Mathias Glatz1, Berthold Stöger1, Daniel Himmelbauer1, Luis F Veiros2, Karl Kirchner1. 1. Institute of Applied Synthetic Chemistry and X-Ray Center, Vienna University of Technology, Getreidemarkt 9, A-1060 Vienna, Austria. 2. Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais No. 1, 1049-001 Lisboa, Portugal.
Abstract
Several hydride Mn(I) and Re(I) PNP pincer complexes were applied as catalysts for the homogeneous chemoselective hydrogenation of aldehydes. Among these, [Mn(PNP-iPr)(CO)2(H)] was found to be one of the most efficient base metal catalysts for this process and represents a rare example which permits the selective hydrogenation of aldehydes in the presence of ketones and other reducible functionalities, such as C=C double bonds, esters, or nitriles. The reaction proceeds at room temperature under base-free conditions with catalyst loadings between 0.1 and 0.05 mol% and a hydrogen pressure of 50 bar (reaching TONs of up to 2000). A mechanism which involves an outer-sphere hydride transfer and reversible PNP ligand deprotonation/protonation is proposed. Analogous isoelectronic and isostructural Re(I) complexes were only poorly active.
Several hydrideMn(I) and Re(I) PNPpincercomplexes were applied as catalysts for the homogeneous chemoselective hydrogenation of aldehydes. Among these, [Mn(PNP-iPr)(CO)2(H)] was found to be one of the most efficient base metalcatalysts for this process and represents a rare example which permits the selective hydrogenation of aldehydes in the presence of ketones and other reducible functionalities, such as C=C double bonds, esters, or nitriles. The reaction proceeds at room temperature under base-free conditions with catalyst loadings between 0.1 and 0.05 mol% and a hydrogen pressure of 50 bar (reaching TONs of up to 2000). A mechanism which involves an outer-spherehydride transfer and reversible PNP ligand deprotonation/protonation is proposed. Analogous isoelectronic and isostructural Re(I) complexes were only poorly active.
One
environmentally friendly and sustainable method to preparealcohols, which are valuable commodities for a large number of fine
and bulk chemicals, is the catalytichydrogenation of carbonyl compounds
with dihydrogen.[1] Over the years, many
highly efficient and active homogeneous catalysts based on precious
but also non-precious metals have been described for this purpose
(Scheme ).[2] Especially catalysts which reveal full selectivity
for aldehydes over ketones and/or alkenes[3,4] are
of practical importance for the synthesis of flavors,[5] fragrances,[3] and pharmaceuticals.[6]
Scheme 1
Well-Defined Catalysts for the Chemoselective
Hydrogenation of Aldehydes
In the past couple of years, the development and advancement
of
hydrogenation catalysts based on earth-abundant, inexpensive non-precious
metals experienced tremendous progress.[7] In particular, iron- and manganese-based catalysts turned out to
be highly active for the hydrogenation of carbonyl compounds, imines,
and nitriles (Scheme ).[8−11] In the case of manganese, however, most hydrogenations proceed at
relatively high catalyst loadings and elevated temperatures and, in
addition, require large amounts of strong bases as additives. As yet,
only iron-based systems proved to be reasonably chemoselective for
the reduction of aldehydes, as shown in Scheme .[12−14] We recently described the application
of [Fe(PNPMe-iPr)(CO)(H)(Br)] and [Fe(PNPMe-iPr)(H)2(CO)] as highly active
catalysts for the homogeneous hydrogenation of aldehydes (Scheme ).[15,16]
Scheme 2
Manganese Catalysts for the Hydrogenation of Ketones and Aldehydes
In this paper, we describe
an experimental and theoretical investigation
of the chemoselective hydrogenation of aldehydes with dihydrogen using
several hydrideMn(I) and Re(I) PNPpincercomplexes as catalysts
(Scheme ). To the
best of our knowledge, this is the first example of an efficient manganese-based
selective hydrogenation of aldehydes which proceeds under mild and
base-free conditions with low catalyst loadings. It has to be noted
that Repincercomplexes have rarely been used in (de)hydrogenation
catalysis.[17,18]
Scheme 3
PNP Pincer Complexes
Tested as Catalysts for the Hydrogenation of
Aldehydes (R = iPr) and Structural View of Re1 Showing 30% Thermal Ellipsoids
The reaction of [M(CO)5X] (M = Mn, X = Br; M = Re, X
= Cl) with the respective PNPpincer ligands in dioxane at elevated
temperatures afforded the neutral biscarbonyl complexes [M(PNP)(CO)2X] (1–5) (Scheme ). Treatment of these intermediates
with Na[HBEt3] (1.1 equiv) in toluene afforded complexes Mn1, Mn2, Mn3, Re1,
and Re2. The synthesis of Mn1 and Mn2 was already reported previously.[19] All new complexes could be isolated in 77–95% isolated yields
and were fully characterized by a combination of elemental analysis, 1H, 13C{1H}, and 31P{1H} NMR, and IR spectroscopy (see Supporting Information (SI)). In addition, the molecular structure of Re1 was determined by X-ray crystallography (Scheme , bottom left).
Scheme 4
Synthesis
of Hydride Mn(I) and Re(I) PNP Pincer Complexes
The catalytic performance of Mn1, Mn2, Mn3, Re1, and Re2 was then
investigated for the hydrogenation of aldehydes. The experiments were
performed in EtOH as solvent using 4-fluorobenzaldehyde as model substrate
to find the most active catalyst and optimal hydrogenation reaction
conditions (Table ). No reaction took place in aprotic solvents such as THF or toluene
at 50 bar H2, a catalyst loading of 1.0 mol%, and a reaction
time of 18 h. In the absence of dihydrogen, the hydrogenation of 4-fluorobenzaldehyde
to yield 4-fluorobenzyl alcohol was not observed—no reaction
took place. Thus, a possible transfer-hydrogenation mechanism in EtOHcould be excluded. It has to be further emphasized that ketones, e.g.,
acetophenone and 4-fluoroacetophenone, did not react with any of the
catalysts tested under the same reaction conditions described below.
Table 1
Hydrogenation of 4-Fluorobenzaldehyde
with Several Manganese and Rhenium Catalystsa
entry
cat.
solvent
S/C
P (bar)
t (h)
conversion
(%)b
TON
1
Mn1
THF
1000
50
18
2
Mn1
toluene
1000
50
18
3
Mn1
EtOH
1000
30
1
54
540
4
Mn1
EtOH
1000
30
4
>99
1000
5
Mn1
EtOH
2000
50
18
>99
2000
6c
Mn1
EtOH
20000
50
48
52
10400
7
Mn2
EtOH
100
50
18
8
Mn3
EtOH
100
50
18
21
21
9
Re1
EtOH
100
50
18
86
86
10d
Re1
EtOH
100
50
18
95
95
11
Re2
EtOH
100
50
18
76
76
Reaction conditions: catalysts (0.4–20.0
μmol), 4-fluorobenzaldehyde (2.0 mmol), EtOH (4 mL), 50 bar
H2, 25 °C.
Determined by 19F NMR
spectroscopy.
In the presence
of DBU (1.2 μmol,
3 equiv).
Performed at 50
°C.
Reaction conditions: catalysts (0.4–20.0
μmol), 4-fluorobenzaldehyde (2.0 mmol), EtOH (4 mL), 50 bar
H2, 25 °C.Determined by 19F NMR
spectroscopy.In the presence
of DBU (1.2 μmol,
3 equiv).Performed at 50
°C.When Mn1 (0.1 mol%) was used as catalyst, complete
conversion was observed after 4 h under a hydrogen pressure of 30
bar (Table , entry
4). By lowering the catalyst loading to 0.05 mol%, quantitative conversion
was achieved after 18 h at a hydrogen pressure of 50 bar (Table , entry 5). If the
reaction was performed in the presence of 3 equiv of DBU (1,8-diaza-bicyclo[5.4.0]undec-7-ene)
as external base, 4-fluorobenzyl alcohol was obtained in 52% yield
after 48 h under a hydrogen pressure of 50 bar and a catalyst loading
of 0.005 mol% (Table , entry 6). This corresponds to a turnover number (TON) of 10400.
Complexes Mn2 and Mn3 showed no or poor
reactivity, even with a catalyst loading of 1 mol% (Table , entries 7 and 8). Surprisingly,
the Re(I) complexes Re1 and Re2 with 1 mol%
catalyst loadings were poorly active, affording only 45 and 76%, respectively,
of 4-fluorobenzyl alcohol (Table , entries 9 and 11). At 50 °C, 4-fluorobenzyl
alcohol was obtained in 95% yield (Table , entry 10).Once Mn1 was
determined to be the most active catalyst
and its general applicability proved, various substrates were been
tested to establish scope and limitations (Table ). The catalytic experiments wereconducted
in the presence of 0.1–0.05 mol% of catalyst at 25 °C
and 50 bar hydrogen pressure, for a reaction time of 18 h, without
addition of any additives. The best results could be obtained for
aromatic aldehydes bearing electron-withdrawing halogen substituents
as well as electron-donating groups such as 4-anisaldehyde and 4-tolylaldehyde
on the phenyl ring (Table , A1–A5) wherecatalyst loadings of 0.05
mol% were employed. Heteroaromatic substrates as well as aliphaticaldehydescould be reduced quantitatively under the same reaction
conditions but with a catalyst loading of 0.1 mol% (Table , A6–A17). Substrates with conjugated and non-conjugated C=C double
bonds were also selectively hydrogenated. For instance, citronellal
or lyral, which are used in the flavor and fragrance industry (Table , A14–17), as well as the morechallenging α,β-unsaturated
substrate cinnamaldehyde (Table , A12) were not hydrogenated. In order
to investigate the catalyst’s selectivity toward substrates
with other unsaturated functionalities which can be easily hydrogenated,
additional studies werecarried out. Competitive experiments werecarried out using equimolar mixtures of 4-fluorobenzaldehyde and the
respective co-substrates at a catalyst-to-substrate ratio of 1:1000
with respect to the aldehyde. These studies showed that ketones, esters,
alkynes, and nitrile groups were not hydrogenated. Moreover, these
functionalities also did not interfere with the hydrogenation of the
aldehyde moieties.
Table 2
Hydrogenation of Aldehydes A1–A17 with Catalyst Mn1a,b
Reaction conditions: A1–A5 (1.0 μmol,
0.05 mol% Mn1), A6–A17 (2.0 μmol,
0.1 mol% Mn1), aldehyde (2 mmol),
EtOH (4 mL), 50 bar H2, 25 °C, 18 h.
Yields (in parentheses) based on
integration of 1H spectra using mesitylene as internal
standard.
Reaction conditions: A1–A5 (1.0 μmol,
0.05 mol% Mn1), A6–A17 (2.0 μmol,
0.1 mol% Mn1), aldehyde (2 mmol),
EtOH (4 mL), 50 bar H2, 25 °C, 18 h.Yields (in parentheses) based on
integration of 1H spectra using mesitylene as internal
standard.Stoichiometric
experiments show that Mn1reacts readily
with aldehydes, even in aprotic solvents such as benzene or THF. The
addition of 1 equiv of 4-fluorobenzaldehyde to a solution of the Mn(I)hydrideMn1 in C6D6 revealed the
formation of a new but minor manganese species (Scheme ). The concentration of this compound did
not change over time but grew with increasing amount of added substrate.
Thus, addition of up to 20 equiv of aldehyde was required to observe
complete conversion of the manganese hydridecomplex. The new compound
was tentatively assigned as the alkoxidecomplex 6, generated
by insertion of the aldehyde into the metal hydride bond of Mn1. Compound 6 could not be isolated and exhibited
singlet resonances at 115.8 and 140.9 ppm in the 19F{1H} and 31P{1H} NMR spectra, respectively
(free 4-fluorobenzyl alcohol exhibits a singlet at 116.1 ppm in the 19F{1H} NMR spectrum). In the IR spectrum, 6 displays the expected two signals of the symmetric and asymmetricCO stretching frequency at 1925 and 1848 cm–1 (cf.
1873 and 1790 cm–1 in Mn1). However,
no further reaction took place when a benzene (or THF) solution of
the in situ-generated alkoxidecomplex 6 was exposed to dihydrogen. There was also no catalyticreaction
if a 3:1 mixture of THF/EtOH was used. Accordingly, EtOH as solvent
is not required for the insertion step but obviously plays a crucial
role in the subsequent dihydrogen activation step. Moreover, Mn1 did not react with 4-fluoroacetophenone in both aprotic
and protic solvents.
Scheme 5
Reaction of Mn1 with 4-Fluorobenzaldehyde
and 4-Fluoroacetophenone
in C6D6
The reaction mechanism was explored in detail by means
of DFT calculations.[20] Benzaldehyde was
taken as substrate and Mn1 (A in the calculations)
as active catalyst.
An explicit ethanol molecule (solvent) was considered, providing a
proton shuttle and H-bond stabilization of the intermediates. Two
different paths wereconsidered, as shown in a simplified manner in Scheme . The more likely
one proceeds via participation of the acidic N–H bond of the
PNP ligand in a bifunctional mechanism (path I). This is supported
by the fact that catalyst Mn2, bearing NMe linkers, is
catalytically inactive and Mn3, featuring CH2 linkers which are less acidic than the NH linkers in Mn1, is only poorly active (Table , entry 8).
Scheme 6
Simplified Catalytic Cycles for Benzaldehyde
Hydrogenation with Mn1
Free energies in kcal/mol
are referred to A (Mn1 + EtOH + benzaldehyde);
transition state energies are given in italics; R = iPr).
Simplified Catalytic Cycles for Benzaldehyde
Hydrogenation with Mn1
Free energies in kcal/mol
arereferred to A (Mn1 + EtOH + benzaldehyde);
transition state energies are given in italics; R = iPr).A reasonable mechanism has been established
by means of DFT calculations.
The free energy profile for path I is depicted in Figure . The first step is the attack
of the hydride ligand in complex A to the carbonyl C-atom
of a free benzaldehyde molecule. The result is intermediate B, a species with the resulting alkoxide weakly bonded to
the metal by one C–H bond. This is a fairly easy step with
a barrier of 11 kcal/mol and a free energy balance of ΔG = 6 kcal/mol, indicating that B is less stable
than the initial reactants. The alkoxide in B can easily
leave the metal following a dissociative path, through intermediate C. From here, the alkoxide may coordinate the metal by the
O-atom, forming D through an easy process involving proton
exchange with the solvent (SI, Figure S1).
Importantly, the alkoxidecomplex D is 13 kcal/mol more
stable than the initial reagents and represents the catalyst resting
state. Naturally, therecan be proton exchange between the solvent,
EtOH, and benzyl alkoxide. Thus, the subsequent species may be either
one. Following the profile in Figure , the coordinated alkoxide in D is protonated
by the N–H proton of the PNP arm, with assistance of the ethanol
molecule, from D to E. This process has
a barrier of 13 kcal/mol and is endergonic, with ΔG = 7 kcal/mol. Intermediate F is 3 kcal/mol more stable
than the reactants and features a dearomatized PNP ligand. The HOMO
and LUMO of complex F are depicted in Figure . The HOMO corresponds to the
ligand π-system, with a significant contribution of the lone
pair of the deprotonated N-atom. The LUMO is essentially metal z2 pointing toward the empty coordination position.
Figure 1
Free energy
profile calculated for the hydrogenation of benzaldehyde
catalyzed by the hydride complex Awith ligand
N–H bond participation. Free energies (kcal/mol) are
referred to the initial reactants (A), and relevant distances
(Å) are presented.
Figure 2
HOMO and LUMO of deprotonated Mn1 (F in
calculations).
Free energy
profile calculated for the hydrogenation of benzaldehydecatalyzed by the hydridecomplex Awith ligand
N–H bond participation. Free energies (kcal/mol) arereferred to the initial reactants (A), and relevant distances
(Å) are presented.HOMO and LUMO of deprotonated Mn1 (F in
calculations).The reaction continues
along the profile represented in Figure , F′ being equivalent
to F with a different relative orientation
of the three molecules. Exchange of benzyl alcohol by one H2 molecule produces intermediate G. Dihydrogencoordination
is facile, with a barrier of only 1 kcal/mol (TS) in a clearly exergonic step, ΔG = −9 kcal/mol. The resulting intermediate H is an η2-H2complex, which is
14 kcal/mol more stable than the initial reagents. Rearrangement of
the H-bond network between the H2complex and the nearby
ethanol molecule changes H into I. In the
final step, there is splitting of the H–H bond with re-protonation
of the PNP N-atom and regeneration of the hydride ligand in J, corresponding to the initial reactant A and
an ethanol molecule. The last step is exergonic, with J being 25 kcal/mol more stable than A. Despite the presence
of an ethanol molecule acting as a proton shuttle, the associated
barrier is significant (ΔG⧧ = 21 kcal/mol). The highest barrier along path I is 25 kcal/mol,
corresponding to the difference between intermediate H, the most stable one, and transition state TS.
Figure 3
Free energy profile calculated for the hydrogenation of
benzaldehyde
catalyzed by the hydride complex A in a bifunctional
mechanism with ligand N–H bond participation. The free energy values (kcal/mol) are referred to the initial reactants
(A), and relevant distances (Å) are presented.
Free energy profile calculated for the hydrogenation of
benzaldehydecatalyzed by the hydridecomplex A in a bifunctional
mechanism with ligand N–H bond participation. The free energy values (kcal/mol) arereferred to the initial reactants
(A), and relevant distances (Å) are presented.For comparison, the first step
of the mechanism was also calculated
for acetophenone as substrate. The barrier for the attack of the hydride
ligand in complex A to the carbonyl C-atom of a free
acetophenone molecule is significantly higher than the one calculated
for benzaldehyde (18 vs 11 kcal/mol, respectively; SI, Figure S2). This trend is in accordance with the fact
that ketones are not hydrogenated under the same reaction conditions.
The remarkable substrate selectivity was recently also explained by
the relative stability of alkoxide intermediates formed upon aldehyde
insertion into the metal–H bond in the case of related ironPNPpincercomplexes based on DFT calculations.[21] It has to be noted that the related Mn(I) PNPpincercomplex
[Mn(PNP-iPr)(CO)3]Br (Scheme ) was shown to act as a catalyst
for the hydrogenation of ketones but at a catalyst loading of 5 mol%,
a temperature of 130 °C in the presence of 10 mol% base, and
a hydrogen pressure of 50 bar in toluene as solvent.[10d]The alternative mechanism (path II) shares the first
part in Figure until
formation
of the cationic intermediate C. Following the profile
represented in Figure , addition of H2 to C yields intermediate K. From here, coordination of dihydrogen is easy, with a barrier
of merely 1 kcal/mol (TS) in
an exergonic step (ΔG = −8 kcal/mol).
The difference between the two mechanisms is that while hereH2coordinates to complex [Mn(PNP)(CO)2]+, producing the cationicdihydrogencomplex [Mn(PNP)(η2-H2)(CO)2]+, in path I that
process occurs with the neutral metallic species [Mn(PNP′)(CO)2], featuring a deprotonated PNP ligand (PNP′), and
yields the corresponding neutral H2complex: [Mn(PNP′)(η2-H2)(CO)2]. The mechanism proceeds from L with protonation of the free alkoxide by means of the coordinated
H2. The associated barrier (TS) is negligible (1 kcal/mol), and the resulting species (M) is 13 kcal/mol more stable than A. The highest
barrier in path II is 26 kcal/mol, measured between the O-coordinated
alkoxidecomplex D and the highest following transition
state TS. This is the transition
state associated with H2coordination and formation of
the dihydrogencomplex in L. It has to be noted that
the same reaction pathway was recently established for the chemoselective
hydrogenation of aldehydescatalyzed by [Fe(PNPMe-iPr)(CO)(H)(Br)], wheremetal–ligand cooperation
was not possible due NMe linkers.[15]
Figure 4
Free energy
profile calculated for the hydrogenation of benzaldehyde
catalyzed by the hydride complex Awithout ligand
N–H bond participation. The free energy values (kcal/mol)
are referred to the initial reactants (A), and relevant
distances (Å) are presented.
Free energy
profile calculated for the hydrogenation of benzaldehydecatalyzed by the hydridecomplex Awithout ligand
N–H bond participation. The free energy values (kcal/mol)
arereferred to the initial reactants (A), and relevant
distances (Å) are presented.In path I, alkoxide protonation is accomplished by the N–H
proton in the PNP ligand, yielding a metallic fragment with a dearomatized
PNP ligand. This corresponds to a bifunctional mechanism with participation
of the PNP ligand that is further regenerated by the coordinated H2 molecule. In path II, there is no participation of the PNP
ligand, and alkoxide protonation is made directly by the coordinated
H2 molecule. The difference between the highest barriers
calculated for the two mechanisms is only 1 kcal/mol (25 kcal/mol
for path I, 26 kcal/mol for path II); thus, in principle, both could
occur under the experimental conditions. If entropy corrections for
non-standard conditions areconsidered, the total barrier for path
I rises to 26.5 kcal/mol due to the lower molecularity of TS when compared to TS and to the reaction conditions. This makes path I
slightly less favorable than path II.
Conclusion
Several
hydrideMn(I) and Re(I) PNPpincercomplexes were prepared
and tested as catalysts for the homogeneous chemoselective hydrogenation
of aldehydes. [Mn(PNP-iPr)(CO)2(H)] (Mn1), based on the 2,6-diaminopyridine scaffold, where the
PiPr2 moieties of the PNP ligandconnect
to the pyridine ring via NH linkers, was found to be the most efficient
catalyst for this process. The reaction is highly chemoselective also
in the presence of other functional groups which can be hydrogenated,
such as ketones, esters, alkynes, olefins, nitriles, and α,β-unsaturated
double bonds. The low catalyst loadings (0.1–0.05 mol%), mild
and base-free reaction conditions (25 °C, 50 bar H2), and broad applicability make this catalyst interesting for the
syntheses of fine and bulk chemicals. The catalysis works also with
lower catalyst loadings (0.005 mol%) but requires then the addition
of an external base. Based on experimental and computational studies,
a bifunctional mechanism with participation of the PNP ligand (deprotonation/protonation)
is proposed. An alternative mechanism without participation of the
PNP ligandcannot be fully dismissed but seems to be less likely.
Surprisingly, analogous isoelectronic and isostructural Re(I) complexes
turned out to be only poorly active.
Authors: Robert Langer; Mark A Iron; Leonid Konstantinovski; Yael Diskin-Posner; Gregory Leitus; Yehoshoa Ben-David; David Milstein Journal: Chemistry Date: 2012-04-24 Impact factor: 5.236
Authors: Wenjun Yang; Tejas Y Kalavalapalli; Annika M Krieger; Taras A Khvorost; Ivan Yu Chernyshov; Manuela Weber; Evgeny A Uslamin; Evgeny A Pidko; Georgy A Filonenko Journal: J Am Chem Soc Date: 2022-04-27 Impact factor: 16.383
Authors: Wenjun Yang; Ivan Yu Chernyshov; Robin K A van Schendel; Manuela Weber; Christian Müller; Georgy A Filonenko; Evgeny A Pidko Journal: Nat Commun Date: 2021-01-04 Impact factor: 14.919