Pyridylidene-Mediated Dihydrogen Activation Coupled with Catalytic Imine Reduction

Article abstract of DOI:10.1002/anie.201503233

In recent years, dihydrogen activation at non-metallic centers has received increasing attention. A system in which dihydrogen is trapped by a pyridylidene intermediate that is generated from a pyridinium salt and a base is now reported. The dihydropyridi

Full text of DOI:10.1002/anie.201503233

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201503233
German Edition: DOI: 10.1002/ange.201503233
Dihydrogen Activation
Pyridylidene-Mediated Dihydrogen Activation Coupled with Catalytic
Imine Reduction**
Johanna Auth, Jaroslav Padevet, Pablo Mauleón,* and Andreas Pfaltz*
Dedicated to Professor Albert Eschenmoser on the occasion of his 90th birthday
Abstract: In recent years, dihydrogen activation at non-
metallic centers has received increasing attention. A system in
which dihydrogen is trapped by a pyridylidene intermediate
that is generated from a pyridinium salt and a base is now
reported. The dihydropyridine formed in this process can act as
reducing agent towards organic electrophiles. By coupling the
hydrogen-activation step with subsequent hydride transfer
from the dihydropyridine to an imine, a catalytic process was
established. Treatment of the N-phenylimine of phenyl tri-
fluoromethyl ketone with 5–20 mol% of N-mesityl-3,5-bis(2,6-
dimethylphenyl)pyridinium triflate and 0.3–1.0 equivalents of
LiN(SiMe3)2 under 50 bar of hydrogen gas resulted in high
conversion into the corresponding amine.
C
atalytic hydrogenation with H2 as the reductant is of
central importance in organic synthesis. In general, a transi-
tion-metal catalyst is required for H2 activation, dissociation
À
Scheme 1. H2 activation by non-metallic systems.
of the H H bond, and transfer of the hydrogen atoms to the
substrate. However, there are examples of transition-metal-
free systems for H2 activation. In 1964, Walling and Bollyky
showed that tBuOK catalyzed the hydrogenation of benzo-
phenone at high temperature and pressure through a mech-
anism that involves heterolytic splitting of H2 (Scheme 1).[1]
More recently, Bertrand and co-workers reported a stoichio-
metric reaction of H2 with a singlet (alkyl)(amino)carbene
leading to an amine.[2] A breakthrough in the development of
non-metallic catalytic systems for H2 activation was achieved
by the groups of Stephan and Erker, based on the concept of
frustrated Lewis pairs (FLPs).[3] FLPs that consist of a per-
fluorinated aryl borane and a sterically demanding base
cannot form a stable Lewis acid–base complex owing to steric
constraints and were shown to reversibly split H2 and catalyze
the hydrogenation of various functional groups.[3,4] Although
a range of phosphorus, nitrogen, and carbon Lewis bases have
been used, the choice of the Lewis acid component appears to
be restricted to highly reactive fluorinated aryl boranes, which
have shown limited functional-group compatibility and
require glove-box or Schlenk techniques. Therefore, it
would be desirable to explore other types of potential hydride
acceptors to generalize the concept of FLP-mediated H2
activation and to enhance its scope.
The base-induced H2 splitting and reduction of benzo-
phenone (Scheme 1)[1] suggests that carbon-centered electro-
philes might serve as Lewis acid components in FLPs. Among
possible carbon-centered hydride acceptors, pyridinium salts
seem particularly attractive because the corresponding dihy-
dropyridines resulting from H2 activation may act as reducing
agents towards organic electrophiles,[5] offering the possibility
to couple H2 activation with subsequent hydride transfer to an
organic substrate. The reaction could then be rendered
catalytic in the pyridinium salt as in NADH-mediated
biological redox processes.
[*] J. Auth, Dr. J. Padevet, Prof. Dr. A. Pfaltz
Department of Chemistry, University of Basel
St. Johanns-Ring 19, 4056 Basel (Switzerland)
E-mail: andreas.pfaltz@unibas.ch
Dr. P. Mauleón
Department of Organic Chemistry
Universidad Autónoma de Madrid
c/Fco. Tomµs y Valiente 7, Cantoblanco 28049, Madrid (Spain)
E-mail: pablo.mauleon@uam.es
A pyridinium salt is expected to be a stronger hydride
acceptor than benzophenone and should therefore react
under milder conditions with a weaker base through a mech-
anism as shown in Scheme 1 (pathway a). An analogous
process based on a formamidinium cation as the hydride
acceptor had been proposed for a hydrogenase from meth-
anogenic archaea that was originally thought to be metal-
free.[6] Although later on the enzyme was shown to contain an
iron atom that was involved in H2 activation,[7] the results
[**] Financial support from the Swiss National Science Foundation is
gratefully acknowledged. P.M. thanks the European Union for
a Marie Curie Intra European Fellowship (IEF) and a Marie Curie
Career Integration Grant (CIG).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201503233.
9542
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 9542 –9545

Angewandte
Chemie
À
from DFT studies indicated that a hypothetical H H bond
ester was formed by release of H2. This process, which finds
cleaving process of this type would be energetically feasible.[6]
An alternative mechanism that may be considered for H2
activation by pyridinium salts involves deprotonation to
precedent in the hydride extrusion from orthoamides
reported by Erhardt and Wuest,[11] is the reversal of the
proposed H2 activation by pyridinium salts, implying that such
a pathway is, in principle, feasible based on the concept of
microscopic reversibility.
After completion of these studies, Clark and Ingleson
reported successful H2 activation experiments with N-meth-
ylacridinium salt 14 (Scheme 3).[12] Using 2,6-lutidine as the
a highly reactive pyridylidene,[8] which inserts into the H H
À
bond in analogy to aminocarbene 3[2] (Scheme 1, pathway b).
In contrast to the amine produced from an aminocarbene, the
resulting dihydropyridine could serve as a hydride donor, and
therefore, a catalytic hydrogenation process via a pyridylidene
intermediate would be conceivable. To examine the feasibility
of these concepts, we synthesized various pyridinium salts and
investigated their reactivity towards H2 in the presence of
a base. Herein, we report the results of our studies that have
led to a pyridinium-based system for H2 activation and
catalytic imine reduction.
First experiments were carried out on hydroxymethyl-
substituted pyridinium salt 9 (Scheme 2a). DFT calculations
Scheme 3. H2 activation with acridinium salt 14 and a base, reported
by Clark and Ingleson.[12] X=tetra(3,5-dichlorophenyl)borate.
base under 4 bar of H2, they achieved 98% conversion into
dihydroacridine 15 after 23 hours at 1008C. They also
achieved the in situ reduction of the N-tert-butylimine of
benzaldehyde under these conditions. However, only 25%
conversion into the corresponding amine was observed after
72 hours, and consequently, a catalytic version of this process
has not been described. Nevertheless, these results demon-
strate that carbon-centered Lewis acids with an electron-
deficient pyridinium core may act as hydride acceptors in
FLPs.
In view of our failed H2 activation attempts based on
pathway a in Scheme 1, we turned our attention to the
alternative pathway b, involving a pyridylidene intermediate.
After initial unsuccessful experiments with N-aryl Hantzsch
esters and nicotinamide derivatives, we chose 1,3,5-triarylpyr-
idinium salt 16 for further investigations because Itami and
co-workers had shown that it forms pyridylidene 17 in the
presence of a strong base (Scheme 4).[13]
Scheme 2. Initial studies on H2 activation by pyridinium salts and
bases. Tf =trifluoromethanesulfonyl.
suggested that the alkoxide pyridinium zwitterion 10 could
react with H2 through a concerted, strongly exothermic
process leading to the corresponding 1,4-dihydropyridine
with an activation barrier of approximately 13 kcalmolÀ1.[9]
Unfortunately, subsequent experiments failed despite the low
barrier predicted by DFT calculations. When pyridinium salt
9 or related hydroxyalkyl-substituted derivatives were treated
with base under H2 atmosphere (up to 100 bar), only
decomposition products were observed with no evidence of
H2 activation.
Scheme 4. Proposed H2 activation by pyridylidene derivatives. R=2,6-
Attempts to reduce pyridinium salts in an intermolecular
process with H2 and various external bases, such as LiN-
(SiMe3)2, 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), phos-
phazenes, phenolates, or N-ethyldiisopropylamine (DIPEA)
also failed. Although dihydropyridine 12 was formed in 28%
yield upon treatment of 13 with DIPEA in toluene at 1008C
under 35 bar of H2, deuteration experiments with D2 revealed
that this product did not arise from H2 activation.[9] Based on
literature precedence,[10] we assume that the observed reduc-
tion occurred through nucleophilic addition of DIPEA (or 13
deprotonated at one of the methyl groups) and subsequent
hydride transfer from the resulting dihydropyridine to
another molecule of 13.
Me2C6H3.
DFT calculations predicted a relatively low activation
energy of 20 kcalmolÀ1 for the concerted addition of H2 to the
carbenoid center of pyridylidene 17, suggesting that the
proposed reaction would be feasible within a chemically
reasonable temperature range.[9] The calculated transition
state (Figure 1) closely resembles that reported for the
analogous reaction of aminocarbene 3 (Scheme 1).[2]
Pyridinium salt 16 was deprotonated by treatment with
LiN(SiMe3)2 (2.25 equiv) in THF or THF/toluene at room
temperature for 20 min, as described by Itami and co-
workers.[13] After subsequent stirring under H2 atmosphere
(1–100 bar) at elevated temperature, traces of 1,2-dihydro-
pyridine 18 were detected. In pure toluene, conversion into 18
However, one notable observation was made in this
context: When 1,4-dihydropyridine 12 was treated with
CF3SO3H (Scheme 2b) the aromatized N-phenyl Hantzsch
Angew. Chem. Int. Ed. 2015, 54, 9542 –9545
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9543

.
Angewandte
Communications
Figure 1. Calculated transition state for the reaction of H2 with
pyridylidene 17.
Figure 2. 2H NMR spectrum for the D2 splitting experiment; CHCl3
increased to 13–33% (Scheme 5). In all experiments, com-
pound 19 was obtained as the major product. Its formation
from pyridylidene by intramolecular activation of one of the
arylmethyl groups had already been observed by Itami and
co-workers.[13]
with CDCl3 (7.26 ppm) was used as the internal standard.
1
full conversion (H2: 5 h; D2: 24 h). H NMR analysis con-
firmed the incorporation of deuterium at the methylene
group of 18, but the measured deuterium content was only
68%. As D2 with a purity of 99.98% was used, we assume that
the observed hydrogen incorporation results from hydride
transfer between dihydropyridine and the pyridinium salt
under the reaction conditions or from base-induced H/D
exchange of H2.[14]
In view of the well-documented reactivity of dihydropyr-
idines as hydride donors,[5] we next explored the possibility to
couple H2 activation with in situ hydride transfer to an
organic substrate. Imine 20 was chosen for these experiments
as it does not contain acidic protons and is expected to be
a good hydride acceptor owing to the electron-withdrawing
CF3 group. When a 1:1 mixture of imine 20 and pyridinium
salt 16 was subjected to the optimized conditions for H2
activation for five hours, 46% conversion into amine 21 was
observed, along with 80% of a 70:30 mixture of dihydropyr-
idine 18 and pyridinium salt 16 (Table 1, entry 1). Extension
of the reaction time to 13 hours resulted in 95% conversion
into the amine (entry 2). Notably, 75% of dihydropyridine 18
and no pyridinium salt 16 were detected, implying that more
than one equivalent of H2 had been consumed, and 18, acting
as a hydride donor, had been almost fully regenerated. These
findings suggest that a catalytic process based on this reaction
might be possible.
Scheme 5. Initial H2 activation studies. R=2,6-Me2C6H3.
To suppress this undesired reaction, the H2 pressure was
applied immediately after the addition of LiN(SiMe3)2
(2.25 equiv). Using this procedure, conversion into 1,2-
dihydropyridine 18 reached 71% after five hours at 508C in
toluene under 10 bar of H2. Furthermore, raising the pressure
to 50 bar led to full conversion (Scheme 6).
Scheme 6. H2 activation under the optimized conditions. R=2,6-
Me2C6H3.
Indeed, the reaction with one equivalent of base but only
20 mol% of pyridinium salt 16 led to full reduction of the
imine (entry 3), affording amine 21 in 82% yield after column
chromatography. Moreover, reducing the catalyst loading to
5 mol% still resulted in 85% conversion into amine 21
(entry 4). In a control experiment without pyridinium salt 16,
no amine was formed (entry 5), ruling out base-induced
heterolytic H2 splitting as in the reduction of benzophenone
with KOtBu (see Scheme 1).[1] Initial attempts to reduce the
amount of base below one equivalent resulted in a drastic
drop in conversion under the standard conditions. However,
at elevated temperatures, the amine could be obtained in high
yields with substoichiometric amounts of base (entries 6–8).
With 0.5 equivalents of LiHMDS and 0.2 equivalents of
pyridinium salt 16 at 808C, the amine was formed in almost
Deuteration experiments confirmed the role of H2 as the
reductant: When the reaction was conducted under 50 bar of
D2 at 508C for five hours, clear evidence for deuterium
incorporation at the C2 position of dihydropyridine 18 was
2
obtained. The H NMR spectrum (Figure 2) showed a reso-
nance at 4.3 ppm in accordance with the chemical shift of the
1
CH2 group of authentic 18 in the H NMR spectrum. The
reaction with D2 proved to be slower than that with H2, as
indicated by the presence of 40% of the starting material.
Extending the reaction time to 24 hours resulted in full
consumption of 16. The kinetic isotope effect kH/kD predicted
by DFT calculations for this reaction is 3.2 at 323 K,[9] which is
in qualitative agreement with the reaction times required for
9544 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 9542 –9545

Angewandte
Chemie
Table 1: H2 activation coupled with imine reduction.[a]
a hydride donor, dihydropyridine 18 can act as a reducing
agent towards an imine. Therefore, a catalytic process
becomes possible through H2 activation with subsequent
hydride transfer (Scheme 7). Our findings and the N-meth-
ylacridinium-mediated heterolytic H2 splitting reported by
Clark and Ingleson[12] demonstrate the potential of pyridi-
nium salts for H2 activation and catalytic hydrogenation.
Variations of the electronic and steric properties of pyridi-
nium salt and base will provide a means for the further
development of catalytic systems of this type.
Entry 16
LiHMDS
t
T
21[a]
[h] [8C] [%]
16+18[d] 18/16
[%]
[equiv] [equiv]
1
1
2.25
2.25
1
5
13
5
50
50
46
80
75
70:30
>97:3
>97:3
>97:3
–
2
1
95[b]
3
4
5
0.2
0.05
–
0.2
0.2
0.2
50 100 (82)[c] 95
1
5
5
50
50
85[b]
0
95
–
2.25
0.5
0.3
0.3
Keywords: catalytic hydrogenation · dihydrogen activation ·
NADH analogues · pyridinium salts · pyridylidenes
6[e]
7[e]
8[e]
5
5
5
80
80
97
76
69
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
100
How to cite: Angew. Chem. Int. Ed. 2015, 54, 9542–9545
Angew. Chem. 2015, 127, 9678–9681
20 (1 equiv) used for all entries, R=2,6-Me2C6H3. [a] Conversion based
on quantitative 19F NMR analysis of the crude reaction mixture.
[b] Formation of side products detected. [c] Yield of pure isolated
product. [d] Based on the starting amount of 16; determined by 1H NMR
analysis using 21 as an internal standard. [e] Experiments with
substoichiometric amounts of base were less reproducible than those
with 1 equivalent of LiHMDS. In four runs with 0.3 equivalents of
LiHMDS at 808C, 21 was obtained in 24, 28, 57, and 76% yield. n.d.: not
determined.
[1] a) C. Walling, L. Bollyky, J. Am. Chem. Soc. 1964, 86, 3750; for
a mechanistic study, see: b) A. Berkessel, T. J. S. Schubert, T. N.
Müller, J. Am. Chem. Soc. 2002, 124, 8693.
[2] G. D. Frey, V. Lavallo, B. Donnadieu, W. W. Schoeller, G.
Bertrand, Science 2007, 316, 439.
[3] a) G. C. Welch, R. R. San Juan, J. D. Masuda, D. W. Stephan,
Science 2006, 314, 1124; b) G. C. Welch, D. W. Stephan, J. Am.
Chem. Soc. 2007, 129, 1880; c) P. Spies, G. Erker, G. Kehr, K.
Bergander, R. Frçhlich, S. Grimme, D. W. Stephan, Chem.
Commun. 2007, 5072.
quantitative yield. With only 0.3 equivalents of LiHMDS and
0.2 equivalents of 16, up to 76% conversion into the amine
was observed, implying that the lithiated amine resulting from
hydride transfer is sufficiently basic to deprotonate the
pyridinium salt and regenerate the active pyridylidene
intermediate. This was confirmed by an NMR experiment,
which demonstrated that upon deprotonation of amine 21
with one equivalent of LiHMDS and subsequent addition of
16, pyridylidene 17 was formed as evidenced by trapping with
S8.[9] Based on these results and DFT calculations,[9] a catalytic
cycle for the observed imine reduction is proposed in
Scheme 7.[15,16]
In conclusion, pyridylidene 17 generated from pyridinium
salt 16 has been shown to be an efficient H2 trapping agent.
The observed addition of H2 to the carbenoid center of 17 has
precedent in the reaction of aminocarbene 3 described by
Bertrand and co-workers (Scheme 1).[2] However, in contrast
to the amine 4 produced from 3, which lacks reactivity as
[4] For reviews, see: a) D. W. Stephan, G. Erker, Angew. Chem. Int.
Ed. 2010, 49, 46; Angew. Chem. 2010, 122, 50; b) J. Paradies,
Angew. Chem. Int. Ed. 2014, 53, 3552; Angew. Chem. 2014, 126,
3624; c) D. W. Stephan, Acc. Chem. Res. 2015, 48, 306.
[5] A. McSkimming, S. B. Colbran, Chem. Soc. Rev. 2013, 42, 5439.
[6] J. H. Teles, S. Brode, A. Berkessel, J. Am. Chem. Soc. 1998, 120,
1345.
[7] E. J. Lyon, S. Shima, G. Buurman, S. Chowdhuri, A. Batschauer,
K. Steinbach, R. K. Thauer, Eur. J. Biochem. 2004, 271, 195.
[8] O. B. Schuster, L. Yang, H. G. Raubenheimer, M. Albrecht,
Chem. Rev. 2009, 109, 3445.
[9] See the Supporting Information.
[10] a) A. Ohno, S. Ushida, S. Oka, Bull. Chem. Soc. Jpn. 1984, 57,
506; b) Y. Ohnishi, M. Kitami, Tetrahedron Lett. 1978, 19, 4035;
c) Y. Ohnishi, Tetrahedron Lett. 1977, 18, 2109.
[11] J. M. Erhardt, J. D. Wuest, J. Am. Chem. Soc. 1980, 102, 6363.
[12] E. R. Clark, M. J. Ingleson, Angew. Chem. Int. Ed. 2014, 53,
11306; Angew. Chem. 2014, 126, 11488.
[13] K. Hata, Y. Segawa, K. Itami, Chem. Commun. 2012, 48, 6642.
[14] E. Buncel, B. C. Menon, J. Phys. Chem. 1987, 91, 3879.
[15] The Li ions introduced with the base may play a role in the
catalytic cycle by coordinating to the imine, thereby facilitating
hydride transfer. This would explain why phosphazene bases did
not induce H2 activation, although they deprotonated the
pyridinium salt as shown by trapping experiments with S8.
[16] In principle, it should be possible to start the catalytic cycle from
dihydropyridine 18 without using an external base. However,
preliminary experiments along these lines with and without
LiOTf as an additive have been unsuccessful thus far.
Received: April 9, 2015
Revised: June 1, 2015
Published online: June 26, 2015
Scheme 7. Proposed catalytic cycle. R=2,6-Me2C6H3.
Angew. Chem. Int. Ed. 2015, 54, 9542 –9545
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9545