Photocatalytic hydrodefluorination: Facile access to partially fluorinated aromatics

Article abstract of DOI:10.1021/ja500031m

Polyfluorinated aromatics are essential to materials science as well as the pharmaceutical and agrochemical industries and yet are often difficult to access. This Communication describes a photocatalytic hydrodefluorination approach which begins with easily accessible perfluoroarenes and selectively reduces the C-F bonds. The method allows facile access to a number of partially fluorinated arenes and takes place with unprecedented catalytic activity using a safe and inexpensive amine as the reductant.

Full text of DOI:10.1021/ja500031m

Communication
pubs.acs.org/JACS
Photocatalytic Hydrodefluorination: Facile Access to Partially
Fluorinated Aromatics
Sameera M. Senaweera,‡ Anuradha Singh,‡ and Jimmie D. Weaver*
Department of Chemistry, Oklahoma State University, 107 Physical Science, Stillwater, Oklahoma 74078, United States
S
* Supporting Information
Scheme 1. Comparison of Current and Proposed Strategy to
Partially Fluorinated Arenes
ABSTRACT: Polyfluorinated aromatics are essential to
materials science as well as the pharmaceutical and
agrochemical industries and yet are often difficult to
access. This Communication describes a photocatalytic
hydrodefluorination approach which begins with easily
accessible perfluoroarenes and selectively reduces the C−F
bonds. The method allows facile access to a number of
partially fluorinated arenes and takes place with
unprecedented catalytic activity using a safe and
inexpensive amine as the reductant.
artially fluorinated arenes are highly prized molecules in
Ppharmaceutical and agricultural chemistry, and in materials
science. A recent report found that 20−25% of all compounds
in the drug pipeline contained fluorine,1 including top selling
drugs such as Diflunisal (I, Figure 1), Januvia (II), and antiviral
Emtricitabine (IV). Fluorinated arenes are commonly found in
agrochemicals (Teflubenzuron III), and fluorination is
frequently used to modulate the properties of liquid crystals
(V).
catalysis of readily available perfluoroarenes (iii) as a new
strategy to access partially fluorinated arenes.
Realization of a simple and robust hydrodefluorination
(HDF) methodology has been a struggle.4 HDF chemistry is
inherently difficult for a number of reasons, including the
relative inertness of the C−F bond to many insertion processes
which, if successful, face problematic strong catalyst−F bonds,
which often hamper catalyst turnover acting as a thermo-
dynamic sink in the catalytic cycle. In attempts to develop a
robust HDF catalytic system, a number of different types of
catalysts have been investigated, including catalysts based on
metal−hydrido complexes,5 Rh,6 Ni,7 Pd,8 and Au.9 Further-
more, most methods use extremely fluorophilic reductants,
such as pyrophoric Al-hydrides10 or expensive Si-hydrides11 to
help alleviate this problem. We speculated that if we used a
catalyst incapable of formation of M−F bonds, we might be
able to circumvent these issues altogether and facilitate HDF
with high catalytic turnover number (TON).
Perfluoroarenes have a lowered LUMO compared to their
hydrocarbon analogs12 and offer a convenient mode of
activation. It is established that addition of an electron into
the LUMO13 generates a radical anion14 which undergoes
fluoride extrusion15 to generate a radical which abstracts an H-
atom.16 We posited that photoredox catalysiswhich has not
been explored for HDF of fluorinated arenes17might have a
distinct advantage over traditional methods in generating both
the requisite radical anion as well as the H-atom source (eq 1,
Scheme 2).16a,d,e,18 Specifically, we speculated that tris[2-
Figure 1. Some important partially fluorinated pharamaceuticals,
agrochemicals and a liquid crystal.
Despite the value of these compounds their syntheses are
often lengthy and rely on harsh methods such as the Balz−
Schiemann reaction or the Halex process. In recognition of the
need for more facile access to partially fluorinated arenes,
catalytic fluorination of prefunctionalized arenes2 and directed
C−H fluorinations3 have received considerable attention.
Arguably, these methods are not well suited for the synthesis
of polyfluorinated arenes because the starting materials would
require multiple prefunctionalizations (i, Scheme 1) or uniquely
arranged directing groups (ii). Herein we present photoredox
Received: January 2, 2014
Published: February 18, 2014
© 2014 American Chemical Society
3002
dx.doi.org/10.1021/ja500031m | J. Am. Chem. Soc. 2014, 136, 3002−3005

Journal of the American Chemical Society
Communication
Scheme 2. Potential Mechanism
Table 2. Scope of the Photocatalytic HDF
phenylpyridinato-C2,N]iridium(III), [Ir(ppy)3] could be used
to facilitate outersphere electron transfer to generate a
perfluororadical anion (B, Scheme 2) which results in
fragmentation to give radical C19 and a fluoride. Finally, C
abstracts an H-atom from the amine radical cation20 generated
in situ giving D and an iminium.21 Because the catalyst is a
robust 18-electron complex and coordinatively saturated with
3-bidentate ligands, it avoids the problematic M−F bond which
prevents decomposition and allows facile turnover. Addition-
ally, we speculated this might alleviate the need for a
fluorophilic reductant such as Al- or Si-hydride and allow us
to use a simple and safer aliphatic amine as a reductant.
Thus, we began our investigation with using Ir(ppy)3 (Ir(II)/
(III), V = −2.20)22 and pentafluoro pyridine (V = −2.12)23
which provides a slight overpotential. We were pleased to find
full conversion to 2,3,5,6-tetrafluoropyridine (2a) after 24 h
(entry 1, Table 1). Remarkably, Ru(bpy)3Cl2 and Ru(bpm)3Cl2
Table 1. Optimization of Reaction Conditions
entry
modification of conditions
none
Ru(bpy)3 instead of Ir(ppy)3
conversion
a
time, h
24
1
100%
a
2
37%
24
a
3
Ru(bpm)3 instead of Ir(ppy)3 2−3%
without degassing
24
a
a
d
f
b
c
19F NMR yield. 65 °C, 72 h. From 3-Cl tetrafluoropyridine.
4
0%
24
e
a
5
no catalyst or in dark
1.2 equiv of amine
2.0 equiv of amine
0.05 mol% catalyst
0.125 mol% catalyst
0.25 mol% catalyst
0.075 M substrate conc
0.10 M substrate conc
0.15 M substrate conc
0.20 M substrate conc
0%
24
Contains 10% di-HDF product. Contains 9% regioisomer.
Incomplete conversion: 2c, 57% conv.; 2g, 70% conv.; 2t, 60%
a
a
6
32/100%
35/100%
0.5/22.5
0.5/22.5
0.5/1
0.5/1
0.5/1
0.5/6
0.5/6
0.5/6
0.5/6
0.5/6
g
conv. 36 h.
7
b
8
5/12%
b
9
25/43%
tolerates a number of functional groups including CF3 (2b, 2n,
and 2r), ketones (2d), esters (2e, 2i, and 2q), nitriles (2h, 2m),
oxazoles (2j), and aliphatic amines26 (2v). The inherent
selectivity can be overcome by starting with a chloride (2l vs
2a). Perfluoroarenes substituted with electron releasing amino,
alcohol, or thiol groups proved to be sluggish or unselective
substrates. However, in conjunction with appropriate functional
groups the sluggish substrates underwent facile reaction. For
instance, symmetrical ether (2g) provides the mono-HDF
product. Alternatively, electronic modulation of substrates can
activate them toward HDF. Thiols become active after
formation of the thiocarbonate (2o) and similarly amines are
activated by acylation (2k). Other ring functionality can also
serve to activate substrates with electron releasing groups (2m,
2n, 2p, 2q, and 2r). Dibromide 2g demonstrates the propensity
to undergo HDF rather than C−Br reduction.27 Highly
fluorinated phosphines, which are an important class of ligand
for several catalytic processes,28 also undergo HDF allowing
facile fine-tuning of the electronics. Fused arenes, which are
important in materials science applications,29 undergo smooth
HDF (2t, 2u, and 2v).
b
10
11
12
13
14
15
47/82%
b
b
b
b
b
0.015/0.050 mmol 2a
0.018/0.063 mmol 2a
0.017/0.044 mmol 2a
0.022/0.058 mmol 2a
0.032/0.070 mmol 2a
0.35 M substrate conc
a
b
Determined by GC. Determined by 19F NMR.
[bpm = 2,2′-bipyrimidine] catalysts sluggishly afforded product
(entries 2 and 3), despite a significant underpotential (V =
−1.33, −0.91).24 Control studies demonstrated the necessity of
both catalyst and light (entry 5). The rate of the reaction did
not appear to depend greatly on the equivalents of amine
(entries 6 and 7) and reached completion using just 1.2 equiv
of amine. There appears to be a rate dependency on catalyst
concentration (entries 8−10).25 Peculiarly, the reaction seems
to have a pseudo-zero-order rate dependency on the substrate
(entries 11−15).25 Ultimately, using 0.1 M substrate, 1.1−3.3
equiv of Hunig’s base, and 0.5 mol% photocatalyst, we began to
evaluate the substrate scope.
Under these conditions a number of perfluoroarenes
underwent smooth HDF (Table 2). Importantly, the reaction
̈
Under these conditions several substrates can undergo di- or
even tri-HDF events to afford products that would be difficult
3003
dx.doi.org/10.1021/ja500031m | J. Am. Chem. Soc. 2014, 136, 3002−3005

Journal of the American Chemical Society
Communication
Table 3. Di- and Tri-HDF
a
b
19F NMR yield. 42 h.
Figure 2. HDF in a flow reactor.
to access by other means. Conveniently, in most cases30 the
difference in rate constants for mono- and di-reduction is
substantial and selectivity can be controlled by the equivalents
of amine used as well as close monitoring of the reaction.
Octafluorotoluene undergoes ortho,para di-reduction, leaving
the benzylic fluorines intact (3b, Table 3). While decafluoro-
biphenyl gives symmetric 4,4′ reduction in excellent yield (3w).
Interestingly, removal of the first fluorine of 1r occurs ortho to
the CF3 group, the second HDF event occurs meta to give a
2,5-di-H arene (3r). While the regioselectivity of the HDF
event is primarily dictated by the electronics of the
perfluorarene,16a,31 the rate difference in the di-HDF of the
methyl (3e) and tert-butyl esters (3i) is substantial and suggests
that there is a steric contribution.14b,35 Finally, perfluorotri-
phenyl phosphine undergoes smooth conversion to the tri-
HDF product (3x) which allows direct synthesis from
commercially available phosphine.
chiller through the condenser. Pentafluoropyridine (1a, Figure
2) underwent smooth conversion, producing 2a at a rate of
0.27 mmol/h (978 mg/24 h). Octafluoronapthalene (1c),
which is a sluggish substrate that never reaches full conversion
under standard conditions was also subjected to the flow
reactor. While full conversion was never reached even in under
flow conditions, we did see a 3.75× rate enhancement in
product formation (2c, flow vs batch), which is likely due, in
part, to the diminished path length of the light.33,34 These
results suggest the method is convertible to flow and larger
quantities of HDF product could be obtained when desired.
In summary, we have introduced photocatalytic HDF as a
viable method to access polyfluorinated arenes and an
alternative to current HDF strategies in which we demonstrated
the importance of avoiding the metal-F bond for catalytic
turnover. One desirable outcome is the use of safer and less
expensive amine reductant than the traditional Al- and Si-H’s.
Importantly, we have demonstrated that with a single catalyst
we can perform mono-HDF, di-HDF, and even tri-HDF on a
number of different types of arenes to give facile access to
polyfluorinated aromatic rings, a task for which selective
fluorination is poorly suited. Finally, we have demonstrated that
the chemistry is amenable to flow methods, suggesting that it
could be used to make sizable quantities of polyfluorinated
arenes in a convenient, safe, and cost-effective manner. Given
the ease with which the requisite perfluoroarenes can be
prepared via SNAr chemistry, we believe this method has great
potential to help in a number of research areas.
We next investigated the robustness of the photocatalyst by
evaluating the TON using pentafluoropyridine (1a, eq 2).
Aliquots (7 total) of 1a and Hunig’s base were repeatedly
̈
added over the 4 day period.25 To our delight, the catalyst
proved remarkably resilient and never ceased to produce
product. Eventually the rate of a byproduct formation began to
increase at which time an internal standard was added and a
yield of 82% was obtained to give an unprecedented TON of
22,550.4,10,32
ASSOCIATED CONTENT
* Supporting Information
Experimental details and characterization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
HDF products are important building blocks for a number of
different applications. As such, obtaining larger quantities of
these products will be necessary. Consequently, we wanted to
investigate the feasibility of our method to be converted to a
flow method. Thus, utilizing a homemade flow reactor,
AUTHOR INFORMATION
Corresponding Author
jimmie.weaver@okstate.edu
■
́
essentially like the one previously described by Gagne (Figure
2).33 The reaction mixture was pumped by an HPLC pump
into perfluoroalkoxy tubing which was wrapped around a
condenser.25 Within the condenser were outward facing blue
LED strips. The temperature was controlled by passing a
mixture of water and ethylene glycol via a recirculating heater/
Author Contributions
‡S.M.S. and A.S. contributed equally.
Notes
The authors declare no competing financial interest.
3004
dx.doi.org/10.1021/ja500031m | J. Am. Chem. Soc. 2014, 136, 3002−3005

Journal of the American Chemical Society
Communication
(19) Experiments suggesting the intermediacy of a radical were
performed; see SI for details.
ACKNOWLEDGMENTS
■
This work was supported by startup funds from Oklahoma
State University. We are grateful to AAG for reading and
providing feedback for this manuscript.
(20) Using d-MeCN gives no deuterium incorporation into the
product while use of deuterium-labeled dicyclohexyl-d5-ethylamine
gives 32% deuterium incorporation into the mono-HDF product.
These results suggest that either α-C-H of the amine radical cation can
serve as the H-atom source. See SI for details.
(21) We believe iminium ions are initially formed and serve as the
fluoride counterion. However, we have not been able to isolate these
salts. Under rigorously anhydrous conditions we have observed
bis(diisopropylammonium) hexafluorosilicate precipitating from sol-
ution when reactions are run in borosilicate tubes. Presumably this
comes from fluoride etching silicon from the tube walls and hydrolysis
of the ethyl group of the iminium. In reactions run under less rigorous
conditions, we believe advantageous water leads to hydrolysis of the
iminium prior to silicon etching to afford secondary amine HF salts.
See SI for details and crystal structure.
REFERENCES
■
(1) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Chem. Soc.
Rev. 2008, 37, 320.
(2) (a) Mazzotti, A. R.; Campbell, M. G.; Tang, P.; Murphy, J. M.;
Ritter, T. J. Am. Chem. Soc. 2013, 135, 14012. (b) Lee, H. G.; Milner,
P. J.; Buchwald, S. L. Org. Lett. 2013, 15, 5602. (c) Ichiishi, N.; Canty,
A. J.; Yates, B. F.; Sanford, M. S. Org. Lett. 2013, 15, 5134. (d) Fier, P.
S.; Luo, J.; Hartwig, J. F. J. Am. Chem. Soc. 2013, 135, 2552. (e) Ye, Y.;
Schimler, S. D.; Hanley, P. S.; Sanford, M. S. J. Am. Chem. Soc. 2013,
135, 16292. (f) Tang, P.; Furuya, T.; Ritter, T. J. Am. Chem. Soc. 2010,
132, 12150. (g) Fier, P. S.; Hartwig, J. F. J. Am. Chem. Soc. 2012, 134,
10795.
(22) Flamigni, L.; Barbieri, A.; Sabatini, C.; Ventura, B.; Barigelletti,
F. Top. Curr. Chem. 2007, 281, 143.
(23) Beletskaya, I. P.; Artamkina, G. A.; Mil’chenko, A. Y.; Sazonov,
P. K.; Shtern, M. M. J. Phys. Org. Chem. 1996, 9, 319.
(24) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.;
von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85.
(3) (a) Wang, X.; Mei, T.-S.; Yu, J.-Q. J. Am. Chem. Soc. 2009, 131,
7520. (b) Chan, K. S. L.; Wasa, M.; Wang, X.; Yu, J.-Q. Angew. Chem.,
Int. Ed. 2011, 50, 9081. (c) Truong, T.; Klimovica, K.; Daugulis, O. J.
Am. Chem. Soc. 2013, 135, 9342.
(25) For details, see the Supporting Information.
(4) Lentz, D.; Braun, T.; Kuehnel, M. F. Angew. Chem., Int. Ed. 2013,
(26) Conceivably, access to the less reactive ortho position is possible
by using the amino group as a temporary blocking group which adds
with similar preferences. After the HDF reaction is complete, the
amino group can be converted to the fluorine via diazotization.
(27) A small amount ∼7% of hydrodebromination product was
observed. This was observed previously: Nguyen, J. D.; D’Amato, E.
M.; Narayanam, J. M. R.; Stephenson, C. R. J. Nature Chem. 2012, 4,
854.
52, 3328.
(5) (a) Edelbach, B. L.; Fazlur Rahman, A. K.; Lachicotte, R. J.; Jones,
W. D. Organometallics 1999, 18, 3170. (b) Arndt, P.; Spannenberg, A.;
Baumann, W.; Burlakov, V. V.; Rosenthal, U.; Becke, S.; Weiss, T.
Organometallics 2004, 23, 4792.
(6) (a) Aizenberg, M.; Milstein, D. J. Am. Chem. Soc. 1995, 117, 8674.
(b) Aizenberg, M.; Milstein, D. Science 1994, 265, 359.
(7) (a) Archibald, S. J.; Braun, T.; Gaunt, J. A.; Hobson, J. E.; Perutz,
(28) (a) K. Morita, D.; A. David, S.; Tumas, W.; K. Morita, D.; R.
Pesiri, D.; H. Glaze, W. Chem. Commun. 1998, 1397. (b) Hirai, T.;
Hamasaki, A.; Nakamura, A.; Tokunaga, M. Org. Lett. 2009, 11, 5510.
(29) For a recent example, see; Zhang, J.; Chen, W.; Rojas, A. J.;
Jucov, E. V.; Timofeeva, T. V.; Parker, T. C.; Barlow, S.; Marder, S. R.
J. Am. Chem. Soc. 2013, 135, 16376.
(30) Of the substrates tried, the rate differences between the first and
sequential reductions were sufficiently different to be able to isolate the
desired HDF product. Two exceptions were tri(pentafluorophenyl)
phosphine (1x) and the decafluorobiphenyl (1w).
(31) We believe the addition of the electron occurs at the LUMO
which resembles primarily π* and that fragmentation then occurs
upon crossover to the lowest lying C−F σ* orbital. For a better
explanation of this process, see; ref 16a. In the case of 1l (Ar-Cl),
presumably the C−Cl σ* is lower in energy and hence fragmention
occurs at this carbon. Alternatively, it is possible that the C−Cl σ*
orbital is low enough in energy that it is the LUMO in which case the
π* orbital may not be involved at all.
R. N. J. Chem. Soc., Dalton Trans. 2000, 2013. (b) Fischer, P.; Gotz, K.;
Eichhorn, A.; Radius, U. Organometallics 2012, 31, 1374.
̈
(8) Breyer, D.; Braun, T.; Klaring, P. Organometallics 2012, 31, 1417.
̈
(9) Zhan, J.-H.; Lv, H.; Yu, Y.; Zhang, J.-L. Adv. Synth. Catal. 2012,
354, 1529.
(10) For an example of an Al-H reductant, see: Jager-Fiedler, U.;
̈
Klahn, M.; Arndt, P.; Baumann, W.; Spannenberg, A.; Burlakov, V. V.;
Rosenthal, U. J. Mol. Catal. A: Chem. 2007, 261, 184.
(11) For an example of an R3Si-H reductant, see: Reade, S. P.;
Mahon, M. F.; Whittlesey, M. K. J. Am. Chem. Soc. 2009, 131, 1847.
(12) Gant, K. S.; Christophorou, L. G. J. Chem. Phys. 1976, 65, 2977.
(13) (a) Clot, E.; Eisenstein, O.; Jasim, N.; Macgregor, S. A.;
McGrady, J. E.; Perutz, R. N. Acc. Chem. Res. 2011, 44, 333. (b) Sun,
A. D.; Love, J. A. Dalton Trans. 2010, 39, 10362. (c) Edelbach, B. L.;
Kraft, B. M.; Jones, W. D. J. Am. Chem. Soc. 1999, 121, 10327.
(d) Kiplinger, J. L.; Richmond, T. G. J. Am. Chem. Soc. 1996, 118,
1805. (e) Kiplinger, J. L.; Richmond, T. G.; Osterberg, C. E. Chem.
Rev. 1994, 94, 373.
(14) (a) Yim, M. B.; Wood, D. E. J. Am. Chem. Soc. 1976, 98, 2053.
(b) Shchegoleva, L. N.; Bilkis, I. I.; Schastnev, P. V. Chem. Phys. 1983,
82, 343. (c) Shoute, L. C. T.; Mittal, J. P. J. Phys. Chem. 1993, 97,
379−384.
(15) (a) Rossi, R. A. Acc. Chem. Res. 1982, 15, 164. (b) Freeman, P.
K.; Srinivasa, R. J. Org. Chem. 1987, 52, 252.
(16) (a) Selivanova, G. A.; Reshetov, A. V.; Beregovaya, I. V.;
Vasil’eva, N. V.; Bagryanskaya, I. Y.; Shteingarts, V. D. J. Fluorine Chem.
2012, 137, 64. (b) Laev, S. S.; Gurskaya, L. Y.; Selivanova, G. A.;
Beregovaya, I. V.; Shchegoleva, L. N.; Vasil’eva, N. V.; Shakirov, M.
M.; Shteingarts, V. D. Eur. J. Org. Chem. 2007, 2007, 306. (c) Jones, W.
D. Dalton Trans. 2003, 2003, 3991. (d) Krasnov, B. I.; Platonov, V. E.
Russ. J. Org. Chem. 2001, 37, 517. (e) Laev, S. S.; Shteingarts, V. D. J.
Fluorine Chem. 1998, 91, 21. (f) Krasnov, V. I.; Platonov, V. E.;
Beregovaya, I. V.; Shohegoleva, L. N. Tetrahedron 1997, 53, 1797.
(17) For an example of a photochemical defluorination of an
electron-rich monofluoroarene, see; Dichiarante, V.; Fagnoni, M.;
Albini, A. Green Chem. 2009, 11, 942.
(32) Lv, H.; Cai, Y.-B.; Zhang, J.-L. Angew. Chem., Int. Ed. 2013, 52,
3203.
(33) Andrews, R. S.; Becker, J. J.; Gagne,
2012, 51, 4140.
́
M. R. Angew. Chem., Int. Ed.
(34) (a) Tucker, J. W.; Zhang, Y.; Jamison, T. F.; Stephenson, C. R. J.
Angew. Chem., Int. Ed. 2012, 51, 4144. (b) Neumann, M.; Zeitler, K.
Org. Lett. 2012, 14, 2658. (c) Lu, H.; Schmidt, M. A.; Jensen, K. F. Lab
Chip 2001, 1, 22.
(35) The structure of the radical anion is not planar. Consequently,
substitutions can destabilize the structure of the radical anion.
Destabilization of the radical anion would be expected to increase
the barrier to formation of the requisite radical anion. For a more
thorough study of the structure of the radical anion of hexafluor-
obenzene, see ref 14b.
(18) Laev, S. S.; Shteingarts, V. D. J. Fluorine Chem. 1999, 96, 175.
3005
dx.doi.org/10.1021/ja500031m | J. Am. Chem. Soc. 2014, 136, 3002−3005