Synthesis of N,N-dimethylformamide from carbon dioxide in aqueous biphasic solvent systems

Article abstract of DOI:10.1016/j.apcata.2017.03.037

This paper presents the homogeneous catalyzed hydrogenation of carbon dioxide to N,N-dimethylformamide (DMF) with an in-situ generated ruthenium catalyst based on RuCl₃?×?H₂O and the phosphine ligand 2,2′-bis(diphenylphosphinomethyl)-1,1′-biphenyl (BISBI). Investigations showed that the complex formation of an active species requires the presence of an amine. The catalyst was recycled by immobilization in a nonpolar alcoholic solvent while the formed product was extracted in-situ into the aqueous phase. The self-assembling reaction system showed stability for 10 recycling runs without a significant loss of activity resulting in an average yield of 31% DMF at 40?bar and 140?°C without the occurrence of any byproducts. Furthermore, a combination of the developed reaction system with ternary amines enables the application of wash amine solutions as carbon dioxide carrier.

Full text of DOI:10.1016/j.apcata.2017.03.037

Accepted Manuscript
Title: Synthesis of N,N-dimethylformamide from carbon
dioxide in aqueous biphasic solvent systems
Authors: R. Kuhlmann, S. Schmitz, K. Haßmann, A. Pru¨llage,
A. Behr
PII:
S0926-860X(17)30146-1
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.03.037
APCATA 16196
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
2-2-2017
24-3-2017
27-3-2017
Please cite this article as: R.Kuhlmann, S.Schmitz, K.Haßmann, A.Pru¨llage, A.Behr,
Synthesis of N,N-dimethylformamide from carbon dioxide in aqueous biphasic solvent
systems, Applied Catalysis A, Generalhttp://dx.doi.org/10.1016/j.apcata.2017.03.037
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Synthesis of N,N-dimethylformamide from carbon dioxide in
aqueous biphasic solvent systems
R. Kuhlmanna, S. Schmitza, K. Haßmanna, A. Prüllagea, A. Behra*
a Department of Biochemical and Chemical Engineering, Chair of Technical Chemistry, Technische
Universität Dortmund, Emil-Figgestr. 66, 44227 Dortmund, Germany
e-mail: arno.behr@bci.tu-dortmund.de
Graphical Abstract
<InlineImage1>
Highlights




CO2 is hydrogenated to formate and condensed with an amine to yield formamides
The immobilization of a homogeneous catalyst in nonpolar solvents is investigated
Catalyst complex formation and product separation are performed in-situ
Several recycling runs demonstrate the concept and the catalyst stability towards air
Abstract
This paper presents the homogeneous catalyzed hydrogenation of carbon dioxide to N,N-
dimethylformamide (DMF) with an in-situ generated ruthenium catalyst based on RuCl3 x H2O and
the phosphine ligand 2,2’-bis(diphenylphosphinomethyl)-1,1’-biphenyl (BISBI). Investigations showed
that the complex formation of an active species requires the presence of an amine. The catalyst was
recycled by immobilization in a nonpolar alcoholic solvent while the formed product was extracted
in-situ into the aqueous phase. The self-assembling reaction system showed stability for 10 recycling
runs without a significant loss of activity resulting in an average yield of 31% DMF at 40 bar and 140
°C without the occurrence of any byproducts. Furthermore, a combination of the developed reaction
system with ternary amines enables the application of wash amine solutions as carbon dioxide
carrier.
Keywords: Carbon dioxide; hydrogenation; formamide; homogeneous catalysis; catalyst recycling
1. Introduction

The utilization of carbon dioxide as universal C1-building block is of high ecological and economical
interest and investigated for several decades [1–6]. Especially the hydrogenation of carbon dioxide
was researched in detail since it allows to broaden the product diversity due to the possible change
of the carbon atom’s oxidation state [7–10]. Interesting hydrogenation products are formic acid or
formic acid derivatives like formamides. Especially N,N-dimethylformamide (DMF) can be used as
intermediate, e.g. for the synthesis of isocyanates, or directly as excellent solvent [11–14]. The first
synthesis of N,N-dimethylformamide by carbon dioxide hydrogenation was performed in 1970 by
Haynes et al. with different metal complexes [15]. The hydrogenation is reported to accompany with
a formate salt as intermediate (Figure 1) [16,17]. Especially, the condensation of formic acid salts
requires higher pressures or temperatures to yield formamides with better turnover numbers [18].
Nevertheless, recent developments of complex tailor-made catalysts lead to high activities at milder
reaction conditions and show that even non precious metals like iron and cobalt can be applied
[19,20].
The industrial formamide synthesis consists of the alcoholate catalyzed carbonylation of amines with
carbon monoxide. In order to establish a competitive carbon dioxide based process, a high recycling
efficiency of the more expensive hydrogenation catalyst is indispensable for a technical application.
However, the long-time stability of most developed homogeneous catalyst systems has not been
proven yet. Only Ding et al. recently reported a recycling concept for a homogeneous catalyst where
the product DMF is vacuum distilled out of the ruthenium catalyst solution resulting in 12 batch
recycling runs without a significant loss in activity but limiting this concept to low boiling products
[21]. The separation of catalyst and synthesized formamide is at all a difficult task because
formamides are generally excellent solvents for a broad variation of components [11].
An attractive method for a homogeneous catalyst recycling is the immobilization of the dissolved
catalyst in a liquid solvent since the catalyst stays in an active environment [22]. For an efficient
immobilization, the catalyst should be insoluble in the product phase while the catalyst-solvent is not
soluble in the product phase. A general approach to this problem is the application of a catalyst and
solvent that have a maximum polarity difference to the formed product so that a second phase is
automatically formed. However, an insolubility of the substrate inside the catalytic liquid phase leads
to mass transfer problems so that a compromise between reactivity and recyclability has to be made
in a liquid liquid two phase system. Nevertheless, due to its efficiency this recycling concept has
already found applications in industrial processes like the Ruhrchemie-Rhône-Poulenc process and
the SHOP-process where the products form a separate phase [23,24]. In these applications the
catalysts are immobilized in polar solvents, since products like aldehydes or α-olefins are rather
nonpolar.
Based on this idea we developed a liquid-liquid biphasic reaction system for the synthesis of N,N-
dimethylformamide as model formamide product. Unlike in the industrial examples, the catalyst has
to be immobilized in a nonpolar solvent as many formamides are polar aprotic solvents [11]. In order
to guarantee a sufficient catalyst separation, water is used as product extraction agent since it is
formed as byproduct in the condensation step anyway and has to be removed. The general
application of this concept should be verified by the utilization of an in-situ generated catalyst system
consisting of the polar precursor ruthenium(III)chloride hydrate and
2,2’-bis(diphenylphosphinomethyl)-1,1’-biphenyl (BISBI) as commercial nonpolar phosphine ligand.
Desired is the formation of a catalyst complex [Ru] that stays in the organic solvent phase (step I in
Figure 2) and catalyzes the hydrogenation reaction selectively (step II, Figure 2).
2. Experimental

2.1. Materials
All applied solvents and the dimethylamine solution (40 wt.% in water) were purchased from Acros
Organics in standard purity. Ruthenium(III)chloride monohydrate (99.9% on metal basis, 38-40 wt.%
ruthenium) and BISBI (99%) were purchased from ABCR. All chemicals were used as delivered
without any further purification.
2.2. Catalytic experiments and sampling
At first, the desired amount of ruthenium precursor and ligand were weighted on a precision scale
and transferred into the reaction vessel. The solvent and subsequently the aqueous amine-solution
were added directly into the autoclave with a precision of 0.01 g. Afterwards, the reactor was sealed
and the gas phase was flushed with argon. The reactor was pressurized with the desired hydrogen
and carbon dioxide pressures and the reaction was started by heating to the desired temperature. In
case of a perturbation experiment, two hours of preforming duration were performed, before the
actual reaction was started with all components.
After the desired reaction time the reactor was cooled to ambient temperature and degassed, if not
other stated. Subsequently, the biphasic reaction mixture was split in a tempered separating funnel
and each phase was weighted and analyzed. In the case of a recycling experiment, the organic phase
was returned to the reactor without any sampling of this phase.
Two different reactor types were used which were made out of stainless steel 316 L. The first reactor
type contained a volume of 22 ml and is stirred via a magnetic bar while the second reactor,
purchased from Parr Instrument GmbH, held a total volume of 300 ml and was stirred via a
magnetically coupled stirrer motor. Both reactor types were electrically heated with a temperature
deviation below 2 K.
2.3. Analysis
The amount of synthesized DMF was measured gas chromatographically (Agilent 6890) by a thermal
conductivity detector. The GC was equipped with an Innowax column (30 m, 250 µm, 0.25 µm) and
dimethyl sulfoxide was used as internal standard. The turn-over-number (TONi) of carbon dioxide to
a component i and the degree of extraction were calculated with the following equations (1) and (2).
n
푖
TONi=
(1)
(2)
n
ruthenium
Ei(%) = n
i,aqueous phase
n
i,total
The measurement of formic acid was performed by HPLC (Hitachi Lachrom Elite) with an UV/Vis
detector at 207 nm. Acidic water (pH 2.2 adjusted by H3PO4) and acetonitrile at a volume ratio of
98:2 were used as mobile phase with a total flow rate of 1.5 ml/min over a Zorbax XDB-C8 stationary
phase (150 mm, 4.6 mm, 5 µm) with propionic acid as internal standard. The column temperature
was set to 413 K. The amount of ruthenium and phosphorous compounds were measured with ICP-
AES (Iris Intrepid from Thermo Elemental) and according standard solutions as reference. The
leaching of a catalyst component was calculated via equation (3).
n
Ru,aqueous phase
( )
Ru-Loss % =
(3)
n
Ru,applied
3. Results and Discussion

3.1. Design of a biphasic solvent system
In homogeneous catalysis, the first step for the development of an efficient catalyst immobilization in
a biphasic solvent system is the identification of suitable solvents. The investigated polar substrate
dimethylamine (DMA) is already applied as 40 wt.% aqueous solution so that a suitable nonpolar
solvent for the catalyst immobilization is necessary. In order to balance out a sufficient solubility of
the substrate dimethylamine in the catalytic phase and a good recyclability of the ruthenium catalyst,
several nonpolar solvents with different water solubility were investigated in the ruthenium
catalyzed hydrogenation reaction (Table 1).
Table 1: Variation of the nonpolar solvent
Entry Solvent
Water solubility (g/l) XDMA (%) YDMF (%) EDMF (%) Ru-Loss (wt.%)
1
2
3
4
5
6
Octane
Anisole
Di-n-butylether
Toluene
2-Ethylhexan-1-ol 0.6
Benzyl alcohol 39
0.001
0.14
0.3
5
9
9
10
22
21
5
9
9
10
22
21
98
64
94
78
55
21
10
15
86
19
9
0.52
3
Reaction conditions: Vreactor = 22 ml, Vliquid = 5 ml, TR = 120 °C, cRu. = 2 mmol/l cBISBI = 2 mmol/l, cDMA = 3.9 mol/l, pCO /H .
=
2
2
20/20 bar, tR = 16 h, VDMA/H O : Vorganic solvent = 1:1, U = 500 1/min
2
The screening revealed that very water insoluble compounds, as aliphatic hydrocarbons are not
suitable as catalyst solvents because they do not dissolve the substrate and additionally lead to a
precipitation of ruthenium. A minimal water solubility increases the yield of DMF significantly.
Especially alcohols form a stable catalytic phase with the highest activities. A reason for the higher
activity might be that the alcohol acts as reducing agent for the ruthenium(III)chloride hydrate to
form an active ruthenium(II) species. General procedures for ex-situ syntheses of ruthenium
complexes with RuCl3 x H2O utilize alcohols like ethanol as reducing agent whereby the corresponding
aldehydes are formed [25,26]. Traces of aldehydes were not observed since the ruthenium species is
highly active for the aldehyde reduction to alcohols (see supporting info for details).
A higher water solubility of benzyl alcohol does not benefit the reaction anymore and leads to an
unsatisfactory product distribution where 79 wt.% of the formed DMF stays in the catalytic phase. In
conclusion, benzyl alcohol does not fulfill the separation between catalyst and product anymore even
though the catalyst leaching is with 3 wt.% of ruthenium minimal. That is why 2-ethylhexan-1-ol
delivers the best combination of reactivity, extraction efficiency and catalyst leaching.
3.2. Investigation of the in-situ catalyst complex formation
Further investigations with 2-ethylhexan-1-ol (EH) as solvent were made regarding the in-situ
generated catalyst system in order to understand and optimize the catalyst formation. Literature
known synthesized catalyst complexes start of a pre-coordinated ruthenium-phosphine ligand
complex which can contain different ligand to precursor ratios [8,17,27]. In our case both catalyst
components are initially located in different solvent phases. The applied hydrophilic ruthenium(III)-
precursor RuCl3 x H2O has to be transferred into the nonpolar solvent. A ruthenium to ligand
screening was performed in order to find an optimum ratio of both components (Figure 3).
Reaction conditions: Precursor: RuCl3∙xH2O, Ligand: BISBI, Vreactor = 22 ml, Vliquid = 5 ml, TR = 120 °C, cDMA = 3.8 mol/, pCO /pH2.
=
2
20/20 bar, tR = 16 h, VDMA-solution:VEH = 1:1, U = 500 1/min

The results of these experiments show that a small excess over a 1:1 ratio of either precursor or
ligand (yellow bars) leads to complexes with the highest activities. This indicates that the active
species only requires about one mole bidentate phosphine ligand per mole ruthenium, which
matches investigations of Li et al. [28]. The small excess of one component could be necessary to
fully utilize the amount of the other component. A bigger excess of ruthenium or ligand can lead to
an enhanced formation of inactive or less active catalytic species. While a higher excess of ligand
might shield the metal center from the substrate, an excess of ruthenium could lead to a formation
of multinuclear ruthenium complexes (Figure 4) [29]. The formation of ruthenium clusters might
occur since the initially applied ruthenium trichloride hydrate usually comes as a chloro-bridged
crystal [30,31] that has to be broken down into single ruthenium complexes. Thus, much higher
concentrations of ruthenium could lead to an increasing probability of the formation of clusters.
The application of a ruthenium : ligand ratio of 1.25:1 does not lead to a higher leaching since more
than 90 wt.% of the ruthenium are transferred into the nonpolar alcoholic phase and additionally
result in highest activities.
Further investigations were made with a small ruthenium excess referring to the complex formation.
Perturbation experiments were performed in which one or more components were missing for 2 h
under reaction conditions (preforming conditions). After this duration, the reaction was induced by
adding the remaining components. This should demonstrate which other components, except
precursor and ligand, take a role in the complex formation, as the formation of unreactive catalyst
complexes should be minimized.
Table 2: Results of the perturbation experiments with additional 2 h of preforming duration
Preforming conditions
Entry
XDMA (%)
YDMF (%)
TONDMF
TONFormate
Amine
CO2
H2
1
2
3
4
5
no preforming
14
1
14
1
218
25
192
< 10
141
223
< 10
-
+
+
-
+
-
+
+
-
9
9
144
254
21
+
+
17
1
17
1
-
Reaction conditions: Precursor: RuCl3∙xH2O, Ligand: BISBI, cKat. = 2.5 mmol/L, cLig. = 2.0 mmol/L, pCO /pH2.= 20/20 bar, TR =
2
120 °C, cDMA = 3.8 mol/L, mDMA-solution:mEH = 1:1, Vreactor = 300 mL, Vliquid = 75 mL, U = 700 1/min, tpreforming = 2 h ,tR = 5 h
The perturbation experiments show that the amine plays a decisive role in the formation of the
catalytic species. The initial absence of amine leads to a formation of black ruthenium precipitates
which are inactive for the hydrogenation reaction and could not be reconverted into a catalytic form

even though amine was added after the preforming duration (Table 2, Entry 2+5). Most likely is the
reduction of the ruthenium chloride with the alcohols to ruthenium nanoparticles (Figure 5, pathway
I) [32]. Since the amine can work as nitrogen-ligand, it could lead to a stabilization of ruthenium
intermediates that are afterwards converted to the active species with BISBI (Figure 5, pathway II).
Moreover, carbon dioxide has a minor positive effect on the complex formation whereas the absence
of hydrogen has no effect at all. Due to the negligible increase in product yield with regard to the
added preforming duration, further investigations were made without an additional preforming.
3.3.
Optimization of the reaction system
In order to increase the activity of the reaction system, the reaction conditions were varied. A closer
look on the product composition reveals that a significant amount of the intermediate formate is still
in the solution (Table 2, Entries 1+4). This indicates that the condensation step for the amide-
formation is slower under the reaction conditions of 120 °C. A temperature screening confirmed that
the condensation of the formate salt benefits from higher temperatures (Figure 6).
Reaction conditions: Precursor: RuCl3∙xH2O, Ligand: BISBI, cRu = 2.5 mmol/L, cBISBI = 2.0 mmol/L, pCO /pH2.= 20/20
2
bar, cDMA = 3.8 mol/L, mDMA-solution:mEH = 1:1, Vreactor = 300 mL, Vliquid = 75 mL, U = 700 1/min, tR = 5 h
Especially an increase of the temperature to 140 °C leads to a much faster formation of DMF whereas
a lower temperature primary results in the intermediate. The ruthenium leaching was with 3 wt.%
lower than in previous investigations. Higher reaction temperatures might convert more of the
ruthenium trichloride into a nonpolar catalytic species. Noteworthy is, that the catalyst is stable up to
160 °C for at least one reaction sequence. Only at 180 °C, a white precipitate was observed which
indicates a ligand decomposition. The yield time plots in Fig. 6 show that a maximum yield of 30 –
35% can be achieved under these conditions before the reaction rate turns towards zero. A high
initial consumption of reaction gas could be observed since the pressure decreased to 30 bar in the
first hour. Several pressure variations show a linear dependency between applied pressure and
product yield so that a pressure reduction is not reasonable (see supporting information for details).
3.4. Recycling of the homogeneous catalyst
After we gained a better understanding of the self-assembling catalyst system and optimized
reaction conditions resulting in a yield of 32% after 5 h, recycling experiments were performed with a
reaction temperature of 140 °C. After each batch reaction, the nonpolar alcoholic phase was
separated from the aqueous product phase and reused without any further addition of precursor,
ligand or solvent. Only aqueous amine solution was added after each batch run.
Reaction conditions: Precursor: RuCl3∙xH2O, Ligand: BISBI, cRu = 2.5 mmol/L, cBISBI = 2.0 mmol/L, pCO /pH2.= 20/20 bar, cDMA
=
2
3.8 mol/L, mDMA-solution:m2-Ethylhexan-1-ol = 1:1, Vreactor = 300 mL, Vliquid = 75 mL, U = 700 1/min, TR = 140 °C, tR = 5 h, tseparation = 1
h, Tseparation = 10 °C *) Tseparation = 25 °C
An analysis of the aqueous liquid revealed a stability of the reaction system of at least 10 runs with
an average yield of 31% DMF (Figure 7). An optimized separation temperature of 10 °C lead to an
enhanced product extraction degree of 65%. The loss of catalyst could be reduced to about 1.5 wt.%
in the 3rd-5th run. However, an increased separation temperature of 25 °C in the last 5 runs shows

that especially the ruthenium leaching is temperature dependent so that a low separation
temperature should be favored. Noteworthy, the evolved catalyst system is stable towards oxygen
since no extra precautions towards air were made during the separation steps. The amount of lost
alcohol is with less than 0.3 wt.% insignificantly low. After ten runs only 1.6 wt.% of 2-ethylhexan-1-ol
were lost while a total turnover number of 4942 could be achieved.
3.5. Application of ternary amines as carbon capture components
The developed biphasic reaction system allows further process intensification. In industry, the added
gaseous carbon dioxide is usually captured by an absorption process in a gas scrubber [33]. By
applying aqueous wash amine solutions, e.g. in coal-fired power plants, CO2 can be absorbed out of
flue gas in sufficient purities [34,35]. A combination of carbon dioxide capture and reaction would
save the energy, which is usually required for the wash amine regeneration. Since primary and
secondary amines can be converted into formamides as well, some ternary amines with different
solubility and functionality were investigated for compatibility.
Table 3: Reaction results with an addition of a ternary amine
Entry Added ternary Amine
Water solubility (g/l)
YDMF (%) Ru-Loss (wt.%) P-Loss (wt.%)
1
2
3
4
none
-
32
31
34
36
3
5
4
5
3
5
7
6
2-Dibutylaminoethanol
N,N-Dimethylbutylamine
2-Dimethylaminoethanol
11.6
37
limitless
Reaction conditions: Precursor: RuCl3∙xH2O, Ligand: BISBI, cRu = 2.5 mmol/L, cBISBI = 2.0 mmol/L, pCO /pH2.= 20/20 bar, cDMA
=
2
3.8 mol/L, cternary amine = 2 mol/L, mDMA-Solution:mEH = 1:1, Vreactor = 300 mL, Vliquid = 75 mL, U = 700 1/min, TR = 140 °C, tR = 5 h
All three added amines result in a comparable reactivity with a yield of DMF between 31 - 36% and
no byproducts were observed in the reaction with the additional amines. The catalyst leaching did
not increase significantly, no matter which polarity the amine has. Thus, a general applicability of
wash amine solutions could be demonstrated.
4. Conclusions
An efficient catalyst recycling concept for the homogeneous catalyzed hydrogenation of carbon
dioxide to N,N-dimethylformamide with an in-situ generated catalyst system consisting of
commercial available components like ruthenium(III)chloride hydrate has been developed. By
applying a liquid-liquid biphasic reaction system comprising 2-ethylhexan-1-ol as catalyst solvent and
water as product extraction agent over 97 wt.% of the catalyst could be recycled.
The reaction system showed a stable activity for at least ten runs resulting in an average yield of 31%
DMF and a total turnover number of 4942. Investigations on the in-situ formation of the catalyst
complex revealed that the amine has a significant influence on the catalyst formation and that a
small excess of ligand or precursor over a 1:1 ratio leads to higher activities. Furthermore, a higher
reaction temperature of 140 °C results in a faster condensation of the formate intermediate. The
developed reaction setup allows further process intensification since the aqueous solution could be
combined with wash amines. This enables the possibility for a combination of a gas scrubber and a
reactor so that carbon dioxide can be converted out of flue gas without further purification steps.

5. Acknowledgements
This research did not receive any specific grant from funding agencies in the public, commercial, or
not-for-profit sectors.

References
[1] A. Behr, S. Neuberg, Erdöl Erdgas Kohle 125 (2009) 367–374.
[2] M. Aresta, A. Dibenedetto, A. Angelini, Chem. Rev. 114 (2014) 1709–1742.
[3] M. Cokoja, C. Bruckmeier, B. Rieger, W.A. Herrmann, F.E. Kühn, Angew. Chem. 123
(2011) 8662–8690.
[4] A. Dibenedetto, A. Angelini, P. Stufano, J. Chem. Technol. Biotechnol. 89 (2014) 334–
353.
[5] M. Hölscher, C. Gürtler, W. Keim, T.E. Müller, M. Peters, W. Leitner, Z. Naturforsch. B
67b (2012) 961–975.
[6] A. Behr, Angew. Chem. 100 (1988) 681–698.
[7] J. Klankermayer, S. Wesselbaum, K. Beydoun, W. Leitner, Angew. Chem. 128 (2016)
7416–7467.
[8] J.L. Drake, C.M. Manna, J.A. Byers, Organometallics 32 (2013) 6891–6894.
[9] P.G. Jessop, F. Joó, C.-C. Tai, Coord. Chem. Rev. 248 (2004) 2425–2442.
[10] C.A. Huff, M.S. Sanford, ACS Catal. 3 (2013) 2412–2416.
[11] H. Bipp, H. Kieczka, (2011), Formamides in: Ullmann's Encyclopedia of Industrial
Chemistry, Wiley-VCH, Weinheim.
[12] S. Ding, N. Jiao, Angew. Chem. 124 (2012) 9360–9371.
[13] T. Mattke, E. Ströfer, J. Leschinski, R. Abdallah, A. Franzke, M. Bock, WO2011067369
A1 (2009), BASF SE.
[14] K. Halbritter, M. Eggersdorfer, K. Brenner, EP0106038 B1 (1982), BASF
Aktiengesellschaft.
[15] P. Haynes, L.H. Slaugh, J.F. Kohnle, Tetrahedron Lett. 11 (1970) 365–368.
[16] P.G. Jessop, Y. Hsiao, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 116 (1994) 8851–8852.
[17] A. Behr, P. Ebbinghaus, F. Naendrup, Chem. Eng. Technol. 27 (2004) 495–501.
[18] P.G. Jessop, T. Ikariya, R. Noyori, Chem. Rev. 95 (1995) 259–272.
[19] C. Ziebart, C. Federsel, P. Anbarasan, R. Jackstell, W. Baumann, A. Spannenberg, M.
Beller, J. Am. Chem. Soc. 134 (2012) 20701–20704.

[20] C. Federsel, C. Ziebart, R. Jackstell, W. Baumann, M. Beller, Chem. Eur. J. 18 (2012)
72–75.
[21] L. Zhang, Z. Han, X. Zhao, Z. Wang, K. Ding, Angew. Chem. 127 (2015) 6284–6287.
[22] A. Behr, P. Neubert, Applied homogeneous catalysis, Wiley-VCH, Weinheim, 2012.
[23] C.W. Kohlpaintner, R.W. Fischer, B. Cornils, Appl. Catal., A 221 (2001) 219–225.
[24] W. Keim, Angew. Chem. Int. Ed. 52 (2013) 12492–12496.
[25] M. Bressan, P. Rigo, Inorg. Chem. 14 (1975) 2286–2288.
[26] J. Chatt, R.G. Hayter, J. Chem. Soc. (1961) 896–904.
[27] C. Federsel, R. Jackstell, M. Beller, Angew. Chem. 122 (2010) 6392–6395.
[28] R.-X. Li, K.-C. Tin, N.-B. Wong, T.C. Mak, Z.-Y. Zhang, X.-J. Li, J. Organomet. Chem.
557 (1998) 207–212.
[29] K. Mauthner, D. Kalt, C. Slugovc, K. Mereiter, R. Schmid, K. Kirchner, Monatsh. Chem.
128 (1997) 533–540.
[30] J.M. Fletcher, W.E. Gardner, A.C. Fox, G. Topping, J. Chem. Soc. A 0 (1967) 1038–
1045.
[31] R.D. Johnson, S.C. Williams, A.A. Haghighirad, J. Singleton, V. Zapf, P. Manuel, I.I.
Mazin, Y. Li, H.O. Jeschke, R. Valentí, R. Coldea, Phys. Rev. B 92 (2015).
[32] X. Yan, H. Liu, K.Y. Liew, J. Mater. Chem. 11 (2001) 3387–3391.
[33] G.T. Rochelle, Science 325 (2009) 1652–1654.
[34] B.P. Spigarelli, S.K. Kawatra, J. CO2 Util. 1 (2013) 69–87.
[35] H. Yang, Z. Xu, M. Fan, R. Gupta, R.B. Slimane, A.E. Bland, I. Wright, J. Environ. Sci.
20 (2008) 14–27.

<InlineImage2>
Figure 1: Hydrogenation of carbon dioxide to formamides
<InlineImage3>
Figure 2: Conceptual design of the reaction system

<InlineImage4>
Figure 3: Variation of precursor and ligand concentration

<InlineImage5>
Figure 4: Proposed influence of different precursor to ligand ratios. Chlorine and hydrogen omitted for clarity.
<InlineImage6>
Figure 5: Hypothesis of the ruthenium trichloride solvation and stabilization

<InlineImage7>
Figure 6: Variation of the reaction temperature
<InlineImage8>
Figure 7: Recycling of the homogeneous catalyst