Alcohol amination with ammonia catalyzed by an acridine-based ruthenium pincer complex: A mechanistic study

Article abstract of DOI:10.1021/ja409368a

The mechanistic course of the amination of alcohols with ammonia catalyzed by a structurally modified congener of Milstein's well-defined acridine-based PNP-pincer Ru complex has been investigated both experimentally and by DFT calculations. Several key Ru intermediates have been isolated and characterized. The detailed analysis of a series of possible catalytic pathways (e.g., with and without metal-ligand cooperation, inner- and outer-sphere mechanisms) leads us to conclude that the most favorable pathway for this catalyst does not require metal-ligand cooperation.

Full text of DOI:10.1021/ja409368a

Article
pubs.acs.org/JACS
Alcohol Amination with Ammonia Catalyzed by an Acridine-Based
Ruthenium Pincer Complex: A Mechanistic Study
Xuan Ye,† Philipp N. Plessow,†,‡ Marion K. Brinks,§ Mathias Schelwies,§ Thomas Schaub,§
Frank Rominger,∥ Rocco Paciello,§ Michael Limbach,†,§ and Peter Hofmann*,†,∥
†Catalysis Research Laboratory (CaRLa), Im Neuenheimer Feld 584, D-69120 Heidelberg, Germany
‡Quantum Chemistry, BASF SE, Carl-Bosch-Straße 38, D-67056 Ludwigshafen, Germany
§Synthesis & Homogeneous Catalysis, BASF SE, Carl-Bosch-Straße 38, D-67056 Ludwigshafen, Germany
∥Organisch-Chemisches Institut, Ruprecht-Karls-Universitat Heidelberg, Im Neuenheimer Feld 270, D-69120 Heidelberg, Germany
̈
S
* Supporting Information
ABSTRACT: The mechanistic course of the amination of alcohols with
ammonia catalyzed by a structurally modified congener of Milstein’s well-
defined acridine-based PNP-pincer Ru complex has been investigated both
experimentally and by DFT calculations. Several key Ru intermediates have
been isolated and characterized. The detailed analysis of a series of possible
catalytic pathways (e.g., with and without metal−ligand cooperation, inner-
and outer-sphere mechanisms) leads us to conclude that the most favorable
pathway for this catalyst does not require metal−ligand cooperation.
INTRODUCTION
■
The direct synthesis of amines from alcohols via a hydrogen-
borrowing mechanism is atom-economical and environmentally
benign, with water as the only stoichiometric side product. The
commonly accepted mechanistic pathway is a sequence of
alcohol oxidation, aldehyde amine condensation, and imine
reduction (Figure 1).1
A variety of transition-metal-based homogeneous catalysts
(e.g., based on Ru,2 Ir,3 Cu,4 Fe,5 and Pd6) have been
developed for this reaction, and they even outperform some of
the state-of-the-art heterogeneous catalysts for alcohol amina-
tion with NH37 with respect to substrate scope, selectivity, and
activity.1a,8 Among those, Ru-based metal precursor/ligand
combinations C and D (Figure 1)9 and especially Milstein’s
well-defined acridine-based PNP-pincer Ru complex A9a are
powerful catalysts with unprecedented reactivity and selectivity
Figure 1. Ruthenium-based homogeneous catalysts or catalyst systems
A−D for the amination of alcohols with NH3 (bottom) and the basic
mechanistic pathway (top).
for primary amines employing ammonia as the nitrogen source.
This exceptional performance of A has been attributed to the
acridine-based PNP-pincer ligand framework, which cooperates
with the metal via dearomatization of the acridine backbone,10
as demonstrated by Milstein and co-workers in stoichiometric
studies of A with H2 and NH3.11,12 This reversible
aromatization−dearomatization is also considered a pivotal
mechanistic feature for a broader range of pyridine-based pincer
complexes13,14 in numerous catalytic transformations other
than alcohol amination,9a including hydrogenation of carbo-
nates, dehydrogenative homocoupling of alcohols or alcohols
and amines, and even water splitting.15,16 Nevertheless, detailed
mechanistic studies of acridine-based systems are scarce,11 and
the impact of the pincer ligand framework on the selectivity
toward primary amines remains speculative. The present study
aims to address those open questions.
Received: September 13, 2013
Published: March 31, 2014
© 2014 American Chemical Society
5923
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929

Journal of the American Chemical Society
Article
RESULTS AND DISCUSSION
Scheme 1. Solid-State (X-ray) Molecular Structures of
a
■
Catalyst Precusor B, Its Isomer B′, and Complex E
Experimental Investigation. In analogy to a synthetic
procedure described by Milstein and co-workers,9a we
synthesized the new complex B with cyclohexyl substituents
on the P atoms in 74% isolated yield from (PPh3)3Ru(H)-
(Cl)(CO) and 4,5-bis(dicyclohexylphosphinomethyl)acridine9a
(70 °C, 2 h; cf. the reaction shown at the top of Table 1). In a
Table 1. Synthesis of Acridine-Based Ru Pincer Complex B
and Tests of Catalysts A−D in the Amination of 1-Octanol
a
with NH3
a
b
entry
cat.
T (°C)
conv. (%)
sel. (%)
1
2
3
4
5
6
A
A
B
B
C
D
135
155
135
155
155
155
37
99
73
99
20
16
85
95
91
95
32
69
a
Ellipsoids are drawn at 50% propability, and H atoms have been
omitted for clarity. B: P−Ru−N, 91.92(5)/92.08(5)°; H−Ru−P1,
88.9(11)°; H−Ru−N, 173.9(11)°; Ru−CO, 1.801(3) Å; Ru−P,
2.316(8)/2.3073(8) Å; Ru−N, 2.488(2) Å; Ru−Cl, 2.4792(8) Å;
Ru−H, 1.45(3) Å. B′: P−Ru−N, 90.73(4)/89.66(4)°; H−Ru−Cl,
172.3(7)°; OC−Ru−P1, 92.34(7)°; Ru−CO, 1.796(2) Å; Ru−P,
2.3145(5)/2.3264(6) Å; Ru−N, 2.4389(17) Å; Ru−Cl, 2.5195(6) Å;
Ru−H, 1.507(19) Å. E: P2−Ru−N, 87.11(12)°; P2−Ru−Cl,
145.25(6)°; N−Ru−Cl, 83.91(11)°; P−Ru−P, 114.32(6)°; Ru−CO,
1.805(7) Å; Ru−P, 2.3121(15)/2.2466(16) Å; Ru−N, 2.249(4) Å;
Ru−Cl, 2.4609(15) Å; Ru−C, 2.208(5) Å.
a
[1-octanol] = 1 M, 0.1 mol % catalyst, p(NH3) = 35−40 bar, toluene
(50 mL), 12 h, 160 mL stainless-steel autoclave. Selectivity in 1-
octylamine.
b
benchmark reaction, namely, the amination of 1-octanol with
NH3, the modified catalyst B exhibited reactivity and selectivity
comparable to those of Milstein’s prototype A and superior to
those of the other Ru-based two-component catalysts C and D
(Table 1).9
Red crystals of B were obtained from a toluene/hexane
mixture, and its molecular structure in the crystal revealed a
pseudo-octahedral geometry of the Ru center with an elongated
Ru−N bond in comparison with A16 (2.488 vs 2.479 Å; cf.
Scheme 1). Noteworthy is the trans geometry of the acridyl N
and the hydride ligand, which contrasts the cis geometry
reported by Milstein and co-workers in their system.16
Interestingly, crystals of complex B′, a stereoisomer of complex
B that features the cis geometry of the acridyl N and the
hydride ligand as reported for A, were obtained upon cooling of
a toluene solution of B and 1-hexylamine that had been heated
at 80 °C for 16 h (Ru−N = 2.439 Å; Scheme 1).
only 2 h at room temperature. Concurrently, the new Ru
complex F with a dearomatized acridine backbone formed from
B in a clean reaction (Scheme 2).17 The crystal structure of F
revealed that the Ru sits in the center of a pseudotrigonal
bipyramid and that the central ring of the dearomatized
acridine ligand adopts a boat-shaped structure. The Ru−N
Scheme 2. Stoichiometric Alcohol Oxidation by Catalyst
Precursor B in the Presence of Base and Solid-State (X-ray)
Molecular Structure of F
a
When the recrystallization of B was carried out under air,
black/brown crystals of the new complex E were obtained, and
its crystal structure revealed that one of the benzylic C−H units
was cyclometalated by the Ru center. In comparison to the
non-cyclometalated complexes B and B′, the Ru−N bond
distance was significantly shortened (2.245 Å; cf. Scheme 1).
This cyclometalation reaction under mild oxidative conditions
implies that the benzylic protons of the PNP-pincer ligands are
susceptible to C−H activation by the Ru center.
The stoichiometric reaction of B with a primary alcohol (e.g.,
benzyl alcohol, 1-octanol, each 2−4 equiv) without an external
base did not yield alcohol oxidation products according to
NMR analysis (toluene-d8, 70 °C, 16 h). In the presence of
NaOtBu or NaH (4−14 equiv), benzyl benzoate, the product of
a formal Claisen−Tishchenko reaction, was obtained within
a
Ellipsoids are drawn at 50% probability, and H atoms have been
omitted for clarity. F: P−Ru−N, 91.38(8)/91.40(8)°; N−Ru−CO,
153.29(14)°; P−Ru−P, 155.02(3)°; Ru−CO, 1.814(4) Å; Ru−P,
2.3196(10)/2.3219(9) Å; Ru−N, 2.109(3) Å; Ru−H, 1.49(4) Å.
5924
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929

Journal of the American Chemical Society
Article
bond is significantly shortened in comparison with that in B
(2.109 vs 2.488 Å) upon dearomatization and is in the range of
the Ru−N bond in the iPr derivative observed by Milstein
(2.109 vs 2.171 Å).11 With the deuterated substrate benzyl
alcohol-α,α-d2, deuterium scrambling into the hydride (H1),
benzylic C−H (H2), and methylene C−H (H4, H5) positions of
F was observed, even at room temperature (Scheme 2).18
Scrambling including the phosphine substituents has also been
observed previously in the reaction of A with D2 and KOH.11
Heating a mixture of F (0.5 equiv) and imine 1 to 100 °C for
2 days predominantly isomerized the imine’s double bond.
Only minor amounts of benzylamine (3) were formed, as
shown by 1H NMR analysis (1:2:3 = 29:57:14; Scheme 3). The
Scheme 4. Reaction of Alcohol/Ammonia and Aldehyde
under the Influence of Complex H
Ru species. The yield of benzylamine exceeded the maximum
yield possible for a stoichiometric reductive amination by a
factor of 2, which points to a catalytic reaction. The reductants
in this stoichiometric redox transformation are proposed to be
H and the in situ-generated benzylidene imine, which is
oxidized to PhCN. The same reaction, but now in the presence
of 1-octanol (0.8 equiv), gave benzylamine and octylamine in
greater than stoichiometric yield (68 and 52%, respectively; cf.
Scheme 4) in addition to multiple byproducts (i.e., heptyl
nitrile, octyl benzoate, N-benzylidene-1-phenylmethanamine,
N-benzylidene-1-heptylmethanamine). The higher yield of 5
(44 vs 68%) implies that the alcohol facilitates the reductive
amination of aldehyde by H.
a
Scheme 3. Intermediate F as a Reductant
When B (20 mol %) was heated in a mixture of 1-octanol
and NH3 (6 equiv) at 150 °C for 16 h, complex H was
identified as the major organometallic species (30% by 31P
NMR spectroscopy) in addition to the expected amine/imine
mixture. A series of control experiments was performed (0.1
mol % B, 135−140 °C, 12 h, toluene or THF, autoclave) to
prove that primary and secondary amines do not interconvert
under the given reaction conditions. Amination of 1-octanol
with NH3 (6 equiv) in the presence of 1-octylamine (1 equiv)
yielded negligible amounts of dioctylamine, and only 50% of
the initial 1-octanol was converted, indicating a moderate
inhibitory effect of the amine product on the reaction rate.
Amination of 1-octanol with NH3 (12 equiv) in the presence of
hexanal (1 equiv) gave a mixture of 1-octylamine and 1-
hexylamine (39 and 19%, respectively) besides the correspond-
ing nitriles. The reaction of solely 1-octylamine under our
standard conditions afforded only traces of dioctylamine
(∼1%), and the same experiment with dioctylamine and NH3
yielded only traces of 1-octylamine (<1%).
To summarize, catalyst precursor B readily oxidizes primary
alcohols to esters in the presence of base at room temperature
and is concurrently reduced to Ru hydride complex F featuring
a dearomatized acridine pincer ligand. Complex F is capable of
reducing imines to amines at elevated temperature, and
reduction of imines by F is facilitated by the addition of an
alcohol. In the presence of ammonia, complex F is rapidly
converted to the NH3-coordinated Ru hydride complex H at
room temperature. Complex H promotes the reductive
amination of aldehydes with ammonia to generate primary
amines, and catalytic turnover can be achieved at elevated
temperature with both complex H and the in situ-generated
imine (RCH2NH) as sacrificial reducing agents. Complex H
has also been characterized as the major Ru species in the
reaction mixture of alcohol amination with ammonia catalyzed
by precursor B. The primary amine selectivity in catalytic
alcohol amination is sensitive to the ammonia pressure. The
amine product from alcohol amination shows a moderate
inhibitory effect on the turnover rate of this reaction. Lastly,
a
R = p-OMe-C6H4, Ph = phenyl, Bn = benzyl. G: P−Ru−N, 88.74(8)/
89.29(8)°; H−Ru−P1, 72.5(16)°; H−Ru−N, 89.6(16)°; Ru−CO,
1.835(5)/1.945(7) Å; Ru−P, 2.3254(11)/2.3380(11) Å; Ru−N,
2.238(3) Å; Ru−H, 1.34(4) Å.
conversion of F was high (70%), but its further fate remained
unclear. When the same reaction was repeated in the presence
of benzyl alcohol (1.8 equiv), the yield of 3 increased
significantly within 3 days (52 vs 14%). Benzyl benzoate (4)
was the final organic oxidation product, and it was formed in
22% yield (based on benzyl alcohol). Besides unreacted F (18%
yield), the bis(carbonyl) Ru hydride complex G (17% yield
based on F) was observed in the product mixture.19 The second
CO ligand might originate from decarbonylation of benzyl
alcohol.20 The sum of complexes F and G in the final solution
accounted for ca. 35% of the total amount of added Ru,
indicating the above-mentioned decomposition of F. Thus, F
can serve as a reductant in the transformation of an imine to an
amine.
Both benzaldehyde and NH3 already reacted with F at room
temperature, leading to either the rapid disproportionation of
benzaldehyde to give 4 or the rapid coordination of NH3 to the
Ru center to afford H (Scheme 3).21 Heating a mixture of
benzaldehyde, complex H (0.2 equiv), and NH3 (20 equiv) at
135 °C for 12 h resulted in the formation of benzylamine (by
1H NMR analysis) along with multiple organic side products
(e.g., benzonitrile, benzamide, benzyl benzoate, and N-
benzylidene-1-phenylmethanamine; cf. Scheme 4). The major-
ity of H (80%) decomposed to give a mixture of unidentified
5925
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929

Journal of the American Chemical Society
Article
primary and secondary amines do not interconvert under the
mentioned experimental conditions.
Scheme 6. Stereoisomers of Pentacoordinate Ru−H
Complex F and the Corresponding NH3-Coordinated Ru−H
a
Computational Investigation. We carried out computa-
tional studies to rationalize these experimental findings. All of
the geometries were optimized at the BP86/def2-SV(P)22 level
of density functional theory (DFT) with the corresponding
effective core potential for ruthenium.23 Calculations were
carried out using the resolution-of-identity approximation and
appropriate auxiliary basis functions24 with TURBOMOLE.25
Gas-phase free energies were obtained within the usual rigid-
rotator, harmonic-oscillator approximation for 150 °C and 50
bar.26 Absolute energies were obtained as single-point energies
at the M06/def2-TZVP//BP86/def2-SV(P)27 level of theory
with GAMESS-US.28 Solvent corrections to reaction free
energies, obtained with COSMO-RS,29 were typically <1.2
kcal/mol. Therefore, only gas-phase free-energies are discussed.
The influence of solvation was investigated by computing
solvation free energies at 150 °C and 50 bar with COSMO-RS
using the parametrization for BP86/def-TZVP. Pure toluene
and a toluene/methanol mixture with 90 mol % toluene were
studied as solvents.
Complex H and the Isomerization Pathway
a
Calculated free energies (ΔG⧧, ΔG) are shown in kcal/mol.
the most stable stereoisomer. H-mer was chosen as the
reference point (GH‑mer = 0 vs GF‑mer = 2.1 kcal/mol).31,32
The calculated structural features of F-mer correlate well with
its experimentally obtained crystal structure (cf. Scheme 2).
The isomerization from F-mer to F′-mer involves flipping of
the dearomatized acridine ring from one side of the PNP plane
to the other, and the most favorable computed isomerization
pathway consists of sequential NH3 coordination to form H-
mer, isomerization to give H′-mer via flipping of the ligand
backbone, and NH3 dissociation. The highest barrier in this
stepwise isomerization arises from the change in the orientation
of the dearomatized acridine backbone in H-mer (GTS1 = 27.0
kcal/mol). Although the stereoisomers F′-mer and F′-fac are
both higher in energy than F-mer (GF′‑mer = 8.5 kcal/mol and
Initiation. Our proposed mechanism for dearomatization of
B to give F in both the stoichiometric reaction with alkoxide
and alcohol (Scheme 2) and catalytic alcohol amination is
shown in Scheme 5. The alcohol is deprotonated by either the
Scheme 5. Proposed Mechanism for Initial Dearomatization
a
of B-mer To Give F-mer
G
F′‑fac = 12.0 kcal/mol vs GF‑mer = 2.1 kcal/mol), they could be
viable intermediates in the imine reduction step because of the
high reaction temperature of 150 °C.
a
Calculated free energies (ΔG⧧, ΔG) are shown in kcal/mol.
Mechanisms involving the fac geometry turned out to be the
most favorable and will now be discussed in detail. Other
investigated mechanisms and those involving the mer geometry
can be found in the Supporting Information. The formation of
Ru amide M3′-fac from F′-fac via the inner-sphere imine
insertion was not only thermodynamically favored (ΔG =
−11.6 kcal/mol; cf. Figure 2) but also had a low overall kinetic
barrier (1.8 kcal/mol). The orientation of the dearomatized
acridine central ring in Ru amide M3′-fac places the methylene
alkoxide or ammonia. Dissociation of chloride and coordination
of the alkoxide gives the complex M20-mer. Dearomatization
can then occur by hydride transfer to the C9 position of the
acridine ring, which requires only a relatively low barrier (ΔG⧧
= 19.2 kcal/mol).
Dearomatization induced by dihydrogen involving a direct
hydride transfer from Ru to C9 was studied computationally by
Milstein and co-workers.11 As the computed barrier found in
that work is significantly higher (ΔG⧧ = 25.2 kcal/mol), we
think it is less relevant in this context. This is also in agreement
with the difference in experimental conditions: Dearomatiza-
tion with alkoxide/alcohol requires 2 h at 25 °C, while
dearomatization with KOH/H2 was reported after 48 h in
refluxing toluene.11
Impact of Stereoisomers and Imine Reduction by
Complex F. Formaldimine (CH2NH) was used as the
substrate in the computational investigations of the imine
reduction step. We took into account different starting
coordination geometries of the reductant F, inner-sphere
versus outer-sphere mechanisms, and mechanisms with and
without metal−ligand cooperation (for a detailed discussion,
see the Supporting Information). As the dearomatized acridine-
based PNP-pincer ligand can adapt both mer and fac
coordination modes (Scheme 6),30 four stereoisomers are
possible for complex F (F-fac, F-mer, F′-mer, and F′-fac) and
for the corresponding NH3-coordinated complex H (H-fac, H-
mer, H′-mer, and H′-fac). Among those, F-mer turned out to be
Figure 2. (a) Ligand-assisted reduction vs (b) inner-sphere alcohol-
assisted reduction of methanimine by F′-fac. Calculated free energies
(ΔG⧧, ΔG) are shown in kcal/mol.
5926
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929

Journal of the American Chemical Society
Article
moiety in close proximity to the amide and renders metal−
ligand cooperation with the dearomatized acridine ring by
intramolecular protonation a reasonable pathway. However, the
direct, intermolecular protonation of Ru amide M3′-fac by
alcohol is energetically more favorable.33 Rearomatization of
the acridine backbone is therefore likely to occur only in the
stoichiometric imine reduction that leads to decomposition of
the Ru complex.11,12 In the presence of alcohol, as in the
catalytic reaction, direct protonation is certainly more favorable.
This also applies to the corresponding mechanisms in the mer
geometry, where intermolecular protonation by the alcohol
competes with intramolecular protonation by the benzylic
position. To access intermediate F′-fac, an initial kinetic barrier
of 27.0 kcal/mol must be overcome. Nevertheless, this kinetic
barrier is much lower than the overall barrier associated with
imine reduction by intermediate F-mer (33.7 kcal/mol).
Therefore, we postulate that the Ru hydride intermediate F′-
fac represents the active reducing species in the catalytic alcohol
amination.
Figure 3. Computed free energy diagram for catalytic methanol
amination with NH3.
Alcohol Oxidation and Regeneration of the Reducing
Species F′-fac and the Catalytic Cycle. Alcohol oxidation is
analogous but reverse in direction to imine reduction and can
therefore proceed via the same mechanisms. The potential
pathways for the alcohol to aldehyde oxidation and
regeneration of the active reducing species F′-fac, namely, an
inner-sphere β-hydride elimination mechanism and an outer-
sphere concerted hydrogen transfer mechanism, are nearly
isoenergetic (ΔΔG⧧ = 0.6 kcal/mol; cf. Scheme 7). Therefore,
both pathways could be operative under catalytic alcohol
amination conditions. Eisenstein and co-workers arrived at a
similar conclusion.34
Scheme 8. Proposed Catalytic Cycle for Alcohol Amination
with NH3 Catalyzed by B
Scheme 7. Inner- and Outer-Sphere Mechanistic Pathways
for the Formation of Formaldehyde and the Regeneration of
a
F′-fac
high kinetic barrier. Complex F′-fac reduces the imine, the
product of condensation of the aldehyde with NH3 (step i), via
sequential imine coordination and insertion into the Ru−H
bond, formation of a pentacoordinated Ru amide intermediate,
protonation of the amide moiety by the alcohol, and
dissociation of the primary amine product (steps ii−vi).
The regeneration of the reducing species F′-fac and the
formation of the aldehyde can proceed either through inner-
sphere β-hydride elimination (steps vii−ix) or outer-sphere
concerted hydrogen transfer (steps x−xi). The reaction of a
Ru−NH2 amide with an alcohol to give a Ru hydride, NH3, and
an aldehyde is known in the literature.33 The formation of the
pentacoordinate Ru amide (step iv) is both the turnover-limiting
a
Calculated free energies (ΔG⧧, ΔG) are shown in kcal/mol.
The favored pathways for catalyst initiation, imine reduction,
and alcohol oxidation sum up to the free energy diagram for the
catalytic alcohol amination reaction shown in Figure 3. Once
intermediate F′-fac is generated via the high-barrier initiation
step (27.0 kcal/mol), it promotes a surprisingly efficient
catalytic cycle with an overall kinetic barrier of only 1.8 kcal/
mol (GTS11 − GF′‑fac = 1.8 kcal/mol).
In detail, in the presence of NH3 as the base, catalyst
precursor B stoichiometrically oxidizes alcohol to aldehyde and
affords Ru hydride complex H-mer via dearomatization of the
acridine ligand (Scheme 8). The active reducing species F′-fac
is generated from H-mer via a stepwise initiation process with a
and selectivity-determining step of the catalytic cycle (GTS11
=
13.8 kcal/mol). This formal H2-borrowing redox cycle neither
requires a change in the formal oxidation state of ruthenium
nor metal−ligand cooperation. This is supported by the
5927
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929

Journal of the American Chemical Society
Article
following experimental findings: (1) the complex H-mer is the
major Ru species in the final reaction mixture of catalytic
alcohol amination with NH3; (2) the primary amine product
inhibits the catalytic cycle, as it competes with the imine for the
empty coordination site in reducing species F′-fac; and (3)
complex H-mer catalyzes the reductive amination of benzalde-
hyde with NH3 in the absence of alcohol.
Origin of the Selectivity toward the Primary Amine.
Although a N-substituted imine generated from the con-
densation of an aldehyde with the primary amine product could
also react with the reducing species F′-fac to afford the
corresponding secondary amine, the overall kinetic barrier of
the inner-sphere, alcohol-assisted imine reduction is signifi-
cantly higher than that of methylamine formation (11.8 vs 1.8
kcal/mol, respectively; Figure 4). Although the stability of N-
can be achieved at elevated temperature with both complex H
and the in situ-generated imine serving as reducing agents. H is
also the major Ru species in the reaction mixture of alcohol
amination with NH3 catalyzed by B. Various stereoisomers of
the pentacoordinated Ru hydride complex F exist as a result of
the flexible coordination modes of the anionic dearomatized
PNP-pincer ligand. Among the stereoisomers of F, F′-fac
turned out to be the most active species in imine reduction.
There is a quite high barrier for the formation of F′-fac.
However, once generated, intermediate F′-fac catalyzes the
conversion of the alcohol to the primary amine with a
surprisingly low overall barrier for turnover. The turnover- and
selectivity-determining step of the catalytic cycle is the
reduction of the imine by F′-fac, which consists of stepwise
formation of the pentacoordinated Ru amide intermediate and
subsequent protonation of the Ru amide by an alcohol. The
results from this study on structure and function provide an
important basis for further insights into the mechanism of
similar catalysts.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, additional schemes, characterization
data, and details of the calculations. This material is available
free of charge via the Internet at http://pubs.acs.org. CCDC
913746 (B), 913747 (B′), 913748 (E), 913749 (F), and
913750 (G) contain the supplementary crystallographic data
for this paper. These data can be obtained free of charge from
The Cambridge Crystallographic Data Centre via www.ccdc.
cam.ac.uk/data_request/cif.
AUTHOR INFORMATION
Corresponding Author
ph@oci.uni-heidelberg.de
■
Notes
The authors declare no competing financial interest.
Figure 4. Computed free energy diagrams for the formation of
dimethylamine (blue) and methylamine (black) via inner-sphere
alcohol-assisted imine reduction by F′-fac. Calculated energies (ΔG⧧,
ΔG) are shown in kcal/mol.
ACKNOWLEDGMENTS
X.Y., P.N.P., M.L., and P.H. work at CaRLa of Heidelberg
University, which is cofinanced by Heidelberg University, the
■
state of Baden-Wurttemberg, and BASF SE. Support from these
̈
institutions is gratefully acknowledged.
methylimine is higher than that of methanimine by 6.0 kcal/
mol,35 the formation of dimethylamine is still disfavored by 3.9
kcal/mol relative to methylamine. This 3.9 kcal/mol difference
in the overall kinetic barrier correlates well with the exceptional
primary amine selectivity of B observed experimentally. Since
the selectivity for primary amine depends directly on the
concentrations of the primary and secondary imines, the
mechanism explains the observed dependence of the selectivity
on the pressure of NH3.
REFERENCES
■
(1) For reviews of alcohol amination via the H2-borrowing method,
see: (a) Bahn, S.; Imm, S.; Neubert, L.; Zhang, M.; Neumann, H.;
̈
Beller, M. ChemCatChem 2011, 3, 1853−1864. (b) Watson, A. J. A.;
Williams, J. M. J. Science 2010, 329, 635−636. (c) Dobereiner, G. E.;
Crabtree, R. H. Chem. Rev. 2010, 110, 681−703. (d) Nixon, T. D.;
Whittlesey, M. K.; Williams, J. M. J. Dalton Trans. 2009, 753−762.
(e) Guillena, G.; Ramon
1641.
́
, D. J.; Yus, M. Chem. Rev. 2010, 110, 1611−
SUMMARY AND CONCLUSIONS
(2) For lead references on Ru-catalyzed alcohol amination, see:
(a) Hamid, M. H. S. A.; Allen, C. L.; Lamb, G. W.; Maxwell, A. C.;
Maytum, H. C.; Watson, A. J. A.; Williams, J. M. J. J. Am. Chem. Soc.
2009, 131, 1766−1774. (b) Hollmann, D.; Tillack, A.; Michalik, D.;
Jackstell, R.; Beller, M. Chem.Asian J. 2007, 2, 403−410. For a
recent mechanistic study based on the Xantphos ligand, see:
(c) Pingen, D.; Lutz, M.; Vogt, D. Organometallics 2014,
DOI: 10.1021/om4011998.
(3) For lead references on Ir-catalyzed alcohol amination, see:
(a) Ohta, H.; Yuyama, Y.; Uozumi, Y.; Yamade, Y. M. A. Org. Lett.
2011, 13, 3892−3895. (b) Kawahara, R.; Fujita, K.-i.; Yamaguchi, R. J.
Am. Chem. Soc. 2010, 132, 15108−15111. (c) Blank, B.; Madalska, M.;
■
We have shown that catalyst precursor B readily oxidizes
primary alcohols to esters in the presence of base at room
temperature and is concurrently reduced to the Ru hydride
complex F. The acridine-based PNP-pincer ligand plays a vital
role in this transformation, as one hydrogen atom is stored on
the acridine backbone of the anionic dearomatized PNP-pincer
ligand framework. Complex F is capable of reducing imines to
amines at elevated temperature, which is facilitated by an
alcohol. In the presence of NH3, F rapidly reacts to give the
NH3-coordinated Ru hydride complex H. Catalytic turnover
5928
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929

Journal of the American Chemical Society
Article
Kempe, R. Adv. Synth. Catal. 2008, 350, 749−758. (d) Fujita, K. I.;
Fujii, T.; Yamaguchi, R. Org. Lett. 2004, 6, 3525−3528. (e) Cami-
Kobeci, G.; Slatford, P. A.; Whittlesey, M. K.; Williams, J. M. J. Bioorg.
(19) Complex G can be independently synthesized by exposing a
toluene solution of F to a CO/H2 atmosphere (rt, <5 min). See the
Supporting Information.
(20) Decarbonylation of benzyl alcohol by Ru(Xantphos)(PPh3)
(CO)H2 led to the formation of Ru(Xantphos)(CO)2H2. See: Ledger,
A. E. W.; Slatford, P. A.; Lowe, J. P.; Mahon, M. F.; Whittlesey, M. K.;
Williams, J. M. J. Dalton Trans. 2009, 716−722.
Med. Chem. Lett. 2005, 15, 535−537. (f) Wetzel, A.; Wockel, S.;
̈
Schelwies, M.; Brinks, M. K.; Rominger, F.; Hofmann, P.; Limbach, M.
Org. Lett. 2013, 5, 266−269.
(4) For lead references on Cu-catalyzed alcohol amination, see:
(21) In contrast to complex F, the hexacoordinated complex H does
not promote the dimerization of aldehyde at room temperature.
(22) (a) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
3297−3305. (b) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3100.
(c) Dirac, P. A. M. Proc. R. Soc. London, Ser. A 1929, 123, 714−733.
(d) Slater, J. C. Phys. Rev. 1951, 81, 385−390. (e) Vosko, S. H.; Wilk,
L.; Nusair, M. Can. J. Phys. 1980, 58, 1200−1211. (f) Perdew, J. P.;
Wang, Y. Phys. Rev. B 1992, 45, 13244−13249. (g) Perdew, J. P. Phys.
Rev. B 1986, 33, 8822−8824.
(a) Shi, F.; Tse, M. K.; Cui, X.; Gordes, D.; Michalik, D.; Thurow, K.;
̈
Deng, Y.; Beller, M. Angew. Chem., Int. Ed. 2009, 48, 5912−5915.
(b) Likhar, P. R.; Arundhathi, R.; Kantam, M. L.; Prathima, P. S. Eur. J.
Org. Chem. 2009, 5383−5389.
(5) For a lead reference on Fe-catalyzed alcohol amination, see:
Martínez, R.; Ramon
2181.
́
, D. J.; Yus, M. Org. Biomol. Chem. 2009, 7, 2176−
(6) For a lead reference on Pd-catalyzed alcohol amination, see:
Kwon, M. S.; Kim, S.; Park, S.; Bosco, W.; Chidrala, R. K.; Park, J. J.
Org. Chem. 2009, 74, 2877−2879.
(7) (a) Hayes, K. S. Appl. Catal., A 2001, 221, 187−195.
(b) Klinkenberg, J. L.; Hartwig, J. F. Angew. Chem., Int. Ed. 2011,
50, 86−95.
(23) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
̈
Theor. Chim. Acta 1990, 77, 123−141.
(24) Weigend, F. Phys. Chem. Chem. Phys. 2006, 8, 1057−1065.
(25) TURBOMOLE version 6.3 2011, a development of the
University of Karlsruhe and Forschungszentrum Karlsruhe GmbH,
1989−2007, TURBOMOLE GmbH, since 2007; available from
http://www.turbomole.com.
(8) Lawrence, S. A. Amines: Synthesis, Properties and Applications;
Cambridge University Press: Cambridge, U.K., 2005.
(9) For Ru-catalyed alcohol amination with NH3, see: (a) Gunana-
than, C.; Milstein, D. Angew. Chem., Int. Ed. 2008, 47, 8661−8664.
(26) In general, gas-phase free energies can be expected to
overestimate entropic contributions arising from changes in particle
number compared with free energies in solution. Since no mechanistic
hypothesis was ruled out solely on the basis of entropic effects, this is
not a severe issue.
(27) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241.
(28) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.;
Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.;
Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A., Jr. J. Comput.
Chem. 1993, 14, 1347−1363.
(29) (a) Eckert, F.; Klamt, A. COSMOtherm, version C3.0, release
12.01; COSMOlogic GmbH & Co. KG: Leverkusen, Germany, 2012.
(b) Eckert, F.; Klamt, A. AIChE J. 2002, 48, 369−385.
(30) For the crystal structure of the fac-coordinated (dearomatized
acridine PNP)Ru(CO)(NH3)Cl complex, see ref 11.
(31) Our calculations revealed that in complexes F and H the cis-N,H
geometry is generally favored over the trans-N,H configuration.
(32) Structural optimization of F-fac led to activation of the benzylic
C−H bond by the Ru center.
(33) The amide moiety in Ru amide complexes is exceptional basic.
See: Fulton, J. R.; Sklenak, S.; Bouwkamp, M. W.; Bergman, R. G. J.
Am. Chem. Soc. 2002, 124, 4722−4737.
(b) Imm, S.; Bahn, S.; Neubert, L.; Neumann, H.; Beller, M. Angew.
̈
Chem., Int. Ed. 2010, 49, 8126−8129. (c) Pingen, D.; Muller, C.; Vogt,
D. Angew. Chem., Int. Ed. 2010, 49, 8130−8133. (d) Gunanathan, C.;
Milstein, D. Science 2013, 341, 249−259.
(10) For examples of aromatization−dearomatization in related
systems, see: (a) Schwartsburd, L.; Iron, M. A.; Konstantinovski, L.;
Diskin-Posner, Y.; Leitus, G.; Shimon, L. J. W.; Milstein, D.
Organometallics 2010, 29, 3817−3827. (b) Khaskin, E.; Iron, M. A.;
Shimon, L. J. W.; Zhang, J.; Milstein, D. J. Am. Chem. Soc. 2010, 132,
8542−8543. (c) Feller, M.; Ben-Ari, E.; Iron, M. A.; Diskin-Posner, Y.;
Leitus, G.; Shimon, L. J. W.; Konstantinovski, L.; Milstein, D. Inorg.
Chem. 2010, 49, 1615−1625. (d) van der Vlugt, J. I.; Reek, J. N. H.
Angew. Chem., Int. Ed. 2009, 48, 8832−8846. (e) Ben-Ari, E.; Leitus,
G.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc. 2006, 128, 15390−
15391.
(11) Gunanathan, C.; Gnanaprakasam, B.; Iron, M. A.; Shimon, L. J.
W.; Milstein, D. J. Am. Chem. Soc. 2010, 132, 14763−14765.
(12) For a review of the noninnocent behavior of PCP- and PCN-
pincer ligands in transition-metal complexes, see: Poverenov, E.;
Milstein, D. Top. Organomet. Chem. 2013, 40, 21−48.
(34) Nova, A.; Balcells, D.; Schley, N. D.; Dobereiner, G. E.;
Crabtree, R. H.; Eisenstein, O. Organometallics 2010, 29, 6548−6558.
(35) The free energy difference of 6.0 kcal/mol between CH2
NCH3 and CH2NH was obtained by comparing the ΔG values for
their formations from a mixture of formaldehyde and equal
concentrations of CH3NH2 and NH3. See the Supporting Information.
(13) For evidence of metal−ligand cooperation via aromatization−
dearomatization in pyridine-based PNP- and PNN-pincer metal
complexes, see: (a) Balaraman, E.; Gunanathan, C.; Zhang, J.;
Shimon, L. J. W.; Milstein, D. Nat. Chem. 2011, 3, 609−614.
(b) Kohl, S. W.; Weiner, L.; Schwartsburd, L.; Konstantinovski, L.;
Shimon, L. J. W.; Ben-David, Y.; Iron, M. A.; Milstein, D. Science 2009,
324, 74−77.
(14) For computational studies of metal−ligand cooperation via
aromatization−dearomatization of pyridine-based PNP- and PNN-
pincer metal complexes, see: (a) Yang, X.; Hall, M. B. J. Am. Chem. Soc.
2010, 132, 120−130. (b) Li, H.; Wang, X.; Huang, F.; Lu, G.; Jiang, J.;
Wang, Z.-X. Organometallics 2011, 30, 5233−5247. (c) Zeng, G.; Li, S.
Inorg. Chem. 2011, 50, 10572−10580. (d) Li, H.; Wen, M.; Wang, Z.-
X. Inorg. Chem. 2012, 51, 5716−5727.
(15) For a review of metal−ligand cooperation via aromatization−
dearomatization, see: Gunanathan, C.; Milstein, D. Acc. Chem. Res.
2011, 44, 588−602.
(16) Gunanathan, C.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
2009, 131, 3146−3147.
1
(17) A time course study of this redox reaction by in situ H NMR
spectroscopy showed no additional Ru intermediate during this
transformation.
1
2
(18) See the Supporting Information for H and D NMR spectra
and corresponding analysis.
5929
dx.doi.org/10.1021/ja409368a | J. Am. Chem. Soc. 2014, 136, 5923−5929