C(sp3)-F bond activation of CF3-substituted anilines with catalytically generated silicon cations: Spectroscopic evidence for a hydride-bridged Ru-S dimer in the catalytic cycle

Article abstract of DOI:10.1021/ja311398j

Heterolytic splitting of the Si-H bond mediated by a Ru-S bond forms a sulfur-stabilized silicon cation that is sufficiently electrophilic to abstract fluoride from CF₃ groups attached to selected anilines. The ability of the Ru-H complex, generated in the cooperative activation step, to intramolecularly transfer its hydride to the intermediate carbenium ion (stabilized in the form of a cationic thioether complex) is markedly dependent on the electronic nature of its phosphine ligand. An electron-deficient phosphine thwarts the reduction step but, based on the Ru-S catalyst, half of an equivalent of an added alkoxide not only facilitates but also accelerates the catalysis. The intriguing effect is rationalized by the formation of a hydride-bridged Ru-S dimer that was detected by ¹H NMR spectroscopy. A refined catalytic cycle is proposed.

Full text of DOI:10.1021/ja311398j

Communication
pubs.acs.org/JACS
C(sp3)−F Bond Activation of CF3‑Substituted Anilines with
Catalytically Generated Silicon Cations: Spectroscopic Evidence for a
Hydride-Bridged Ru−S Dimer in the Catalytic Cycle
Timo Stahl, Hendrik F. T. Klare, and Martin Oestreich*
Institut fur Chemie, Technische Universitat Berlin, Strasse des 17. Juni 115, 10623 Berlin, Germany
̈
̈
S
* Supporting Information
Scheme 1. Tethered Ruthenium Thiolate Complexes and
Cooperative Si−H Bond Activation [ArF = 3,5-
Bis(trifluoromethyl)phenyl]
ABSTRACT: Heterolytic splitting of the Si−H bond
mediated by a Ru−S bond forms a sulfur-stabilized silicon
cation that is sufficiently electrophilic to abstract fluoride
from CF3 groups attached to selected anilines. The ability
of the Ru−H complex, generated in the cooperative
activation step, to intramolecularly transfer its hydride to
the intermediate carbenium ion (stabilized in the form of a
cationic thioether complex) is markedly dependent on the
electronic nature of its phosphine ligand. An electron-
deficient phosphine thwarts the reduction step but, based
on the Ru-S catalyst, half of an equivalent of an added
alkoxide not only facilitates but also accelerates the
catalysis. The intriguing effect is rationalized by the
formation of a hydride-bridged Ru−S dimer that was
1
detected by H NMR spectroscopy. A refined catalytic
Anilines emerged as suitable substrates from exploratory
experiments. We were delighted to find that the para-CF3-
substituted aniline 2a was hydrodefluorinated with full
conversion in 48 h at ambient temperature (Table 1, entry
cycle is proposed.
hile the C−F bond is one of the thermodynamically
W
strongest covalent single bonds, it is heterolytically
activated by exceptionally potent, cationic1,2 or neutral3 main
group element Lewis acids. Cationic silicon compounds in
particular possess a high affinity to fluoride and are well-suited
for fluoride abstraction4−6 as independently shown by the
Table 1. Regioisomeric Anilines in the Ru−S-Catalyzed
a
Hydrodefluorination with Silanes
laboratories of Ozerov1a,c and Muller.1b While these methods
̈
usually rely on the preformation of an inter-1a,c or intra-
molecularly1b stabilized cationic silicon catalyst, we recently
disclosed a new way to catalytically generate a strong silicon
electrophile.7a Inspired by the related H−H bond activation
with late transition metal complexes containing a polar M−S
bond,8 we demonstrated that the tethered Ru−S complex 1a9
with an electron-rich phosphine ligand (Scheme 1, upper) is
able to activate Si−H bonds to produce a transition metal
hydride and a sulfur-stabilized silicon cation10 (Scheme 1,
lower).7 Using that catalyst, this new methodology enables C-3-
selective Friedel−Crafts-type silylation of indoles7a as well as
direct dehydrogenative formation of silyl enol ethers from
ketones.7b Both protocols do not require any added base, and
dihydrogen is liberated as the sole byproduct. We now report
on the hydrodefluorination of CF3 groups catalyzed by a
member of the family of those Ru−S complexes 1.
Unexpectedly, added alkoxide was found to accelerate the
catalysis, and we provide a mechanistic rationale and propose a
entry
aniline w/
CF3 group
aniline w/
CH3 group
T
t
conv.
b
[°C]
[h]
[%]
1
2
3
para (2a)
meta (2b)
ortho (2c)
para (3a)
meta (3b)
ortho (3c)
20
100
100
48
72
72
>99
0
77
a
All reactions were performed according to the General Procedure 4.
Determined by GLC analysis using tetracosane as internal standard.
b
1). A catalyst loading of 10 mol % of 1a appears reasonable,
considering that the catalysis must pass through three cycles for
complete transformation of a CF3 into a CH3 group. The use of
regioisomeric anilines demonstrated a strong dependence on
both the steric and electronic situation. The meta-CF3-
substituted aniline 2b was not even transformed at high
temperature and prolonged reaction time (entry 2), indicating
1
refined catalytic cycle based on the H NMR spectroscopic
detection of intermediate mono- and dimeric Ru−H com-
plexes.
Received: November 20, 2012
Published: January 11, 2013
© 2013 American Chemical Society
1248
dx.doi.org/10.1021/ja311398j | J. Am. Chem. Soc. 2013, 135, 1248−1251

Journal of the American Chemical Society
Communication
the importance of the electron-donating effect of the Et2N
group. The ortho-CF3-substituted aniline 2c, an example of a
good electronic but a poor steric situation, also afforded
incomplete conversion after 72 h at 100 °C (entry 3). Partially
fluorinated intermediates (with CHF2 or CH2F groups) were
not observed, even in reactions with incomplete conversion.
This observation supports a mechanism involving carbenium
ions as a result of the fluoride abstraction since its activation
barrier is supposed to decrease in the order CF3 > CHF2 >
CH2F.
Et3P, Table 2, entry 1). We, therefore, tested several
phosphines in the in situ protocol. While the reactivity of
Table 2. Variation of the Phosphine Ligand in In Situ-
Formed or Preformed Ru−S Complexes (cf. Scheme 3 for
a
Catalyst Formation)
Referring to previous mechanistic insight7a,9 and supported
by experimental observations (see the Supporting Information
(SI) for details), we conceived a basic catalytic cycle for the
Ru−S-catalyzed hydrodefluorination with its elementary steps
(Scheme 2). Si−H bond activation (1 → 4) is followed by
c
in situ-formed
b
preformed
b
entry
phosphine (complex)
conv. [%]
conv. [%]
1
2
3
4
5
Et3P (1a)
67
26
82
17
0
Me3P (1b)
Ph2MeP (1c)
Ph3P (1d)
20
Scheme 2. Elementary Steps of the Ru−S-Catalyzed
Hydrodefluorination with Silanes (BArF − as Counteranion
>99
0
4
d
(p-FC6H4)3P (1e)
>99
0
Omitted for Clarity)
a,b
c
For footnotes a and b, see Table 1. Complex 1 is formed in situ by
using [(R3P)Ru(SDmp)Cl] (6, 10 mol %) and NaBArF (10 mol %).
4
d
Full conversion after 18 h.
complexes 1b (Me3P) and 1c (Ph2MeP) was poor (entries 2
and 3), we were delighted to see that 1d9 (Ph3P) resulted in full
conversion after 24 h (entry 4). An even more reactive complex
formed with para-fluorinated aryl groups at the phosphorus
atom; full conversion was obtained with 1e at ambient
temperature in 18 h (entry 5). Perfluorinated triarylphosphine
(C6F5)3P was too electron-deficient to act as a ligand. To
further verify these results, we prepared and characterized
complex 1e separately. Surprisingly, preformed 1e was inactive
(entry 5). The same was true for preformed complexes 1c and
1d (entries 3 and 4). Conversely, preformed 1b behaves
similarly to its in situ-formed counterpart (entry 2). These
unexcepted findings indicated a subtle influence of the
electronic nature of the phosphine on catalyst formation.
A closer look at the ruthenium complex synthesis provided
an explanation for the striking reactivity difference of in situ-
formed and preformed 1 with electron-deficient phosphines
(Scheme 3). For electron-donating phosphine ligands, chloride
abstraction from the corresponding Ru−Cl complex 6 with
fluoride abstraction (4 → 5), forming a fluorosilane and a
cationic thioether complex. Complexes similar to 5 arising from
C(sp3)−X bond activation (X = Br and I) by the Ru−S bond
had already been crystallographically characterized by Ohki,
Tatsumi, and co-workers.9 The 18e Ru−H complex 5 then
transfers its hydride intramolecularly to afford the hydrocarbon,
thereby regenerating the 16e Ru−S complex 1 (5 → 1). We
assume fluoride abstraction to be rate-determining in the initial
CF3 → CHF2 hydrodefluorination, and this is in accordance
with the fact that partially fluorinated intermediates with CHF2
and CH2F groups were not detected.
NaBArF is quantitative at ambient temperature (6 → 1). The
4
situation changes with electron-withdrawing phosphine ligands
since these impede the formation of cationic complexes 1 with
a vacant coordination site. When monitoring the chloride
abstraction by 31P NMR spectroscopy (in CD2Cl2 and/or C6D6
at 20 °C, selected data for the latter in Scheme 3), we observed
an additional resonance signal that we assign to the phosphorus
atoms of a chloride-bridged ruthenium dimer (1 + 6 → 7). This
dimer formation appears plausible as it stabilizes these
otherwise disfavored 16e complexes 1. “Free” complexes 1
with electron-deficient phosphines are obtained at higher
reaction temperatures with full conversion.
The insight that remaining 6 is critical in lending stabilization
to 1 also serves to rationalize the activity of 7 (in situ
formation) and inactivity of 1 (preformation) in the hydro-
defluorination. The intramolecular hydride transfer that
regenerates the coordinatively unsaturated Ru−S complex (cf.
5 → 1, Scheme 2) becomes rate-determining with electron-
withdrawing phosphine ligands. For the in situ protocol, 6 will
assist the Ru-to-C hydride transfer through coordination to
released 1 (5 + 6 → 7, Scheme 4, upper), thereby closing the
An obvious starting point for improving the catalyst
performance is the variation of the phosphine ligand. We
anticipated though that the effect would be antagonistic in the
essential steps of the cycle, fluoride abstraction (4 → 5) and
hydride transfer (5 → 1). The above results were obtained with
Et3P as ligand where the latter step is likely to be favored.
Changing the phosphine ligand from electron-rich to -poor, this
step ought to be disfavored because it destabilizes the cationic
16e Ru−S complex 1. Conversely, the same effect might be
beneficial to the Lewis acidity of the silicon electrophile in 4.
For a systematic screening of phosphine ligands, in situ
formation of the coordinatively unsaturated Ru−S catalysts 1
from the corresponding chloride complexes by treatment with
NaBArF would avoid the capricious small-scale preparation of
4
these reactive 16e complexes. We quickly found out that results
are similar with in situ-formed and preformed catalyst 1a (with
1249
dx.doi.org/10.1021/ja311398j | J. Am. Chem. Soc. 2013, 135, 1248−1251

Journal of the American Chemical Society
Communication
dimer 9e derived from complex 1e by careful 1H NMR analysis
(in C6D6 at 20 °C). A resonance signal at −17.1 ppm is a
strong indication for the existence of 9e; the nonbridged
hydride complex 8e was detected at −8.6 ppm. A similar
hydride-bridged dimer with a chemical shift of −17.8 ppm was
reported by Shvo et al.11
Scheme 3. Dichotomy in the Catalyst Formation: Electron-
Donating vs Electron-Withdrawing Phosphine Ligands (Ar =
p-FC6H4)
a
In view of these findings, the catalytic cycle must be extended
for the case where electron-withdrawing phosphine ligands are
used in combination with an alkoxide additive (Scheme 5).
Scheme 5. Refined Catalytic Cycle for Ru−S Complexes with
Electron-Withdrawing Phosphine Ligands: Remarkable
Effect of Alkoxides as Additive (BArF − as Counteranion
4
Omitted for Clarity)
a
All reactions were performed according to the General Procedure 2.
Scheme 4. Catalyst Dimerization through Chloride and
Hydride Bridging: Increasing the Hydride Donor Strength
(Ar = p-FC6H4)
Again, the catalysis commences with cooperative Si−H bond
activation at the Ru−S bond (1 → 4). As long as the additive is
present, the silicon cation will transfer to the alkoxide, yielding
an unreactive silyl ether along with Ru−H complex 8 and a
weakly Lewis acidic countercation M+ (usually Na+ or K+).
That controlled quench of the silicon electrophile secures a
steady concentration of 8. After full consumption of the
alkoxide, complex 4 is finally available for fluoride abstraction
(4 → 5). In the presence of 8, Ru−H-assisted hydride transfer
is now facile, accompanied by the formation of the hydride-
bridged dimer 9. Subsequent Si−H bond activation involving 9
regenerates 4 and releases the facilitator 8.
catalyic cycle. It is that step that cannot occur in the absence of
any stabilizing donor, and Ru−H complexes 5 with electron-
deficient phosphine ligands mark a dead end of the catalytic
cycle.
With the understanding of the additive effect, we finally
optimized the hydrodefluorination protocol by variation of the
silane−additive−solvent combination (see the SI for details).
Using Ph2MeSiH and NaOMe in n-hexane for the hydro-
defluorination of para-CF3-substituted aniline 2a resulted in full
conversion after just 6 h at ambient temperature (Table 3, entry
1). As expected, 2b (meta) and 2c (ortho) were far less reactive
but 2b now showed little conversion and 2c was even fully
hydrodefluorinated in 72 h at 60 °C (entries 2 and 3). Excellent
conversion was obtained for CF3-disubstituted 2d at 100 °C
(entry 4) but chemoselective hydrodefluorination of the para-
CF3 group was not possible. Aniline 2e with a free NH2 group
also reacted, subsequent to dehydrogenative N-silylation12
(entry 5). Current efforts are directed toward the expansion of
the substrate scope.
The activating role of remaining Ru−Cl complex 6 through
formation of chloride-bridged dimer 7 prompted us to consider
hydride bridging to bring about the same net effect (5 + 8 → 9,
Scheme 4, lower). We hypothesized that deliberate generation
of monomeric Ru−H complex 8 would render the formation of
hydride-bridged dimer 9 possible. We assumed that sub-
stoichiometric amounts (based on catalyst) of alkoxide or
hydroxide additives would sequester the silicon electrophile
from intermediate 4 (cf. Scheme 2), partially yielding complex
8 required for subsequent stabilization of the vacant
coordination site in 1 (5 + 8 → 9, Scheme 4, lower). A
screening of various additives in the reactions with the inactive
preformed complexes 1c−1e (cf. Table 2, entries 3−5) revealed
a dramatic effect on catalyst performance (see the SI for
details). The catalyses proceeded smoothly at accelerated
reaction rate. In turn, stoichiometric amounts of the additive
thwarted the hydrodefluorination completely. We were then
able to verify the formation of the proposed hydride-bridged
To summarize, we disclosed a heterolytic C(sp3)−F bond
cleavage of activated CF3 groups using a catalytically generated
silicon electrophile. The catalysis is likely to proceed through
“sulfur-stabilized” silylium and carbenium ions, and the tethered
1250
dx.doi.org/10.1021/ja311398j | J. Am. Chem. Soc. 2013, 135, 1248−1251

Journal of the American Chemical Society
Communication
professorship. We thank Dr. Sebastian Kemper (TU Berlin) for
expert advice with the NMR measurements.
Table 3. Ru−S-Catalyzed Hydrodefluorination in the
a
Presence of Alkoxide Additive
REFERENCES
■
(1) For silicon-based systems, see: C(sp3)−F: (a) Scott, V. J.;
C
̧
elenligil-C
̧
etin, R.; Ozerov, O. V. J. Am. Chem. Soc. 2005, 127, 2852.
(b) Panisch, R.; Bolte, M.; Muller, T. J. Am. Chem. Soc. 2006, 128,
̈
9676. (c) Douvris, C.; Ozerov, O. V. Science 2008, 321, 1188.
C(sp2)−F: (d) Duttwyler, S.; Douvris, C.; Fackler, N. L. P.; Tham, F.
S.; Reed, C. A.; Baldridge, K. K.; Siegel, J. S. Angew. Chem., Int. Ed.
2010, 49, 7519. (e) Allemann, O.; Duttwyler, S.; Romanato, P.;
Baldrigde, K. K.; Siegel, J. S. Science 2011, 332, 574.
(2) For aluminum-based systems, see: (a) Vol’plin, M. E.; Shev-
chenko, N. V.; Bolestova, G. I.; Zeifman, Y. V.; Fialkov, Y. A.; Parnes,
N. A. Mendeleev Commun. 1991, 1, 118. (b) Terao, J.; Begum, S. A.;
Shinohara, Y.; Tomita, M.; Naitoh, Y.; Kambe, N. Chem. Commun.
2007, 855 [C(sp3)−F]. (c) Klahn, M.; Fischer, C.; Spannenberg, A.;
Rosenthal, U.; Krossing, I. Tetrahedron Lett. 2007, 48, 8900 [C(sp3)−F
and C(sp2)−F]. (d) Gu, W.; Haneline, M. R.; Douvris, C.; Ozerov, O.
V. J. Am. Chem. Soc. 2009, 131, 11203 [C(sp3)−F].
(3) For boron-based systems, see: C(sp3)−F: (a) Hirano, K.; Fujita,
K.; Yorimitsu, H.; Shinokubo, H.; Oshima, K. Tetrahedron Lett. 2004,
45, 2555. (b) Caputo, C. B.; Stephan, D. W. Organometallics 2012, 31,
27.
(4) For reviews of silylium ion catalysis including C−F bond
activation, see: (a) Klare, H. F. T.; Oestreich, M. Dalton Trans. 2010,
39, 9176. (b) Muller, T. Adv. Organomet. Chem. 2005, 53, 155. For
̈
highlight articles, see: (c) Meier, G.; Braun, T. Angew. Chem., Int. Ed.
2009, 48, 1546. (d) Schulz, A.; Villinger, A. Angew. Chem., Int. Ed.
2012, 51, 4526.
(5) For selected authoritative reviews on C−F bond activation, see:
(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) Amii, H.; Uneyama, K. Chem.
Rev. 2009, 109, 2119. (d) Kiplinger, J. L.; Richmond, T. G.; Osterberg,
C. E. Chem. Rev. 1994, 94, 373.
(6) For selected examples of hydrodefluorination using silanes as a
hydride source, see: (a) Yang, J.; Brookhart, M. J. Am. Chem. Soc. 2007,
129, 12656. (b) Reade, S. P.; Mahon, M. F.; Whittlesey, M. K. J. Am.
Chem. Soc. 2009, 131, 1847. (c) Kuehnel, M. F.; Holstein, P.; Kliche,
a
All reactions were performed according to the General Procedure 5.
b
Determined by GLC analysis using tetracosane as internal standard.
c
d
7.0 equiv of Ph2MeSiH were used. 5.0 equiv of Ph2MeSiH were
used.
thiolate ligand in the Ru−S catalyst 1 is instrumental in that.
Remarkably, regeneration of coordinatively unsaturated 16e
complex 1 proved difficult when electron-withdrawing
phosphine ligands were used. That was overcome by the
deliberate addition of a silicon cation scavenger (alkoxides or
hydroxide) to form a Ru−H (co)catalyst that facilitates the
crucial intramolecular Ru-to-C hydride transfer by formation of
a hydride-bridged dimeric complex.
M.; Kruger, J.; Matthies, S.; Nitsch, D.; Schutt, J.; Sparenberg, M.;
̈
Lentz, D. Chem.Eur. J. 2012, 18, 10701.
(7) (a) Klare, H. F. T.; Oestreich, M.; Ito, J.-i.; Nishiyama, H.; Ohki,
Y.; Tatsumi, K. J. Am. Chem. Soc. 2011, 133, 3312. (b) Konigs, C. D.
̈
F.; Klare, H. F. T.; Ohki, Y.; Tatsumi, K.; Oestreich, M. Org. Lett.
2012, 14, 2842.
(8) Ohki, Y.; Sakamoto, M.; Tatsumi, K. J. Am. Chem. Soc. 2008, 130,
11610.
(9) Ohki, Y.; Takikawa, Y.; Sadohara, H.; Kesenheimer, C.; Engen-
dahl, B.; Kapatina, E.; Tatsumi, K. Chem.Asian J. 2008, 3, 1625.
(10) Prakash, G. K. S.; Bae, C.; Wang, Q.; Rasul, G.; Olah, G. A. J.
Org. Chem. 2000, 65, 7646.
(11) Shvo, Y.; Czarkie, D.; Rahamim, Y.; Chodosh, D. F. J. Am.
Chem. Soc. 1986, 108, 7400.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details, characterization data, as well as H, 11B,
1
13C, 19F, 29Si, and 31P NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
(12) Konigs, C. D. F.; Muller, M. F.; Aiguabella, N.; Klare, H. F. T.;
̈
̈
Oestreich, M. Chem. Commun. 2013, DOI: 10.1039/C3CC38900F.
AUTHOR INFORMATION
Corresponding Author
martin.oestreich@tu-berlin.de
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by the Deutsche Forschungs-
gemeinschaft (Oe 249/8-1) and the Fonds der Chemischen
Industrie (predoctoral fellowship to T.S., 2011−2013). M.O. is
indebted to the Einstein Foundation (Berlin) for an endowed
1251
dx.doi.org/10.1021/ja311398j | J. Am. Chem. Soc. 2013, 135, 1248−1251