
molecules, including structurally complex
natural products, acyl fluorides, and
fluorinated amino acids.
expensive photocatalyst tetrabutylammo-
nium decatungstate (TBADT). This con-
venient reaction provides direct access to
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!