A polycomponent metal-catalyzed aliphatic, allylic, and benzylic fluorination

Full text of DOI:10.1002/anie.201203642

Angewandte
Chemie
DOI: 10.1002/anie.201203642
Synthetic Methods
A Polycomponent Metal-Catalyzed Aliphatic, Allylic, and Benzylic
Fluorination**
Steven Bloom, Cody Ross Pitts, David Curtin Miller, Nathan Haselton, Maxwell Gargiulo Holl,
Ellen Urheim, and Thomas Lectka*
The selective incorporation of fluorine into organic molecules
has advanced in dramatic ways over the last 30 years.[1]
Although arene[2] and alkyne[3] fluorination using metal
catalysis has received much attention, the corresponding
methods for metal-catalyzed alkane fluorination remain only
a promising goal [Eq. (1)].[4] To date, the most notable
18% yield and in good selectivity [8:1 with respect to 2-
fluoroadamantane (3)]. It should be noted that in the absence
of a metal catalyst, the reaction produced no fluorinated
products under the specified reaction conditions. At this
point, a number of other copper(I) salts were screened (CuBr,
CuCl, CuClO4), but CuI proved to be the most effective. For
example, CuCl afforded only trace amounts of product,
whereas CuClO4 resulted in a complex mixture of highly
fluorinated adamantane-based products in variable quanti-
ties.
Our group has shown that reactions employing Selectfluor
concomitant with a metal cocatalyst can be accelerated by the
addition of a phase-transfer catalyst as a solubilizing agent
and presumed metal counteranion exchanger.[10] To this end,
we added KB(C6F5)4 (10 mol%) to our reaction and were
gratified to find a substantial increase in the rate of formation
and yield of 2 (Table 1).[11] In a parallel effort toward making
the reaction more homogeneous, we also sought the use of
a ligand to bring CuI into solution more effectively and to
modulate its reactivity. Bis(imine)-derived ligands are well
established in copper catalysis, so we tried them at the
outset.[12] For example, addition of N,N-bis(phenylmethy-
lene)-1,2-ethanediamine (BPMED; 10 mol%) to our reac-
tion afforded further increases in the yields of 2. We then
examined the effects of time, temperature, solvent, and the
equivalency of Selectfluor on the reaction. Immediately, we
methods for alkane fluorination[5] involve the use of stoichio-
metric quantities of difficult-to-handle or indiscriminate
reagents such as elemental fluorine,[6] cobalt trifluoride
(polyfluorination),[7] or the potentially explosive cesium
fluoroxysulfate.[8] We were particularly interested in the
question of whether alkane fluorination could indeed be
selectively catalyzed under mild reaction conditions, thus
perhaps opening up a realm of fluorination catalysis in
addition to the much-studied aromatic fluorination. What is
more, a mild procedure using safe, commercially available
alternatives would be complementary to pioneering methods,
and perhaps be applicable to closely related allylic and
benzylic substrates. Herein, we report the fluorination of
a series of aliphatic, benzylic, and allylic substrates using
a polycomponent catalytic system involving commercially
available Selectfluor, the putative radical precursor N-
hydroxyphthalimide (NHPI), an anionic phase-transfer cata-
lyst (KB(C6F5)4), and a copper(I) bisimine complex.[9,10]
For the purposes of this study, we chose first to focus our
efforts on the well-characterized adamantane system 1 and its
fluorinated derivatives. Initial catalyst screening employed
a variety of transition-metal salts, Selectfluor, and adaman-
tane in dry MeCN (1 mL) and was stirred for 24 h at room
temperature. Most notably, 10 mol% of CuI turned out to be
a competent lead, thus yielding 1-fluoroadamantane (2) in
Table 1: A test substrate: Screening of conditions for adamantane
fluorination
Catalyst
Solvent
Ligand
T [8C]
2/3
2
Yield [%]
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
–
–
25
25
25
25
25
25
25
–
0[a,b]
CuI
CuI
CuI
CuI
CuI
CuI
8:1
7:1
5:1
–
3:1
8.4:1
18[a,b]
28[a,b]
35[a]
BPMED
BPMED
BPMED
BPMED
BPMED
[*] S. Bloom, C. R. Pitts, D. C. Miller, N. Haselton, M. G. Holl,
E. Urheim, Prof. T. Lectka
trace[a]
12[a]
Department of Chemistry, Johns Hopkins University
3400 N. Charles Street, Baltimore, MD 21218 (USA)
E-mail: lectka@jhu.edu
75[a]
Homepage: http://www.jhu.edu/chem/lectka/Home.html
[a] Yield without KB(C6F5)4. [b] MeCN was not degassed. [c] Yield after 1 .
[d] Yield after 3 h. Otherwise yields were determined after 24 h by
19F NMR spectroscopy using 2-fluorobenzonitrile as an internal stan-
dard, and also by column chromatography. Final entry indicates yield of
isolated product.
[**] T.L. thanks the PRF-ACS and NSF (CHE 113175) for support, as well
as Heather Neu, Dr. Cathy Moore, and Ryan Peterson for their help.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201203642.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 2: Optimization of fluorination conditions for a secondary alka-
found the reaction proceeded well when MeCN was used as
a solvent, and the yield of the fluorinated product depended
significantly on reaction time. In this respect, longer reaction
times were shown to give lower yields of 2 and an increased
formation of the corresponding acetamide 4 [Eq. (2)].
ne.[a]
Entry
Catalyst
Cocatalyst
Cocatalyst
Yield [%]
1
2
3
4
(BPMED)CuI
(BPMED)CuI
(BPMED)CuI
(BPMED)CuI
–
–
–
26
44
63
72
KB(C6F5)4
KB(C6F5)4
KB(C6F5)4
NHPI
NHPI[b]
Furthermore, heating the reaction proved deleterious,
whereas cooling of the reaction mixture to 08C remarkably
yielded increased quantities of 3 over the anticipated 2
(40%). Performing the reaction for 3 hours at room temper-
ature provided for optimal yields of 2 (75%; Table 1).
Although a detailed mechanistic study is forthcoming,
a few observations point to the putative participation of
radicals[5] (either free or metal-based) or single-electron
transfer (SET) during fluorination: 1) yields in the strict
absence of O2 are much higher than in its presence;[13]
2) interference from the MeCN solvent is minimal (at least
during the initial fluorination), and consistent with its sluggish
reaction with free radicals.[14] 3) finally, there is precedent for
Selectfluor engaging in SET chemistry.[15] In contrast, bare
fluoro radicals are unlikely to be major participants in the
optimized reaction; we would expect them to abstract H
atoms with virtually equal facility from both tertiary and
secondary alkyl sites in adamantane.[16] One final piece of
evidence in support of the involvement of radicals may be
discerned through the use of a radical trapping agent such as
2,2,6,6-tetramethylpiperidin-1-yloxyl (TEMPO). When the
reaction is performed under optimized reaction conditions
using a stoichiometric amount of adamantane and TEMPO,
only trace amounts of 2, 3, and 4 are evident.
The depletion of 2 involving a putative SN1 solvolysis is
likewise expected based on the formation of the correspond-
ing acetamide after work-up [Eq. (2)].[17] Adamantyl cations
are well-established intermediates in solvolysis reactions.[18]
The higher stability of the 1-adamantyl cation relative to the
2-isomer would explain the relative depletion of product 2
versus 3 during longer reaction times. For example, as
calculated at the B3LYP/6-311 ++ G** level of theory, the
1-adamantyl cation is more stable than the 2-isomer by almost
11.3 kcalmolÀ1.[19]
[a] All reactions were performed at reflux for 2 h and yields were
determined by 19F NMR spectroscopy using 3-chlorobenzotrifluoride as
an internal standard and isolation of the product by chromatography.
[b] Reaction was performed with KI (10 mol%) as additive.
the rate of the reaction, N-hydroxyphthalimide (5; 10 mol%),
which is known to form the phthalimide N-oxyl (PINO)
radical 6 in situ in the presence of redox active metals
[Eq. (3)],[20] was examined as a possible cocatalyst. Interest-
ingly enough, addition of NHPI provided additional increases
in the yield of 9 (63%; entry 3), and along with the additive
KI, which forms the putative cuprate(I) complex
(10 mol%), the optimal yield of 9 (72%; entry 4 was
obtained.[21] When longer reflux times were employed, 1,1-
difluorocyclododecane began to form in appreciable amounts
(18% after 24 h).
7
By using the optimized reaction conditions of Table 2, we
explored other cycloalkanes, as they each give rise to one
distinct monofluorinated product (Table 3). Medium sized
rings such as cycloheptane (10), cyclooctane (12), and cyclo-
decane (14) worked as well (entries 5–7). Given the fact that
an excess of cycloalkane is not used, it is remarkable that
more polyfluorination is not observed. In fact, reaction
conditions can be found under which the monofluoride is
virtually the exclusive fluorinated product. In contrast,
extended reaction times for 12 and 14 lead to diminished
yields. It is clear that the products 13 and 15 undergo a slow
solvolysis reaction, an unsurprising observation given the
demonstrable release of Prelog strain[22] during SN1 reactions
of 8- and 10-membered ring systems, and that reaction times
in MeCN reflect the susceptibility of substrates to solvolysis.
Straight-chain substrates such as n-dodecane (20) give rise
to a virtual 1:1:1:1:1 mixture of monofluorinated products in
63% yield, even though the necessity of 1.2 equivalents of KI
at reflux may reflect their less reactive nature (Table 3,
entry 9).[23] Allylic substrates proved to be interesting in their
We turned our attention to an investigation of the scope of
the reaction. A variety of aliphatic, allylic, and benzylic
substrates were investigated. Unfortunately, upon initial
screening, the reaction conditions optimized for reactive
adamantane yielded only trace amounts of the desired
fluorinated products on a variety of substrates. At this
point, we chose to focus our efforts on a less-reactive model
substrate such as cyclododecane (8; Table 2). We found that
heating the reaction increased the yield of 9 greatly (43%),
likely a result from increased reaction rate, and as the product
is a secondary fluoride, solvolysis in MeCN was not such
a serious problem. In an effort to improve both the yield and
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
tities of allylic fluorides.[24] Under catalytic conditions at room
Table 3: Catalytic substrate fluorinations.
temperature, as demonstrated herein, the allylic fluorides
predominate to the virtual exclusion of the fluoroacetamides
(entries 10 and 11). This result would seem to bolster the case
for a different (non-electrophilic) mechanistic pathway as
well.
Entry
Substrate
Product
Yield [%] t [h] T [8C]
1
75[c]
40[c]
72[a]
3
3
2
1
25
0
A benzylic substrate, ethylbenzene (26), fluorinated to
provide a-fluoroethylbenzene (27; Table 3, entry 12) in 28%
yield (56% based on recovered substrate). Once again, this
product is unlikely to form by a strictly electrophilic process
under these reaction conditions. Although the yield of this
reaction is modest, only very minor amounts of ring
fluorinated products were observed. Especially intriguing
results are shown in entries 13–15, in which oxygen-contain-
ing substrates (esters) fluorinate productively. For example, n-
hexyl acetate (28) fluorinates predominately on the 3- and 5-
positions of the hexyl chain (72%; 81% total fluorination,
entry 13), whereas dihydrocoumarin (30) reacts at its benzylic
position to afford product 31, which is the equivalent of
a conjugate addition of fluoride to coumarin (entry 14).[25] In
contrast, the lactone 32 fluorinates exclusively on its side
chain (entry 15).
2
3
81
81
4
5
66[b]
41[b]
0.5 81
0.5 81
6
7
47[b]
52[b]
2
1
25
81
At this point we undertook a preliminary UV/Vis study of
components of the catalytic system. The main observations
include: 1) BPMED + CuI affords a spectrum consistent with
a CuI complex; 2) addition of KI maintains the oxidation state
of CuI; 3) addition of Selectfluor gives rise to weak bands
indicative of CuII and they disappear rapidly, concomitant
8
33[b]
9
63[b]
53[c]
2
81
25
with the appearance of a prominent I3 band.[26] In turn,
À
À
[27]
addition of NHPI results in consumption of I3
.
Presum-
10
24
ably, CuI is regenerated as well. Thus, along with electron
transfer to Selectfluor from the cuprate complex, the addition
of KI may aid in the production of PINO radicals and in the
regeneration of the catalyst, thus allowing less reactive
substrates to convert smoothly. Further investigations will
prove essential to the elucidation of a reaction mechanism
and additional study will address the utility of the reaction
and its application to complex substrates and natural prod-
ucts.
11
42[c] (88) 24
28[c] (70) 24
25
25
12
13
56[b] (81) 1.5 81
14
47[b]
3
25
Experimental Section
Representative procedure for the syntheses of fluorinated alkanes:
An oven-dried, 10 mL round bottom flask equipped with a stir bar
was placed under an atmosphere of N2. KB(C6F5)4 (22.0 mg,
0.025 mmol, 0.1 equiv), Selectfluor (195 mg, 0.55 mmol, 2.2 equiv),
copper(I) iodide (5.0 mg, 0.025 mmol, 0.1 equiv), N,N-bis(phenyl-
methylene)-1,2-ethanediamine (6.0 mg, 0.025 mmol, 0.1 equiv), N-
hydroxyphthalimide (4.0 mg, 0.025 mmol, 0.1 equiv), and potassium
iodide (50.0 mg, 0.3 mmol, 1.2 equiv) were then added, followed by
degassed MeCN (3.0 mL). The reaction mixture was then stirred for
10 min. Under a stream on N2, cyclododecane (8; 42.0 mg, 0.25 mmol,
1.0 equiv) was added neat and the mixture heated to reflux for 2 h.
The reaction was monitored by 19F NMR at 30 min intervals. Final
yields were determined either by 19F NMR spectroscopy using 3-
chlorobenzotrifluoride as an internal standard or column chromato-
graphy on silica (either method was in good agreement). Note that
this procedure (omitting the addition of KI) also applies to the allylic
and benzylic substrates (Table 3).
15
16
62[b]
55[b]
5
3
81
81
[a] 10 mol% KI. [b] 1.2 equiv KI. [c] No KI. Yields within parentheses for
entries 11–13 are based on recovered starting materials. All reactions
were monitored by 19F NMR spectroscopy and yields determined using 3-
chlorobenzotrifluoride as an internal standard and/or isolation of the
product by column chromatography. Compounds 2, 3, 9, 15, 29, 33, and
35 were isolated.
own way. For example, a-methylstyrenes 22 and 24 are known
to fluorinate in MeCN to form fluoroacetamides (under so-
called electrophilic conditions) admixed with variable quan-
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
[11] Toste et al. have recently developed a system involving an
asymmetric phase-transfer-catalyzed fluorination of silyl ethers:
V. Rauniyar, A. D. Lackner, G. L. Hamilton, F. D. Toste, Science
2011, 334, 1681 – 1683.
Received: May 10, 2012
Revised: July 20, 2012
Published online: && &&, &&&&
[12] CuI·bisimine complexes are known: A. Toth, C. Floriani, M.
Pasquali, A. Chiesi-Villa, A. Gaetani-Manfredotti, C. Guastini,
Inorg. Chem. 1985, 24, 648 – 653.
Keywords: alkanes · copper · fluorine · phase-transfer catalysis ·
synthetic methods
.
[13] a) K. Okamura, Y. Takahashi, T. Miyashi, J. Phys. Chem. 1995,
99, 16925 – 16931; b) F. Wilkinson, Pure Appl. Chem. 1997, 69,
851 – 856. In contrast, this observation may also be explained by
suppression of the formation of catalytically deactivated Cu-O2
complexes; see: c) N. Kitajima, Y. Moro-oka, Chem. Rev. 1994,
94, 737 – 757; d) Z. Tyeklar, R. Jacobson, N. Wei, N. N. Murthy, J.
Zubieta, K. D. Karlin, J. Am. Chem. Soc. 1993, 115, 2677 – 2689.
[14] P. S. Engel, W. K. Lee, G. E. Marschke, H. J. Shine, J. Org. Chem.
1987, 52, 2813 – 2817.
[15] a) P. T. Nyffeler, S. G. Durꢁn, M. D. Burkhart, S. P. Vincent,
C. H. Wong, Angew. Chem. 2005, 117, 196 – 217; Angew. Chem.
Int. Ed. 2005, 44, 192 – 212; b) X. Zhang, H. Wang, Y. Guo, Rapid
Commun. Mass Spectrom. 2006, 20, 1877 – 1882; c) Y. A. Sergu-
chev, M. V. Ponomarenko, L. F. Lourie, A. A. Fokin, J. Phys.
Org. Chem. 2011, 24, 407 – 413.
[1] a) S. Purser, P. R. Moore, S. Swallow, V. Gouverneur, Chem. Soc.
Rev. 2008, 37, 320 – 330; b) D. Cahard, X. Xu, S. Couve-
Bonnaire, X. Pannecoucke, Chem. Soc. Rev. 2010, 39, 558 – 568.
[2] a) D. Watson, S. Mingjuan, G. Teverovskiy, Y. Zhang, G. F. Jorge,
T. Kinzel, S. L. Buchwald, Science 2009, 325, 1661 – 1664; b) T.
Furuya, J. E. M. N. Klein, T. Ritter, Synthesis 2010, 1804 – 1821;
c) P. Anbarasan, H. Neumann, M. Beller, Angew. Chem. 2010,
122, 2265 – 2268; Angew. Chem. Int. Ed. 2010, 49, 2219 – 2222;
d) Y. Ye, S. Lee, M. S. Sanford, Org. Lett. 2011, 13, 5464 – 5467;
e) R. N. Loy, M. S. Sanford, Org. Lett. 2011, 13, 2548 – 2551.
[3] a) J. A. Akana, K. X. Bhattacharyya, P. Mꢀller, J. P. Sadighi, J.
Am. Chem. Soc. 2007, 129, 7736 – 7737; b) B. C. Gorske, C. T.
Mbofana, S. J. Miller, Org. Lett. 2009, 11, 4318 – 4321.
À
[16] M. L. Poutsma in Free Radicals, Vol. II (Ed.: J. K. Kochi), Wiley,
New York, 1973.
[17] Q. Michaudel, D. Thevenet, P. S. Baran, J. Am. Chem. Soc. 2012,
134, 2547 – 2550.
[18] P. von R. Schleyer, R. C. Fort, W. E. Watts, M. B. Comisarow,
G. H. Olah, J. Am. Chem. Soc. 1964, 86, 4195 – 4197.
[19] Geometry optimizations were performed using the Spartanꢂ06
program, Wavefunction, Inc.
[4] Considerable progress has been made in aliphatic C H bond
À
activation; see: “C H activation”: H. M. Davies, A. R. Dick,
Top. Curr. Chem. 2010, 303 – 346; and for a representative
example, see: M. S. Chen, M. C. White, Science 2007, 318, 783 –
787.
[5] Fluorination of alkane derived free radicals has been recently
observed: M. Rueda-Becerril, C. C. Sazepin, J. C. T. Leung, T.
Okbinoglu, P. Kennepohl, J.-F. Paquins, G. M. Sammis, J. Am.
Chem. Soc. 2012, 134, 4026 – 4029.
[20] a) S. Coseri, Catal. Rev. Sci. Eng. 2009, 51, 218 – 292; b) S. Coseri,
´
Eur. J. Org. Chem. 2007, 1725 – 1729; c) B. Orlinska, I. Roma-
nowska, Cent. Eur. J. Chem. 2011, 9, 670 – 676.
[6] a) W. T. Miller, A. L. Dittman, J. Am. Chem. Soc. 1956, 78,
2793 – 2797; b) W. T. Miller, D. D. Koch, F. W. McLafferty, J. Am.
Chem. Soc. 1956, 78, 4992 – 4995; c) W. T. Miller, S. D. Koch, J.
Am. Chem. Soc. 1957, 79, 3084 – 3089; d) D. H. R. Barton, R. H.
Hesse, R. F. Markwell, M. M. Pechet, H. T. Toh, J. Am. Chem.
Soc. 1976, 98, 3034 – 3035; e) S. Rozen, Acc. Chem. Res. 1988, 21,
307 – 312.
[7] a) R. W. Fowler, W. B. Burford, J. M. Hamilton, R. G. Sweet,
C. E. Weber, J. S. Kasper, I. Litant, Preparation, Properties and
Technology of Fluorine and Organic Fluoro Compounds,
McGraw Hill, New York, 1951, pp. 349 – 371; b) B. D. Joyner, J.
Fluorine Chem. 1986, 33, 337 – 346; c) J. Burdon, J. C. Creasey,
L. D. Proctor, R. G. Plevey, J. R. N. Yeoman, J. Chem. Soc.
Perkin Trans. 2 1991, 445 – 447.
[8] a) G. G. Furin, New Fluorinating Agents in Organic Synthesis,
Springer, Berlin, 1989, pp.135 – 168; b) M. Zupan, S. Stavber,
Tetrahedron 1989, 45, 2737 – 2742.
[9] For our interest in polyfunctional catalysis, see: a) D. H. Paull,
M. T. Scerba, E. Alden-Danforth, L. R. Widger, T. Lectka, J.
Am. Chem. Soc. 2008, 130, 17260 – 17261; b) J. Erb, D. H. Paull,
L. Belding, T. Dudding, T. Lectka, J. Am. Chem. Soc. 2011, 133,
7536 – 7546.
[21] It is known that aqueous solutions of CuI and KI react to form
À
a host of copper halide species in which the ate complex CuI2
predominates: K. Endo, K. Yamamoto, K. Deguchi, K. Mat-
sushita, Bull. Chem. Soc. Jpn. 1987, 60, 2803 – 2807.
[22] Y. I. Golꢃdfarb, L. I. Belenꢃkii, Russ. Chem. Rev. 1960, 29, 214 –
235.
[23] A chart based on substrate reactivity under given reaction
conditions was recommended and is provided in the Supporting
Information.
[24] J. S. Yadav, B. V. Subba Reddy, D. Narasimha Chary, D. Chan-
drakanth, Tetrahedron Lett. 2009, 50, 1136 – 1138.
[25] Compound 31 is prone to dehydrofluorination over time.
[26] Similarly, unstable CuI2 rapidly disproportionates to give
a mixture of CuI and soluble I2 in acetonitrile; see: Y.-L.
Wang, X.-B. Wang, X.-P. Xing, F. Wie, J. Li, L.-S. Wang, J. Phys.
Chem. A 2010, 114, 11244 – 11251.
[27] Selectfluor is also known to react with I3Àand IÀ: a) R. E. Banks,
M. K. Besheesh, S. N. Mohialdin-Khaffaf, I. Sharif, J. Chem. Soc.
Perkin Trans. 1 1996, 2069 – 2076; b) S. Stavber, P. Kralj, M.
Zupan, Synlett 2002, 4, 598 – 600.
[10] S. Bloom, M. T. Scerba, J. Erb, T. Lectka, Org. Lett. 2011, 13,
5068 – 5071.
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Synthetic Methods
S. Bloom, C. R. Pitts, D. C. Miller,
N. Haselton, M. G. Holl, E. Urheim,
T. Lectka*
&&&& — &&&&
A Polycomponent Metal-Catalyzed
Aliphatic, Allylic, and Benzylic
Fluorination
A group effort: Reported is the title
per(I) bis(imine). The catalyst system
reaction using a polycomponent catalytic
system involving commercially available
Selectfluor, a putative radical precursor
N-hydroxyphthalimide, an anionic phase-
transfer catalyst (KB(C6F5)4), and a cop-
formed leads to monofluorinated com-
pounds selectively (see example) without
the necessity for an excess of the alkane
substrate.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!