Ligand effects on the hydrogenation of biomass-inspired substrates with bifunctional Ru, Ir, and Rh complexes

Article abstract of DOI:10.1002/cssc.201300363

We herein report on the application and structural investigation of a new set of complexes that contain bidentate N-heterocyclic carbenes (NHCs) and primary amine moieties of the type [M(arene)Cl(L)] [M=Ru, Ir, or Rh; arene=p-cymene or pentamethylcyclopentadienyl; L=1-(2-aminophenyl)-3-(n-alkyl) imidazol-2-ylidine]. These complexes were tested and compared in the hydrogenation of acetophenone with hydrogen. Structural variations in the chelate ring size of the heteroditopic ligand revealed that smaller chelate ring sizes in combination with ring conjugation in the ligand are beneficial for the activity of this type of catalyst, favoring an inner-sphere coordination pathway. Additionally, increasing the steric bulk of the alkyl substituent on the NHC aided the reaction, showing almost no induction period and formation of a more active catalyst for the n-butyl complex relative to complexes with smaller Me and Et substituents. As is common in hydrogenation reactions, the activity of the complexes decreases in the order Ru>Ir>Rh. The application of [Ru(p-cym)Cl(L)]PF₆, which outperforms its reported analogues, has been successfully extended to the hydrogenation of more challenging biomass-inspired substrates. Adding value to biomass: The influence of structural variations of N-heterocyclic carbene-amine motif in half-sandwich Ru, Ir, and Rh complexes on the hydrogenation of biomass-inspired substrates is reported. Through this study insight is gained for the future rational design of hydrogenation catalysts for the conversion of biomass into useful/added-value compounds. Copyright

Full text of DOI:10.1002/cssc.201300363

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201300363
Ligand Effects on the Hydrogenation of Biomass-Inspired
Substrates with Bifunctional Ru, Ir, and Rh Complexes
Eveline Jansen,[a] Linda S. Jongbloed,[a] Dorette S. Tromp,[a] Martin Lutz,[b] Bas de Bruin,[a] and
Cornelis J. Elsevier*[a]
We herein report on the application and structural investiga-
tion of a new set of complexes that contain bidentate N-het-
erocyclic carbenes (NHCs) and primary amine moieties of the
type [M(arene)Cl(L)] [M=Ru, Ir, or Rh; arene=p-cymene or
pentamethylcyclopentadienyl; L=1-(2-aminophenyl)-3-(n-alky-
l)imidazol-2-ylidine]. These complexes were tested and com-
pared in the hydrogenation of acetophenone with hydrogen.
Structural variations in the chelate ring size of the heteroditop-
ic ligand revealed that smaller chelate ring sizes in combina-
tion with ring conjugation in the ligand are beneficial for the
activity of this type of catalyst, favoring an inner-sphere coordi-
nation pathway. Additionally, increasing the steric bulk of the
alkyl substituent on the NHC aided the reaction, showing
almost no induction period and formation of a more active
catalyst for the n-butyl complex relative to complexes with
smaller Me and Et substituents. As is common in hydrogena-
tion reactions, the activity of the complexes decreases in the
order Ru>Ir>Rh. The application of [Ru(p-cym)Cl(L)]PF6, which
outperforms its reported analogues, has been successfully ex-
tended to the hydrogenation of more challenging biomass-in-
spired substrates.
Introduction
Environmental issues and depletion of fossil-fuel reserves are
stimulating society to develop renewable sources of energy
and materials:[1–3] sources that are either abundantly available
or replaceable. The current feedstock of choice when it comes
to finding a renewable carbon source is biomass (i.e., indirectly
CO2). In addition to converting biomass into alternative fuels,
this feedstock can be used to render many other processes
sustainable. In light of these developments a set of key inter-
mediates (platform molecules) originating from processed bio-
mass has been defined.[4] These can be transformed into
a large selection of valuable products. To achieve this, methods
for the re- and defunctionalization of biogenic substrates have
to be developed and optimized, creating a toolkit filled with
selective catalytic conversions that a chemist can use to create
the desired product.[5,6] For the pharmaceutical industry this
predominantly means development of homogeneously cata-
lyzed conversions because the mild conditions associated with
these processes characteristically allow for the retention of nat-
ural complexity or incorporation of difficult functionalities in
a molecule.
N-Heterocyclic carbenes (NHCs) have already found wide-
spread use in many catalytic conversions, including those con-
cerning biomass. The research groups of Marr,[7] Crabtree,[8] Ya-
maguchi,[9] and Peris[10] produced metal–NHC complexes that
were shown to be applicable for the conversion of platform
chemicals or biomass-inspired substrates. The hydrogenation
of oxygenates, such as ketones, esters, and similarly functional-
ized compounds using molecular hydrogen, is an attractive
green process,[11] and an attractive method to convert biomass
into useful chemicals. Especially, the discovery of the NH-effect
by Noyori et al.[12] provided a breakthrough in these hydroge-
nation reactions. This bifunctional mechanism is proposed to
proceed in the outer coordination sphere, where an NH
moiety activates, for example, a carbonyl-containing substrate
through a NH···O=C hydrogen bond, which preorganizes and
facilitates the attack of the carbonyl group by the metal hy-
dride in a pericyclic six-membered transition state.[13,14] Since
that discovery, bifunctional catalysts have played a key role in
several pioneering reports on polar bond hydrogenation reac-
tions, most notably by the research groups of Noyori,[15]
Morris,[16] Clarke,[17] and Milstein.[18]
[a] E. Jansen, L. S. Jongbloed, Dr. D. S. Tromp, Prof. Dr. B. de Bruin,
Prof. Dr. C. J. Elsevier
van ‘t Hoff Institute for Molecular Chemistry
University of Amsterdam
In our design of new ligands for carbonyl-group hydrogena-
tion reactions, we decided to combine a NHC moiety, provid-
ing favorable donating properties widely used in hydrogena-
tion reactions,[19] with an amine functionality possibly allowing
bifunctional substrate activation.[20,21] In general, the use of li-
gands that contain NHC and primary amine groups in coordi-
nation chemistry is thus far poorly developed and only a few
transition-metal-based catalysts based on this motif have been
Science Park 904, 1098 XH Amsterdam (The Netherlands)
E-mail: c.j.elsevier@uva.nl
[b] Dr. M. Lutz
Crystal and Structural Chemistry
Bijvoet Center for Biomolecular Research
Utrecht University
Padualaan 8, 3584 CH Utrecht (The Netherlands)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201300363.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1737

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
developed to date.[22–25] The field is, however, rapidly develop-
ing due to the promising results obtained with this ligand scaf-
fold thus far. Hence, to develop new tools for efficient conver-
sion of biomass-derived platform molecules and model sub-
strates thereof, we synthesized a set of new NHC–amine li-
gands and used these to prepare a set of group 8 and 9 transi-
tion-metal complexes of general structure I (Figure 1).
alytic activity in H2-hydrogenation of ketones in more detail.
In general, the activity of a catalyst is closely related to its
structure: a small structural change can bring about large dif-
ferences in activity and selectivity. To this end, we studied the
influence of the chelate ring size and the size of the alkyl sub-
stituent on the ligand and we also varied the metal.
The influence of ligand variations on the performance of
[Ru(p-cym)(L)] complexes was investigated first. Syn-
theses of related [RuCp*(L)] complexes proved trou-
blesome (vide infra), confirming the synthetic prob-
lems reported by Cross et al.[23] Additionally, ligand
effects and structure–activity relations were more
easily detected for somewhat slower catalysts, which
was an additional reason why we focused on [Ru-
(p-cym)(L)] rather than [RuCp*(L)] complexes in our
studies. For comparative investigations with iridium
and rhodium we synthesized the Cp* analogues be-
cause the p-cym precursor was not accessible. To
evaluate the potential of the NHC–amine motif, we
Figure 1. General structures of half-sandwich complexes synthesized herein (I), and by
investigated the performance of these types of cata-
lysts on more challenging compounds: biomass-
inspired model substrates, such as cinnamaldehyde,
and actual platform chemicals, such as levulinic acid
(LA) and hydroxymethyl furfural (HMF). Additionally, esters
such as methyl butyrate (MB) or dimethyl oxalate (DMO) were
tested. In doing so, we initiated broadening of the substrate
scope of these NHC–amine complexes to include these plat-
form chemicals. These substrates are of particular interest in
the fine-chemical and pharmaceutical sector because they are
all key building blocks or precursors for sustainable chemicals,
materials, and biofuels.
the groups of Cross (II)[22] and Morris (III)[25–28]
.
An analogous NHC–amine complex (Figure 1, structure III)
was previously reported by Morris et al.,[25–28] but differs on
two distinct points: the chelate ring that is formed is larger
and the electronic structure lacks the conjugation with the
amine function present in our ligand. They performed exten-
sive DFT studies on the performance of the catalyst in, among
others, the H2-hydrogenation of ketones, which laid a firm
foundation in understanding the basics of the reaction. During
the course of our investigation Cross et al.[22] reported a similar
complex (Figure 1, structure II), but until now they only tested
transfer hydrogenation reactions.[26–28]
Morris et al.[27] showed that, based on DFT calculations, Results and Discussion
a [Ru(p-cym)(L)] complex (p-cym=p-cymene, L=NHC–amine-
Synthesis and characterization of ligands and catalysts
type ligand) in the H2-hydrogenation of acetophenone favors
an inner-sphere bifunctional mechanism, whereas a pentame-
thylcyclopentadienyl (Cp*) substituent facilitates a much faster
The ligands used were prepared according to the procedures
shown in Scheme 1. 2-(2-Imidazolyl)aniline was obtained by an
Ulmann-type coupling followed by hydrogenation of the nitro
group. The imidazolium ligand precursors (1–4) were then ob-
tained in good yields by N-alkylation in a sealed tube with the
respective iodo- or bromoalkyl reagent. These were subse-
quently used to prepare the ruthenium complexes 6 and 9–11
(Scheme 1).
[Ru(p-cym)Cl(L)]PF6 (6) was prepared from the isolated silver
carbene precursor (1a), followed by reaction with the
[Ru(p-cym)Cl2]2 precursor in the presence of a large excess of
KPF6 (route A). [Ru(p-cym)(ligand)]I complexes 9–11 were ob-
tained in a one-pot reaction (route B) from 2–4, but this route
resulted in lower yields. The analogous IrCp* and RhCp* com-
plexes were obtained in a similar manner following route A,
using [IrCp*Cl]2 and [RhCp*Cl2]2 as the metal precursors and
yielding 7 and 8, respectively. From complexes 6–8, crystals
that were suitable for crystallographic analysis were grown.[35]
Complex 6 is shown in Figure 3 and complex 7, which is iso-
structural with 8 and differs only in the disordered solvent
molecules, is shown in Figure 4.
Figure 2. Principle of carbonyl hydrogenation by H2. Two possible transition
states for hydride transfer are shown: outer-sphere bifunctional mechanism
(left) and direct inner-sphere mechanism (right).
outer-sphere bifunctional mechanism (Figure 2). In the H2-hy-
drogenation of acetophenone turnover frequencies (TOFs) of
213 and 10800 hÀ1 for [Ru(p-cym)(L)] and [RuCp*(L)], respec-
tively, have been reported.[26,27] Building on this work, we inves-
tigated the effect of structural variations of the catalyst on cat-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1738

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Scheme 1. Synthesis of the imidazolium ligands M–half-sandwich complexes through Route A and B.
Figure 4. Displacement ellipsoid plot of [IrCp*Cl(L)]PF6] (7) in the crystal,
drawn at the 50% probability level. CÀH hydrogen atoms, PF6 anion, and
disordered CH2Cl2 molecule are omitted for clarity.
Figure 3. Displacement ellipsoid plot of [Ru(p-cym)Cl(L)]PF6 (6) in the crystal,
drawn at the 50% probability level. CÀH hydrogen atoms and PF6 anion are
omitted for clarity. Only the major conformation of the disordered n-butyl
group is shown.
[RuCp*Cl2]n, and [RuCp*Cl(cod)] (cod=1,5-cyclooctadiene), re-
sulted in unidentifiable mixtures of products. Apparently, the
RuCp* complex with a 6-membered NHC–aniline chelate ring
was difficult to prepare and there are no reports of the synthe-
sis of this or similar complexes. A likely reason for this behavior
is the intrinsic reactivity of ruthenium towards the aromatic
aniline ring. When stirring ligand 5 and the [RuCp*(m3-Cl)]4 pre-
Several attempts to synthesize the analogous [RuCp*(L)]Cl
complex have proven unsuccessful. Unfortunately, using either
NaH, KOtBu, Ag2O, or sodium hexamethyldisilazide (NaHMDS)
as the base or using the appropriate silver complex for trans-
metallation of several metal precursors, such as [RuCp*(m3-Cl)]4,
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1739

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
cursor in THF in the absence of
base, fine yellow precipitate
of the RuÀN bond, an important factor when an inner-sphere
reaction mechanism requires dissociation of the amine. The
carbene carbon of 6 resonates slightly downfield in 13C NMR
spectra relative to 13 and 14, indicating a more electron-rich
carbene. That correlates to the shorter RuÀC bond length, sug-
gesting a stronger Ru–C interaction. A more electron-rich car-
bene means a more strongly donating NHC moiety, which is
likely beneficial for hydrogenation. Using the aniline ligand
with a conjugated system also renders the NHC group more
electron rich. Assuming that the [Ru(p-cym)] complexes follow
an inner-sphere route, the amine probably needs to dissociate
from the metal center. This is promoted by a more strained
system with a small chelating ring. The 13C NMR carbene sig-
nals of complexes 9–11, which differ in their alkyl substituents
appear slightly further downfield at 178.21 (nBu), 176.62 (Et),
and 179.09 ppm (Me).
a
formed. NMR experiments indicat-
ed formation of complex 12
(Figure 5). The chemical shifts of
the aromatic ring of the ligand as
detected in both 1H and 13C NMR
spectra (around 5.5 and 80 ppm,
respectively) were comparable to
those reported for the aniline
ligand in [RuCp*(h6-aniline)]Cl.[29]
The large downfield chemical shift
Figure 5. Proposed structure
of RuCp* compound.
of the amine protons in the 1H NMR spectrum of 12 to
6.95 ppm, showing only coupling to itself in (1H–1H) correlation
spectra (COSY) and no coupling to a carbon in (13C–1H) hetero-
nuclear single quantum coherence (HSQC) NMR spectra, sug-
gested that the amine protons became more acidic relative to
the free ligand due to resonance of the amine lone pair with
the p system of the aromatic ring. This readily explains the ob-
served and undesired coordination mode. The RuCp* analogue
containing a benzylic amine (general structure III) does not
share this resonance structure, and hence, in contrast to 12,[30]
this complex is readily synthesized in its desired NHC coordina-
tion mode.
Catalytic hydrogenation experiments
We next evaluated the catalytic activity of complex 6 and com-
pared the results with the activity reported for 13 and 14.
As a benchmark reaction, the catalytic hydrogenation of ace-
tophenone to 1-phenylethanol with H2 in the presence of
Structural parameters of the catalysts
We first analyzed the structural and spectroscopic differences
between 6 (containing an aryl-derived six-membered chelate
ring), and the analogous complexes 13 (containing an aliphatic
six-membered chelate ring) and 14 (containing a benzylic
seven-membered chelate ring) synthesized by Morris et al.[24,25]
Several important properties related to the chelate ring size of
these complexes are summarized in Table 1.
Scheme 2. Benchmark catalytic hydrogenation of acetophenone with H2.
KOtBu in THF was investigated (Scheme 2, Table 2). The TOFs
increase with decreasing chelate ring size: 14<13<6 (Table 2,
entries 6, 4, and 1, respectively). Additionally, increasing the
substrate loading linearly increases the TOF (Table 2, entries 1–
3, 548 vs. 1202 vs. 3609 hÀ1). This activity is unprecedented for
this type of half-sandwich NHC–amine complexes in this reac-
tion. A linear dependence of the TOF with respect to substrate
concentration was also observed for complex 13 (Table 2, en-
tries 4 and 5, from 298 to 595 hÀ1). Curiously, the op-
posite has been reported for complex 14: an increase
in substrate loading results in a decrease of the
TOF.[26] To further investigate the effect of the chelate
ring size, we turned our attention to the iridium and
rhodium complexes 7 and 8 and the seven-mem-
bered-ring chelate iridium complex 15.
The influence of the nBu group of 6 relative to the Me sub-
stituent in 13 and 14 is evaluated further on (vide infra).
The CÀRuÀN bite angle in 6 is 79.62(6)8, which is significant-
ly smaller than that in 13 and 14 (87.0 and 91.988, respective-
ly).[24,25] In the geometry of these complexes, the optimal angle
of the chelate ring would be 908, implying that complex 6 is
somewhat strained. That causes a nonoptimal orbital overlap
Table 1. Selected structural and spectroscopic parameters.[a]
Although the iridium systems are less active than
the ruthenium catalyst, indeed a similar relation be-
Complex Bond length [ꢁ]
Bond angle [8]
13C NMR shift of
tween the TOF and the chelate ring effect was
observed for the IrCp* complexes 7 and 15 (Table 3,
entries 8 and 9). For the even less active rhodium
complex 8 (Table 3, entries 10 and 11) we have no
comparison for the larger seven-membered-ring
compound. Besides Cross et al.,[23] another related
rhodium example containing a secondary amino-
RuÀCcarbene RuÀN
RuÀCl
carbeneÀRuÀN carbene[b] [ppm]
6
13
14
2.0394(15) 2.1399(14) 2.3973(4)
79.62(6)
2.4140(15) 87.0(2)
2.4180(13) 91.98(17)
177.2
174
175.1
2.041(6)
2.092(5)
2.114(4)
2.146(4)
[a] Structural parameters of 6 obtained from the X-ray crystal structure. Structural pa-
rameters of 13 and 14 taken from Refs. [25] and [24]. [b] Carbene shift in CD2Cl2.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1740

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Table 2. H2 hydrogenation of acetophenone with Ru complexes of varia-
Table 4. H2-hydrogenation of acetophenone with complexes with varying
ble chelating ring size.
alkyl substituents.
Entry[a]
Complex
C/B/S[b]
Reaction
time [h]
Conversion[c]
[%]
TOF[d]
[hÀ1
Entry[a]
Complex
C/B/S[b]
Conversion[c]
[%]
TOF[d]
[hÀ1
]
]
1
2
3
4
5
6
7
6
6
6
13
13
14
14
1/8/100
1/8/200
1/8/600
1/8/200
1/8/600
1/8/200
1/8/600
0.5
0.5
0.5
1
2
2
99
99
99
99
99
99
57
548
1202
3609
298
595
213
12
13
14
9 (R=Me)
10 (R=Et)
11 (R=nBu)
1/8/200
1/8/200
1/8/200
99
99
99
909
915
1543
[a] Reactions (Scheme 2) were carried out at 25 bar H2 and at 508C, in
THF (15 mL) using KOtBu as base. [b] C/B/S=catalyst/base/substrate
ratio. [c] Conversions were determined by GC analysis using p-xylene as
an internal standard after 0.5 h reaction. [d] Determined from the steep-
est part of the plot (between 40–60% conversion) and as an average of
two runs.
2.5
122
[a] Reactions (Scheme 2) were carried out at 25 bar H2 and at 508C in THF
(15 mL) using KOtBu as base. [b] C/B/S=catalyst/base/substrate ratio.
[c] Conversions were determined by GC analysis using p-xylene as an in-
ternal standard. [d] Determined from the steepest part of the plot (be-
tween 40–60% conversion). Details for complexes 13[25] and 14[26] were
taken from the literature.
which the amine remains coordinated and the substrate does
not have to coordinate to the metal.[28]
We further evaluated the influence of the N-alkyl substituent
on the ligand by testing complexes 9–11 in the H2 hydrogena-
tion of acetophenone (Scheme 2, Table 4). Surprisingly, a sub-
stantial increase of the TOF was observed on going from Me/
Et complexes (9/10) to the nBu complex 11 (Table 4).
Table 3. H2 hydrogenation of acetophenone with iridium and rhodium
complexes of variable chelating ring size.
Entry[a] Complex
C/B/S[b] Reaction Conversion[c] TOF[d]
time [h] [%]
[hÀ1
]
This effect is primarily caused by differences in catalyst acti-
8
9
7
1/8/200
1/8/200
4
95
190
3
98
154
10
11[f]
8
8
1/8/200
1/8/200
0
3
0
98
n.d.[e]
n.d.[e]
[a] Reactions (Scheme 2) were carried out at 25 bar H2 and at 508C in THF
(15 mL) using KOtBu as base. [b] C/B/S=catalyst/base/substrate ratio.
[c] Conversions were determined by GC analysis using p-xylene as an in-
ternal standard. [d] Determined from the steepest part of the plot (be-
tween 40–60% conversion). [e] Not determined. [f] PH =50 bar. Details for
2
complex 15 were taken from the literature.[28]
tethered NHC, [Rh(cod)(L)] [L=1-Mes-3-(2-(Mes-NH)ethyl)imida-
zolium, Mes=2,4,6-trimethylphenyl], was reported by Fryzuk
et al.[31] This complex did not show any activity in hydrogena-
tion of polar bonds at all and only alkene hydrogenation was
reported. Complex 8 thus seems to be the first Rh/NHC–amine
complex reported to be active in the direct hydrogenation of
ketones.
Figure 6. Performance in time of Ru(p-cym)Cl(L) complexes in the H2-hydro-
genation of acetophenone with varying alkyl substituents on the ligand:
R=Me (red triangles), Et (green circles), nBu (blue squares).
vation rates. Inspection of the reaction profile reveals that the
induction period for all complexes is different (Figure 6). At t=
0 the reaction mixture containing deprotonated complex and
substrate is pressurized with H2, forming the active species
(Scheme 3). Going from nBu to Me the induction period be-
comes much longer, clearly pointing to differences in the rates
of formation of the active species in each case. The larger bulk
of the nBu group with respect to Me and Et substituents possi-
bly influences the spatial positioning of the ligand in such
a way that the amine is tilted into a position closer to the
metal center, rendering the complex better suited for activa-
tion of the incoming dihydrogen reagent and thus leading to
more efficient catalyst activation (Scheme 3).
The increased TOF for complexes containing ligand 1 can be
explained by two contributing effects. (1) The smaller chelating
ring size can account for a less tightly bound amine because
the system is more strained. The aniline is also less basic than
the benzyl amine and both factors could cause the amine to
de-coordinate more readily, thereby promoting coordination of
the substrate. (2) Additionally, the conjugated system renders
the NHC more electron rich, thus making it a stronger donor.
The first effect should be much less pronounced for the iridi-
um systems because these Cp*-containing complexes most
likely follow an (alcohol-assisted) outer-sphere mechanism, in
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1741

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
a color change from brown to
green was observed, showing
that the conditions become
acidic enough to protonate the
amino functionality to form a dif-
ferent active species.[32] Increas-
ing both temperature and pres-
sure does improve conversion
(Table 5, entry 1b), but increas-
ing substrate loading (Table 5,
Scheme 3. Catalyst activation by amine deprotonation followed by (bifunctional) dihydrogen activation involving
heterolytic splitting of H2 over the M–amine bond leading to the proposed activated species.
Alternatively, there is a possibility that the ketone substrate
coordinates to the deprotonated complex before formation of
the active hydride species. The bulk of the nBu group might
help to push away the ketone substrate to allow the formation
of the active hydride species by H2.
The RuCp*Cl(h6L) species (12) was also tested in a catalysis
run. In absence of a coordinating electron-donating NHC the
catalytic activity of this Ru sandwich species is low, converting
84% of acetophenone in 17 h at 508C and 25 bar H2.
entries 1c and b) does not. Applying methyl levulinate in hy-
drogenation also showed selective conversion of the ketone
functionality (Table 5, entry 2). Again, higher temperatures and
pressures than those used in the benchmark reaction with ace-
tophenone were required, resulting in 86% conversion over-
night. The lower activity might be attributed to competition
between coordination of the ketone and the ester functionali-
ty. Linear ester functionalities, such as those in MB, are not
converted by 6 (Table 5, entry 3). However, the conjungated
diester DMO, in which one ester functionality activates the
other one for hydrogenation, is converted slowly (29% conver-
sion in 24 h) to methyl 2-hydroxyacetate (Table 5, entry 4). The
highly functionalized aldehyde HMF is converted to the bishy-
droxy compound furan-2,5-diyldimethanol in 2 h at 708C and
50 bar H2 (Table 5, entry 5). Application of the a,b-unsaturated
compound cinnamaldehyde is a good measure for the selectiv-
ity of the catalyst. After 6 h at an elevated temperature and
pressure, a roughly 50:50 prod-
Conversion of platform chemicals and model compounds
thereof
To further investigate the performance of 6, several relevant
types of biomass-inspired substrates have been subjected to
hydrogenation. More challenging and with a higher degree of
functionalization than acetophenone, these substrates are
uct distribution of the hydrogen-
Table 5. Biomass-inspired substrates tested in H2 hydrogenation using 6 as catalyst.
ated ketone and the completely
hydrogenated product was de-
tected. The latter is probably
formed by hydrogenation of the
hydrocinnamaldehyde intermedi-
ate, which is only present in
trace amounts throughout the
entire reaction (Table 5, entry 6).
Entry[a]
Substrate
Conditions
T [8C] PH [bar]
Product
Reaction
time [h]
Conversion[b]
[%]
2
1a
1b
1c[c]
1d[d]
50
70
70
70
25
50
50
50
17
17
17
17
15
85
15
35
2
70
80
50
80
17
17
86
0
3[e]
Conclusions
We have shown that half-sand-
wich complexes with an amino-
tethered NHC, most notably
ruthenium complexes, show
good activity in hydrogenation
of polar functionalities. Aceto-
phenone is hydrogenated by
4
5
80
70
80
50
24
2
29
99
6
24
47/2/38
8/0/90
6
70
50
[a] Reactions were carried out at 25 bar H2 and at 508C in THF (15 mL) using KOtBu as base and complex 6 as
catalyst; C/B/S=1/8/200. [b] Conversions were determined by GC analysis using p-xylene as an internal stan-
dard. [c] C/B/S=1/8/2000 in THF (2 mL). [d] C/B/S=1/8/1000. [e] C/B/S/=1/8/600.
[Ru(p-cym)Cl(L)]A
(6,
9–11)
which most probably proceeds
through an inner-sphere mecha-
nism, showing excellent activity
a good measure of the capacities of this type of NHC–amine-
containing catalysts for application in biomass (Table 5). Using
LA as substrate only minor conversion to the ring-closed prod-
uct g-valerolactone was observed (Table 5, entry 1a) probably
due to incompatibility with the amine functionality of the cata-
lyst. Upon mixing the deprotonated complex with this acid,
with TOFs as high as 3600 hÀ1 for complex 6. We are also the
first to report a Rh/NHC–amine species of this type to be
active in the direct hydrogenation of ketones. The activity of
this type of catalyst is greatly influenced by the type of chelat-
ing ligand and the N-substituent on the NHC. A conjugating,
small chelate ring is beneficial for the activity. Also, a larger
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1742

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
C23H32ClN3Ir: m/z calcd 578.1907 (100%) [M–PF6À]+, found
578.1907.
alkyl substituent on the NHC promotes the formation of the
active species and thereby greatly reduces the induction
period. The applicability of complex 6 is successfully extended
to key biomass-derived substrates. Functionalized aldehydes,
ketones, and activated diesters have been converted readily at
moderate pressures and temperatures (50 bar H2, 708C).
Our study has provided more thorough knowledge concern-
ing the workings of NHC–amine complexes of Ru, Ir, and Rh in
direct hydrogenation of polar functionalities, which will aid the
future rational design of hydrogenation catalysts for the con-
version of biomass into useful/added-value compounds.
[RhCp*Cl(nBuÀCNHCÀNH2)] PF6 (8): [RhCp*Cl2]2 (0.095 mmol, 59 mg),
KPF6 (175 mg, 0.95 mmol), and 1a (77 mg, 0.19 mmol) were weigh-
ed in a Schlenk tube and dissolved in CH2Cl2 (10 mL). The solution
was stirred at RT for 24 h under the exclusion of light. The mixture
was then filtered over a pad of Celite, and the solvent was evapo-
rated in vacuo (to a volume of 1 mL) and precipitated with Et2O.
The complex was purified by dissolution in a THF/pentane mixture
(8 mL; 1:3 v/v) and cooling to 48C to obtain 8 in a crystalline form
1
in 62% yield. H NMR (300 MHz, CD2Cl2): d=7.73 (d, J=2.2 Hz, 1H,
=CH), 7.57–7.42 (m, 4H, HAr), 7.45 (d, J=2.4 Hz, 1H, =CH), 5.19 (bs,
2H, NH2), 4.26 (m, 1H, NCH2), 4.02 (m, 1H, NCH2), 1.87 (m, 2H, CH2),
1.42 (s, 17H, Cp*ÀCH3Àand CH2), 1.02 ppm (t, J=7.4 Hz, 2H, CH3);
13C NMR (75 MHz, CD2Cl2): d=173.02 (d, J=53 Hz, N2CÀRh), 133.12,
132.11, 128.46, 127.57, 123.63, 122.32, 121.84, 120.98, 97.62 (d, J=
98, h5-C5Me5), 51.04, 32.79, 20.03, 13.57, 8.34 ppm; 19F NMR
(282 MHz, CD2Cl2): d=À72.19 ppm (d, J=711.9 Hz); FAB+-MS:
C23H32ClN3Rh: m/z calcd 488.1340 (100%) [M–PF6À]+, found
488.1335; elemental analysis calcd (%) for C23H32ClF6N3PRh: C 43.58,
H 5.0, N 6.63, found: C 42.15, H 4.97, N, 6.49.
Experimental Section
General remarks, detailed experimental procedures (to obtain the
reported ligands 1–5, the silver carbene 1a in Scheme 1, and com-
plex 12), and crystallographic data of complexes 6–8 can be found
in the Supporting Information.
[Ru(p-cym)Cl(RÀCNHCÀNH2)][I] (9–11) (based on a modified proce-
dure from Ref. [23]): [Ru(p-cym)Cl2]2 (97 mg, 0.16 mmol), the appro-
priate imidazolium salt (2–4) (0.32 mmol), and Ag2O (37 mg,
0.16 mmol) were weighed in a Schlenk tube, dissolved in CH2Cl2
(15 mL), and stirred under absence of light at 338C overnight. The
solvent was then removed in vacuo; KI (0.53 gr, 3.17 mmol) was
added to the residue, and the mixture was redissolved in acetone
(20 mL) and stirred at reflux for 1 h. The solvent was removed in
vacuo, the crude product redissolved in CH2Cl2 (20 mL), filtered
over Celite pad, and concentrated under vacuum. The complex
was purified by repeatedly dissolving in a CH2Cl2/Et2O mixture and
cooling to 48C to obtain the pure complex.
Complex synthesis
[Ru(p-cym)Cl(nBuÀCNHCÀNH2)]PF6 (6):
A
solution of 1a (0.27 g,
solution of
0.66 mmol) in CH2Cl2 (15 mL) was added to
a
[RuCl2(p-cym)]2 (0.19 g, 0.30 mmol) and KPF6 (0.36 g, 1.98 mmol) in
CH2Cl2 (20 mL), and the mixture was stirred at RT for 3 h, during
which the solution became dark green with grey silver halides sus-
pended in it. The mixture was filtered over a pad of Celite and the
volume of filtrate was reduced to 3 mL. Pentane (20 mL) was
added, which resulted in a precipitation of a dark green solid.
Repeated precipitation from CH2Cl2 with Et2O furnished the desired
1
light green complex (0.19 g, 96%). H NMR (300 MHz, CD2Cl2): d=
7.74 (d, J=6.7 Hz, 1H, =CH), 7.65 (d, J=2.2 Hz, 1H, =CH), 7.42 (m,
4H, HAr), 6.52 (d, J=8.9 Hz, 1H, NH2), 5.83 (d, J=5.9 Hz, 1H, HAr p-
cym), 5.71 (d, J=6.1 Hz, 1H, HAr p-cym), 5.09 (m, 2H, HAr p-cym),
4.38 (d, J=10.6 Hz, 1H, NH2), 4.29 (t, J=8.1 Hz, 2H, NCH2), 2.05–
1.80 (m, 3H, CH2 and CH p-cym), 1.74 (s, 3H, CH3 p-cym), 1.52 (m,
2H, CH2), 1.05 (t, J=7.3 Hz, 3H, CH3), 0.93 ppm (dd, J=16.6, 6.9 Hz,
6H, CH3 iPr); 13C NMR (75 MHz, CD2Cl2): d=177.22, 134.26, 132.06,
128.80, 127.68, 123.71, 121.75, 120.14, 109.37, 101.81, 89.06, 83.89,
83.73, 81.96, 51.52, 33.01, 30.78, 23.33, 20.19, 19.85, 17.83, 13.62,
0.74 ppm; 31P NMR (121 MHz, CD2Cl2): d=À144.06 ppm (septet, J=
711.9 Hz); 19F NMR (282 MHz, CD2Cl2): d=À70.14 ppm (d, J=
711.1 Hz); FAB+–MS (CH2Cl2) (FAB: fast atom bombardment mass
spectrometry): calcd for C23H31ClN3Ru+PF6À: m/z calcd 486.1253
(100%) [M–PF6À]+, found 486.1248.
Spectral data for 9 (R=Me): Complex 9 was obtained as a dark
brown crystalline solid in 33% yield. H NMR (300 MHz, CD2Cl2): d=
1
10.08 (d, J=9.9 Hz, 1H, NH2), 8.60 (d, J=6.8 Hz, 1H, HAr), 7.62 (s,
1H, =CH), 7.36 (s, 1H, HAr), 7.42–7.18 (m, 3H, HAr), 6.24 (d, J=
5.4 Hz, 1H, HAr p-cym), 5.85 (d, J=5.7 Hz, 1H, HAr p-cym), 5.32 (d,
J=5.4 Hz, 1H, HAr p-cym), 5.19 (d, J=5.7 Hz, 1H, HAr p-cym), 4.06 (s,
1H, NH2), 4.01 (s, 3H, CNH3), 1.90 (m, 1H, CH p-cym), 1.77 (s, 3H,
CH3 p-cym), 0.86 ppm (dd, J=19.8, 6.8 Hz, 6H, CH3 iPr); 13C NMR
(75 MHz, CD2Cl2): d=179.09, 155.58, 155.44, 135.61, 132.53, 128.26,
126.85, 125.78, 123.72, 120.98, 119.41, 107.37, 102.08, 89.87, 83.89,
83.11, 30.62, 23.35, 20.15, 18.00 ppm.
Spectral data for 10 (R=Et). Complex 10 was obtained as a brown
microcrystalline powder in 29% yield. 1H NMR (300 MHz, CD2Cl2):
d=8.81–8.63 (m, 1H, HAr), 8.56 (d, J=11.2 Hz, 1H, NH2), 7.77 (d, J=
2.2 Hz, 1H, =CH), 7.46 (d, J=2.1 Hz, 1H, =CH), 7.37 (dt, J=6.8,
4.0 Hz, 3H, HAr), 6.31 (d, J=6.0 Hz, 1H, HAr p-cym), 5.79 (d, J=
6.0 Hz, 1H, HAr p-cym), 5.22 (dd, J=10.5, 6.1 Hz, 2H, HAr p-cym),
4.33 (ddq, J=35.5, 14.6, 7.4 Hz, 2H, NCH2), 4.17 (d, J=11.3 Hz, 1H,
NH2), 2.14 (m, 1H, CH p-cym) 2.09 (s, 3H, CH3 p-cym), 1.58 (t, J=
7.3 Hz, 3H, CH3), 0.92 ppm (dd, J=6.9, 4.1 Hz, 6H, CH3 iPr);
13C NMR (75 MHz, CD2Cl2): d=176.62, 135.80, 132.57, 128.22,
127.28, 123.13, 122.39, 121.57, 120.53, 110.48, 102.22, 89.51, 84.02,
82.76, 82.22, 48.61, 31.50, 23.67, 20.44, 19.75, 15.69 ppm.
[IrCp*Cl(nBuÀCNHCÀNH2)] PF6 (7): [IrCp*Cl]2 (0.25 mmol, 199 mg) and
KPF6 (1.25 mmol, 230 mg) in 8 mL CH2Cl2 was added to a solution
of 1a (201 mg, 0.25 mmol) in CH2Cl2 (7 mL) and stirred at RT under
the exclusion of light for 2 h. The resulting reaction mixture was fil-
tered over a pad of Celite, and the solvent was evaporated in va-
cuo. The complex was purified by dissolution in a CH2Cl2/Et2O mix-
ture (7 mL; 2:5 v/v) and cooling to 48C to obtain 7 in a crystalline
form in 51% yield. 1H NMR (300 MHz, CD2Cl2): d=7.70 (d, J=
2.2 Hz, 1H, =CH), 7.54 (m, 2H, HAr), 7.49–7.40 (m, 2H, HAr), 7.38 (d,
J=2.2 Hz, 1H, =CH), 6.12 (bs, 2H, NH2), 4.26 (m, 1H, NCH2), 3.97
(m, 1H, NCH2), 1.88 (m, 2H, CH2), 1.46 (s, 17H, Cp*ÀCH3 and CH2),
1.01 ppm (t, J=7.4 Hz, 2H, CH3); 13C NMR (75 MHz, CD2Cl2): d=
160.97, 133.81, 132.54, 128.12, 127.89, 122.87, 122.82, 121.84,
121.66, 119.71, 90.13, 50.81, 33.21, 20.03, 13.59, 8.13 ppm; 19F NMR
(282 MHz, CD2Cl2): d=À72.06 ppm (d, J=711.7 Hz); FAB+–MS:
Spectral data for 11 (R=nBu). Complex 11 was obtained in a dark
brown microcrystalline solid in 46% yield. 1H NMR (300 MHz,
CD2Cl2): d=10.14 (bs, 1H, NH2), 8.57 (s, 1H, =CH), 7.72 (s, 1H, =CH),
7.36 (m, 4H, HAr), 6.16 (d, J=5.5 Hz, 1H, HAr p-cym), 5.81 (d, J=
5.9 Hz, 1H, HAr p-cym), 5.28 (d, J=5.5 Hz, 1H, HAr p-cym), 5.18 (d,
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1743

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[6] F. M. A. Geilen, B. Engendahl, M. Hçlscher, J. Klankermayer, W. Leitner, J.
Am. Chem. Soc. 2011, 133, 14349–14358.
[7] S. D. Lacroix, A. Pennycook, S. Liu, T. T. Eisenhart, A. C. Marr, Catal. Sci.
Technol. 2012, 2, 288.
[8] D. Gnanamgari, E. L. O. Sauer, N. D. Schley, C. Butler, C. D. Incarvito, R. H.
Crabtree, Organometallics 2009, 28, 321–325.
[9] F. Hanasaka, K. Fujita, R. Yamaguchi, Organometallics 2006, 25, 4643–
4647.
J=5.8 Hz, 1H, HAr p-cym), 4.31 (t, J=7.1 Hz, 2H, NCH2), 4.01 (bs, J=
9.1 Hz, 1H, NH2), 1.91 (m, 3H, CH2 +CH p-cym), 1.50 (m, J=7.4 Hz,
2H, CH2), 1.03 (t, J=7.3 Hz, 3H, CH3), 0.87 ppm (dd, J=25.8, 6.9 Hz,
6H, CH3 iPr); 13C NMR (75 MHz, CD2Cl2): d=178.21, 135.78, 132.57,
128.10, 126.90, 123.58, 123.14, 121.36, 120.18, 107.75, 101.25,
89.54, 84.20, 83.89, 51.48, 33.10, 30.58, 23.37, 20.23, 17.90,
13.70 ppm.
[10] A. Azua, J. A. Mata, E. Peris, F. Lamaty, J. Martinez, E. Colacino, Organo-
metallics 2012, 31, 3911–3919.
[11] The Handbook of Homogeneous Hydrogenation (Eds.: J. G. de Vries, C. J.
Elsevier), Wiley-VCH, Weinheim, 2006.
Hydrogenation experiments
Catalyst (0.014 mmol) and KOtBu (0.12 mmol) were weighed in
a Schlenk flask, dissolved in THF (15 mL), and let to stir for 30 min
to ensure formation of the deprotonated complex. The appropriate
amount of substrate and p-xylene (internal standard, 1.4 mmol)
were then added. A homebuilt stainless-steel autoclave with
a volume of 200 mL was flushed three times with N2, and the mix-
ture was inserted into the autoclave under a flow of N2. The auto-
clave was then heated until the desired temperature (50–808C)
was reached, flushed three times with H2, and filled with the ap-
propriate pressure of H2 (25–80 bar H2). If the reaction was moni-
tored with regard to time, samples were taken through a sample
port (first aliquot was discarded). After the reaction, the autoclave
was cooled to RT with an ice bath and H2 was vented carefully. The
products were determined by GC analysis with p-xylene as internal
standard, using a Thermo Scientific Trace GC Ultra system with
a Restek RTX-200 (30 m, 0.25 mm inner diameter) capillary column
with a split injecting method.
[12] S. Hashiguchi, J. Takehara, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 1995,
117, 7562–7563.
[13] A. Hadzovic, D. Song, C. M. MacLaughlin, R. H. Morris, Organometallics
2007, 26, 5987–5999.
[14] R. Noyori, C. A. Sandoval, K. MuÇiz, T. Ohkuma, Philos. Trans. R. Soc. A
2005, 363, 901–912; discussion 1035–1040.
[15] M. Yamakawa, H. Ito, R. Noyori, J. Am. Chem. Soc. 2000, 122, 1466–
1478.
[16] W. W. N. O, R. H. Morris, ACS Catal. 2013, 3, 32–40.
[17] M. L. Clarke, Catal. Sci. Technol. 2012, 2, 2418–2423.
[18] D. Milstein, Top. Catal. 2010, 53, 915–923.
[19] S. Dꢂez-Gonzꢃlez, N. Marion, S. P. Nolan, Chem. Rev. 2009, 109, 3612–
3676.
[20] M. Watanabe, Y. Kashiwame, S. Kuwata, T. Ikariya, Eur. J. Inorg. Chem.
2012, 504–511.
[21] T. Ikariya, I. D. Gridnev, Chem. Rec. 2009, 9, 106–123.
[22] W. B. Cross, C. G. Daly, R. L. Ackerman, I. R. George, K. Singh, Dalton
Trans. 2011, 40, 495–505.
[23] W. B. Cross, C. G. Daly, Y. Boutadla, K. Singh, Dalton Trans. 2011, 40,
9722–9730.
[24] W. W. N. O, A. J. Lough, R. H. Morris, Organometallics 2009, 28, 6755–
6761.
Acknowledgements
[25] H. Ohara, W. W. N. O, A. J. Lough, R. H. Morris, Dalton Trans. 2012, 41,
8797–8808.
[26] W. W. N. O, A. J. Lough, R. H. Morris, Organometallics 2011, 30, 1236–
1252.
[27] W. W. N. O, A. J. Lough, R. H. Morris, Organometallics 2012, 31, 2137–
2151.
[28] W. W. N. O, A. J. Lough, R. H. Morris, Organometallics 2012, 31, 2152–
2165.
[29] R. M. Fairchild, K. T. Holman, Organometallics 2007, 26, 3049–3053.
[30] W. W. N. O, A. J. Lough, R. H. Morris, Chem. Commun. 2010, 46, 8240–
8242.
[31] H. Jong, B. O. Patrick, M. D. Fryzuk, Can. J. Chem. 2008, 86, 803–810.
[32] C. A. Sandoval, T. Ohkuma, N. Utsumi, K. Tsutsumi, K. Murata, R. Noyori,
Chem. Asian J. 2006, 1, 102–110.
[33] A. J. Blake, B. a. J. Clark, H. McNab, C. C. Sommerville, J. Chem. Soc.
Perkin Trans. 1 1997, 1605–1608.
[34] P. J. Fagan, M. D. Ward, J. C. Calabrese, J. Am. Chem. Soc. 1989, 111,
1698–1719.
This research has been performed within the framework of the
CatchBio program. The authors gratefully acknowledge the sup-
port of the Smart Mix Program of the Netherlands Ministry of
Economic Affairs and the Netherlands Ministry of Education, Cul-
ture, and Science. The X-ray diffractometer has been financed by
the Netherlands Organization for Scientific Research (NWO).
Keywords: bifunctional ligand · biomass · hydrogenation ·
rhodium · ruthenium
[1] A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney,
C. A. Eckert, W. J. Frederick, Jr., J. P. Hallett, D. J. Leak, C. L. Liotta, J. R.
Mielenz, R. Murphy, R. Templer, T. Tschaplinski, Science 2006, 311, 484–
489.
[2] G. W. Huber, A. Corma, Angew. Chem. 2007, 119, 7320–7338; Angew.
Chem. Int. Ed. 2007, 46, 7184–7201.
[3] J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. 2007, 119,
7298–7318; Angew. Chem. Int. Ed. 2007, 46, 7164–7183.
[4] T. Werpy, G. Petersen, J. E. Holladay, A. Aden, J. J. Bozell, J. F. White, D.
Johnson, A. Manheim, Top Value Added Chemicals from Biomass, Vol. I,
http://www1.eere.energy.gov/bioenergy/pdfs/35523.pdf, 2004.
[5] F. M. A. Geilen, B. Engendahl, A. Harwardt, W. Marquardt, J. Klankermay-
er, W. Leitner, Angew. Chem. 2010, 122, 5642–5646; Angew. Chem. Int.
Ed. 2010, 49, 5510–5514.
[35] CCDC 931590 (compound 6), CCDC 931591 (compound 7), and
CCDC 931592 (compound 8) contain the supplementary crystallograph-
ic 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.
Received: April 19, 2013
Revised: June 21, 2013
Part of a Special Issue on ”CatchBio“. To view the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cssc.v6.9/issuetoc
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 1737 – 1744 1744