Oxidative aliphatic C-H fluorination with fluoride ion catalyzed by a manganese porphyrin

Article abstract of DOI:10.1126/science.1222327

Despite the growing importance of fluorinated organic compounds in drug development, there are no direct protocols for the fluorination of aliphatic C-H bonds using conveniently handled fluoride salts. We have discovered that a manganese porphyrin complex catalyzes alkyl fluorination by fluoride ion under mild conditions in conjunction with stoichiometric oxidation by iodosylbenzene. Simple alkanes, terpenoids, and even steroids were selectively fluorinated at otherwise inaccessible sites in 50 to 60% yield. Decalin was fluorinated predominantly at the C2 and C3 methylene positions. Bornyl acetate was converted to exo-5-fluoro-bornyl acetate, and 5a-androstan-17-one was fluorinated selectively in the A ring. Mechanistic analysis suggests that the regioselectivity for C-H bond cleavage is directed by an oxomanganese(V) catalytic intermediate followed by F delivery via an unusual manganese(IV) fluoride that has been isolated and structurally characterized.

Full text of DOI:10.1126/science.1222327

Oxidative Aliphatic C-H Fluorination with Fluoride Ion Catalyzed by a
Manganese Porphyrin
Wei Liu et al.
Science 337, 1322 (2012);
DOI: 10.1126/science.1222327
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.
Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.
The following resources related to this article are available online at
www.sciencemag.org (this information is current as of September 13, 2012 ):
Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/337/6100/1322.full.html
Supporting Online Material can be found at:
http://www.sciencemag.org/content/suppl/2012/09/12/337.6100.1322.DC1.html
This article cites 54 articles, 9 of which can be accessed free:
http://www.sciencemag.org/content/337/6100/1322.full.html#ref-list-1
This article appears in the following subject collections:
Chemistry
http://www.sciencemag.org/cgi/collection/chemistry
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2012 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.

REPORTS
P22590-N16 and the Wittgenstein Prize, and the European
University Research Initiative (MURI), and the Army Research
Office–MURI on Atomtronics.
Supplementary Text
Figs. S1 to S5
References (34–58)
Union through the integrating project AQUTE, Siemens Austria,
and the City of Vienna. T.K. and E.D. thank the Army
Research Office for funding from the Defense Advanced
Research Projects Agency Optical Lattice Emulator program,
Harvard-MIT CUA, NSF grant DMR-07-05472, Air Force Office
of Scientific Research Quantum Simulation Multidisciplinary
Supplementary Materials
21 May 2012; accepted 21 August 2012
Published online 6 September 2012;
10.1126/science.1224953
www.sciencemag.org/cgi/content/full/science.1224953/DC1
Materials and Methods
metalloenzymes such as SyrB2 that generate an
oxo-chloroferrate intermediate (25). Thus, could
a manganese fluoride capture substrate radicals
generated by oxomanganese-mediated H abstrac-
tion? We report here a successful manganese-
catalyzed oxidative C-H bond fluorination using
fluoride ion.
We have found that a variety of simple al-
kanes and substituted alkanes, as well as larger
natural-product molecules, can be fluorinated ef-
fectively in the presence of catalytic amounts of
the bulky manganese porphyrin Mn(TMP)Cl (1).
This oxidative aliphatic fluorination reaction is
driven by iodosylbenzene as the oxo-transfer agent,
using silver fluoride/tetrabutylammonium fluo-
ride trihydrate as the fluoride source, both in stoi-
chiometric excess (26). Ultra-dry conditions are
not required. Results for the initial exploratory
reactions of a panel of simple substrates are pres-
ented in Table 1. Cycloalkanes afforded mono-
fluorinated products in ~50% yield. Typically,
conversions were ~70%, with small amounts
(15 to 20%) of alcohols and ketones also being
Oxidative Aliphatic C-H Fluorination
with Fluoride Ion Catalyzed by a
Manganese Porphyrin
Wei Liu,1 Xiongyi Huang,1 Mu-Jeng Cheng,2 Robert J. Nielsen,2
William A. Goddard III,2 John T. Groves1*
Despite the growing importance of fluorinated organic compounds in drug development,
there are no direct protocols for the fluorination of aliphatic C-H bonds using conveniently
handled fluoride salts. We have discovered that a manganese porphyrin complex catalyzes alkyl
fluorination by fluoride ion under mild conditions in conjunction with stoichiometric oxidation
by iodosylbenzene. Simple alkanes, terpenoids, and even steroids were selectively fluorinated
at otherwise inaccessible sites in 50 to 60% yield. Decalin was fluorinated predominantly at the
C2 and C3 methylene positions. Bornyl acetate was converted to exo-5-fluoro-bornyl acetate,
and 5a-androstan-17-one was fluorinated selectively in the A ring. Mechanistic analysis suggests
that the regioselectivity for C-H bond cleavage is directed by an oxomanganese(V) catalytic
intermediate followed by F delivery via an unusual manganese(IV) fluoride that has been
isolated and structurally characterized.
iochemistry manifests numerous highly by Rozen, Sandford, and Chambers (15–17), are produced. No products were detected in control
selective transformations of aliphatic C-H viable and very useful methods. Metal-catalyzed experiments that omitted the manganese por-
bonds into alcohols, halides, and olefins direct benzylic C-H fluorination has been achieved phyrin or iodosylbenzene, whereas a ~2:1 ratio
B
catalyzed by reactive metal-oxo intermediates within recently with palladium catalysts and an “F+” of oxygenated to fluorinated products was formed
enzymes (1, 2). A notable exception is aliphatic source (18). Also, advances in organocatalytic in the absence of tetrabutylammonium fluo-
C-H bond fluorination: The only fluorinase en- fluorination have enabled enantioselective for- ride. Only oxygenated products were formed
zymes yet characterized form C-F bonds by mation of C-F bonds adjacent to a carbonyl without silver fluoride (27). There were negli-
nucleophilic displacement at the preactivated C group (19) or to an alcohol, in the latter case via gible amounts of difluorides produced at this
center of S-adenosylmethionine (3, 4). There are the ring-opening of epoxides with a fluoride nu- level of conversion, probably due to the electron
no catalytic ways to selectively and directly in- cleophile (20). A chemo-enzymatic fluorination deficiency of the products induced by the F atom
corporate fluoride ions into unreactive sp3 C-H strategy via initial P450-mediated hydroxylation (28). A preliminary investigation of the substrate
bonds. Yet there is an enormous impetus today to (21) and chemical routes involving initial de- scope led to the results shown in Table 1 (entries
place F at such inaccessible sites in biomolecules carboxylation (22, 23) have also been reported. 7 to 12). A range of substituted molecules, in-
and drug candidates. Fluorination of drugs can Despite this impressive progress, a method for cluding ester, tertiary alcohol, ketone, and amide
block sites of phase I metabolism by cytochrome the selective and efficient incorporation of F at substituents, proved to be good substrates for
P450 enzymes as well as improving target-binding unactivated C-H sites within a target molecule fluorination with catalyst 1. Fluorination of meth-
affinities (5). Further, the incorporation of 18F using fluoride ion is singularly absent from the yl cyclohexylcarboxylate (entry 7) and methyl cy-
into biomolecules can allow direct imaging of repertoire of chemical synthesis.
metabolic activity and drug targets using the ex- The paradoxical challenge of achieving selec- as the major products. Monosubstituted five- and
quisite sensitivity of positron emission tomog- tive aliphatic fluorination lies in discovering a seven-membered cycloalkanes (entries 9, 10, and
raphy (6, 7). catalyst that is both sufficiently reactive and pre- 12) were fluorinated exclusively at the C3 and C4
clohexanol (entry 8) afforded trans-C3 fluorides
Although strategies for aromatic fluorination dictably selective to transform these ubiquitous positions, respectively, suggesting subtle stereo-
developed over the past 5 years have provided yet unactivated sp3 C-H bonds in the presence of electronic effects on the selectivity of this reaction.
access to complex aryl fluorides (8–11), there has other common functional groups. Such a catalyst
Having demonstrated that it is possible to re-
been a notable lack of recent progress in the cat- would be particularly useful if it could use flu- direct manganese-catalyzed hydroxylation to flu-
alytic fluorination of aliphatic C-H bonds (12–14). oride ion under convenient laboratory conditions. orination, we next aimed to apply this reaction to
Direct C-H fluorination with elemental F and A strategy was suggested to us by our recent larger molecules. The reaction of trans-decalin
electrophilic fluorinating agents, as developed discovery of a C-H chlorination protocol using under the same conditions afforded methylene
manganese porphyrin catalysts and hypochlorite monofluorination products with a 3.5-to-1 prefer-
1Department of Chemistry, Frick Chemistry Laboratory, Princeton
ion (24). We considered that site-selective flu- ence for C2 over C1 in an overall 51% yield and
orination might be achieved if a suitable fluoride a 75% conversion (Fig. 1A). Very high methyl-
source could be found that redirected the usual ene regioselectivity was observed for this sub-
biomimetic O rebound scenario (2) to metal- strate (>95%), similar to that observed for the
University, Princeton, NJ 08544, USA. 2Department of Chemis-
try, Materials and Process Simulation Center (MC 139-74),
California Institute of Technology, Pasadena, CA 91125, USA.
*To whom correspondence should be addressed. E-mail:
jtgroves@princeton.edu
bound fluoride, in the manner of halogenating manganese-catalyzed chlorination reaction we have
1322
14 SEPTEMBER 2012 VOL 337 SCIENCE www.sciencemag.org

REPORTS
recently reported (24), suggesting that a similar
F-substituted steroids, such as dexamethasone spite the complexity of the molecule, only the C2
reactive oxo- or dioxo-manganese(V) intermediate and fluasterone, have been found to be beneficial and C3 positions in the A ring were fluorinated
(29) is responsible for the H abstraction step in in blocking metabolic pathways (33–35), and in an overall yield of 55% (78% of the product
both reactions.
18F-fluorodihydrotestosterone has shown prom- distribution at 70% conversion, with minor amounts
Likewise, sclareolide, a plant-derived terpenoid ise as a radiotracer for imaging prostate cancer in of oxygenated products). The products of the re-
with antifungal and cytotoxic activities, afforded men (36). Because a direct, late-stage steroid flu- action could be readily separated by column
C2 and C3 methylene-fluorinated products in an orination protocol could greatly facilitate such chromatography and structurally assigned by the
overall 58% yield (Fig. 1B). C2-fluorination was applications, we sought to apply this manganese- diagnostic 19F–nuclear magnetic resonance (NMR)
favored by nearly 3:1, probably because of the catalyzed fluorination reaction to simple steroids. spectrum and the characteristic proton J-couplings
steric hindrance of the gem-dimethyl groups at We examined the fluorination of 5a-androstan- (figs. S19 to S22). A 5:1 a/b diastereoselectivity
C4. The products could be separated chromato- 17-one, which contains 28 unactivated sp3 C-H was observed for both the C2 and C3 positions,
graphically. C2 selectivity has been observed for bonds (Fig. 1C). Analysis of this molecule sug- probably reflecting the steric effect of the axial
this substrate by Baran and Eschenmoser for a gested that the carbonyl group would electron- methyl group at C10.
Rh-catalyzed amination (30, 31) and by White ically deactivate ring D. Rings B and C are
The reaction of bornyl acetate afforded a 55%
and Chen in a Fe(pdp)/H2O2-mediated oxidation sterically hindered, leaving the methylene groups yield of a single product, exo-5-fluoro-bornyl ac-
(32). In contrast, reaction of this molecule using of the A ring as the most likely sites for H ab- etate (Fig. 1D). The characterization of this pro-
Selectfluor (17) afforded an intractable mixture.
straction. Consistent with this analysis, and de- duct was based on C-H correlation NMR and
19F-NMR spectroscopy (figs. S27 to S30) (37).
We anticipated that the C5 position of camphor
Table 1. Manganese porphyrin–catalyzed fluorination of simple molecules. Reactions were run for 6 to
8 hours at 50°C under N2 in 3:1 CH3CN/CH2Cl2 solvent, 1.5 mmol substrate, 4.5 mmol silver fluoride, 6 to
8 mole % catalyst, 0.3 equivalent of tetrabutylammonium fluoride (TBAF) trihydrate, and 6 to 8 equiv-
alents of iodosylbenzene. Yields were determined by integration of gas chromatography traces using
naphthalene as the internal standard. Typical conversions were 70%. Unless otherwise noted, all major
fluorination products were isolated as single compounds.
would also be accessible, in analogy to the selec-
tivity of P450cam (CYP101) (38). However,
treating camphor under the standard fluorination
conditions resulted in 95% recovered starting ma-
terial. We attribute the low reactivity in this case
to the electron-withdrawing carbonyl group,
which apparently deactivates the entire molecule
toward fluorination, as with the monofluoride
products. These results highlight the subtle elec-
tronic effects on both the reactivity and selectivity
of the fluorination reaction.
We suggest the catalytic cycle shown in Fig.
2A for this manganese porphyrin–catalyzed
fluorination, although there are numerous aspects
of these transformations that will require further
elucidation. Oxidation of the resting Mn(TMP)F
catalyst, formed in situ, would afford a reactive
oxomanganese(V) species (29), O=MnV(TMP)F,
which then abstracts a substrate H atom to
produce a C-centered radical and a HO-MnIV-F
rebound intermediate. Fluoride binding to sepa-
rately prepared MnIV(O)(TMP) was indicated by
an ultraviolet (UV) spectral shift (423 to 427 nm) that
we assign to the formation of [MnIV(O)(F)(TMP)]–,
in analogy to the well-characterized coordination
of hydroxide to MnIV(O) (39).
The key step in forming the fluorinated products
is capture of the incipient substrate radicals either by
HO-MnIV-F or a trans-difluoro-manganese(IV)
species. There is no precedent for such a F
atom transfer. In this important regard, the flu-
orination reaction differs from the manganese/
hypochlorite chlorinating system we have de-
scribed (24). Chloride ion is rapidly and re-
versibly oxidized to hypochlorite by oxoMnV
porphyrins (40). Although HOF is known (15),
there is no evidence that fluoride is oxidized in
that way under these conditions. The importance
of the hypochlorite in the Mn/–OCl case is illus-
trated by the observation of C-H bromination in
the presence of hypobromite, even with a large
excess of chloride ion present. We attribute the
unusual methylene selectivity observed in both
the fluorination and chlorination reactions to stereo-
electronically enforced steric clashes between the
substrate and the approaching oxoMnV catalyst
Entry
1
Substrate
Fluorination product Entry
Substrate
Major fluorination
product
Minor
sites
7
C4 14%
2, 49%
8, 46% dr=6:1
2
8
C4 12%
C3 11%
3, 51%
9, 44% dr=8:1
10, 42%
3
4
9
4, 55%
10
C2 <2%
5, 53%
1:1.4
11, 51% dr=1.5:1
12b, C3
27%
cis/trans=
2:1
5
6
11
12a, 30%
cis/trans=1:1
6, 49%
exo: endo=5.7
12
C3 9%
7, 2:1
13, 49% dr=1.6:1
*Rearranged product identified by the characteristic m-(CH2F) peak in the mass spectrum.
†Isolated as diastereomers.
www.sciencemag.org SCIENCE VOL 337 14 SEPTEMBER 2012
1323

REPORTS
(Fig. 2B). The lowest unoccupied molecular or-
A
B
Methylene selectivity
bitals in a low-spin, d2 oxoMnV complex are ex-
pected to be the two orthogonal Mn-O p* orbitals,
which would direct the approach of the scissile C-H
bond into a bent p*-approach trajectory (29, 41).
We conducted a number of experiments to
examine this mechanistic hypothesis. Initial C-H
hydroxylation was ruled out by controls showing
that cyclohexanol was oxidized to cyclohexanone
under these conditions. No cyclohexylfluoride
was detected. Also, the hydroxyl group of 1-
methylcyclohexanol is stable to the reaction condi-
tions (Table 1, entry 8). Deuterium kinetic isotope
effects (KIEs) were evaluated by the reaction of a
1:1 mixture of cyclohexane and cyclohexane-d12,
producing an intermolecular competitive KIE
of 6.1. A similar value (5.7) was observed with a
mixture of ethylbenzene and ethylbenzene-d10.
The large KIE indicates that C-H bond cleavage
is the rate-limiting step in the reaction, consistent
with typical manganese porphyrin–catalyzed hy-
droxylation reactions. Furthermore, reaction of
norcarane, a diagnostic radical clock substrate (2),
afforded 2-fluoronorcaranes and a significant
amount of the rearranged fluorinated product, 3-
fluoromethylcyclohexene (7), which is indicative
of a C radical ring-opening process (Table 1, entry
6). The 2:1 ratio of these cyclopropylcarbinyl and
homoallyl fluorides indicates a short radical life-
time of 2.5 ns and rapid trapping of the substrate
radicals, given the ring-opening rate constant for
the 2-norcaranyl radical of 2 × 108 M−1 s−1 (42).
We have shown that the chlorination of norcarane
with t-butyl hypochlorite, also involving diffus-
ing C radical intermediates, gave a similar ratio of
rearranged and unrearranged products. Further,
the yields of alkyl fluorides were reduced when
the reactions were run in air, indicating substrate
radical trapping by O2.
H
H
Mn(TMP)Cl (6 mol%)
AgF (3 equiv.) TBAF (0.3 equiv.)
F
PhIO (6 equiv.)
F
H
51%
3.5
:
1
26 unactivated sp3 C-H bonds
Mn(TMP)Cl (12 mol%)
AgF (3 equiv.) TBAF (0.3 equiv.)
O
O
Me
O
Me
H
Me
16%
C3-fluoride
/ =7.8
H
F
2
3
O
PhIO (10 equiv.)
Me
H
/ =3.1
42%
O
Selective A ring fluorination
C
Me
Me
O
O
Mn(TMP)Cl (8 mol%)
Me
Me
D
AgF (3 equiv.) TBAF (0.3 equiv.)
C
F
PhIO (8 equiv.)
A
B
H
H
F
H
/ =6.2
32%
/ =4.5
23%
Fig. 1. Manganese porphyrin–
catalyzed selective C-H fluorina-
tions. Yields were determined by
the integration of gas chroma-
tography traces using naphtha-
lene as the internal standard. (A)
P450-like selectivity
D
Mn(TMP)Cl (8 mol%)
AgF (3 equiv.) TBAF (0.3 equiv.)
H
PhIO (8 equiv.)
F
OAc
OAc
57%
Methylene-selective fluorination of trans-decalin. (B) Selective fluorination of sclareolide. (C) Selective
A-ring fluorination of 5a-androstan-17-one. (D) Selective 5-exo-fluorination of bornyl acetate.
A
B
approach
H
MnV
L
The identification of trans-difluoroMnIV(TMP)
as the likely fluorinating agent was made possi-
ble by its isolation and structural characteriza-
tion. We were able to obtain pure crystals of the
MnIV(TMP)F2 by treating MnIV(TMP)Cl2 (43)
with excess AgF. The molecular structure of this
compound showed two axially bound fluoride
ions with F-MnIV-F bond lengths of 1.7931(17)
and 1.7968(16) Å (Fig. 2C and tables S1 to S5).
These distances are very close to those of di-
ammonium hexafluoromanganate(IV), the only
other fluoromanganese(IV) species to be struc-
turally characterized to date (44).
C
MnIV(TMP)F2
MnIV(TMP)F2
Fig. 2. (A) Posited catalytic cycle for manganese porphyrin–catalyzed C-H fluorination reactions. (B)
Inferred stereoelectronics for H abstraction. (C) X-ray crystal structure of trans-MnIV(TMP)F2 drawn at 50%
probability of the electron density. Highlighted atoms are F (yellow), Mn (magenta), and N (blue) (H atoms
are omitted for clarity).
We found that stoichiometric amounts of
MnIV(TMP)F2 could replace silver fluoride in a
single-turnover C-H fluorination of cyclooctane
using Mn(TMP)Cl and iodosylbenzene. A 43%
yield of cyclooctyl fluoride was obtained based ical trapping experiments are probably due to the radicals and AgF might also be considered, the
on added MnIV(TMP)F2. Thermal decomposi- falling concentration of the manganese(IV) di- reaction between AgF and phenethyl radicals gen-
tion of azo-bis-a-phenylethane to generate the fluoride under these conditions. Crucial roles erated in situ from azo-bis-a-phenylethane af-
phenethyl radical in the presence of MnIV(TMP) for silver fluoride in this scenario under cata- forded only trace amounts of fluorinated products.
F2 led to a 41% yield of 1-fluoroethylbenzene. lytic conditions are first to convert the added
We have explored the potential energy land-
These observations indicate that after initial H Mn(TMP)Cl to the manganese(III) fluoride form scape and electronic structures of the intermedi-
abstraction, MnIV(TMP)F2 can trap the substrate of the catalyst and then to replenish the inventory ates and transition states proposed in Fig. 2A
radicals in the fluorine delivery step (Fig. 2A). of manganese(IV) fluoride during turnover (26, 27). using density functional theory (DFT) computa-
The moderate fluorination yields from these rad- Although a direct reaction between the substrate tions and a polarizable continuum solvation model.
1324
14 SEPTEMBER 2012 VOL 337 SCIENCE www.sciencemag.org

REPORTS
22. M. Rueda-Becerril et al., J. Am. Chem. Soc. 134, 4026
A
B
(2012).
23. F. Yin, Z. Wang, Z. Li, C. Li, J. Am. Chem. Soc. 134,
H
10401 (2012).
2.48 Å
1.79 Å
24. W. Liu, J. T. Groves, J. Am. Chem. Soc. 132, 12847 (2010).
25. M. L. Matthews et al., Proc. Natl. Acad. Sci. U.S.A. 106,
17723 (2009).
178o
H
6.5
3.1
26. The excess of iodosylbenzene typically used in
metalloporphyrin oxidations is due to the competing
disproportionation of this reagent, which produces
unreactive iodoxybenzene. The requirement for excess
fluoride ion appears to derive from the stoichiometry
of the fluorination reaction, which also produces
hydroxide ions. AgF converts Mn-OH to Mn-F species
and Ag2O.
1.77 Å
27. The need for both AgF and tetrabutyammonium
fluoride apparently derives from the limited solubility
of AgF in the reaction medium and the need for a
higher fluoride ion concentration than can be maintained
by AgF alone. The UV-visible lmax observed for
(TMP)MnIII-Cl (1) (475 nm) changed immediately to
that of a mixture of (TMP)MnIII-F (453 nm) and
[(TMP)MnIII(F)2]– (440 nm) under the reaction conditions.
28. The high selectivity for monofluorination, the low
reactivity of C-H bonds near carbonyl groups, and the
limited reactivity of the solvents as well as the
tetrabutylammonium ion seem to reflect a very strong
polar effect in the C-H bond cleavage step in this
reaction.
Fig. 3. Energy landscape for F rebound to the
cyclohexyl radical. (A) Schematic depiction of F
atom abstraction from X-MnIV-F by the cyclohexyl
radical in the equatorial configuration and the
influence of the axial ligand (X = OH and F) on the
F abstraction potential energy surface (enthalpies
H
-36.2
-43.4
in kilocalories per mole at 298 K). Bond distances shown are calculated for X = F. (B) Frontier orbital
depiction of the transition state (TS) for F transfer.
29. N. Jin, M. Ibrahim, T. G. Spiro, J. T. Groves, J. Am. Chem.
Soc. 129, 12416 (2007).
30. T. Newhouse, P. S. Baran, Angew. Chem. Int. Ed. 50,
3362 (2011).
We found that F atom transfer from Mn(THP)F2 is fluoride ion, we anticipate the potential ap-
to a cyclohexyl radical in the equatorial config- plication of these techniques to the incorpora-
uration was predicted to occur with a surprisingly tion of 18F into a wide variety of biomolecules
low activation barrier of only 3 kcal/mol (Fig. 3), and synthetic building blocks. Moreover, the
which is very similar to the O rebound barrier for isolation, structural characterization, and reac-
hydroxylation reactions catalyzed by oxomanga- tivity of the trans-difluoromanganese(IV) por-
nese porphyrins. A slightly higher transition state phyrin, MnIV(TMP)F2, suggest the existence of
was located for the delivery of F to a cyclohexyl a rich chemistry of such transition-metal fluorides
radical in an axial configuration (4.2 kcal/mol). for the delivery of F substituents.
Further, the calculated barrier for F transfer was
31. K. Chen, A. Eschenmoser, P. S. Baran, Angew. Chem.
Int. Ed. 48, 9705 (2009).
32. M. S. Chen, M. C. White, Science 327, 566 (2010).
33. J. P. Bégué, D. Bonnet-Delpon, J. Fluor. Chem. 127, 992
(2006).
34. N. Houstis, E. D. Rosen, E. S. Lander, Nature 440, 944
(2006).
35. R. I. Scheinman, P. C. Cogswell, A. K. Lofquist,
A. S. Baldwin Jr., Science 270, 283 (1995).
36. P. B. Zanzonico et al., J. Nucl. Med. 45, 1966 (2004).
37. L. F. Lourie et al., J. Fluor. Chem. 127, 377 (2006).
38. I. Schlichting et al., Science 287, 1615 (2000).
39. J. T. Groves, M. K. Stern, J. Am. Chem. Soc. 110, 8628
(1988).
~3 kcal/mol lower for the trans-difluoroMnIV
References and Notes
species (Fig. 3, X = F) than for the analogous
hydroxy-fluoride (Fig. 3, X = OH), implicating a
much faster reaction rate for the difluoride. Con-
1. J. C. Lewis, P. S. Coelho, F. H. Arnold, Chem. Soc. Rev.
40, 2003 (2011).
40. T. P. Umile, D. Wang, J. T. Groves, Inorg. Chem. 50,
2. J. T. Groves, J. Inorg. Biochem. 100, 434 (2006).
10353 (2011).
3. A. S. Eustáquio, D. O’Hagan, B. S. Moore, J. Nat. Prod.
73, 378 (2010).
sistent with this low barrier for F transfer, the
transition state is very early in the reaction trajec-
tory, showing an exceedingly long C-F distance of
2.48 Å and a Mn-F distance that is only very slightly
elongated from the starting manganese(IV) di-
fluoride (TS in Fig. 3 and fig. S31). The visible
spectrum of the reaction mixture was complex,
apparently due to the presence of several forms of
the catalyst during turnover. However, the good
yield of 1-fluoroethylbenzene from the generation
of phenethyl radical in the presence of Mn(TMP)F2
provides experimental support for these computa-
tional predictions that manganese(IV) fluorides
of this type are excellent radical fluorinating
agents.
We are encouraged by the promising initial
results described here for the selective fluorina-
tion of simple hydrocarbons, substituted cyclic
molecules, terpenoids, and steroid derivatives.
The yields are sufficiently high and the tech-
niques sufficiently simple that the reaction can be
performed without specialized apparatus or com-
plicated precautions, other than the normal care
that should be taken whenever strong oxidants or
fluoride-containing reagents are used. Given that
the source of F in this one-step, one-pot protocol
41. N. Jin, D. E. Lahaye, J. T. Groves, Inorg. Chem. 49, 11516
(2010).
4. D. O’Hagan, C. Schaffrath, S. L. Cobb, J. T. G. Hamilton,
C. D. Murphy, Nature 416, 279 (2002).
42. K. Auclair, Z. B. Hu, D. M. Little, P. R. Ortiz De Montellano,
J. T. Groves, J. Am. Chem. Soc. 124, 6020 (2002).
43. L. Kaustov, M. E. Tal, A. I. Shames, Z. Gross, Inorg. Chem.
36, 3503 (1997).
5. K. Müller, C. Faeh, F. Diederich, Science 317, 1881 (2007).
6. D. A. Harki, N. Satyamurthy, D. B. Stout, M. E. Phelps,
P. B. Dervan, Proc. Natl. Acad. Sci. U.S.A. 105, 13039
(2008).
44. S. Kaskel, J. Strähle, Z. Anorg. Allg. Chem. 623, 1259 (1997).
7. S. M. Ametamey, M. Honer, P. A. Schubiger, Chem. Rev.
Acknowledgments: We are grateful to NSF (grants
CHE-0616633 and CHE-1148597 to J.T.G.) for support of
this research. DFT calculations, hydrocarbon C-H activation,
and the CalTech-Princeton collaboration were supported
by the Center for Catalytic Hydrocarbon Functionalization,
an Energy Frontier Research Center, U.S. Department of
Energy, Office of Science, Basic Energy Sciences, under
award no. DE-SC0001298 to W.A.G. and J.T.G. A patent
application has been filed through Princeton University on
methods and catalysts presented in this paper. We thank
S. Semproni for assistance with the x-ray crystal structure of
MnIV(TMP)F2, data for which are available free of charge
from the Cambridge Crystallographic Data Centre under
reference no. CCDC 893742.
108, 1501 (2008).
8. T. Furuya, A. S. Kamlet, T. Ritter, Nature 473, 470 (2011).
9. P. P. Tang, T. Furuya, T. Ritter, J. Am. Chem. Soc. 132,
12150 (2010).
10. D. A. Watson et al., Science 325, 1661 (2009).
11. E. Lee et al., Science 334, 639 (2011).
12. P. Herrmann, J. Kvicala, V. Pouzar, H. Chodounska,
Collect. Czech. Chem. Commun. 73, 1825 (2008).
13. S. Rozen, C. Gal, J. Org. Chem. 52, 2769 (1987).
14. S. Stavber, M. Zupan, Tetrahedron 45, 2737 (1989).
15. S. Rozen, Eur. J. Org. Chem. 2005, 2433 (2005).
16. G. Sandford, J. Fluor. Chem. 128, 90 (2007).
17. R. D. Chambers, A. M. Kenwright, M. Parsons,
G. Sandford, J. S. Moilliet, J. Chem. Soc. Perkin Trans. 1,
2190 (2002).
Supplementary Materials
www.sciencemag.org/cgi/content/full/337/6100/1322/DC1
Materials and Methods
Figs. S1 to S31
Tables S1 to S6
18. K. L. Hull, W. Q. Anani, M. S. Sanford, J. Am. Chem. Soc.
128, 7134 (2006).
19. Y. J. Zhao, Y. H. Pan, S. B. D. Sim, C. H. Tan,
Org. Biomol. Chem. 10, 479 (2012).
20. J. A. Kalow, A. G. Doyle, J. Am. Chem. Soc. 132, 3268
References (45–58)
(2010).
21. A. Rentmeister, F. H. Arnold, R. Fasan, Nat. Chem. Biol.
5, 26 (2009).
23 March 2012; accepted 26 July 2012
10.1126/science.1222327
www.sciencemag.org SCIENCE VOL 337 14 SEPTEMBER 2012
1325