Fluorine transfer to alkyl radicals

Article abstract of DOI:10.1021/ja211679v

The development of new synthetic technologies for the selective fluorination of organic compounds has increased with the escalating importance of fluorine-containing pharmaceuticals. Traditional methods potentially applicable to drug synthesis rely on the use of ionic forms of fluorine (F ^- or F⁺). Radical methods, while potentially attractive as a complementary approach, are hindered by a paucity of safe sources of atomic fluorine (F^?). A new approach to alkyl fluorination has been developed that utilizes the reagent N-fluorobenzenesulfonimide as a fluorine transfer agent to alkyl radicals. This approach is successful for a broad range of alkyl radicals, including primary, secondary, tertiary, benzylic, and heteroatom-stabilized radicals. Furthermore, calculations reveal that fluorine-containing ionic reagents are likely candidates for further expansion of this approach to polar reaction media. The use of these reagents in alkyl radical fluorination has the potential to enable powerful new transformations that otherwise would take multiple synthetic steps.

Full text of DOI:10.1021/ja211679v

Communication
pubs.acs.org/JACS
Fluorine Transfer to Alkyl Radicals
Montserrat Rueda-Becerril,‡ Claire Chatalova Sazepin,‡ Joe C. T. Leung,‡ Tulin Okbinoglu,‡
Pierre Kennepohl,‡ Jean-Francois Paquin,§ and Glenn M. Sammis*,‡
̧
‡Chemistry Department, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia V6T 1Z1, Canada
§Canada Research Chair in Organic and Medicinal Chemistry, Dep
artement de chimie, 1045 avenue de la Medecine, Pavillon
ec G1V 0A6, Canada
́
́
́ ́ ́
Alexandre-Vachon, Universite Laval, Quebec, Queb
S
* Supporting Information
fluorobenzenesulfonimide (NFSI),8 Selectfluor® reagent (Air
Products and Chemicals, Inc.),9,10 and N-fluoropyridinium salts
(NFPY).11,12
ABSTRACT: The development of new synthetic tech-
nologies for the selective fluorination of organic
compounds has increased with the escalating importance
of fluorine-containing pharmaceuticals. Traditional meth-
ods potentially applicable to drug synthesis rely on the use
of ionic forms of fluorine (F− or F+). Radical methods,
while potentially attractive as a complementary approach,
are hindered by a paucity of safe sources of atomic fluorine
(F•). A new approach to alkyl fluorination has been
developed that utilizes the reagent N-fluorobenzenesulfo-
nimide as a fluorine transfer agent to alkyl radicals. This
approach is successful for a broad range of alkyl radicals,
including primary, secondary, tertiary, benzylic, and
heteroatom-stabilized radicals. Furthermore, calculations
reveal that fluorine-containing ionic reagents are likely
candidates for further expansion of this approach to polar
reaction media. The use of these reagents in alkyl radical
fluorination has the potential to enable powerful new
transformations that otherwise would take multiple
synthetic steps.
It occurred to us that a technique for the fluorination of
carbon radicals, especially aliphatic ones, would be a powerful
tool that would nicely complement methodologies based on the
above reagents. Unfortunately, there are very few sources of
atomic fluorine (F•) that might be serviceable in such a
transformation. Molecular fluorine is a definite possibility,13 but
its notoriously uncontrollable reactivity and the hazards
associated with its use overshadow its application in preparative
operations.3 Similarly, fluoroxytrifluoromethane (trifluorometh-
yl hypofluorite) is a potential source of atomic fluorine, but it is
similar to molecular fluorine in its reactivity, along with the
associated hazards.14 At this time, the sole selective source of
fluorine atoms appears to be the expensive noble gas
compound, XeF2,15 although perfluoroalkanes may behave
similarly, at least toward aryl radicals.16,17
Literature data10 on Selectfluor imply that it, as well as its
congeners, may be suitable as radical transfer agents. Computa-
tional studies18 supported the foregoing surmise. Calculated
N−F homolytic bond dissociation energies (Table 1, DNF)
indicate that Selectfluor and NFSI share very similar homolytic
bond dissociation properties that are essentially independent of
the dielectric properties of the medium.
The greater electronegativity of the formally cationic
nitrogen in Selectfluor results in a more positively charged
(and thus electrophilic) fluorine atom (qF) than for NFSI.
While NFPY has a comparatively larger DNF, which is likely due
to poor delocalization of the unpaired electron in the resultant
nitrogen-centered radical, the N−F bond should be sufficiently
weak to react with most alkyl radicals. Thus, the use of these
reagents in alkyl radical fluorination has the potential to
revolutionize organic radical fluorination and enable powerful
new transformations that otherwise would take multiple
synthetic steps.19
he observation that fluorine incorporation into drugs
T
often leads to significantly improved medicinal properties
has revolutionized the development of pharmaceuticals.1−5
Indeed, there were no fluorinated medicaments on the market
in 1957, while today more than 20% of medicinal compounds
contain fluorine.6,7 As a consequence, synthetic technologies for
the selective fluorination of organic substrates are receiving
increasing attention.2,3 Virtually all contemporary fluorination
methods potentially applicable to drug synthesis rely on the use
of ionic forms of the halogen, i.e., on sources of F− or F+
(Figure 1). Notable among the latter are reagents such as N-
Experimental verification of our hypothesis focused on the
interaction of NFSI, which is soluble in organic media suitable
for conducting radical reactions, with lauroyl peroxide, a well-
known source of alkyl radicals.20,21 Heating a solution of lauroyl
peroxide and NFSI in benzene-d6 afforded 1-fluoroundecane
(2) in 20% yield (Scheme 1, eq 1). The majority of the
remaining mass balance was accounted for in the formation of
Received: December 14, 2011
Published: February 10, 2012
Figure 1. Synthetic approaches for the incorporation of fluorine.
© 2012 American Chemical Society
4026
dx.doi.org/10.1021/ja211679v | J. Am. Chem. Soc. 2012, 134, 4026−4029

Journal of the American Chemical Society
Communication
Table 1. Comparison of DFT-Calculated Properties of the N−F bond in NFSI, Selectfluor® Reagent, and NFPY Species
a
DFT calculations were performed using ORCA wth the VWN/BPP86 functional and the available TZV/P basis set for all atoms. Molecular
geometries were optimized using the COSMO solvation model as implemented in ORCA with a variety of different solvents ranging from n-hexane
b
to acetonitrile. Additional details are provided in the SI. Computational results for the cationic reagents without counterions are shown in this table.
Results from calculations with appropriate counterions differ slightly due to ion pairing effects; these differences do not affect the overall conclusions
herein. For each species, the N−F bond distance (rNF in pm), Mayer bond order (BONF), and bond dissociation energy (DNF in kcal/mol) are
computed in four different solvents: hexane, tetrahydrofuran, acetonitrile, and water. In addition, the Loewdin charge on the fluorine atom (qF) is
given.
Scheme 1. Decarboxylative Radical Fluorination of Lauroyl
Peroxide
(entry 3).23 It is noteworthy that the less reactive secondary
radical is fluorinated more efficiently. This increase in yield is
consistent with previously reported slower homodimerization
rates of bulky aliphatic radicals.24−26 Not only is the yield of
secondary fluoride 7 synthetically useful, but the decarbox-
ylative fluorination of a secondary position could enable an
enantioselective variant of the reaction, perhaps by the use of a
chiral fluorine transfer agent.27 Thermolysis of tertiary radical
precursor28 8 afforded fluorinated product 9 in 98% yield
(entry 4). This result is important as tertiary fluorides are
difficult to access using nucleophilic fluorine sources. Perester
10, which homolyzes to a benzylic radical, afforded the desired
fluorine-containing product 11 in 45% yield (entry 5). The
yield of fluorinated product 11 is slightly lower than was
observed with fluorinated products 7 and 9 (entries 3 and 4)
due to product degradation under the reaction conditions.
We next examined radical fluorodecarboxylations α to
heteroatoms (Table 2, entries 6−8). Fluorination was
successfully achieved α to both oxygen (entries 6 and 7) and
nitrogen (entry 8). Similar to the fluorination of perester 10,
the slightly lower yields of fluorinated product are due to
product degradation under reaction conditions.29 For some
peresters, such as 12, the yield of fluorinated product could be
increased by changing the solvent from benzene to acetonitrile
(entries 6 and 7). In a preliminary investigation, Selectfluor was
successfully utilized in the radical fluorodecarboxylation of
perester 12 to provide fluoride 13 in 52% yield. Under the same
conditions, NFPY provided no fluorinated product. Overall,
these results demonstrate that even tertiary, benzylic, and
heteroatom-stabilized radicals are sufficiently reactive to
undergo a fluorine transfer reaction with NFSI.
a
a
The reported yields are the average GC yields of three trials and are
based on the formation of two undecyl radicals for every molecule of
lauroyl peroxide.
the expected radical recombination product (n-C22H46). To
further confirm that this process involves radical intermediates,
the reaction was carried out under photolytic conditions
(Scheme 1, eq 2). Indeed, irradiation of a benzene-d6 solution
of lauroyl peroxide and NFSI with UV light (300 nm) provided
a comparable 34% yield of fluorinated product.
These proof-of-concept experiments represent the first
examples of fluorine transfer from an organic reagent to an
alkyl radical.22 For the approach to be broadly applicable, it
needs to be viable for less reactive secondary, tertiary, and
benzylic radicals, and radicals α to heteroatoms. Studies in that
sense centered on the thermolysis of the peresters shown in
Table 2 (entries 2−8). As a point of comparison with lauroyl
peroxide and diacylperoxide 3 (entry 1), perester 5 should
homolytically fragment to a primary radical with reactivity
comparable to that of a 1-undecyl radical. Indeed, thermolysis
of 5 provided fluorine-containing product 4 in 24% yield (entry
2), which is comparable to what was observed with both lauroyl
peroxide and diacylperoxide 3. Thermolysis of the secondary
radical precursor 6 provided fluorinated product 7 in 98% yield
The fast reaction times needed to perform the radical
fluorodecarboxylation make it well-suited for application to
PET imaging. As the half-life of 18F is ∼2 h,6 any methodology
that incorporates this radioisotope into an organic molecule
must be a relatively fast process. Isotopically enriched NFSI can
be prepared, and with the exception of primary substrates
4027
dx.doi.org/10.1021/ja211679v | J. Am. Chem. Soc. 2012, 134, 4026−4029

Journal of the American Chemical Society
Communication
stabilized radicals. Furthermore, Selectfluor and related ionic
reagents are likely candidates for further expansion of this
approach to polar, or even aqueous, reaction media. Current
efforts are focused on expanding the range of potential radical
fluorinations from these and other traditional “electrophilic”
fluorinating agents. We believe that careful and systematic
choice of appropriate fluorinating agents will allow a rapid
expansion of both the scope and utility of this reaction.
Table 2. Substitution Patterns Compatible with
Decarboxylative Radical Fluorination
ASSOCIATED CONTENT
■
S
* Supporting Information
1
Full experimental details and H, 13C, and 19F NMR spectra.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
gsammis@chem.ubc.ca
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the University of British Columbia
(UBC), the Universite
Engineering Research Council of Canada (NSERC), the
Canada Research Chair program, a doctoral fellowship from
́
Laval, the Natural Sciences and
́
Consejo Nacional de Ciencia y Tecnologia to M.R.-B., a
doctoral fellowship from UBC to C.C.S., and a doctoral
fellowship NSERC to J.C.T.L. We also thank Prof. Marco
Ciufolini for scientific discussions and advice, Prof. Jennifer
Love and Alex Dauth for assistance with the GC, Prof. David
Chen and Roxana G. Jayo-Benavides for use of laboratory space
and analytical equipment, Dr. Maria Ezhova for assistance with
NMR, Natalie E. Campbell for work on the preparation of
fluoroalkane standards, and Cameron J. Reive for work on the
preparation of perester starting materials.
a
Conditions: 5 equiv of NFSI, 0.21 M benzene-d6, 110 °C.
b
Conditions: 5 equiv of NFSI, 0.1 M acetonitrile-d3, 110 °C. See SI
c
for a discussion on solvent effects in these transformations. Yield
determined by GC and reported as an average of three runs. Isolated
d
e
yield after distillation. See SI for details. The NMR yield in
f
acetonitrile-d3 is 86%. Yield determined by 1H NMR using
ethyltrifluoroacetate as an internal standard and reported as an
average of three runs.
REFERENCES
■
(Table 2, entries 1 and 2), all of the substrates have sufficiently
fast rates to allow for 18F incorporation.
(1) Muller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881−1886.
̈
(2) Hagmann, W. K. J. Med. Chem. 2008, 51, 4359−4369.
(3) Kirk, K. L. Org. Process Res. Dev. 2008, 12, 305−3251.
(4) Isanbor, C.; O’Hagan, D. J. J. Fluorine Chem. 2006, 127, 303−319.
(5) Kirk, K. L. J. Fluorine Chem. 2006, 127, 1013−1029.
(6) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Chem. Soc.
Rev. 2008, 37, 320−330.
We next explored this new radical fluorodecarboxylation in
the context of the larger substrate, cholic acid derivative 16.
Formation of the perester from 16, followed by treatment with
3 equiv of NFSI in acetonitrile, provided the desired fluorine-
containing product 17 in 68% NMR yield and 54% isolated
yield (Scheme 2).
(7) Thayer, A. M. Chem. Eng. News 2006, 84 (23), 15−24.
(8) Differding, E.; Ofner, H. Synlett 1991, 187−189.
(9) Banks, R. E.; Mohialdin-Khaffa, S. N.; Lal, G. S.; Sharif, I.; Syvret,
R. G. J. Chem. Soc., Chem. Commun. 1992, 8, 595−596.
Scheme 2. Fluorination of Cholic Acid Derivative 16
́
(10) Nyffleler, P. T.; Duron, S. G.; Burkart, M. D.; Vincent, S. P.;
Wong, C.-H. Angew. Chem., Int. Ed. 2004, 43, 192−212.
(11) Umemoto, T.; Tomita, K. Tetrahedron Lett. 1986, 27, 3271−
3274.
(12) Strekowski, L.; Kiselyov, A. S. Chem. Soc. Rev. 2005, 34, 1031−
1037.
(13) Rozen, S. Acc. Chem. Res. 1988, 21, 307−312.
(14) Hiyama, T. Organofluorine Compounds: Chemistry and
Applications; Yamamoto, H., Ed.; Springer: Berlin, 2000.
(15) Tius, M. A. Tetrahedron 1995, 51, 6605−6634.
(16) Yamada, S.; Gavryushin, A.; Knochel, P. Angew. Chem., Int. Ed.
2010, 49, 2215−2218.
The successful utilization of NFSI as a fluorine transfer
reagent with alkyl radicals represents a new paradigm for
synthetic fluorination. This approach works for a broad range of
alkyl radicals, including tertiary, benzylic, and heteroatom-
(17) Dobele, M.; Vanderheiden, S.; Jung, N.; Brase, S. Angew. Chem.,
̈
̈
Int. Ed. 2010, 49, 5986−5988.
4028
dx.doi.org/10.1021/ja211679v | J. Am. Chem. Soc. 2012, 134, 4026−4029

Journal of the American Chemical Society
Communication
(18) Neese, F. ORCA, An Ab Initio, Density Functional and
Semiempirical Program Package, v. 2.80; Universitat Bonn: Bonn,
̈
Germany, 2011.
(19) Rentmeister, A.; Arnold, F. H.; Fasan, R. Nat. Chem. Biol. 2009,
5, 26−28.
(20) Sanchez, J.; Myers, T. N. In Kirk-Othmer Encyclopedia of
Chemical Technology, 5th ed.; Seidel, A., Ed.; John Wiley: Hoboken,
NJ, 2007; Vol. 18, pp 34−41.
(21) Ryzhkov, L. R. J. Org. Chem. 1996, 61, 2801−2808.
(22) There are two likely mechanisms that may be occurring in this
radical fluorination process. The first is a concerted fluorine atom
transfer from NFSI to the alkyl radical. The second is an initial SET
from the alkyl radical to NFSI, followed fluoride transfer to the newly
generated alkyl cation.
(23) NFSI is converted to dibenzenesulfonamide during the reaction.
See SI for details.
(24) For studies on the pair production and cage reactions with
diacyl peroxide decomposition, see: Sheldon, R. A.; Kochi, J. K. J. Am.
Chem. Soc. 1970, 92, 4395−4404.
(25) The increased yield of the secondary substrate is also consistent
with an SET mechanism (see ref 22).
(26) Ma, J.-A.; Cahard, D. Chem. Rev. 2008, 108, PR1−PR43.
(27) Ernst, J. A.; Thankachan, C.; Tidwell, T. T. J. Org. Chem. 1974,
39, 3614−3615.
(28) Teare, H.; Robins, E. G.; Årstad, E.; Luthra, S. K.; Gouverneur,
V. Chem. Commun. 2007, 2330−2332.
(29) Fluorides, such as 13, are known to be unstable and prone to
hydrolysis. See the following study for further details: Ochiai, M.;
Yoshimura, A.; Miyamoto, K. Tetrahedron Lett. 2009, 50, 4792−4795.
4029
dx.doi.org/10.1021/ja211679v | J. Am. Chem. Soc. 2012, 134, 4026−4029