An improved catalyst system for the Pd-catalyzed fluorination of (hetero)aryl triflates

Article abstract of DOI:10.1021/ol402859k

The stable Pd(0) species [(1,5-cyclooctadiene)(L·Pd)₂] (L = AdBrettPhos) has been prepared and successfully evaluated as a precatalyst for the fluorination of aryl triflates derived from biologically active and heteroaryl phenols, challenging substrates for our previously reported catalyst system. Additionally, this precatalyst activates at room temperature under neutral conditions, generates 1,5-cyclooctadiene as the only byproduct, and leads to overall cleaner reaction profiles.

Full text of DOI:10.1021/ol402859k

ORGANIC
LETTERS
2013
Vol. 15, No. 21
5602–5605
An Improved Catalyst System for the
Pd-Catalyzed Fluorination
of (Hetero)Aryl Triflates
Hong Geun Lee,† Phillip J. Milner,† and Stephen L. Buchwald*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge,
Massachusetts 02139, United States
sbuchwal@mit.edu
Received October 3, 2013
ABSTRACT
The stable Pd(0) species [(1,5-cyclooctadiene)(L Pd)2] (L = AdBrettPhos) has been prepared and successfully evaluated as a precatalyst for the
3
fluorination of aryl triflates derived from biologically active and heteroaryl phenols, challenging substrates for our previously reported catalyst system.
Additionally, this precatalyst activates at room temperature under neutral conditions, generates 1,5-cyclooctadiene as the only byproduct, and leads to
overall cleaner reaction profiles.
Fluorination of aromatic rings is a widely used strat-
egy for modifying the biological activities of potential
pharmaceutical and agrochemical agents.1 In addition,
18F-substituted compounds are important radiotracers
for positron emission tomography (PET).2 Aryl fluorides
are typically installed early in a target molecule’s synthe-
sis using the harsh BalzꢀSchiemann reaction, making
the synthesis of 18F-radiotracers and highly functionalized
fluorinated materials difficult. Although a number of
methods for electrophilic aryl fluorination with Ag,3 Pd,4
and Cu5 catalysts, and without added transition metals,6
have been developed to address this need, these reac-
tions typically do not tolerate easily oxidizable functional
groups such as tertiary amines and electron-rich hetero-
cycles, result in 5ꢀ50% reduction of the starting material,
and/or require the synthesis of unstable or toxic organo-
metallic reagents. The direct transformation of aryl
(pseudo)halides to aryl fluorides using a metal fluoride
salt is a promising alternative to electrophilic fluorination
in terms of generality and practicality7 that has received
less attention than electrophilic fluorination methods.8
To this end, we reported the successful coupling of aryl
triflates with CsF using a Pd catalyst based on the bulkyl
biaryl phosphine ligand tBuBrettPhos (1) (Figure 1).9,10
† These authors contributed equally.
(1) (a) Purse, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Chem.
€
Soc. Rev. 2008, 37, 320. (b) Muller, K.; Faeh, C.; Diederich, F. Science
2007, 317, 1881. (c) Kirk, K. L. Org. Process Res. Dev. 2008, 12, 305.
(2) (a) Lee, E.; Hooker, J. M.; Ritter, T. J. Am. Chem. Soc. 2012, 134,
17456. (b) Lee, E.; Kamlet, A. S.; Choi, D. C.; Hooker, J. M.; Ritter, T.
Science 2011, 334, 639. (c) Miller, P. W.; Long, N. J.; Vilar, R.; Gee,
A. D. Angew. Chem., Int. Ed. 2008, 47, 8998.
(3) (a) Tang, P.; Ritter, T. Tetrahedron 2011, 67, 449. (b) Tang, P.;
Furuya, T.; Ritter, T. J. Am. Chem. Soc. 2010, 132, 12150. (c) Furuya, T.;
Ritter, T. Org. Lett. 2009, 11, 2860. (d) Furuya, T.; Strom, A. E.; Ritter,
T. J. Am. Chem. Soc. 2009, 131, 1662.
(5) (a) Ye, Y.; Sanford, M. S. J. Am. Chem. Soc. 2013, 135, 4648. (b)
Fier, P. S.; Luo, J.; Hartwig, J. F. J. Am. Chem. Soc. 2013, 135, 2552.
(6) (a) Anbarasan, P.; Neumann, H.; Beller, M. Angew. Chem., Int.
Ed. 2010, 49, 2219. (b) Yamada, S.; Gavryushin, A.; Knochel, P. Angew.
Chem., Int. Ed. 2010, 49, 2215.
(4) For catalytic examples, see: (a) Mazzotti, M.; Campbell, M.;
Tang, P.; Murphy, J.; Ritter, T. J. Am. Chem. Soc. 2013, 135, 14012. (b)
Chan, K. S. L.; Wasa, M.; Wang, X.; Yu, J.-Q. Angew. Chem., Int. Ed.
2011, 50, 9081. (c) Wang, X.; Mei, T.-S.; Yu, J.-Q. J. Am. Chem. Soc.
2009, 131, 7520. (d) Hull, K. L.; Anani, W. Q.; Sanford, M. S. J. Am.
Chem. Soc. 2006, 128, 7134. For stoichiometric examples, see: (e)
Furuya, T.; Benitez, D.; Tkatchouk, E.; Strom, A. E.; Tang, P.;
Goddard, T., III; Ritter, T. J. Am. Chem. Soc. 2010, 132, 3793. (f) Ball,
N. D.; Sanford, M. S. J. Am. Chem. Soc. 2009, 131, 3796. (g) Furuya, T.;
Ritter, T. J. Am. Chem. Soc. 2008, 130, 10060. (h) Furuya, T.; Kaiser,
H. M.; Ritter, T. Angew. Chem., Int. Ed. 2008, 47, 5993.
(7) Grushin, V. V. Acc. Chem. Res. 2010, 43, 160.
(8) (a) Fier, P. S.; Hartwig, J. F. J. Am. Chem. Soc. 2012, 134, 10795.
ꢀ
(b) Casitas, A.; Canta, M.; Sola,þM.; Costas, M.; Ribas, X. J. Am. Chem.
Soc. 2011, 133, 19386. For Ar2I reagents, see: (c) Ichiishi, N.; Cantry,
A. J.; Yates, B. F.; Sanford, M. S. Org. Lett. 2013, 15, 5134.
€
(9) (a) Noel, T.; Maimone, T. J.; Buchwald, S. L. Angew. Chem., Int.
Ed. 2011, 50, 8900. (b) Watson, D. A.; Su, M.; Teverovskiy, G.; Zhang,
Y.; Garcia-Fortanet, J.; Kinzel, T.; Buchwald, S. L. Science 2009, 325,
1661.
r
10.1021/ol402859k
Published on Web 10/18/2013
2013 American Chemical Society

catalyst based on 2 is likely due to the faster rate of
reductive elimination from PdꢀF intermediates bearing 2
compared to those bearing 1.11 However, the use of
[(cinnamyl)PdCl]2 as the source of active Pd requires
1.5 equiv of 2 relative to Pd to be added and results
in generation of 1 equiv of “Clꢀ”, which participates in a
competitive cross-coupling process to produce the corre-
sponding aryl chloride. In some cases, this side product can
be difficult to separate from the desired aryl fluoride
product. Thus, we evaluated alternative Pd sources in
conjunction with 2. However, other Pd sources, such as
Pd(OAc)2 (Table 1, entry 3), Pd2(dba)3 (Table 1, entry 4),
and Pd(dba)2 (Table 1, entry 5), did not prove as effective
as [(cinnamyl)PdCl]2 for one or both of these substrates.12
In contrast to the use of separate sources of Pd and ligand,
we have found that precatalysts are superior in terms
of convenience, efficiency of catalyst generation, and the
use of only 1 equiv of ligand relative to Pd.13 However,
when our third generation palladacycle precatalyst 314
(Figure 1) was employed in the fluorination of estrone
and 3-quinolinyl triflates, low yields of the desired pro-
ductswereobtained(Table1, entry 6). The low yieldsresult
from arylation of the carbazole generated as the byproduct
of catalyst activation, resulting in an overall formation
of 2 equiv of HF, which is also detrimental to the reaction
Figure 1. Ligands and precatalysts for the Pd-catalyzed fluorina-
tion of aryl triflates.
However, there remains a strong need for the further
development of simple methods for aryl fluorination that
demonstrate a broad substrate scope and clean reaction
profiles.
Table 1. Fluorination with Catalysts Based on 1 or 2a
yield.15 Only when 2 Pd(4-nBuPh)OTf (4) was employed
3
as catalyst could high yields of these aryl fluoride products
be obtained without formation of the aryl chloride side
product or the need for excess 2 relative to Pd (Table 1,
entry 7). Based on these results, a precatalyst bearing 2
thatactivateswithout producing reactiveand/orinhibitory
byproducts, such as chloride, dba, carbazole, or HF, would
be ideal for this transformation.
In pursuit of such a species, we found that simply
combining equimolar quantities of 2 and [(1,5-COD)Pd-
(CH2TMS)2] (COD = cyclooctadiene) togetherinpentane
resultedin precipitation of a paleyellow solid (5) (Figure 2);
due to the lack of oxidant present in this reaction, we
anticipated that 5 was an isolable Pd(0) species. The filtrate
of this reaction mixture contained 53% of the starting
amount of COD16 and minimal quantities of 2, suggesting
that 5 possesses a 2:2:1 ratio of 2/Pd/1,5-COD. However,
assessing the structure of 5 by solution-state NMR proved
difficult due to its insolubility or instability in all tested
organic solvents (see Supporting Information).
a Reaction conditions: Aryl triflate (0.1 mmol), CsF (0.3 mmol),
4% “Pd”, tol (0.1 M), 120 °C, 12 h. 19F NMR yields. b Corresponding
ArCl detected by GC analysis.
Our original catalyst system of [(cinnamyl)PdCl]2/1
facilitates the catalytic fluorination of a variety of aryl
triflates with minimal formation (<5%) of the corre-
sponding reduction product.9b However, this method suf-
fers from poor reactivity with highly electron-rich and
heteroaryl substrates, such as estrone and 3-quinolinyl
triflates (Table 1, entry 1). After extensive investigation,
a catalyst system based on the related ligand AdBrettPhos
(2)11 (Figure 1) was found to be more capable in the
fluorination of these substrates (Table 1, entry 2), though
formation of two regioisomericarylfluorideswas observed
in the case of estrone triflate.9b The effectiveness of a
(12) When 0.1 equiv of dba was added to the reactions in Table 1,
entry 1, an identical yield of fluorodeoxyestrone (66%, R:β = 5:1), but a
diminished yield of 3-fluoroquinoline (40%), was observed. Thus, in the
case of fluorination of heteroaryl triflates, dba likely inhibits the desired
transformation. See: Amatore, C.; Broeker, G.; Jutand, A.; Khalil, F.
J. Am. Chem. Soc. 1997, 119, 5176.
(13) Bruno, N. C.; Tudge, M. T.; Buchwald, S. L. Chem. Sci. 2013,
4, 916and the references cited therein.
(14) Bruno, N. C.; Buchwald, S. L. Org. Lett. 2013, 15, 2879.
(15) In reactions conducted with 1, HF is also generated as a result of
30-arylation of the ligand. Our results suggest that HF is detrimental to
the yield; see: Maimone, T. J.; Milner, P. J.; Kinzel, T.; Zhang, Y.;
Takase, M. K.; Buchwald, S. L. J. Am. Chem. Soc. 2011, 133, 18106.
(16) GC analysis of the filtrate showed 11% of 1,3-COD, 33% 1,5-
COD, and 9% of what is likely 1,4-COD to be present (determined by
assuming an identical response factor that of 1,5-COD). Although
isomerization of alkenes in the presence of Pd(0) has not been reported,
isomerization in the presence of Pd(II) has been thoroughly studied.
(10) Notably, Ritter has recently disclosed a straightforward method
for transition-metal-free direct deoxyfluorination of phenols, but this
methodology requires anhydrous reaction conditions and stochoimetric
quantities of Phenofluor, which is expensive and moisture sensitive. See:
Tang, P.; Wang, W.; Ritter, T. J. Am. Chem. Soc. 2011, 133, 11482.
(11) Su, M.; Buchwald, S. L. Angew. Chem., Int. Ed. 2012, 51, 4710.
Org. Lett., Vol. 15, No. 21, 2013
5603

Scheme 1. Fluorination of Aryl Triflates Derived from
Biologically Active Phenolsa
Figure 2. Synthesis of Pd(0) species 5 from 2 and [(1,5-COD)-
Pd(CH2TMS)2].
Nonetheless, when exposed to 4-nBuPhOTf in toluene-
d8, 5 quantitatively converts to 2 (4-nBuPh)PdOTf (4)
3
in less than 10 min at room temperature under neutral
conditions, generating 0.5 equiv of 1,5-COD relative to Pd
in the process. Significantly, when this experiment was
repeated using 5 that had been exposed to air for 24 h on
1
the benchtop, the H NMR yield of 4 produced was only
diminished to 83%, suggesting that the half-life of 5 in air
is on the order of days. Based on the available structural
information, we propose a C2-symmetric structure for 5
wherein two trigonal-planar Pd(0) centers, each ligated by
a molecule of 2, are coordinated to a double bond of the 1,5-
COD ligand (Figure 2); a similar structure has been pro-
a Isolated yields, average of two runs. Reaction conditions unless
posed for the complex [(1,5-COD)(dippe Pd)2] (dippe =
3
otherwise noted: Aryl triflate (1 mmol), CsF (3 mmol), 5 (1ꢀ3%), tol
1,2-bis(diisopropylphosphino)ethane).17 Despite contain-
ing two reactive tricoordinate Pd(0) centers, 5 is indefinitely
stable at rt in a glovebox or when stored under N2 in a
benchtop desiccator. Collectively, these results suggest that
dissociation of the 1,5-COD ligand from 5 is very facile at rt,
even though 5 itself is relatively stable and easy to handle.
Most importantly, 5 is also an excellent precatalyst for the
fluorination of estrone and 3-quinolinyl triflates (Table 1,
entry 8), providing yields comparable to those obtained
with [(cinnamyl)PdCl]2/2 without generation of aryl chlor-
ide byproducts.18 By GC analysis, the only side products
observed were trace amounts (<5%) of the reduction
product and the corresponding biaryl ethers.9b In light of
these results, the ability of 5 to function as a precatalyst for
the fluorination of aryl triflates derived from heterocyclic
and biologically active phenols, the substrates for which
incorporation of fluorine centers is most important, was
further investigated (Schemes 1 and 2).
b
(0.1 M). Aryl triflate (0.5 mmol), CsF (1.5 mmol), 5 (1.5ꢀ3%), tol
(0.1 M). c Yield when conducted under same reaction conditions using
[(cinnamyl)PdCl]2/1 (Pd/1 = 1:1.5) instead of 5. Aryl triflate (0.1 mmol),
d
CsF (0.3 mmol), 19F NMR yield. Aryl triflate (0.5 mmol), CsF
(1.5 mmol), 5 (1%), cy (0.1 M). e Aryl triflate (0.5 mmol), CsF (3 mmol),
5 (3%), tol (0.1 M). f Aryl triflate (1 mmol), 5 (1.5ꢀ2%), cy (0.1 M).
pharmaceuticals could be accessed, including the antimi-
crobrial chloroxylenol (9),19 the chloretic 4-methylumbel-
liferone (Hymecromone, 11),20 the N-methyl derivative of
the analgesic acetaminophen (13), the antidepressant
(des)venlaflaxine (14),21 and N-methylnonivamide, whose
relative, capsaicin, has a number of promising medicinal
uses (Scheme 1).22 The des-aminofluorinated analog of the
antibacterial dapsone (Aczone)23 could also be prepared
(10). Additionally, the aryl nonaflate of 4-methylumbelli-
ferone could be used in place of the triflate to access 11 in
comparable yield (89%). Regiosiomeric mixtures of pro-
ducts were observed for the fluorinations leading to electron-
rich aryl fluorides 8, 13, 14, and 15 when the reactions were
conducted in toluene. As we have previously reported,9b
switching the solvent to cyclohexane in these cases led to
an increase in regioselectivity for the desired product. Sub-
strates containing free NH or OH groups are not tolerated
because they undergo competitive cross-coupling processes,
Fluorination could be cleanly carried out on the tri-
flate derivatives of various naturally occurring phenols
(Scheme 1). These include the N-methyl derivative of
nonivamide (pseudocapsaicin, 6), the diethyl amide of iso-
vanillin (7), estrone (8), the plant-derived antioxidant methyl
p-coumarate (12), and δ-tocopherol (15). In addition, the
fluorinated derivatives of a number of phenol-containing
(20) Molho, D.; Boschetti, E.; Fontaine, L. Process for stimulating
choleresis. U.S. Patent 3175943, Mar. 30, 1965.
(21) Husbands, G. E. M.; Muth, E. A.; Yardley, J. P. 2-Phenyl-2-(1-
hydroxyycloalkyl or 1-hydroxycycloalk-2-enyl)ethylamine derivatives.
U.S. Patent 4535186A, Aug. 13, 1985.
(22) (a) Mason, L.; Moore, R. A.; Derry, S.; Edwards, J. E.; McQuay,
H. J. Br. Med. J. 2004, 328, 991. (b) Mori, A.; Lehmann, S.; O’Kelly, J.;
Kumagai, T.; Desmond, J. C.; Pervan, M.; McBride, W. H.; Kizaki, M.;
Koeffler, H. P. Cancer Res. 2005, 66, 3222.
(23) Osborne, D. W.. Compositions and methods for topical applica-
tion of therapeutic agents. U.S. Patent 5863560A, Jan 26, 1999.
€
(17) Krause, J.; Bonrath, W.; Porschke, K. R. Organometallics 1992,
11, 1158.
(18) Unlike with dba, when 1,5-COD was added to the reactions in
Table 1, entry 1, comparable yields of both fluorodeoxyestrone (67%,
R:β = 5:1) and 3-fluoroquinoline (73%) to reactions conducted in the
absence of 1,5-COD were observed. Thus, 1,5-COD does not show an
inhibitory effect on these reactions.
(19) Bruch, M. K. Chloroxylenol: a new old antimicrobial. In Hand-
book of Disinfectants and Antiseptics; Ascenzi, J. M., Ed.; Marcel Dekker,
Inc.: New York, 1996; pp 265ꢀ294.
5604
Org. Lett., Vol. 15, No. 21, 2013

confirming that decomposition of these electron-rich aro-
matic systems does not occur under the reaction condi-
tions. As before, when the fluorinations of 17, 20, and
23 were carried out using [(cinnamyl)PdCl]2/1, lower yields
of the desired products were observed in every case
(Scheme 2), although the improvements when switching
to 5 were not dramatic. Other nitrogen-containing hetero-
aryl triflates, such as those derived from 5- membered
heteroaryl phenols, could not be fluorinated under these
conditions. An inhibition study24 of the conversion of
δ-tocopherol triflate to 15 in the presence of various
sp2-hybridized nitrogen-containing additives suggests that
these species inhibit this transformation to some degree,
which may explain why such heteroaryl triflates remain
challenging substrates for this reaction (see Supporting
Information for details).
Scheme 2. Fluorination of Heteroaryl and Heterocycle-
Containing Triflatesa
a Isolated yields, average of two runs. Reaction conditions unless
In conclusion, Pd(0) precatalyst 5 activates to provide
otherwise noted: Aryl triflate (1 mmol), 6 (1ꢀ2%), CsF (3 mmol), tol
the active catalytic species 2 Pd(0) without generation
3
b
(0.1 M). Yield when conducted under same reaction conditions
of reactive byproducts, such as “Clꢀ”, dba, carbazole, or
HF, that accompany the activation of other Pd sources
and represents a significant improvement for this reaction.
As such, 5 is an efficient precatalyst for the fluorination of
electron-rich substrates, as well as heterocycle-containing
aryl triflates, providing the corresponding aryl fluorides
in synthetically useful yields and contaminated only with
easily separated side products. Further investigation of
the structure and reactivity of 5 is currently underway.
Future efforts will also focus on preventing the formation
of regioisomeric aryl fluoride products for electron-rich
substrates, improving the scope with respect to heteroaryl
triflates, and finding nonhygroscopic fluoride sources to
carry out this reaction.25
using [(cinnamyl)PdCl]2/1 (Pd/1 = 1:1.5) instead of 5. Aryl triflate
(0.1 mmol), CsF(0.3 mmol), 19F NMRyield. c 3-Pyridyl triflate (0.1 mmol),
CsF (0.3 mmol), 5 (2%), tol (0.1 M), 19F NMR yield.
although the hindered tertiary carbinol present in
(des)venlaflaxine (14) did not need to be protected. In
every case <5% of the reduction product was observed;
the major side product of these couplings is the correspond-
ing biaryl ethers.9b Notably, when the preparations of
electron-rich aryl fluorides 7, 8, 13, and 15 were attempted
using our previous catalyst system of [(cinnamyl)PdCl]2/1,
lower yields of the desired products were observed in each
case (Scheme 1), demonstrating the superiority of our new
catalyst system. Even in the case of electron-deficient 12, an
improved yield was observed when using 5 in place of
[(cinnamyl)PdCl]2/1 (Scheme 1).
We also examined the application of 5 to the synthesis
ofheteroarylfluorides, which are typicallymoredifficultto
access using transition-metal-mediated fluorination meth-
ods (Scheme 2). Precatalyst5 is competent for the synthesis
of 4- and 3-quinolinyl fluorides alike (16ꢀ17). Although
fluorination of 3-pyridyl triflate to 18 proceeded in low
yield, electron-deficient 3-pyridyl triflates, such as that
corresponding to methyl nicotinate (19), provided some-
what higher yields of the desired aryl fluoride. In addi-
tion, 2,6-disubstituted 3-pyridyl fluorides can be accessed
in good yield using this reaction (20). Lastly, substrates
containing the five-membered heterocycles pyrrole (21),
furan (22), and thiophene (23) could also be fluorinated,
Acknowledgment. Research reported in this publica-
tion was supported by the National Institutes of Health
under Award Number GM46059. The content is solely the
responsibility of the authors and does not necessarily
represent the official views of the National Institutes
of Health. P.J.M. additionally thanks the National Science
Foundation (2010094243) and Amgen for funding. The
NMR spectrometers used for portions of this work
were purchased with funds from the National Science
Foundation (Grants CHE 980861 and DBI 9729592).
We also thank Ms. Elizabeth Kelley (MIT) for assistance
with 2D-NMR experiments and Dr. J. Robb DeBergh
(MIT) for assistance with this manuscript.
Supporting Information Available. Procedural and
spectroscopic data for all compounds are provided. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(24) Collins, K. D.; Glorius, F. Nat. Chem. 2013, 5, 597.
(25) In their recently reported one-pot nonaflation/fluorination of
phenols methodology based on our reaction, Larhed and coworkers found
that their reactions could be set up on the benchtop using CsF dried in a
vacuum oven with the reaction carried out at 180 °C in a microwave reactor.
However, in our hands this procedure could not be replicated when using
The authors declare the following competing financial interest(s):
MIT has patents on the ligands and precatalysts used in this work from
which S.L.B. receives royalty payments.
€
catalyst systems based on 2. See: Wannberg, J.; Wallinder, C.; Unlusoy, M.;
€
Skold, C.; Larhed, M. J. Org. Chem. 2013, 78, 4184.
Org. Lett., Vol. 15, No. 21, 2013
5605