Hydrogenation of carbon dioxide to methanol using a homogeneous ruthenium-Triphos catalyst: From mechanistic investigations to multiphase catalysis

Article abstract of DOI:10.1039/c4sc02087a

The hydrogenation of CO₂ to methanol can be achieved using a single molecular organometallic catalyst. Whereas homogeneous catalysts were previously believed to allow the hydrogenation only via formate esters as stable intermediates, the present mechanistic study demonstrates that the multistep transformation can occur directly on the Ru-Triphos (Triphos = 1,1,1-tris(diphenylphosphinomethyl)ethane) centre. The cationic formate complex [(Triphos)Ru(η²-O₂CH)(S)]⁺ (S = solvent) was identified as the key intermediate, leading to the synthesis of the analogous acetate complex as a robust and stable precursor for the catalytic transformation. A detailed mechanistic study using DFT calculations shows that a sequential series of hydride transfer and protonolysis steps can account for the transformation of CO₂via formate/formic acid to hydroxymethanolate/formaldehyde and finally methanolate/methanol within the coordination sphere of a single Ru-Triphos-fragment. All experimental results of the systematic parameter optimisation are fully consistent with this mechanistic picture. Based on these findings, a biphasic system consisting of H₂O and 2-MTHF was developed, in which the active cationic Ru-complex resides in the organic phase for recycling and methanol is extracted with the aqueous phase. This journal is

Full text of DOI:10.1039/c4sc02087a

View Article Online
View Journal
Chemical
Science
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. Wesselbaum, V.
Moha, M. Meuresch, S. Brosinski, K. M. Thenert, J. Kothe, T. vom Stein, U. Englert, M. Hoelscher, J.
Klankermayer and W. Leitner, Chem. Sci., 2014, DOI: 10.1039/C4SC02087A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/chemicalscience

View Article Online
DOI: 10.1039/C4SC02087A
Page 1 of 10
Chemical Science
Dynamic Article Links ►
Journal Name
Cite this: DOI: 10.1039/c0xx00000x
www.rsc.org/xxxxxx
ARTICLE TYPE
Hydrogenation of Carbon Dioxide to Methanol using a Homogeneous
Ruthenium-Triphos Catalyst: From Mechanistic Investigations to
†
Multiphase Catalysis
a
a
a
a
a
Sebastian Wesselbaum, Verena Moha, Markus Meuresch, Sandra Brosinski, Katharina M. Thenert,
a
a
b
a
a
5
Jens Kothe, Thorsten vom Stein, Ulli Englert, Markus Hölscher, Jürgen Klankermayer, * and Walter
a
Leitner *
Received (in XXX, XXX) Xth XXXXXXXXX 20XX, Accepted Xth XXXXXXXXX 20XX
DOI: 10.1039/b000000x
The hydrogenation of CO to methanol can be achieved by a single molecular organometallic catalyst.
2
1
1
2
0
5
0
Whereas homogeneous catalysts were previously believed to allow the hydrogenation only via formate
esters as stable intermediates, the present mechanistic study demonstrates that the multistep
transformation
can
occur
directly
on
the
Ru-Triphos
(Triphos
=
1,1,1-
2
+
tris(diphenylphosphinomethyl)ethane) centre. The cationic formate complex [(Triphos)Ru( -O CH)(S)]
2
(S = solvent) was identified as key intermediate, leading to the synthesis of the analogous acetate complex
as robust and stable precursor for the catalytic transformation. A detailed mechanistic study using DFT
calculation shows that a sequential series of hydride transfer and protonolysis steps can account for the
transformation of CO2 via formate/formic acid to hydroxymethanolate/formaldehyde and finally
methanolate/methanol within the coordination sphere of a single Ru-Triphos-fragment. All experimental
results of systematic parameter optimisation are fully consistent with this mechanistic picture. Based on
these findings, a biphasic system consisting of H O und 2-MTHF was developed, where the active
2
cationic Ru-complex resides in the organic phase for recycling and methanol is extracted with the
aqueous phase.
Introduction
A certain percentage of CO is added to the feedstock stream to
2
balance the C/H ratio. The heterogeneously catalysed
The depletion of fossil carbon sources together with the
increasing global energy consumption demand alternative ways
for the sustainable production of fuels and chemicals. In this
context, the usage of carbon dioxide (CO ) as alternative carbon
source has seen renewed and increasing interest at the interface of
the chemical and energy sector, as it is a readily available, non-
5
5
6
6
7
0
5
0
5
0
hydrogenation of pure CO to methanol has been implemented
2
2
3
3
5
0
5
capitalising on the specific regional energy and feedstock basis in
17
Iceland, for example. A detailed picture of the elementary steps
2
and the role of the multi-component catalyst material have been
elucidated for the classical Cu/ZnO/Al O systems, mapping out
2
3
the complex series of bond cleavage and bond forming processes
1
-11
toxic by-product of various large scale industries.
Particularly
on the catalyst surface that enable the seemingly simple overall
the effective hydrogenation of carbon dioxide to methanol could
play an important role in supply chains with reduced carbon
footprint economy, as methanol can serve as energy carrier and
versatile basic chemical.
megaton scale from fossil feedstock-based syngas (CO/H2).
18
transformation of CO or CO and hydrogen to methanol.
2
In sharp contrast, the hydrogenation of CO to methanol
2
using a molecularly defined, single-site catalyst has remained
elusive up to now. Tominaga et al. reported the formation of
methanol besides methane and CO from CO2 hydrogenation
using Ru (CO) in the presence of alkaline iodides under harsh
1
2-15
Today methanol is produced on
1
5, 16
These processes utilise heterogeneous catalysts at elevated
temperatures (200‒300 °C) and pressures (50‒100 bar).
3
12
reaction conditions (240 °C, 80 bar). Under these conditions, CO2
was reduced to CO followed by the hydrogenation of the CO to
1
9
methanol and methane. Later, the catalytic formation of
4
4
0
5
methanol from CO was reported with organometallic complexes
a
2
Institut für Technische und Makromolekulare Chemie, RWTH Aachen
2
0
using high energy reduction reagents such as boranes. With
hydrogen, indirect routes via conversion of CO -derived
University, Worringerweg 1, 52074 Aachen, Germany. E-mail:
jklankermayer@itmc.rwth-aachen.de; leitner@itmc.rwth-aachen.de
2
b
Institut für Anorganische Chemie, RWTH Aachen University,
intermediates like organic carbonates, carbamates, formate esters
and ureas were proposed (see Scheme 1, upper pathway, for
formate esters). The viability of this concept was first
Landoltweg 1, 52074 Aachen, Germany.
†
Electronic Supplementary Information (ESI) available: [Experimental
procedures, analytical data]. See DOI: 10.1039/b000000x/
This journal is © The Royal Society of Chemistry [year]
[journal], [year], [vol], 00–00 | 1

View Article Online
DOI: 10.1039/C4SC02087A
Chemical Science
Page 2 of 10
the possibility to separate and recycle these catalysts from the
MeOH/water product mixture in a biphasic aqueous system using
2
-methyl tetrahydrofuran (2-MTHF) as catalyst phase.
Results and Discussion
35
Basic reactivity and identification of the active species
Attempting to identify intermediates of the previously
reported catalytic reaction sequence with catalyst 2,
multinuclear NMR experiments were carried out to monitor the
formation of organometallic species upon stepwise addition of the
2
Scheme 1 Hydrogenation of CO to methanol via formate esters as
23
2
4, 30
2
1
proposed by Milstein et al. and previously shown by Sanford/Huff and
24
Klankermayer/Leitner (upper pathway), and hydrogenation of CO
2
to
methanol without the need for an alcohol additive as shown in the present
report (lower pathway).
4
4
5
5
6
6
0
5
0
5
0
5
required components. Thus,
a solution of the precursor
[
(Triphos)Ru(TMM)] (2) and HNTf2 (1 eq.) in d -THF was
8
demonstrated in the seminal work by Milstein et al., who
pressurised with CO (20 bar at r.t.) and H (60 bar at r.t.), stirred
for 1 h at 140 °C, and then transferred to a NMR tube for
analysis. Unexpectedly, a sharp signal at 3.27 ppm in the H-
NMR spectrum indicated the catalytic formation of MeOH
already in the absence of any alcohol additive with a TON of 35.
Thus, one of the species formed under these conditions must be
able to serve as catalyst for the hydrogenation of CO to
methanol. The P{ H}-NMR spectrum of the clear, yellow
solution obtained under these conditions is depicted in Figure 1,
2
2
developed highly efficient Ruthenium(II) pincer complexes for
2
1, 22
the hydrogenation of these challenging substrates.
Huff and
1
Sanford reported a three step, one pot hydrogenation of CO to
2
5
0
5
0
5
0
methanol via methyl formate as intermediate using a combination
2
3
of the Milstein catalyst with two other catalysts.
Most recently, we were able to show that the sequential
reduction via formate ester for the homogeneous hydrogenation
of CO2 to methanol could be achieved in a fully integrated
reaction with a single molecular catalyst based on Ruthenium as
central metal and the tridentate ligand Triphos (Triphos = 1,1,1-
2
3
1
1
1
1
2
2
3
1
13
the corresponding H, C and 2D-correlation spectra are shown
in the supporting information.
2
4
tris(diphenylphosphinomethyl)ethane). The catalyst could be
Formation of the cationic carbonyl complex
formed in situ from Ru(acac) and Triphos 1 or using the readily
3
+
[
(Triphos)RuH(CO) ] (5) was inferred from the characteristic set
2
accessible ruthenium(II)-complex [(Triphos)Ru(TMM)] 2 (TMM
of a doublet (18.6 ppm, J = 28.7 Hz) and triplet (6.3 ppm,
J = 28.7 Hz) in the P{ H}-NMR. Correlation with the hydride
= trimethylenemethane) as precursor, both in the presence of an
3
1
1
31
2
5-29
acid co-catalyst (Scheme 2).
1
31
signal at  = −6.7 ppm in the [ H, P]-HMBC-NMR spectrum and
ESI-MS analysis further confirmed this assignment. The content
in solution was about 4 % according to the integral ratios in the
In the present report we disclose for the first time the
hydrogenation of CO to methanol using a single molecularly-
2
defined homogeneous catalyst without the need for an alcohol
additive (Scheme 1, lower pathway). This fundamental step
forward was derived from comprehensive mechanistic
investigations concerning the catalyst system 2 which led to the
3
1
1
P{ H}-NMR spectrum. The formation of complex
5
corroborates the assumption of cationic complexes as
catalytically active species. The carbonyl ligands are most likely
formed by decarbonylation of intermediates on the pathway to
2
identification of the cationic formate complex [(Triphos)Ru( -
2
8,32,33
+
methanol.
Supporting this hypothesis,
5
could be
O CH)(S)] (S = solvent) 3 as catalytically active intermediate in
2
synthesised in pure form by stirring complex 2 together with 1
equivalent of HNTf2 in ethyl formate and 60 bar H2 in the
absence of CO2 for 24 hours at 140 °C. Testing the isolated
solution (Scheme 2). Based thereon, the analogous cationic
acetate complex 4 was developed as pre-catalyst (Scheme 2).
These molecular catalysts allow the homogeneously catalysed
formation of methanol using CO and H as the sole feedstock
2
2
with turnover frequencies in the same range as reported for the
active sites of the heterogeneous systems. We also demonstrate
31
1
Fig. 1 P{ H}-NMR spectra (top: at r.t., bottom: at −40 °C) of the reaction
solution after CO hydrogenation to methanol (20 bar CO + 60 bar H
140 °C, 1 h) with catalyst 2 (50 mol) and HNTf (1 eq.) in d -THF
(-H) (Triphos) ],
2
2
2
,
Scheme 2 Catalyst precursors 1, 2 and 4 (S = free coordination site or
2
8
+
solvent) for the CO
2
hydrogenation to methanol and structure of the
(2 mL). 5 = [(Triphos)RuH(CO)
2
] , 6 = [Ru
2
2
2
+
2
+
catalytically active intermediate 3 (S = solvent).
7 = [Ru
2
(Cl)
3
(Triphos)
2
] , 3a = [(Triphos)Ru( -O CH)(THF)] .
2
2
| Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

View Article Online
DOI: 10.1039/C4SC02087A
Page 3 of 10
Chemical Science
complex 5 for its catalytic activity in CO hydrogenation in the
2
absence of alcohol (standard conditions: V(THF) = 2.08 mL,
c(Ru) = 12 mmol/L, 1 eq. of HNTf2, p(CO ) = 20 bar at r.t.,
2
p(H ) = 60 bar at r.t., T = 140°C, t = 24 h) gave only a TON of 4,
2
5
0
5
0
5
0
5
0
5
0
5
identifying the formation of 5 as a possible deactivation pathway.
3
1
1
The sharp singlet at 43.3 ppm in the P{ H}-NMR
spectrum was correlated with a broad hydride signal at -8.7 ppm
1
1
31
in the H-NMR spectrum by [ H, P]-HMBC-NMR. Comparison
with literature data and analysis of the mixture by ESI-MS
1
1
2
2
3
3
4
4
5
5
allowed unambiguous assignment to the dimeric complex [Ru (-
2
2
9
H) (Triphos) ] (6), which formed in about 2 %. Using isolated 6
2
2
in the CO2 hydrogenation reaction under standard conditions
showed no formation of methanol, revealing formation of 6 as a
second major deactivation pathway.
Scheme 3 Formation of the catalytically active formate complex 3a from
The small sharp singlet at 36.7 ppm in the
3
1
1
+
catalyst precursor 2 in presence of 1 eq. HNTf
reaction conditions (upper pathway) and by addition of 1 eq. HNTf
eq. HCO H in THF.
2
and H
2
/CO
2
under
P{ H}-NMR spectrum was assigned to [Ru (Cl) (Triphos) ]
2
3
2
2
and 1
(
7) by comparison with literature data and analysis of the mixture
2
by ESI-MS (6 % according to the integral ratios in the
3
1
1
34
P{ H}-NMR spectrum). A CO hydrogenation reaction under
freshly prepared solution of 3a in 0.5 mL THF at room
standard conditions using the [(Triphos)Ru(TMM)] (2) precursor 60 temperature (d, 42.8 ppm, 2P; t, 29.6 ppm, 1P, J = 42.2 Hz; see
2
but with the addition of 1-butyl-3-methylimidazolium chloride
ESI for details). FT-IR analysis of this solution at room
3
1
1
-1
(
3 eq.) gave only a TON of 1 after 24 hours. The P{ H}-NMR
spectrum of the solution showed the formation of 7, 5 and
(Triphos)RuH(CO)Cl] (18) demonstrating that Ru-Triphos
complexes bearing chloro ligands are again inactive in this 65 hours at room temperature at the expense of a new set of doublet
temperature again showed a CO stretching mode at 1544 cm ,
2
+
consistent with the structure [(Triphos)Ru( -O CH)(MeCN)]
2
[
(3b). Interestingly, the signals of 3b decreased over a period of 5
3
5
transformation.
(47.6 ppm, 2P, J = 20.6 Hz) and triplet (5.5 ppm, 1P,
J = 20.6 Hz). In parallel, the formate signal at 8.7 ppm
disappeared with concomitant formation of an upfield hydride
The main species accounting for over 85 % of the total signal
3
1
1
intensity in the P{ H}-NMR gave rise to a broad singlet at
4.2 ppm, indicating fluxional behaviour at room temperature.
1
4
signal (dt, -5.5 ppm, J = 105.0 Hz, J = 19.3 Hz) in the H-NMR.
Low-temperature NMR at 233 K resulted in splitting into a 70 These NMR-data are consistent with the decarboxylation of 3b to
+
doublet (46.3 ppm, 2P, J = 42.5 Hz) and triplet (43.9 ppm, 1P,
give the literature known complex [(Triphos)Ru(H)(MeCN) ]
2
1
31
41
J = 42.5 Hz). A [ H, P]-HMBC-NMR experiment revealed a
(8) (Scheme 4). Consequently, the formation of the formate
coupling of this signals to a proton signal at 8.7 ppm (bs), which
complex 3 from 2 in presence of HNTf under CO and hydrogen
2
2
3
6-38
1
is well in the range of ruthenium coordinated formate.
A [ H-
C]-HMBC-NMR experiment showed the coupling of that 75 into the analogous solvent-coordinated cationic Ru-hydride
pressure is most plausibly explained via reversible CO -insertion
2
1
3
1
3
proton signal to a singlet at 178.8 ppm in the C-NMR, further
complex as intermediate.
2
3, 39, 40
corroborating the formation of a formate-complex.
No
Ruthenium-formate complexes are well known to be
4
2-44
hydride signals corresponding to this species were detected in the
respective correlation NMR spectra.
intermediates in the CO hydrogenation to formic acid.
In
2
order to probe whether the formate complex 3 is a kinetically
The same formate complex could be generated 80 competent intermediate in the hydrogenation of CO to methanol,
2
independently by adding one equivalent of HNTf to complex 2
a solution of 3a was prepared from 2/HNTf (1:1) and HCO H in
2 2
d -THF, pressurised only with 60 bar H , and heated to 140 °C in
8 2
an external oil bath in a high-pressure NMR tube for 40 minutes
2
in d -THF, followed by the addition of one equivalent of HCO H
8
2
at room temperature. NMR analysis of the crude reaction mixture
at room temperature and 233 K showed the identical set of
(Scheme 5). Indeed, this led to nearly quantitative conversion (ca.
3
1
1
1
signals in the P{ H}-NMR in about 80 % of the total intensity 85 97 %) of the coordinated formate to methanol based on H-NMR
3
1
together with a second, yet unidentified phosphor containing
analysis (see ESI). In the corresponding P-NMR spectra the
species (singlet at 59 ppm, ca 20 % of total intensity), as well as
formation of [Ru (-H) (Triphos) ] (6) in about 44 % was
2
2
2
1
in the H-NMR spectra (see ESI). FT-IR analysis of this solution
observed. Furthermore, in situ high pressure NMR studies using
-
1
at room temperature showed a CO stretching mode at 1543 cm ,
complex 2 directly under turnover conditions (2/HNTf (1:1),
2
2
37-40
a typical value for  -coordinated formate.
Based on these 90 d -THF, T = 80 °C, p(H ) = 60 bar, p(CO ) = 20 bar) revealed
8
2
2
3
7
data and on basis of literature precedence, we assign the
structure of this complex as [(Triphos)Ru( -
O CH)(THF)] (3a), where the weakly bound
that complex 3a was formed right after pressurisation and
2
+
2
solvent molecule THF accounts for the
fluxionality at room temperature (Scheme 3).
This interpretation is supported by the
formation of a non-fluxional formate complex
upon addition of 0.1 mL acetonitrile to the
Scheme 4 Formation of the acetonitrile formate complex 3b from 3a by addition of MeCN to a
solution of 3a in THF and decarboxylation of 3b to hydride complex 8 at room temperature.
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 3

View Article Online
DOI: 10.1039/C4SC02087A
Chemical Science
Page 4 of 10
(
4) with S being a free coordination site or
4
6
weakly bound solvent molecule.
The
formation of dimeric or trimeric species
[(Triphos)Ru(-OAc)] (NTf ) could be
35
x
2 x
excluded by using two structurally different
Scheme 5 Methanol is formed with high yield by hydrogenation of complex 3a.
Triphos derivatives and stirring an equimolar
2
(
12.5 mol) mixture of [(Triphos)Ru( -
remained the major detectable phosphorus containing species
2
OAc)Cl] (9) and [(Triphos-Anisyl)Ru( -OAc)Cl] (10) (Triphos-
Anisyl 1,1,1-tris{bis(4-
methoxyphenyl)phosphinemethyl}ethan) together with AgNTf2
30 mol) in toluene (1.5 mL) at 60 °C for 5 hours. The toluene
was removed in vacuo, the residue dissolved in d -DCM
present in solution throughout the reaction. No hydride signals
that could be related to an active species were observed in the H-
40
=
1
NMR spectra (see ESI). These data indicate that the presumed
cationic hydride intermediate is too short-lived to be observed on
(
5
0
5
0
5
0
2
3
7
the NMR-time-scale, but is converted to the observable formate
31
1
(
0.5 mL), and the mixture analysed by NMR. The P{ H}-NMR
2
+
complex [(Triphos)Ru( -O CH)(THF)] 3a as the resting state
2
45 spectrum at room temperature showed two singlets at 43.5 and
by rapid and reversible CO insertion into the metal-hydride bond
2
2
45.3 ppm in a ratio of nearly 1:1 related to [(Triphos)Ru( -
3
8, 42-44
under turnover conditions.
2
OAc)(S)]NTf2 (4) and [(Triphos-Anisyl)Ru( -OAc)(S)]NTf
2
1
1
2
2
3
Identification of the formate complex 3 as the active
intermediate suggests that the major role of the acid additive in
(
11), respectively. The absence of further signals due to mixed
complexes (e.g. [Ru (Triphos)(Triphos-Anisyl)(-OAc) ](NTf ) )
2
2
2 2
the catalytic system 2/HNTf is the generation of cationic species
47
2
50 supports the monomeric structure of 4 in solution.
as the active site upon reductive removal of the TMM-ligand. In
order to probe this assumption, we decided to start from an
isolated cationic complex as precursor. After numerous
unsuccessful attempts to isolate complex 3 in stable form as solid,
we turned our efforts towards the analogous cationic acetate
The reactivity of the acetate complex 4 under CO2
(
20 bar at r.t.) and H (60 bar at r.t.) pressure was investigated in
2
a HP-NMR experiment (see ESI). After 1.5 hours at 80 °C and
31
1
1
hour at 140 °C, ca. 60 % ( P{ H}-NMR) of 4 was converted to
5
5
the formate complex 3a and ethanol from acetate hydrogenation
2
45
complex. Stirring [(Triphos)Ru( -OAc)Cl] (9) together with
1
was detected by H-NMR in the reaction mixture. Methanol was
one equivalent AgNTf2 for 3 h at 60 °C in THF led to the
indeed observed in the solution with a TON of 5, confirming that
cationic complex 4 was operating as molecularly defined direct
3
1
1
precipitation of AgCl. The P{ H}-NMR spectrum showed the
selective formation of only one sharp singlet at 44.0 ppm,
indicating the formation of a symmetrical complex species. After
filtration of the yellow solution over silica and removal of the
solvent in vacuo a yellow powder was obtained. Characterisation
precursor for the catalytic cycle for CO hydrogenation without
2
60
the need of any acid additive.
1
13
19
Mechanistic pathways on basis of DFT calculations
of the material by H-, C- and F-NMR, FT-IR and by ESI-
2
HRMS revealed the presence of the cationic [(Triphos)Ru( -
OAc)] fragment (see ESI). Crystallisation from dichloromethane
In summary, the experimental results described above
clearly demonstrate that the Ru-Triphos framework is able to act
+
layered with pentane gave yellow single crystals of complex 4a
6
7
7
8
5
0
5
0
as molecular single-site catalyst for the hydrogenation of CO to
2
where the open coordination site was saturated with H O from
2
methanol. All observations are in accordance with a stepwise
reduction of CO2 to methanol via the formate anion in the
coordination sphere of a homogeneous cationic organometallic
complex. Complex 3, which is accessible from different
precursors in the presence or absence of acid co-catalysts,
represents the resting state under turnover conditions.
Consequently, spectroscopic insight into the subsequent reduction
steps cannot be obtained directly. We therefore used DFT
calculations to explore possible reaction pathways for this multi-
step transformation. Based on our previous investigations on the
Ruthenium-Triphos system and on recent work by other groups
on catalytic hydrogenation of CO2 or methanol reforming, a
plausible basic catalytic cycle that reduces carbon dioxide
stepwise through the formic acid and formaldehyde stage to
methanol via the key intermediates I, V, IX, XVIII can be
adventitious traces of water (Fig. 2). Thus, the acetate complex in
2
solution can be formulated as [(Triphos)Ru( -OAc)(S)][NTf ]
2
2
1, 23, 28, 30, 37, 48-51
formulated as shown in Scheme 6.
Starting from a cationic Ruthenium-hydride complex I,
the migratory insertion of CO results in the formation of the
2
spectroscopically observed Ruthenium-formate species V.
Fig. 2 Molecular structure of the cation 4a (S = H
derived from single crystal X-ray diffraction (hydrogen atoms are omitted
for clarity). Some selected bond lengths (Å): Ru-P1 = 2.245(9); Ru-P2 =
2
O) in the solid state as
85 Reaction with one equivalent of hydrogen leads to reduction
beyond the formic acid stage to give the respective Ruthenium-
hydroxymethanolate species IX, which is then transformed to the
Ruthenium-methanolate complex XVIII via intermediate
formation of formaldehyde and consumption of a second
2
.255(3); Ru-P3 = 2.253(0); Ru-O1 = 2.171(2); Ru-O2 = 2.208(6); Ru-O3
2.204(7).
=
4
| Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

View Article Online
DOI: 10.1039/C4SC02087A
Page 5 of 10
Chemical Science
hyper surface. Rotation of the formate species in III about the
Ru-O and the O-C-bond generates complex IV (3.3 kcal/mol),
4
3, 48
which is significantly more stable than III.
The barrier for the
45
dissociation of H from IV is very low (4.2 kcal/mol) and the
2
exchange of H in IV by solvent generates the stable ruthenium
2
formate complex
V (-3.2 kcal/mol), which is also the
experimentally observed resting state complex 3a.
In accord with the work by Sakaki, the proton transfer to the
carbonyl C=O bond in the six-membered transition state is also
energetically favourable in the Ru-Triphos system. The change
50
4
8
of coordination mode of the formate species in V from bidentate
to monodentate with subsequent coordination of H at the vacant
2
coordination site forms VI. The coordinated H2 molecule is
Scheme 6 Basic catalytic cycle for the transformation of CO
at the Ru-Triphos fragment via the formic acid and formaldehyde stage
through the key intermediates I, V, IX, XVIII. P Ru denotes the Triphos-
Ru(+II) fragment comprising additional ligands to fill the coordination
sphere as discussed in the detailed analysis.
2
to methanol
55 cleaved heterolytically via TSVI-VII with a very small barrier of
1.7 kcal/mol to VII (12.5 kcal/mol). At this point of the cycle the
generation of formic acid is completed and the Ru-H unit for
further reduction is regenerated.
3
3
0, 50, 51
2
In contrast to hydrogenation of CO to formic acid, far
equivalent of hydrogen.
In the last step, hydrogenolysis of
6
6
7
7
0
5
0
5
less is known about the reduction beyond the formate level. Only
most recently, Ruthenium and Iridium catalysts have been
reported to facilitate this reaction step via unprecedented catalytic
the Ru-OMe unit requires the third equivalent of H to liberate
the product and closes the cycle by restoring the Ruthenium-
hydride complex I. A plausible structure for complex I as starting
2
3
0, 50, 52, 53
pathways.
In the Ru-Triphos system, the hydride
5
0
5
0
5
0
5
0
point of the calculations is the cationic species
+
transfer on to the coordinated formic acid was found to be
energetically feasible only after the solvent in VII is replaced by
[
(Triphos)Ru(H)(H )(THF)] that is, for example, most likely to
2
be formed from complex 4 upon hydrogenative removal of the
2
8
H generating complex VIII. The exchange of solvent by H is
acetate ligand as the initiating step. The individual steps of the
cycle shown in Scheme 6 were therefore analysed in detail from
this starting point, whereby the reduction steps are composed of
hydride migration/protonolysis events. For clarity we constrain
the discussion here to the energetically most favourable pathways
and some particularly relevant alternatives and refer the reader to
the SI for additional information.
2
2
almost thermoneutral placing VIII at an energy of 14.3 kcal/mol.
The hydride transfer in VIII to the carbon atom of formic acid
has a barrier of 15.5 kcal/mol, placing TSVIII-IX at 29.8
kcal/mol on the hyper surface. This energetically well accessible
reaction step forms the Ruthenium-hydroxymethanolate species
1
1
2
2
3
3
4
(
IX) which is the crucial intermediate for the unique performance
of the Ru-Triphos system in the hydrogenation of CO beyond the
2
formic acid stage. Three other energetically less favoured reaction
pathways for the formation of hydroxymethanolate from CO2,
Formation of hydroxymethanolate via formic acid (I – IX,
Figure 3): The insertion of CO into metal-hydride bonds and
2
including outer sphere attack of CO , were calculated and are
subsequent hydrogenolysis of the metal-formate units has been
subject of numerous experimental and theoretical studies in the
context of formic acid production. The
closest related example to the Triphos-
system are Ru(II)-catalysts bearing
2
shown in the ESI. The next key step is the cleavage of the carbon-
three monodentate phosphine ligands
whose high efficiency for formic acid
production was rationalised in
a
comprehensive theoretical study by the
4
8
group of Sakaki. An analogous route
was therefore investigated for the initial
step of the current system (Figure 3).
In the starting complex I either the
THF molecule or the H2 molecule is
replaced by CO generating complexes
2
II and IIa, respectively. Both
compounds are endergonic with respect
to the reference point I by 8 and 10.2
kcal/mol. The classical hydride centre in
II can subsequently be transferred to the
carbon atom of CO passing transition
2
state TSII-III. The barrier is
appreciably low (11.7 kcal/mol) placing
the TSII-III at 21.2 kcal/mol on the
Fig. 3 Initial steps of the DFT calculated reaction pathways for the hydrogenation of CO to methanol at
2
the cationic Ru-Triphos centre, starting from complex I (S = THF) as the active species. The Triphos
ligand is omitted for clarity.
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 5

View Article Online
DOI: 10.1039/C4SC02087A
Chemical Science
Page 6 of 10
3
4
4
5
5
5
0
5
0
5
medium-assisted
proton
transfer or intramolecular
proton transferwithin the
coordination sphere of the
Ru-centre.
The
lowest
energy pathway (ca. 28
kcal/mol) from the presently
investigated alternatives is
provided by external proton
transfer using carboxylates
as proton shuttle. The
intramolecular
pathways
result in significantly higher
barriers (ca. 40 kcal/mol),
but still provide general
viable alternatives that are
also in line with the
experimental
described below.
results
Hydrogenation
of
formaldehyde to methanol
(XV-I’; Figure 5): Once the
Fig. 4 Calculated reaction pathways generating formaldehyde complex XV via medium-assisted proton transfer
(acetate: gray; water: blue). The Triphos ligand is omitted for clarity, S = THF.
formaldehyde level is reached
in complex XV the solvent
can be replaced again by H to arrive at XVI. The subsequent
oxygen bond leading to formaldehyde on the path to methanol.
6
6
7
7
8
8
9
0
5
0
5
0
5
0
2
migratory transfer of the classical hydride (TSXVI-XVII) is
almost barrierless with 0.7 kcal/mol and leads to the methanolate
complex XVII that is stabilised by an agostic C-H-Ru interaction.
Association of a solvent molecule opens the agostic bond to give
Cleavage of the C-O bond and generation of formaldehyde
(
IX – XV; Figure 4): The conversion between free methanediol
5
0
5
0
5
0
5
0
and formaldehyde has been largely explored.
In the
coordination sphere of IX, the protonolysis of the Ru-O is again
required to initiate this process. Firstly, we considered transfer of
XVIII. The intramolecular proton transfer from the coordinated
54,55
H -molecule through the “-bond metathesis-like”
2
four-
protons generated from the acidic Ru-H units under turnover
2
membered transition state TSXVIII-XXIV has an energy barrier
of 31.5 kcal/mol. Finally, the reaction product methanol
dissociates from the corresponding complex XXIV and for
completeness we calculated the barrier TSXXIV-I’ for H2
association, which is below 10 kcal/mol indicating this process to
be facile. It should be noted that I’ lies 14.1 kcal/mol above the
reference point, indicating the overall reaction to be endergonic
under the boundary conditions of the calculation model. The
inclusion of solvent effects, however, predicts the reaction to be
exergonic, in accordance with the experimental observation and
standard state thermodynamics (see ESI for details).
Similar to the protonolysis of the hydroxymethanolate
complex, carboxylate-assisted proton transfer provides an
alternative low energy pathway. Substitution of the hydrogen
molecule in XVIII by acetic acid is energetically favourable to
give XX and subsequent protonation of the methanolate oxygen
atom via TSXX-XXI has no significant barrier and results in the
acetate-methanol complex XXI (-2.2 kcal/mol). Hydrogen
addition (XXII) and carboxylate assisted heterolytic cleavage of
conditions via the reaction medium (Figure 4). A low energy
pathway (grey profile) was calculated if acetic acid was used as
model for carboxylate units as proton shuttles in the presence of
catalyst precursors 4 (acetate) or 2 (formate) according to the in
situ NMR studies. After de-coordination of the hydroxy group in
the hydroxymethanolate species IX, a molecule of acetic acid
coordinates via the carbonyl oxygen atom forming X in a
practically a thermoneutral event. The acetic acid then protonates
the hydroxy group of the hydroxymethanolate (TSX-XI, 27.1
kcal/mol), with a barrier of 9.3 kcal/mol. Water is loosely
coordinated after the reaction (XI) and cleaved off generating
XII. Dissociation of one of the acetate oxygen atoms of XII,
changing the acetate coordination from bi- to monodentate, and
1
1
2
2
3
association of H generates XIII, which is placed at a height of
2
2
0.3 kcal/mol. The subsequent heterolytic cleavage of H under
2
regeneration of acetic acid is practically barrierless (0.4 kcal/mol)
and product XIV is only marginally more stable than the reactant.
Dissociation of acetic acid and association of solvent generates
XV. An analogous path using water, which is formed
stoichiometrically in the overall hydrogenation sequence, as the
proton shuttle gave a significantly higher barrier of 41.2 kcal/mol
(XIIIa-XIVa, blue profile). In addition to the external proton
transfer, direct protonolysis within the coordination sphere was
also investigated (see ESI).
H to regenerate acetic acid (XXIII) can occur through a six-
2
membered transition state, rendering this more facile than the
direct heterolytic cleavage of the Ru-methanolate unit.
Dissociation of acetic acid leads to XXIV at which point the two
pathways merge again.
Overall, the results of the DFT calculations demonstrate
In essence, the C-O bond cleavage can be achieved from the
hydroxymethanolate intermediate IX through pathways involving
the possibility of a stepwise reduction of CO to methanol in the
2
6
| Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

View Article Online
DOI: 10.1039/C4SC02087A
Page 7 of 10
Chemical Science
The lower TON obtained
when using the acetate complex 4
instead of 2/HNTf2 (1:1) under
otherwise identical conditions can be
explained by the less efficient
initiation with 4 due to more difficult
hydrogenation of the acetate groups to
form the common intermediate 3a
3
4
4
5
5
0
5
0
(
vide supra). This distinct reactivity is
also reflected in catalytic experiments
for the hydrogenation of the
corresponding free acids: Using 2
together with 1 eq. HNTf2, 100
equivalents of formic acid could be
fully converted to methanol at a
hydrogen pressure of 60 bar (at r.t.)
and a reaction temperature of 140 °C
within 24 hours, whereas a reaction
temperature of 180 °C was necessary
for the full conversion acetic acid to
ethanol
2.0 mL
(c(Ru) = 12.5 mmol/L,
THF). The efficient
Fig. 5 Calculated reaction pathways for the hydrogenation of formaldehyde generating methanol via
intramolecular proton transfer (red) and carboxylate assisted proton transfer (acetic acid: grey). The Triphos
ligand is omitted for clarity, S = THF.
5
5
hydrogenation of formic acid under
these conditions is in accordance with
coordination sphere of a single Ru-Triphos centre. The individual
the proposed catalytic cycle shown in Scheme 6. To complete the
picture, the hydrogenation of 100 equivalents of
paraformaldehyde was also assessed and full conversion was
reduction steps occur by migratory transfer of classical Ru-
hydride ligands, exhibiting low to moderate barriers in all cases.
The protonolysis steps of the resulting Ru-O bonds can occur
intramolecularly via heterolytic cleavage of coordinated H2
molecules. External proton transfer assisted by carboxylate
groups present under turnover conditions may lower the
corresponding barriers significantly.
6
0
indeed
(
achieved
with
the
same
catalytic
system
5
c(Ru) = 12.5 mmol/L, 2.0 mL THF, 0.2 mL H O, 60 bar H at
r.t., 140 °C, 24 h). Again, formation of formate complex 3a was
observed in P{ H}-NMR, indicating full reversibility of the
catalytic cycle.
2
2
3
1
1
65
Table 1 Hydrogenation of carbon dioxide to methanol in the absence of
alcohol additive.
1
0
Parameter variation and catalyst recycling in a biphasic
system
[a]
[c]
Entry
Cat
Acid (eq)
T
°C]
[b]
TON
pH2 / pCO2
After demonstrating the principle possibility for the catalytic
[
hydrogenation of CO to methanol in the absence of an alcohol
[bar/bar]
20/60
20/60
20/60
20/60
20/60
20/60
20/60
20/60
20/60
20/60
10/30
30/90
20/80
20/100
20/60
20/60
20/60
2
1
2
3
4
5
6
7
8
9
2
2
4
4
2
2
2
2
2
2
2
2
2
2
2
2
12
HNTf
2
(1.0)
140
140
140
140
140
140
140
120
100
80
140
140
140
140
140
140
140
228
8
additive, the performance of the Ru-Triphos precursor systems 2
and 4 was evaluated further by systematic variation of key
reaction parameters and the results were corroborated for
consistency with the mechanistic proposal (Table 1).
Firstly, the catalyst systems 2 and 4 were compared
under a standard set of reaction conditions (V(THF) = 2.08 mL,
-
-
1
2
2
3
5
0
5
0
165
156
196
181
135
169
67
HNTf
2
HNTf (1.5)
2
(0.5)
HNTf
p-TsOH (1.0)
HNTf (1.0)
2
(2.0)
2
c(Ru) = 12 mmol/L, p(CO ) = 20 bar at r.t., p(H ) = 60 bar at
2
2
HNTf (1.0)
2
r.t., = 140°C, t = 24 h). Using 4 as catalyst precursor for the CO2
hydrogenation in THF gave a TON (turnover number = mmol
MeOH/mmol catalyst) of 165 in the absence of any additives
10
HNTf (1.0)
24
78
2
1
1
1
1
2
3
HNTf
HNTf
HNTf
2
2
2
(1.0)
(1.0)
(1.0)
367
301
348
335
442
256
(Table 1, entry 3). Using an additional 0.5 eq. HNTf2 did not
14
2
HNTf (1.0)
[
d]
show an increased TON (Table 1, entry 4). In contrast precursor 2
showed only a very poor performance in the absence of the acid
additive (Table 1, entry 2). However, a TON of 228 was found
15
HNTf
HNTf
HNTf
2
2
2
(1.0)
(1.0)
(1.0)
[
e]
1
1
6
7
[
d]
when conducting a CO hydrogenation reaction with catalyst 2
[a] Reaction conditions: 25 µmol [Ru], 2.08 mL THF, 24 h; [b] = at room
temperature; [c] TON = mmol MeOH/mmol catalyst; [d] 12.5 µmol [Ru];
[e] 6.3 µmol [Ru].
2
and 1 eq. of HNTf2 (Table 1, entry 1). The need for an acid
70
additive in the case of precursor 2 is consistent with the formation
+
of the cationic species [(Triphos)Ru(H)(H )(S)] , which is the
The lack of activity with catalyst 2 in absence of acid
2
catalytically active species I used as starting point in the
can be directly corroborated with the formation of the neutral
calculated catalytic cycle.
complex [(Triphos)Ru(H) CO] (12), which was observed as the
almost exclusive species present in solution by P{ H}-NMR
2
3
1
1
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 |
7

View Article Online
DOI: 10.1039/C4SC02087A
Chemical Science
Page 8 of 10
2
8
spectroscopy under these conditions (see ESI). The protonation
reactions after different reaction times (Figure 6). The reaction
of this complex with strong protic acids was shown by Zanobini 60 started without any pronounced induction period reaching a TON
et al. to lead to the formation of the complex
of 70 already after 1 hour. This corresponds to an initial turnover
frequency (TOF) of 70 h that is well in the range of the activity
+
-1
[
(Triphos)Ru(CO)(H)(H )] (13), a cationic structure closely
2
5
6
5
0
5
0
5
0
5
0
5
0
5
resembling the active hydride species I inferred above. Using
the isolated complex 12 together with 1 equivalent of HNTf2 in
of the active sites in the state-of-the-art heterogeneous Cu/ZnO-
1
8
based catalysts. Methanol formation continued smoothly to
THF resulted indeed in an active catalyst for the CO2 65 reach a TON of 258 after 16 hours. At this point the pressure in
hydrogenation reaction yielding a TON of 256 after 24 h
Table 1, entry 17) which is about 76 % of the TON obtained
the reactor vessel had dropped from initially 120 bars to 72 bars
due to the consumption of the reactive gases. Therefore, a
reaction was conducted for 32 h where the reactor was re-
pressurised to the initial pressure with p(CO )/p(H ) = 1/3 after
(
1
1
2
2
3
3
4
4
5
5
using the catalyst 2/HNTf under identical conditions (Table 1,
2
entry 15). Again, formation of the formate intermediate 3a was
2
2
3
1
1
observed when the solution was analysed by P{ H}-NMR (see 70 16 hours leading to TON of 478. Finally, a reaction was run for
ESI).
48 hours with re-pressurisation to the initial pressure with
Variation of the amount of HNTf added to complex 2
p(CO )/p(H ) = 1/3 after 16 hours and again after 32 hours,
2
2
2
revealed a maximum of the observed TON at the 1:1 ratio
Table 1, entries 5-6) corresponding to the stoichiometric ratio
yielding a total TON of 603. These experiments clearly indicate
the high stability of the active catalyst resulting in a nearly linear
(
required for the reductive removal of the TMM-ligand leading to 75 increase of the TON under isobaric conditions. The absence of an
I. Using 2 together with 1 eq. p-TsOH instead of HNTf gave a
induction period indicates that the presence of methanol is not
enhancing the rate of catalysis under these conditions. This does
not rule out the possibility that the reaction proceeds partly also
via a cascade reaction involving methyl formate as intermediate
2
lower TON of 135 (Table 1, entry 7) under otherwise identical
conditions. NMR-analysis of the reaction solution after 1 h
reaction time showed the formation of the formate species 3a as
major component in solution in both cases (see ESI). However, 80 once methanol has been formed, as catalyst 2 is able to promote
2
4
[
(Triphos)Ru(p-TsO) ] (14) was present also in about 15 % as
the hydrogenation of alkyl formates to methanol.
2
3
1
1
indicated by a broad singlet at 38.7 ppm in the P{ H}-NMR
spectrum measured at room temperature, which split up into a
triplet ( = 42.0 ppm, J = 47.5 Hz) and doublet ( = 36.1 ppm,
J = 47.5 Hz)
when measured at 233 K in d -THF. This
8
assignment was supported by mass spectrometry (FAB) and
independent generation of 14 by addition of 2 equivalents of p-
TsOH to 2 in THF at room temperature. Thus, the presence of
even weakly-coordinating anions in the reaction mixture hampers
the formation of the formate species 3a, explaining the preferred
choice of HNTf as acid additive.
2
Identifying the system 2/HNTf2 (1:1) as the most
practical and effective catalyst precursor so far, the influence of
some key reaction parameters on the TON after 24 hours reaction
time was assessed. Lowering the catalyst concentration together
with the acid concentration from 12 µmol/mL to 6 µmol/mL and
further to 3 µmol/mL resulted in a significant increase in TON
from 228 to 335 and 442, respectively (Table 1, entries 1 and 15-
Fig. 6 Conversion/time-profile of the hydrogenation of CO
using catalyst 2 (12.5 μmol catalyst 2; 12.5 μmol HNTf , 20 bar CO
0 bar H at r.t.; 140 °C reaction temperature; 2.08 mL THF), as
obtained from batch experiments terminated at the given reaction times.
In the case of the reaction terminated after 32 h the autoclave was re-
pressurised to the initial pressure with p(CO )/p(H ) = 1/3 after 16 h
2 2
(▲). In the case of the reaction terminated after 48 h the autoclave was
re-pressurised after 16 h and again after 32 h (●).
2
to methanol
2
2
+
6
2
1
6). Although final conclusions have to await a detailed kinetic
analysis, the formation of the dimeric complex 6 as part of the
deactivation mechanism is in line with this trend. Decreasing the
reaction temperature to 120 °C, 100 °C and 80 °C resulted in
reduced TONs of 169, 67 and 24 (Table 1, entries 8-10).
Variation of the total pressure while maintaining the
stoichiometric ratio of p(CO )/p(H ) = 1/3 from 40 bar to 80 bar
As all previous results are consistent with the cationic
Ru-Triphos formate complex 3 as resting state in this process, we
explored strategies for the recycling of the catalyst in active form.
Considering the various options for the isolation of the product
MeOH from the homogeneous catalyst, distillation seems an
2
2
and 120 bar resulted in an increase of the obtained TONs from 78
to 228 and 367 (Table 1, entries 1 and 11-12). Using an excess of
8
5
H2 (20 bar CO2 + 80 bar or 100 bar H ) resulted in largely
2
increased TONs of 301 and 348, respectively (Table 1, entries 16-
obvious possibility. However, the hydrogenation of CO yields a
2
1
7). In the latter case about 40 % of the totally available carbon
stoichiometric amount of water, which is the least volatile
component in the THF/MeOH/water product mixture. Thus, it
would accumulate upon stripping of the MeOH product,
ultimately becoming the limiting factor even if the catalyst would
be thermally stable for recycling. Multiphase catalysis offers an
alternative strategy, where the separation is based on differences
in solubility rather than volatility. For the current process, an
feedstock CO was converted to methanol, as calculated from the
2
amount of MeOH formed (8.7 mmol) and the amount of CO2
initially charged (22.1 mmol, determined by weight).
90
A conversion/time profile of the CO hydrogenation to
2
methanol in THF using 2/HNTf2 (1:1, 12.5 mol; otherwise
standard conditions) was mapped out by termination of batch
8
| Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]

View Article Online
DOI: 10.1039/C4SC02087A
Page 9 of 10
Chemical Science
30
2/HNTf could be recycled three times, resulting in
2
a total TON of 769 after 4 cycles. The TON per
cycle was reduced significantly especially between
cycle three and four, but still nearly 50 % of the
initial productivity was retained in this non-
optimised sequence.
35
Conclusions
Fig. 7 Aqueous biphasic system for recycling of the cationic Ru-Triphos catalyst in the
organic 2-MTHF phase and removing the product MeOH in the aqueous phase for
downstream processing.
The results of this study demonstrate for the first
time the hydrogenation of CO to methanol using a
2
single organometallic catalyst in homogenous
aqueous biphasic system was envisaged, where the catalyst is
retained and recycled in an organic phase, whereas the product is
removed in an aqueous phase for downstream processing.
4
4
5
5
6
6
7
7
8
0
5
0
5
0
5
0
5
0
solution without the need for an alcohol additive. The
experimental and theoretical results are consistent with a
mechanistic picture where this unprecedented transformation
occurs at a cationic Triphos-Ru fragment as molecularly-defined
5
7, 58
Substitution of the solvent THF with 2-methyl tetrahydrofuran (2-
MTHF) opens up the possibility for the realisation of such a two
phasic reaction/separation system as 2-MTHF has a miscibility
5
0
5
0
5
2
active site. The cationic formate complex [(Triphos)Ru( -
+
O CH)(S)] (3) (S = solvent) represents the resting state of the
2
2
7
gap with water. All material streams can be recycled internally
in such a process scheme providing also the possibility for
continuous-flow operation (Figure 7).
catalytic cycle under turnover conditions and can be obtained
from various stable and readily available catalyst precursors.
Through a series of hydride transfer and protonolysis steps, the
CO2 reduction can pass through the formic acid and
formaldehyde stage within the coordination sphere of a single
ruthenium centre. The barriers for the proton transfer steps may
be significantly lowered if assisted by the reaction medium. The
active species shows a remarkable stability, with decarbonylation
and dimerisation as potential deactivation mechanisms. Recycling
of the catalyst is possible in the aqueous biphasic system 2-
MTHF/water, opening the possibility for continuous-flow
operation.
1
1
2
2
Assessing the partitioning of methanol in a 2-
MTHF/water biphasic mixture showed that 80 % of 0.2 mL
MeOH can be isolated from 1.0 mL of 2-MTHF using 1.0 mL
water in a single extraction step. Application of 2-MTHF as
solvent for the catalytic reaction under standard conditions
(p(CO ) = 20 bar / p(H ) = 60 bar at r.t., T = 140°C, V(solvent) =
2
2
2
.08 mL) was also found to be possible without any problem for
catalyst 2 (25 µmol complex 2, 2/HNTf = 1:1), albeit with a
2
somewhat lower TON of 186 after 24 h as compared to THF. To
validate the combination of reaction and separation, a reaction
with 2 (12.5 µmol complex 2, 2/HNTf2 = 1:1) in 2-MTHF
The Triphos-Ruthenium system is the very first
homogeneous catalyst to enable this transformation. The facial
(2.0 mL) was terminated after 16 hours, the reaction mixture
coordination of the Triphos ligand imposes
a favorable
extracted by addition of 2.0 mL H O, and the orange catalyst/2-
2
geometrical arrangement for the hydride transfer to carboxylate
units in general (see ESI for a comparison of facial with
MTHF phase recycled to the autoclave after simple decantation.
A small amount of fresh 2-MTHF (0.25 mL) was added to
compensate for any loss of 2-MTHF with the product phase. The
28
meridional arrangement). Furthermore, the heterolytic cleavage
of hydrogen offers low-energy pathways for the protonolysis of
Ru-O units under regeneration of the hydride ligand. Together
with the high thermal stability of Triphos-Ruthenium complexes,
these features seem to play an important role in the reduction of
CO2 beyond the formate level with this catalyst. Further
developments on the basis of the methodological approach of
organometallic chemistry e.g. by systematic ligand variation
based on the current mechanistic hypothesis are likely to produce
even more active and stable systems. Already at this early stage
of the development, turnover frequencies per Ru-centre are in the
same range as for the active sites in traditional heterogeneous
catalysts for methanol synthesis. The possibility to operate under
multiphase conditions provides opportunities to overcome the
limitations in productivity inherent to batch or repetitive batch
operation. Therefore, we believe that the results of this study
provide not only fundamental mechanistic insight into the
activation and transformation of CO and H in organometallic
aqueous layers were analysed for MeOH content by quantitative
1
H-NMR in d -acetone using mesitylene as standard, and only the
6
concentration in the aqueous streams was used for the calculation
of the apparent TON. As seen from Figure 8, the catalyst system
769
8
6
4
2
00
00
00
00
0
6
60
4
69
cycle TON
total TON
TON
247 247
2
22
1
91
110
1
2
3
4
Cycle
2
2
Fig. 8 Recycling of catalyst system 2 (12.5 µmol complex 2,
/HNTf = 1:1) using MTHF (2.0 mL) as solvent and water
2.0 mL) as extracting agent in an aqueous biphasic system. Each
cycle was run for 16 h (20 bar CO + 60 bar H at r.t.; 140 °C
chemistry, but also open promising targets for research at the
interface of molecular and engineering sciences.
2
2
(
2
2
reaction temperature). The TONs obtained per cycle are shown in
dark grey, the total TONs summing up the cycles are shown in
light grey.
Acknowledgements
This work was supported in part by the Cluster of Excellence
This journal is © The Royal Society of Chemistry [year]
Journal Name, [year], [vol], 00–00 | 9

View Article Online
DOI: 10.1039/C4SC02087A
Chemical Science
Page 10 of 10
“
Tailor-Made Fuels from Biomass”, which is funded by the 65 27. F. M. A. Geilen, B. Engendahl, A. Harwardt, W. Marquardt, J.
Klankermayer and W. Leitner, Angew. Chem. Int. Ed., 2010, 49,
Excellence Initiative by the German Federal and State
Governments to promote science and research at German
universities and in part by the project SusChemSys. The project
5
510-5514.
2
8. F. M. Geilen, B. Engendahl, M. Hölscher, J. Klankermayer and W.
Leitner, J. Am. Chem. Soc., 2011, 133, 14349-14358.
5
"Sustainable Chemical Synthesis (SusChemSys)" is co-financed 70 29. J. Coetzee, D. L. Dodds, J. Klankermayer, S. Brosinski, W. Leitner,
A. M. Slawin and D. J. Cole-Hamilton, Chemistry, 2013, 19, 11039-
by the European Regional Development Fund (ERDF) and the
state of North Rhine-Westphalia, Germany, under the Operational
Programme "Regional Competitiveness and Employment" 2007-
1
1050.
3
0. M. Nielsen, E. Alberico, W. Baumann, H. J. Drexler, H. Junge, S.
Gladiali and M. Beller, Nature, 2013.
2
013. Generous allocation of computer time by the Computation
7
8
8
9
9
5
0
5
0
5
31. S. I. Hommeltoft and M. C. Baird, Organometallics, 1986, 5, 190-
1
0
and Communication Centre of RWTH Aachen University is
gratefully acknowledged.
195.
3
3
2. K. Tominaga and Y. Sasaki, J. Mol. Catal. A: Chem., 2004, 220, 159-
65.
1
3. N. Sieffert, R. Reocreux, P. Lorusso, D. J. Cole-Hamilton and M.
Buhl, Chem. Eur. J., 2014, 20, 4141-4155.
34. B. D. Yeomans, D. G. Humphrey and G. A. Heath, J. Chem. Soc.,
Dalton Trans., 1997, 4153-4166.
5. K. M. Sung, S. Huh and M. J. Jun, Polyhedron, 1999, 18, 469-479.
6. O. R. Allen, S. J. Dalgarno, L. D. Field, P. Jensen and A. C. Willis,
Organometallics, 2009, 28, 2385-2390.
7. I. Mellone, M. Peruzzini, L. Rosi, D. Mellmann, H. Junge, M. Beller
and L. Gonsalvi, Dalton Trans., 2013, 42, 2495-2501.
References
3
3
1
2
2
3
3
4
4
5
5
6
5
0
5
0
5
0
5
0
5
0
1. M. Peters, B. Koehler, W. Kuckshinrichs, W. Leitner, P. Markewitz
and T. E. Mueller, ChemSusChem, 2011, 4, 1216-1240.
2
3
.
.
M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann and F. E.
Kühn, Angew. Chem. Int. Ed., 2011, 50, 8510-8537.
M. Aresta and A. Dibenedetto, Dalton Trans., 2007, 2975-2992.
3
3
8. M. K. Whittlesey, R. N. Perutz and M. H. Moore, Organometallics,
4. T. Sakakura, J. C. Choi and H. Yasuda, Chem. Rev., 2007, 107, 2365-
387.
1
996, 15, 5166-5169.
2
39. I. S. Kolomnikov, A. I. Gusev, G. G. Aleksandrov, T. S. Lobeeva, Y.
5
6
.
.
A. Behr, Angew. Chem. Int. Ed., 1988, 27, 661-678.
D. J. Darensbourg and R. A. Kudaroski, Adv. Organomet. Chem.,
T. Struchkov and M. E. Volpin, J. Organomet. Chem., 1973, 59, 349-
3
51.
0. M. G. Bradley, D. A. Roberts and G. L. Geoffroy, J. Am. Chem. Soc.,
981, 103, 379-384.
41. C. Bianchini, A. Meli, S. Moneti and F. Vizza, Organometallics,
998, 17, 2636-2645.
1
983, 22, 129-168.
4
7. P. Braunstein, D. Matt and D. Nobel, Chem. Rev., 1988, 88, 747-764.
1
8
9
1
.
.
D. Walther, Coord. Chem. Rev., 1987, 79, 135-174.
W. Leitner, Coord. Chem. Rev., 1996, 153, 257-284.
1
0. G. Centi, E. A. Quadrelli and S. Perathoner, Energy & Environmental
Science, 2013, 6, 1711-1731.
4
4
4
2. A. D. Getty, C.-C. Tai, J. C. Linehan, P. G. Jessop, M. M. Olmstead
and A. L. Rheingold, Organometallics, 2009, 28, 5466-5477.
3. A. Urakawa, F. Jutz, G. Laurenczy and A. Baiker, Chem. Eur. J.,
2007, 13, 3886-3899.
11. R. Fornika and E. Dinjus, Carbon Dioxide as C1-Building Block", in
Applied Homogeneous Catalysis with Organometallic Compounds
B. Cornils, W.A. Herrmann, Eds), VCH, Weinheim, 1996, 1048-
071.
"
(
1
1
1
1
1
1
1
1
00
05
10
15
4. P. G. Jessop, F. Joo and C.-C. Tai, Coord. Chem. Rev., 2004, 248,
2
425-2442.
1
1
1
2. A. Goeppert, M. Czaun, G. K. S. Prakash and G. A. Olah, Energy
Environ. Sci., 2012, 5, 7833-7853.
3. F. Asinger, Methanol: Chemie- und Energierohstoff, Springer-
Verlag, Berlin, 1986.
4. A. Goeppert, G. K. S. Prakash and G. A. Olah, Beyond Oil and Gas:
The Methanol Economy, WILEY-VCH, Weinheim, 2006.
4
4
5. A. B. Chaplin and P. J. Dyson, Inorg. Chem., 2008, 47, 381-390.
6. M. Hirano, R. Fujimoto, K. Hatagami, N. Komine and S. Komiya,
ChemCatChem, 2013, 5, 1101-1115.
7. J. M. Brown, I. Gridnev and J. Klankermayer, Amplification of
Chirality, 2008, 284, 35-65.
8. Y. Musashi and S. Sakaki, J. Am. Chem. Soc., 2000, 122, 3867-3877.
9. C. Fellay, N. Yan, P. J. Dyson and G. Laurenczy, Chem. Eur. J.,
2009, 15, 3752-3760.
0. L. E. Heim, N. E. Schlörer, J.-H. Choi and M. H. G. Prechtl, Nature
Commun., 2014, 5.
1. X. Yang, ACS Catal., 2014, 4, 1129-1133.
2. A. J. M. Miller, D. M. Heinekey, J. M. Mayer and K. I. Goldberg,
Angew. Chem. Int. Ed., 2013, 52, 3981-3984.
3. S. Savourey, G. Lefevre, J. C. Berthet, P. Thuery, C. Genre and T.
Cantat, Angew. Chem. Int. Ed., 2014.
4. F. Hutschka, A. Dedieu and W. Leitner, Angew. Chem. Int. Ed.,
4
4
4
15. M. Bertau, H. Offermanns, L. Plass, F. Schmidt and H.-J. Wernicke,
Springer, Berlin/Heidelberg,, 2014.
1
6. W. Wang, S. P. Wang, X. B. Ma and J. L. Gong, Chem. Soc. Rev.,
011, 40, 3703-3727.
7. http://www.carbonrecycling.is/ (2014).
5
2
1
5
5
18. M. Behrens, F. Studt, I. Kasatkin, S. Kuhl, M. Havecker, F. Abild-
Pedersen, S. Zander, F. Girgsdies, P. Kurr, B. L. Kniep, M. Tovar, R.
W. Fischer, J. K. Norskov and R. Schlogl, Science, 2012, 336, 893-
5
5
8
97.
1
2
2
9. K. Tominaga, Y. Sasaki, T. Watanabe and M. Saito, Bull. Chem. Soc.
Jpn., 1995, 68, 2837-2842.
0. S. Chakraborty, J. Zhang, J. A. Krause and H. R. Guan, J. Am. Chem.
Soc., 2010, 132, 8872-+.
1. E. Balaraman, C. Gunanathan, J. Zhang, L. J. W. Shimon and D.
Milstein, Nature Chem., 2011, 3, 609-614.
1
995, 34, 1742-1745.
20 55. F. Hutschka, A. Dedieu, M. Eichberger, R. Fornika and W. Leitner, J.
Am. Chem. Soc., 1997, 119, 4432-4443.
5
6. V. I. Bakhmutov, E. V. Bakhmutova, N. V. Belkova, C. Bianchini, L.
M. Epstein, D. Masi, M. Peruzzini, E. S. Shubina, E. V. Vorontsov
and F. Zanobini, Can. J. Chem., 2001, 79, 479-489.
22. E. Balaraman, Y. Ben-David and D. Milstein, Angew. Chem. Int. Ed.,
011, 50, 11702-11705.
3. C. A. Huff and M. S. Sanford, J. Am. Chem. Soc., 2011, 133, 18122-
8125.
2
25 57. G. Verspui, G. Elbertse, F. A. Sheldon, M. A. P. J. Hacking and R. A.
2
2
2
2
Sheldon, Chem. Commun., 2000, 1363-1364.
8. K. Burgemeister, G. Francio, V. H. Gego, L. Greiner, H. Hugl and
1
5
4. S. Wesselbaum, T. vom Stein, J. Klankermayer and W. Leitner,
Angew. Chem. Int. Ed., 2012, 51, 7499-7502.
5. H. T. Teunissen and C. J. Elsevier, Chem. Commun., 1998, 1367-
W. Leitner, Chem. Eurr J., 2007, 13, 2798-2804.
30
1
368.
6. T. vom Stein, T. Weigand, C. Merkens, J. Klankermayer and W.
Leitner, ChemCatChem, 2013, 5, 439-441.
1
0
| Journal Name, [year], [vol], 00–00
This journal is © The Royal Society of Chemistry [year]