Hydrogenation of CO2 to Formate over Ruthenium Immobilized on Solid Molecular Phosphines

Article abstract of DOI:10.1002/cssc.201800413

Formic acid is a promising hydrogen storage medium and can be produced by catalytic hydrogenation of CO₂. Molecular ruthenium complexes immobilized on phosphine polymers have been found to exhibit excellent productivity and selectivity in the catalytic hydrogenation of CO₂ under mild conditions. The polymeric analog of 1,2-bis(diphenylphosphino)ethane exhibited the highest activity and turnover numbers up to 13 170 were obtained in a single run. This catalyst was already active at 40 °C and with a catalyst loading of only 0.0006 mol %. Recycling experiments revealed a loss of activity after the first run, followed by a gradual decrease during the subsequent runs. This is attributed to a change in the catalytically active complex during the hydrogenation reaction. High selectivity towards formate and low leaching were maintained in the absence of CO formation. Based on the catalyst characterization, a mechanism for the CO₂ hydrogenation is proposed.

Full text of DOI:10.1002/cssc.201800413

Accepted Article
Title: Hydrogenation of CO2 to Formate over Ru Immobilized on Solid
Molecular Phosphines
Authors: Anna Kann, Heinrich Hartmann, Astrid Besmehn, Peter
Hausoul, and Regina Palkovits
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201800413
Link to VoR: http://dx.doi.org/10.1002/cssc.201800413
A Journal of
www.chemsuschem.org

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
Hydrogenation of CO2 to Formate over Ru Immobilized on Solid
Molecular Phosphines
Anna Kann,[a] Heinrich Hartmann,[b] Astrid Besmehn,[b] Peter J. C. Hausoul*[a] and Regina Palkovits*[a]
Abstract: Formic acid is a promising hydrogen storage medium
CO2
and can be produced by catalytic hydrogenation of CO2. We
report molecular ruthenium complexes immobilized on phosphine
polymers, presenting excellent productivity and selectivity in the
catalytic hydrogenation of CO2 under mild conditions. The
H2
H2
H2 delivery
H2 storage
polymeric form of dppe exhibited the highest activity and turnover
numbers up to 13,170 were obtained in a single run. This catalyst
was already active at 40 °C and a catalyst loading of only
0.0006 mol%. Recycling experiments revealed a loss of activity
after the first run, followed by a gradual decease in the
subsequent runs. This is attributed to a change of the catalytically
active complex during the hydrogenation reaction. High selectivity
towards formate and low leaching were maintained in the absence
of CO formation. Based on the catalyst characterization, a
mechanism for the CO2 hydrogenation is proposed.
Scheme 1. Proposed circular carbon economy of the hydrogen storage system
based on formic acid and CO2/H2.
Ru and Rh are highly active and selective for the hydrogenation
of CO2 to FA-.[1c-d] E.g. a Ir(III)-PNP-pincer complex proposed by
Tanaka et al. exhibited a remarkable turnover number (TON) of
3,500,000 and a turnover frequency (TOF) of 73,000 h-1.[6] Using
a Ru-PNP-Pincer complex, Filonenko et al. reported a TOF of
1,100,000 h-1. Despite the excellent catalytic performance of
homogeneous catalysts, difficulties in the catalyst separation and
recycling as well as simultaneous decomposition to CO2 and H2
during the purification of the product, are drawbacks for large-
scale production. Here, heterogeneous catalysts show crucial
advantages in terms of stability, separability, recycling and reactor
design.
First approaches focused on bulk metals such as Au, Pd and Ni
as well as supported metal catalysts based on Au, Pd and Ru.[7]
Most recently, Filonenko et al. reported a TON of 1088 for
Au/Al2O3.[8] This demonstrates that supported metal catalysts are
significantly less active. Alternatively, immobilized catalysts have
been proposed to combine the selectivity and activity of metal
complexes with the advantages of heterogeneous catalysts. First
attempts involve grafting Ir or Ru complexes onto polymeric and
oxidic supports.[9] Recently, the group of Hicks reported a
maximum TON of 2700 for Ir complexes grafted onto SBA-15.[10]
Although, promising results were obtained, most of these
immobilized catalysts showed relatively low activity and stability.
Therefore, porous organic polymers have been considered as
support materials owing to large surface areas and tailor-made
chemical structures.[1c] E.g. Yang et al. immobilized RuCl3 onto a
Tröger’s base-derived polymer and achieved a TON of 2254 in
the presence of PPh3.[11] The group of Yoon reported [IrCp*Cl2]2
loaded onto 2,2’-bipyridine-based covalent triazine framework
(bpy-CTF) and a heptazine-based covalent organic framework,
achieving TONs of 5000 and 6400, respectively.[12] Recently, the
same group proposed CTF functionalized with Ir-NHC sites which
Introduction
To avert climate change and reduce CO2 emissions, novel
technologies for a CO2-neutral energy supply are intensively
studied.[1] Electrochemical water splitting is a promising option to
compensate fluctuations and overcapacities in the energy supply
from renewable sources, such as wind and solar energy.[2] H2
produced in this process can be converted into liquid fuels and
chemicals (e.g. formic acid (FA) and/or MeOH), using CO2 as an
abundant C1 building block. As such, a truly sustainable and
circular carbon economy is enabled (Scheme 1).[1a-c, 2c, 3] In this
regard, the direct hydrogenation of CO2 to FA has enjoyed
considerable attention and presents an attractive alternative to
the conventional carbonylation route.[2c, 4] Moret et al. reported
homogeneous [RuCl2(PTA)2] for the synthesis of FA under acidic
conditions, achieving a maximum TON of 475.[5] However, the
greatest challenge in the hydrogenation of CO2 is the high kinetic
and thermodynamic stability of CO2. In the presence of a base,
the equilibrium can be shifted towards the formation of formate
(FA-).[1c, 4a] Particularly, transition metal complexes based on Ir,
[a]
[b]
Anna Kann M.Sc., Dr. Peter J. C. Hausoul, Prof. Dr. Regina
Palkovits
Institut für Technische und Makromolekulare Chemie, RWTH
Aachen University, Worringerweg 2, 52074, Aachen, Germany
E-mail: Kann@itmc.rwth-aachen.de, Palkovits@itmc.rwth-aachen.de
Dr. Heinrich Hartmann, Dr. Astrid Besmehn
Zentralinstitut für Engineering, Elektronik und Analytik ZEA-3:
Analytik, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
Figure 1. Ligand structures of porous phosphine polymers (pTPP, pDPPE,
Figure 3. Solid-state MAS (a) 13C-CP and (b) 31P{1H} NMR spectra of the
pTPPB and pTPrP) and covalent triazine frameworks (CTF-b and bpy-CTF).
polymer pTPPB. Spinning side bands are indicated with *.
resulted in an enhanced TON of 24,300 after 15 h.[13] This shows
that the direct immobilization of molecular complexes onto
functionalized porous polymers leads to improved catalytic
activity and is a promising strategy for catalyst design.[14] The
polymeric analogs of PPh3 (pTPP) and dppe (pDPPE), already
showed high activity in the decomposition of aqueous base-free
FA.[15] So far, phosphine-based polymers have not been reported
for the hydrogenation of CO2 to FA-. Here, we present the
synthesis and characterization of new phosphine-polymers and
highlight the performance of Ir and Ru immobilized on various N-
and P-donor polymers for the direct hydrogenation of CO2 to FA-.
In addition, optimizations of reaction parameters for the best
catalytic system as well as recycling experiments are presented.
The coordination environment of the prepared and recycled
(i.e. polylitiation using n-BuLi, cross-linked with PCl3, and capped
with PhLi). The resulting polymer pTPPB presented similar
characteristics as pTPP and pDPPE (i.e., polydisperse particles
(Figure 2), XRD amorphous (Figure S1) and thermally stable up
to 340 °C (Figure S2)). N2 physisorption reveals a BET surface
area of 290 m2 g-1 (Figure S3), which is slightly higher compared
to pTPP (209 m2 g-1). The solid-state magic angle spinning (SS-
MAS) 13C-CP NMR spectrum (Figure 3a) presents peaks in the
aromatic range, which are assigned to the aromatic linker.[16a,17,18]
Besides this, small peaks are observed in the aliphatic range. The
SS-MAS 31P{1H} NMR spectrum (Figure 3b) shows one single
peak at -3.7 ppm, confirming the proposed structure. The
occurrence of aliphatic signals in the 13C NMR can be explained
by incorporation of n-BuLi into the polymer structure, leading to
PAr3 and P(n-Bu)Ar2 groups. To account for this, it is proposed
that n-BuLi co-precipitates with the lithiated linker and remains
bound during washing.[19]
catalyst was evaluated and
a mechanism for the CO2
hydrogenation to FA- was proposed.
The second polymer, pTPrP was prepared by activation of
1,6-dibromohexane with magnesium and subsequent cross-
linking with PCl3. During synthesis, a gel-like material precipitated
from solution. Any attempt to dry the gel resulted in extreme
shrinking. It is proposed that the high flexibility of the aliphatic
linker enables a complete collapse of the structure. This was
confirmed by N2 physisorption of the dried material, which showed
no measurable porosity.
Results and Discussion
Synthesis and Characterization of Phosphine Polymers
The polymers pTPP and pDPPE (Figure 1) were prepared as
described previously.[15-16] To study the effect of linker topology on
porosity and electronic structure, two additional polymers were
synthesized. The first, based on 1,3,5-tris(4-bromophenyl)-
benzene, was prepared in similar fashion as pTPP
Catalyst Preparation
A series of catalysts based on P- and N-donor containing
polymers loaded with Ru and Ir was screened to assess the
activity of the polymers in the aqueous hydrogenation of CO2 to
FA-. Catalysts were prepared by stirring the polymeric supports
(pTPP pDPPE, pTPPB, bpy-CTF,[12a] CTF-b,[20] Figure 1) with
either [Ru(p-cymene)Cl2]2 or [IrCp*Cl2]2 in methanol, followed by
filtration.[15] ICP-OES of the filtrate confirmed complete (> 99 %)
uptake in all cases. In the case of pTPrP, the purified gel was
impregnated with Ru prior to drying. Here, a metal uptake of only
30 % was confirmed by ICP-OES.
The prepared catalyst Ru@pDPPE was further characterized to
identify the coordination environment. SS-MAS 31P{1H} NMR of
the pure polymer pDPPE presents a peak at -10.6 ppm. In
comparison, the respective catalyst with a maximum loading of
Figure 2: SEM image of pTPPB.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
5000
4000
3000
2000
1000
0
C1s: 284.8 eV
Ru3d3/2: 285.1 eV
Ru3d5/2: 280.9 eV
290
288
286
284
282
280
278
276
Binding energy [eV]
Figure 4. Solid-state MAS 31P{1H} NMR spectra of the polymer pDPPE (blue)
and Ru@pDPPE with a maximum loading of 24 wt% Ru (P/Ru = 2). Spinning
side bands are indicated with *.
Figure 6. XPS spectra of the prepared Ru@pDPPE catalyst (10 wt% Ru).
Figure 5. Proposed bidentate coordination of Ru@pDPPE during the
immobilization of [Ru(p-cymene)Cl2]2 to pDPPE.
24 wt%, corresponding to a P/Ru ratio of 2, shows a complete
shift of this peak to 31.3 ppm (Figure 4). This suggests that all
phosphine centers are accessible for coordination and mainly
bidentate ruthenium complexes are formed, either with one dppe
unit or with two separate phosphines (Figure 5). The tendency of
[Ru(p-cymene)Cl2]2 to form coordinated phosphines and chelating
complexes with dppe in liquid phase was already described by
Chaplin et al..[21] In further X-ray photoelectron spectroscopy
(XPS) measurements, three peaks with binding energies of 284.8
eV (C1s), 285.1 (Ru3d3/2, overlapping with C1s) and 280.9 eV
(Ru3d5/2) were observed (Figure 6). This confirms Ru species with
an oxidation state of +2. In addition, TEM analysis confirmed the
absence of nanoparticle (Figure 7). In N2 physisorption
experiments, the reduction of the BET surface area from 51 to 45
m² g-1 was observed as a result of metal incorporation.
Figure 7. TEM image of the prepared Ru@pDPPE catalyst (225 kHz).
of 3170 was achieved. The highest TON of 7800 was observed
for Ru@pDPPE. For all Ru/phosphine catalysts no CO was
detected. In addition, various Ir-based catalysts were tested. In
the presence of Ir/C, the FA- yield is negligible (TON 30) and
55 ppm CO were detected. For [IrCp*Cl2]2, a TON of 140 was
obtained
without
CO
formation.,
CO [ppm]
Ru/C
267
335
0
135
0
0
0
0
55
0
[Ru(p-cymene)Cl2]2
[Ru(p-cymene)Cl(PPh3)]
Ru@CTF-b
Catalytic Hydrogenation of CO2 to Formate
Ru@pTPP
Ru@pTPPB
Ru@pTPrP[a]
Ru@pDPPE
Ir/C
[IrCp*Cl2]2
Ir@CTF-b
Ir@bpy-CTF
Ir@pDPPE
Catalyst screening reactions were performed using 100 bar
(CO2/H2 = 1:1) at 120 °C (Figure 8). A blank experiment as well
as a run with the polymer pDPPE showed no FA- formation. Ru/C
gave a TON of 780 and gas phase analysis revealed a large
quantity of CO (267 ppm), which is likely formed by dehydration
74
262
60
of FA. For the precatalyst [Ru(p-cymene)Cl2]2,
productivity (TON 2480) and a similar CO impurity was obtained.
For the respective monophosphine complex
[Ru(p-cymene)Cl(PPh3)], a TON of 1600 and no CO was
observed. [Ru(p-cymene)Cl2]2 immobilized on CTF-b
a higher
0
1000 2000 3000 4000 5000 6000 7000 8000 9000
TON
Figure 8. Catalyst screening for the hydrogenation of CO2 to formate.
Conditions: 50 mL Hastelloy autoclave, 10 mL water, 0.0026 mol% Ru, 1 wt%
Ru or Ir (except Ru@C 5 wt%), 120 °C, 100 bar CO2/H2 (1:1), 2.59 M K2CO3
(0.64 eq OH- w.r.t H2), 4 h. TON = molformate per molRu. [a] 0.0061 mol% Ru, 6.05
M K2CO3 (1.52 eq OH- w.r.t H2).
(Ru@CTF-b) performed comparably to Ru/C. In contrast,
Ru@pTPP and Ru@pTPPB are considerably more active, giving
TONs of 4390 and 4290, respectively. Using pTPrP as support
(0.0061 mol% Ru, 6.05 M K2CO3 (1.52 eq OH- w.r.t. H2)), a TON
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
12000
10000
8000
6000
4000
2000
0
12000
10000
8000
6000
4000
2000
0
Table 1. Influence of base on the hydrogenation of CO2 to formate.[a]
Entry
Base
Base
conc.
[M]
eq
TON[b]
Yield
w.r.t
H2 [%]
Yield
w.r.t
OH-
[%]
CO
[ppm]
OH-
w.r.t
H2
1
none
-
-
70
0.4
-
20
2
3.00
4.50
6.75
9.00
4.04
6.05
8.07
8.07
3.33
6.13
8.07
0.38
0.56
0.84
1.13
1.02
1.52
2.02
2.02
0.42
0.76
1.01
4440
6010
6390
530
27.2
36.9
39.2
3.2
73.3
66.1
46.8
2.9
109
177
48
40
60
80 100 120 140 160
Temperature [°C]
0
2
4
6
8
10 12 14 16 18
Time [h]
3
NaOH
4
Figure 9. Influence of reaction time (left, 120°C) and temperature (right, 4 h) on
the hydrogenation of CO2 to formate. Conditions: 50 mL Hastelloy autoclave,
10 mL water, 0.0026 mol% Ru, Ru@pDPPE (1 wt% Ru), 100 bar CO2/H2 (1:1),
2.59 M K2CO3 (0.64 eq OH- w.r.t. H2).
5
5
6
7750
9950
5390
30
47.5
61.0
33.1
0.1
47.5
40.7
16.6
0.1
109
195
33
7
K2CO3
experiments showed a poor reproducibility. A similar behavior
was observed for K2CO3, which enabled a maximum TON of 9950
at 6.05 M. A reaction performed using only H2 pressure (entry 9)
shows that the reaction does not proceed without gaseous CO2.
Therefore, CO2 is essential for the reaction. The highest TON of
10,060 was observed for KHCO3 at 8.07 M; however, also here
the high salt concentration leads to precipitation. Due to a
comparable TON at lower base concentration, K2CO3 was chosen
for further investigations. In all cases, the formation of CO can be
assigned to longer reaction times (24 h) in comparison to 4 h in
the catalyst screening.
8
9[c]
10
11
12
0
6370
7540
10,060
39.1
46.2
61.7
94.8
60.9
61.8
102
85
KHCO3
86
[a] Conditions: 50 mL Hastelloy autoclave, 10 mL water, 0.0061 mol% Ru,
Ru@pDPPE (1 wt% Ru), 120 °C, 100 bar CO2/H2 (1:1), 24 h. [b] TON =
molformate per molRu. [c] 50 bar H2.
Shorter reactions (4h) were carried out with varying K2CO3
concentrations and mol% Ru (at constant Base/Ru) to access the
intrinsic activity of the Ru centers. The data (Table 2) illustrate that
decreasing the base concentration from 6.05 to 2.59 M leads to
an increasing TON from 4300 to 7800, respectively. This suggests
an inhibitory action of the carbonate.
The time dependence was studied using a series of reactions
stopped after different intervals (Figure 9, left). After 30 min, a
TON of 1690 was obtained. Leaving the reaction longer resulted
in a gradual increase of the TON and after 8 h a TON of 9600 was
observed. After this, the reaction rate decreases and a TON of
12,040 (equaling 32% FA- yield w.r.t H2) was observed after 18 h.
This suggests a pressure dependence of the reaction. It is
noteworthy, that for these reactions no CO was detected.
The temperature dependence was studied in a range of 40 to
160 °C (Figure 9, right). The reaction already occurs at 40 °C,
reaching a TON of 120 after 4h. Between 80 °C and 160 °C, the
TON was significantly increased to a maximum of 11,450 at
160 °C. The CO formation was suppressed up to 140 °C,
followed by 172 ppm at elevated temperature.
The immobilization of this complex onto CTF-b and bpy-CTF led
to improved productivities, achieving TONs of 1910 and 1450 with
74 and 262 ppm CO respectively. A TON of 5000 was reported
for Ir@bpy-CTF in an aqueous NEt3 solution under comparable
reaction conditions (120 °C, 80 bar, 2 h).[12] For Ir@pDPPE, a
TON of only 130 was obtained. In summary, Ru supported on
phosphine polymers outperform Ir-based systems and prevent
CO formation. Since Ru@pDPPE gave the best results, it was
further investigated.
Firstly, the influence of base type (e.g. NaOH, KHCO3, K2CO3)
and concentration was studied in 24 h reactions (Table 1). For
NaOH, the TON increases with increasing base concentration
and exhibits a maximum TON of 6390 at 6.75 M, after which the
TON drops to 530 at 9.00 M. The high salt concentration leads to
precipitation and inclusion of the catalyst and repeated
Table 2. Influence of catalyst to substrate ratio for the Ru-catalyzed
hydrogenation of CO2 to formate.[a]
Entry
mol % Ru
Base
conc. [M]
TON[b]
Yield
w.r.t H2
[%]
Yield
w.r.t. OH-
[%]
CO
[ppm]
The pressure dependence was investigated between 40 and
100 bar at constant CO2/H2 ratio and 120 °C (Table 3). At a total
pressure of 40 bar, a TON of 4570 was reached. Increasing the
pressure to 100 bar led to a significant increase of the TON to
7800. Varying the CO2/H2 ratio from 2:1 to 1:2 also increased the
TON from 6000 to 11,250, respectively. This shows that the
reaction strongly dependent on the partial H2 pressure.
1
2
3
0.0061
0.0044
0.0026
6.05
4.33
2.59
4300
6000
7800
26.4
26.3
20.5
17.6
26.3
31.9
0
0
0
In addition, the catalyst loading was varied from 0.0006 mol% to
0.0049 mol% (Table 4, entries 1-4) at 100 bar (CO2/H2 1:1). At
[a] Conditions: 50 mL Hastelloy autoclave, 10 mL water, Ru@pDPPE (1 wt%),
120 °C, 100 bar CO2/H2 (1:1), K2CO3, 4 h. [b] TON = molformate per molRu
.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
Table 3. Pressure variation on the hydrogenation of CO2 to formate.[a]
Table 4. Influence of the catalyst loading and Ru loading on the hydrogenation
of CO2 to formate.[a]
Entry
pCO2 [bar]
pH2 [bar]
ptotal [bar]
TON[b]
Entry
mol % Ru
Ru loading
[wt %]
TON[b]
Yield
w.r.t H2
[%]
Yield
w.r.t.
OH- [%]
1
2
3
4
5
6
20
30
40
50
50
50
20
30
40
50
25
100
40
4570
5990
6790
7800
6000
11,250
60
1
2
3
4
5
6
0.0006
0.0012
0.0026
0.0049
0.0026
0.0026
1
5110
7590
7800
7540
13,170
4750
3.1
4.9
80
1
9.3
14.5
31.9
57.6
53.9
19.4
100
75
1
20.5
37.0
34.6
12.5
1
150
0.5
5
[a] Conditions: 50 mL Hastelloy autoclave, 10 mL water, 0.0026 mol% Ru,
Ru@pDPPE (1 wt%), 120 °C, 2.59 M K2CO3 (0.64 eq OH- w.r.t H2), 4 h. [b]
TON = molformate per molRu
.
[a] Conditions: 50 mL Hastelloy autoclave, 10 mL water, Ru@pDPPE, 120 °C,
100 bar CO2/H2 (1:1), 2.59 M K2CO3 (0.64 eq OH- w.r.t. H2), 4 h. [b] TON =
the lowest catalyst loading of 0.0006 mol%, a TON of 5110 was
obtained. For an increased loading of 0.0012 mol%, the TON
raised to 7590. Interestingly, further increasing the catalyst
loading did not affect the TON. This suggests that the reaction is
externally mass transport limited by H2 diffusion. Thus, the
diffusion of the gases is the rate limiting process of this
transformation. The Ru-loading on the polymer was also varied
from 0.5 wt% to 5 wt% at a catalyst loading of 0.0026 mol%
(Table 4, entries 3, 5-6). Starting from a metal loading of 0.5 wt%,
a maximum TON of 13,170 was obtained. The TON decreases to
7540 at 1 wt% and 4750 at 5 wt%, respectively. Since these
materials are microporous and all P-centers are available for
coordination, it is proposed that with increasing loading an
increasing proportion of metal centers is inactive due to internal
diffusion limitation, resulting in a lower average TON per metal
atom.
molformate per molRu
.
followed by a significant drop of the catalytic productivity in the
second run to a TON of 2300. In further cycles, the TON
decreased gradually, reaching a TON of 1730 in the eighth cycle.
ICP-OES analysis of the reaction solution revealed a catalyst
leaching below 0.06 % Ru up to the fifth run, followed by an
increase to around 4 % Ru in the last runs. In literature, the
recycled catalysts also showed a constant and pronouncing
decrease of catalytic activity, but without a significant drop at the
beginning.[12] Therefore, further characterization of the spent
catalysts was performed. Although, TEM analysis did reveal some
isolated Ru nanoparticles, it still resembles the freshly prepared
material (Figure 11). This illustrates that the Ru- complex remains
largely dispersed on the polymer matrix. XPS experiments
revealed no change in the oxidation state of ruthenium during the
hydrogenation reaction, as the same binding energies as the
freshly prepared catalyst were found (Figure 12). In addition, the
cleavage of chloride ligands was found in the survey spectra
(Figure S4). In SS-MAS 13C NMR, no changes of the polymer
structure of pDPPE were observed (Figure 13a).[15] After catalysis,
the broad peak around 99.4 ppm disappeared and the peaks in
Catalyst recycling was studied for Ru@pDPPE at 120 °C and
100 bar (CO2/H2 1:1) (Figure 10, Table S1). A solvent-change
procedure was applied in which the catalyst was filtered after each
cycle and recharged with the appropriate amount of fresh K2CO3
and water, maintaining a constant ratio of catalyst to base and
solvent. In the first cycle, a TON of 4530 was observed,
the
aliphatic
range
(38.4
–
5000
4000
3000
2000
1000
0
1
2
3
4
5
6
7
8
Recycling step
Figure 10. Solvent-change recycling for the Ru@pDPPE catalyzed
hydrogenation of CO2 to formate. Conditions: 50 mL Hastelloy autoclave,
0.0061 mol% Ru (Start), Ru@pDPPE (1 wt%), 120 °C, 100 bar CO2/H2 (1:1),
2.59 M K2CO3 (0.64 eq OH- w.r.t. H2), 4 h.
Figure 11. STEM images of the spent Ru@pDPPE catalyst after the recycling
experiments (160 kHz).
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
5000
4000
3000
2000
1000
0
C1s: 284.8 eV
2030 cm-1
1967 cm-1
recycled
catalyst
Ru3d3/2: 285.1 eV
Ru3d5/2: 280.9 eV
fresh
catalyst
290
288
286
284
282
280
278
276
pDPPE
Binding energy [eV]
Figure 12. XPS spectra of the Ru@pDPPE catalyst (10 wt% Ru) after the CO2
hydrogenation.
2500
2000
1500
1000
500
wavenumber [cm-1]
Figure 14. ATR-IR spectra of the pure polymer pDPPE, the freshly prepared
Ru@pDPPE catalyst (10 wt% Ru) and the Ru@pDPPE catalyst (10 wt% Ru)
after the CO2 hydrogenation.
27.1 ppm) decreased, suggesting the loss of the p-cymene ligand.
Comparing the 31P{1H} NMR spectra before and after the
hydrogenation reaction (Figure 13b), the major peak at -10.3 ppm
remained while a minor peak at 22.8 ppm appeared. In order to
identify the coordination environment of the complex after
catalysis, ATR-IR analysis was performed. Two broad bands at
1967 and 2030 cm-1 were observed additionally to the bands of
the freshly prepared catalyst (Figure 14). Most likely, these bands
can be assigned to a carbonyl ligand bond to Ru (1967 cm-1) and
Ru hydride species (2030 cm-1).[22] The formation of hydride
species was further confirmed by SS-MAS 1H NMR (Figure S5).
Based on this it is proposed that the steep drop in acitivty after the
first run is most likely caused by changes of the catalytically active
complex.
formula [Ru(PR3)2(H)2(CO)] (3) for the catalyst coordination after
the hydrogenation reaction (Figure 15). At the start of the reaction,
the chloride ligand is most probably exchanged by hydride ligands
(complex 2)[23], which are essential for the catalytic conversion of
CO2. During the consecutive reaction, the p-cymene ligand is
further substituted with CO to obtain complex 3. We propose that
complex 2 is more active in comparison to complex 3, as the
addition of CO was reported to act as poison and retarded the FA
formation.[23] Starting from the hydride complex (2 or 3), we
propose the mechanism described in Figure 15.[24-25] In the first
step, CO2 is inserted into the Ru(II)-H bond, forming the formate
complex 4. Subsequent oxidative addition of the dihydrogen
molecule and reductive elimination of formic acid regenerate the
active complex 3. Simultaneously, formic acid is deprotonated by
the corresponding base.
According to these analytical data, we propose the general
Conclusions
In summary, cross-linked polyphosphines were used as support
materials
for
the
immobilization
of
the
molecular
[Ru(p-cymene)Cl2]2 complex and applied in the catalytic
hydrogenation of CO2 to FA-. An initial catalyst screening
highlighted the excellent catalytic performance of Ru@pDPPE in
terms of TON and suppressing CO formation. A screening of
carbonate/OH bases revealed that K2CO3 gave the best results
Figure 13. Solid-state MAS (a) 13C-CP and (b) 31P{1H} NMR spectra of 10 wt%
Ru@pDPPE before (red) and after one single catalysis (blue). Spinning side
bands are indicated with *.
Figure 15. Proposed mechanism for the catalytic hydrogenation of CO2 to
formate.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
in terms of TON and that CO2 is necessary for the catalysis. The
CO formation was suppressed by short reaction times and the
reaction showed a strong dependence on the partial H2 pressure.
Variation of the temperature revealed increasing productivity up
to 140 °C followed by a steady TON in the presence of CO. Low
catalyst loadings are suggested due to external mass transport
limitations while gas diffusion is rate limiting. Finally, recycling
experiments were performed for eight cycles, presenting a
significant drop after the first one due to structural changes of the
metal complex and gradual decrease afterwards.
Avance III (4 mm MAS rotor, 14 kHz spinning at rt). The chemical
shifts δ are presented in parts per million (ppm) relative to
calibration standards. 31P{1H} spectra were recorded at 8 kHz and
12 kHz to identify spinning side-bands. Scanning transmission
electron microscopy (STEM) was performed on a FEI Tecnai F20
device with HAADF detector. The measurements were conducted
at the GFE, RWTH Aachen. X-ray photoelectron spectroscopy
(XPS) data was performed on
a Phi5000 VersaProbe II
spectrometer (ULVAC-Phi Inc., USA) using Al Kα as the excitation
source (1.486 keV) and operated at 15 kV. The binding energies
were referenced to the C1s line at 284.8 eV. A Vertex 70 with ATR
unit was used to obtain IR spectra. A measuring range of 4500 –
400 cm-1 was chosen. Gas chromatography (GC analysis) of the
gas phase was performed on an Agilent HP6890 system with a 2
m Shin carbon ST 100/120 mesh micropacked column and a TCD
detector. Helium was used as transport gas with a constant flow
rate of 24.6 mL min-1. The temperature was 35 – 150 °C with a
Experimental Section
General Information
All reactions were carried out under argon atmosphere using
standard schlenk techniques. n-Butyllithium (1.6 M in hexane),
4,4’-dibromobiphenyl (98 %), 1,6-dibromohexane (96 %),
dichloro(pentamethylcyclopentadienyl)iridium(III) dimer (96 %),
dichloro(p-cymene)ruthenium(II) dimer, magnesium (≥ 99 %),
potassium bicarbonate (≥ 99.95 %), potassium carbonate
(≥ 99.0 %), ruthenium on carbon (5 wt% Ru), sodium hydroxide
and 1,3,5-tris(4-bromophenyl)benzene (97 %) were purchased
from Sigma Aldrich. 1,2-Bis(dichlorophosphino)ethane (96 %),
Ir@C (1 wt% Ir) and phenyl lithium (1.8 M in dibutyl ether) were
obtained from Alfa Aesar. Phosphorus trichloride (> 99.0 %) was
obtained from Merck. The chemicals were used without further
purification. Anhydrous diethyl ether (99.8 %) was purchased
from AppliChem and anhydrous methanol (99.8 %) from
ChemSolute. Dry THF and n-pentane were obtained from a
solvent purification system. All solvents were degassed and
stored under argon atmosphere prior to use. CTF-b material was
synthesized according to Artz et al.,[20] bpy-CTF according to Park
et al.[12a] and pTPP according to the procedure described by
Hausoul et al.[15,16a]
Scanning electron microscopy (SEM) was performed on a Zeiss
DSM 982 Gemini with a field emission cathode. For energy
dispersive X-ray spectroscopy (EDX) of the surface, the system
Oxford Link ISIS with HPGe-detector and UT-Window was
applied. The measurements were conducted at the
Gemeinschaftslabor für Elektronenmikroskopie (GFE), RWTH
Aachen. Powder X-ray diffraction (XRD) pattern were recorded on
a Siemens D5000 instrument with a Cu Kα radiation (λ =
0.154 nm). The scanning range was set from 3° to 90° 2ϴ with
heating rate of
8
°C min-1. High performance liquid
chromatography (HPLC) analysis was performed on a Shimadzu-
2020 device equipped with an organic acid resin column (S-
chromatographie Service, Langerwehe) maintained at 40 °C. 2
mM trifluoroacetic acid was used as eluent. Inductively-coupled
plasma-optical emission spectroscopy (ICP-OES) was carried out
with Spectroblue. The filtered samples were prepared by
acidification with 20 % HCl solution in a 1:1 ratio.
Polymer Syntheses of pTPP, pDPPE and pTPPB
The syntheses of pTPP, pDPPE and pTPPB were performed
according to the procedure described by Hausoul et al.[15]
n-Butylithium (50 mL, 80.0 mmol, 1.6 M in hexane) was added
dropwise to a stirring solution of arylbromid-linker in diethyl ether
(150 mL) and left to stir for 1 h. The resulting suspension was
filtered over a glass filter and the isolated solid was washed with
diethyl ether (200 mL) followed by n-pentane (200 mL) for 30 min
each. After filtration, the solid was suspended in THF (120 mL)
and a solution of the chlorophosphine in THF (60 mL) was added
dropwise over a period of 1 h using a syringe pump. Subsequently,
phenyl lithium (10 mL, 19.0 mmol, 1.8 M in dibutyl ether) was
added and left to stir over night. The suspension was quenched
by adding methanol (100 mL). After stirring for 15 min, the
suspension was filtered over a glass filter. The isolated solid was
suspended in methanol (200 mL) and was stirred at 55 °C for 3 h.
The solid obtained after filtration was washed with diethyl ether
(200 mL). Subsequently, the suspension was filtered and the solid
was dried in vacuo at 60 °C.
0.02° intervals and
a collection time of 1 s per step.
pTPP: 4,4’-Dibromobiphenyl (6.24 g, 20.0 mmol) and
phosphorous trichloride (1.74 g, 12.7 mmol) resulted in 3.26 g
pTPP (yield 98 %) as pale yellow powder.
pDPPE: 4,4’-Dibromobiphenyl (6.24 g, 20.0 mmol) and 1,2-
bis(dichlorophosphino)ethane (2.26 g, 9.75 mmol) resulted in
3.78 g pDPPE (yield 97 %) as pale yellow powder.
pTPPB: 1,3,5-Tris(4-bromophenyl)benzene (7.22 g, 13.3 mmol)
and phosphorus trichloride (1.74 g, 12.7 mmol) resulted in 4.51 g
pTPPB (yield 97 %) as a pale yellow powder. MAS 13C-CP NMR
(176 MHz, MAS 14 kHz, 25 °C): δ = 149.4, 139.6, 135.4, 75.3,
33.0 ppm. 31P{1H} NMR (283 MHz, 14 kHz spinning, 25 °C): δ = -
2.6 ppm.
Thermogravimetric analyses (TG) and differential scanning
calorimetry (DSC) were carried out using a STA 409 cell from
Netzsch with a heating rate of 5 K min-1 in air up to 1273 K.
Nitrogen physisorption isotherms were measured at 77 K using a
ASAP2000 apparatur from Micrometrics. High purity gas
(nitrogen: 99.999 %) was used for the adsorption measurements.
Prior to the measurements, the samples were outgassed under
vacuum at 393 K for several hours. The BET model was applied
in a range of 0.05 ≤ p/p0 ≤ 0.2 in order to calculate the specific
surface areas. The external surface areas and micropore volumes
were determined using the t-plot method. Solid-state MAS 13C-CP,
31P{1H} and 1H NMR were measured on a 500 MHz BRUKER
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
degassed water (20 mL) was used to wash salts out of the
autoclave. The suspension was filtered over a glass filter and the
remaining catalyst was stirred in degassed water (20 mL) for 30
min. After filtration, the catalyst was dried in vacuo at 60 °C. The
reaction solutions were collected and transferred to a 100 mL
volumetric flask, filtered over a syringe filter and analyzed by
HPLC analysis. In addition, catalyst leaching was determined by
ICP-OES analysis. In the second run, the residual amount of
catalyst was applied and the appropriate amounts of K2CO3 and
water were added, maintaining the same ratio of catalyst to base
and solvent as in the first cycle. The procedure of the first run was
repeated. In total, eight cycles were performed and the catalyst
was analyzed by TEM analysis after the last run.
Polymer Synthesis of pTPrP
Magnesium (1.17 g, 48 mmol) was activated by heating six times
under vacuo. THF (200 mL) was added and the suspension was
cooled down to 0 °C. 1,6-Dibromohexane (3.1 mL, 20 mmol) was
added dropwise and the reaction solution was refluxed for two
days. After filtration over a glass filter, a solution of phosphorus
trichloride (1.37 g, 10 mmol) in THF was added dropwise to the
Grignard reagent and stirred for 3 h at rt. End-capping was
realized with n-butyllithium (11.3 mL, 18 mmol, 1.6 M in hexane)
and further stirring for 1 h. The product was purified by soxhlet
extraction in methanol (700 mL) for 5 days. The solid gel was
transferred in a schlenk flask and approximately half of the
suspension was used for catalyst impregnation, while the residual
solution was dried under vacuo at 60 °C, resulting in 0.1974 g
product as brown solid.
Acknowledgements
Catalyst Preparation
This work was supported in part by the Cluster of Excellence
“Tailor-Made Fuels from Biomass”, which is funded by the
Excellence Initiative by the German Federal and State
Governments to promote science and research at German
universities, and by the German Federal Ministry of Education
and Research (BMBF) within the Kopernikus Project P2X:
Flexible use of renewable resources – exploration, validation
and implementation of ‘Power-to-X’ concepts. A.K. thanks the
German Federal Environmental Foundation for financial support.
The catalyst immobilization of the metal precursor on support
materials was performed by incipient wet impregnation method in
methanol. The polymers (e.g. 1000 mg) were combined with the
appropriate amounts of dichloro(p-cymene)ruthenium(II) dimer
(e.g. 30.3 mg for
1 wt% Ru) or dichloro(pentamethyl-
cyclopentadienyl)iridium(III) dimer (20.5 mg for 1 wt% Ir) and
were dissolved in methanol (100 mL). The suspension was stirred
for three days at 60 °C. The catalysts were filtered over a glass
filter followed by drying in vacuo at 60 °C. The filtrate was
concentrated in vacuo and transferred into a volumetric flask
(10 mL). The solution was analyzed by ICP-OES to determine the
metal uptake.
Keywords: CO2 activation • immobilization • polymer •formic
acid • phosphines
In the case of the polymer pTPrP, the suspension prior to drying
was divided into two parts. One part was dried as described
previously, while the other one was combined with an appropriate
amount of dichloro(p-cymene)ruthenium(II) dimer. Subsequently,
the catalyst preparation procedure was followed as described
above.
[1]
[2]
a) Q. Liu, L. Wu, R. Jackstell, M. Beller, Nat. Commun. 2015, 6, 5933–
5974. b) C. Federsel, R. Jackstell, M. Beller, Angew. Chem. Int. Ed. 2010,
49, 6254–6257. c) G. H. Gunasekar, K. Park, K.-D. Jung, S. Yoon, Inorg.
Chem. Front. 2016, 3, 882–895. d) W. Wang, S. Wang, X. Ma, J. Gong,
Chem. Soc. Rev. 2011, 40, 3703–3727. e) S. Saeidi, N. A. S. Amin, M.
R. Rahimpour, J. CO2 Util. 2014, 5, 66–81.
a) M. Carmo, D. L. Fritz, J. Mergel, D. Stolten, Int. J. Hydrogen Energy
2013, 38, 4901–4934. b) M. Wang, Z. Wang, X. Gong, Z. Guo, Renew.
Sust. Energ. Rev. 2014, 29, 573–588. c) S. Enthaler, J. von Langermann,
T. Schmidt, Energy Environ. Sci. 2010, 3, 1207–1217.
Catalytic Applications
All hydrogenation reactions were performed in 50 mL Hastelloy
autoclaves with cross-stirring bar. For each reaction, the
appropriate amount of catalyst and base were dissolved in
degassed water (10 mL). The autoclave was flushed five times
with 20 bar CO2 followed by pressurizing with the desired
pressure of CO2 and subsequently H2 at ambient temperature.
The autoclaves were placed in a preheated aluminium cone at a
stirring speed of 750 rpm. After the reaction time, the autoclaves
were cooled down in an ice bath. The gas phase was collected in
gas bags for GC analysis. The reaction solution was diluted to 50
mL in a volumetric flask and filtered with a syringe filter. The
sample was analyzed by HPLC.
For recycling experiments, a solvent-change procedure was
applied. For the first run, 50 mg of the catalyst Ru@pDPPE
(1 wt% Ru) and 3.58 g K2CO3 were dissolved in 10 mL of
degassed water. The reaction was carried out as described before
at 100 bar (CO2/H2 1:1) and 120 °C. After cooling the autoclave in
an ice bath, the gas phase was analyzed by GC analysis. The
reaction solution was transferred into a Schlenk flask and
[3]
[4]
a) S. Wesselbaum, U. Hintermair, W. Leitner, Angew. Chem. Int. Ed.
2012, 51, 8585–8588. b) B. Loges, A. Boddien, F. Gärtner, H. Junge, M.
Beller, Top. Catal. 2010, 53, 902–914.
a) J. Klankermayer, S. Wesselbaum, K. Beydoun, W. Leitner, Angew.
Chem. Int. Ed. 2016, 55, 7296–7343. b) F. Joó, ChemSusChem 2008, 1,
805–808.
[5]
[6]
S. Moret, P. J. Dyson, G. Laurenczy, Nat. Commun. 2014, 5, 4017–4023.
a) R. Tanaka, M. Yamashita, K. Nozaki, J. Am. Chem. Soc. 2009, 131,
14168–14169. b) G. A. Filonenko, R. van Putten, E. N. Schulpen, E. J.
M. Hensen, E. A. Pidko, ChemCatChem 2014, 6, 1526–1530.
a) P. R. Upadhyay, V. Srivastava, RSC Adv. 2016, 6, 42297–42306. b)
L. T. M. Nguyen, H. Park, M. Banu, J. Y. Kim, D. H. Youn, G. Magesh,
W. Y. Kim, J. S. Lee, RSC Adv. 2015, 5, 105560–105566. c) N. Liu, R.
Du, W. Li, Adv. Mat. Res. 2013, 821, 1330–1335. d) D. Preti, C. Resta,
S. Squarcialupi, G. Fachinetti, Angew. Chem. Int. Ed. 2011, 50, 12551–
12554. e) C. Hao, S. Wang, M. Li, L. Kang, X. Ma, Catal. today 2011,
160, 184–190. f) M. W. Farlow, H. Adkins, J. Am. Chem. Soc. 1935, 57,
2222–2223.
[7]
[8]
G. A. Filonenko, W. L. Vrijburg, E. J. M. Hensen, E. A. Pidko, J. Catal.,
2016, 343, 97–105.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
[9]
a) Y. Zhang, J. Fei, Y. Yu, X. Zheng, Catal. Commun. 2004, 5, 643–646.
b) A. V. Bavykina, E. Rozhko, M. G. Goesten, T. Wezendonk, B. Seoane,
F. Kapteijn, M. Makkee, J. Gascon, ChemCatChem 2016, 8, 2217–2221.
c) Y. M. Yu, J. H. Fei, Y. P. Zhang, X. M. Zheng, Chin. J. Chem. 2006,
24, 840–844. d) N. D. McNamara, J. C. Hicks, ChemSusChem 2014, 7,
1114–1124. e) J. H. Lee, J. Ryu, J. Y. Kim, S.-W. Nam, J. H. Han, T.-H.
Lim, S. Gautam, K. H. Chae, C. W. Yoon, J. Mater. Chem. A 2014, 2,
9490–9495.
[10] Z. Xu, N. D. McNamara, G. T. Neumann, W. F. Schneider, J. C. Hicks,
ChemCatChem 2013, 5, 1769–1771.
[11] Z.-Z. Yang, H. Zhang, B. Yu, Y. Zhao, G. Ji, Z. Liu, Chem. Commun.
2015, 51, 1271–1274.
[12] a) K. Park, G. H. Gunasekar, N. Prakash, K. D. Jung, S. Yoon,
ChemSusChem 2015, 8, 3410–3413. b) G. G. Hariyanandam, D. Hyun,
P. Natarajan, K.-D. Jung, S. Yoon, Catal. today 2016, 265, 52–55.
[13] G.H. Gunasekar, K. Park, V. Ganesan, K. Lee, N.-K. Kim, K.-D. Jung, S.
Yoon, Chem. Mat. 2017, 29, 6740–6748.
[14] a) Q. Sun, Z. Dai, X.Liu, N. Sheng, F. Deng, X. Meng, F.-S. Xiao, J. Am.
Chem. Soc. 2015, 137, 5204–5209. b) Q. Sun, M. Jiang, Z. Sun, Y. Jin,
S. Pan, L. Wang, X. Meng, W. Chen, Y. Ding, J. Li, F.-S. Xiao, Chem.
Commun. 2014, 50, 11844–11847.
[15] P. J. C. Hausoul, C. Broicher, R. Vegliante, C. Göb, R. Palkovits, Angew.
Chem. Int. Ed. 2016, 55, 5597–5601.
[16] a) P. J. C. Hausoul, T. M. Eggenhuisen, D. Nand, M. Baldus, B. M.
Weckhuysen, R. J. K. Klein Gebbink, P. C. A. Bruijnincx, Catal. Sci.
Technol. 2013, 3, 2571–2579. b) J. Fritsch, F. Drache, G. Nickerl, W.
Böhlmann, S. Kaskel, Microporous Mesoporous Mater. 2013, 172, 167–
173.
[17] M. Stol, D.J.M. Snelders, H. Kooijman, A.L. Spek, G.P.M. van Klink, G.
van Koten, Dalton Trans. 2007, 24, 2589–2593.
[18] I.O. Koshevoy, L. Koskinen, E.S. Smirnova, M. Haukka, T.A. Pakkanen,
A.S. Melnikov, S.P. Tunik, Z. Anorg. Allg. Chem. 2010, 636, 795–802.
[19] M. Wander, P. J. C. Hausoul, L. A. J. M. Sliedregt, B. J. van Steen, G.
van Koten, R. J. M. Klein Gebbink, Organometallics 2009, 28, 4406–
4415.
[20] J. Artz, S. Mallmann, R. Palkovits, ChemSusChem 2015, 8, 672–679.
[21] A.B. Chaplin, C. Fellay, G. Laurenczy, P.J. Dyson, Organometallics 2007,
26, 586–593.
[22] W. Yang, H. Cheng, B. Zhang, Y. Li, T. Liu, M. Lan, Y. Yu, C. Zhang, W.
Lin, S. Fujita, M. Arai, F. Zhao, Green Chem. 2016, 18, 3370–3377.
[23] P.G. Jessop, T. Ikariya, R. Noyori, Nature 1994, 368, 231–233.
[24] J.Z. Zhang, Z. Li, H. Wang, C. Y. Wang, J. Mol. Catal. A:Chem. 1996,
112, 9–14.
[25] a) Y. Ohnishi, Y. Nakao, H.Sato, S. Sakaki, Organometallics 2006, 25,
3352–3363. b) Y. Musashi, S. Sakaki, J.Am.Chem.Soc. 2000, 122,
3867–3877. c) A. Urakawa, F. Jutz, G. Laurenczy, A. Baiker, Chem. Eur.
J. 2007, 13, 3886–3899.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201800413
ChemSusChem
FULL PAPER
COMMUNICATIONS
Heterogeneous catalysts based on the
immobilization of molecular Ru
complexes on phosphine polymers are
reported for the catalytic hydrogenation
of CO2 to formate. Excellent
productivities and selectivities were
Anna Kann, Heinrich Hartmann, Astrid
Besmehn, Peter J. C. Hausoul*, Regina
Palkovits*
O
H
C
H
O
Page No. – Page No.
Hydrogenation of CO2 to Formate over
Ru Immobilized on Solid Molecular
Phosphines
obtained
under
mild
reaction
O
C
conditions.
Ru2+
H
O
K2CO3
This article is protected by copyright. All rights reserved.