Green Chemistry

View Article Online

PAPER View Journal

Efficient synthesis of ethanol by methanol

homologation using CO2 at lower temperature†
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.

Cite this: DOI: 10.1039/c8gc03320j

Ying Wang, a,b Jingjing Zhang,a,b Qingli Qian, *a

Bernard Baffour Asare Bediako,a,b Meng Cui,a,b Guanying Yang,a Jiang Yana,b and
Buxing Han *a,b

Received 23rd October 2018,
Accepted 2nd January 2019
DOI: 10.1039/c8gc03320j
rsc.li/greenchem
CO2 transformation is an important topic in green chemistry, and methanol homologation using CO2 is a
promising route to produce ethanol. In this work, we studied this reaction using different homogeneous
catalytic systems. It was found that a [RuCl2(CO)3]2/Co4(CO)12 bimetallic catalyst using LiI as a promoter
and N-ethyl-2-pyrrolidone (NEP) as the solvent was very effective under mild conditions. The reaction
could proceed efficiently at 160 °C, which is much lower than that reported before. The turnover fre-
quency (TOF) of ethanol based on Ru was as high as 7.5 h−1 and the selectivity of ethanol in total products
could reach 65.0 C-mol%, which are obviously higher than those reported in the literature. Ethanol was
synthesized through cascade catalysis of a reverse water gas shift (RWGS) reaction and methanol homolo-
gation with syngas (CO/H2). The outstanding performance of the catalytic system originated from the
excellent cooperation of the components. The catalyst could be reused at least five times without any
obvious decrease of the catalytic performance. The effect of the solvent on this reaction was studied sys-
tematically. The mechanism was also discussed based on isotope labeling tests.

Introduction reported that ethanol could be produced from paraformalde-

Carbon dioxide (CO2) is a notorious greenhouse gas, and it is also
a nontoxic, cheap and easily available carbon resource. CO2 has
been widely used as a C1 synthon to produce value added chemi-
cals, such as urea, salicylic acid, carboxylic acids, alcohols, hydro-
carbons, cyclic carbonates, amides, etc.1 CO2 hydrogenation is
among the most important issues of CO2 transformation, where
the major achievement focused on the synthesis of C1 products,
i.e., methanol, CO, methane, and formic acid and its derivatives.2
In many cases, C2+ products, such as ethanol, acetic acid, olefins,
and liquid fuels, are more useful. But it is very difficult to efficien-
tly synthesize C2+ products from CO2 and H2 because this involves
controlled CO2 hydrogenation with simultaneous C–C bond for-
mation, which is still a challenge in basic science.3
Currently, ethanol production via CO2 hydrogenation gener-
ally encounters higher reaction temperature, and low reaction
activity and/or selectivity.4 Synthesis of ethanol using some
substrate to react with CO2 and H2 is a feasible way to improve
the catalytic activity and/or selectivity. Recently, it has been
a
Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Colloid,
Interface and Chemical Thermodynamics, CAS Research/Education Center for
Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of
Sciences, Beijing 100190, China. E-mail: qianql@iccas.ac.cn, hanbx@iccas.ac.cn
b
University of Chinese Academy of Sciences, Beijing 100049, China
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c8gc03320j
hyde, CO2 and H2 at 180 °C, where paraformaldehyde was con-
verted into methanol before further transformation.5 Dimethyl
ether could also be used as a substrate to synthesize ethanol
with CO2 and H2 at 180 °C. In this route, the reaction selecti-
vity was improved because of the peculiar reaction pathway.6
Currently, both paraformaldehyde and dimethyl ether are pre-
dominantly manufactured from methanol. As a well known
platform chemical, methanol is a more suitable substrate to
produce ethanol with CO2 and H2. However, at the present
stage only one study of ethanol synthesis from methanol and
CO2/H2 has been reported.7 At 180 °C, the highest TOF of
ethanol based on Ru metal was only 1.1 h−1 and the ethanol
selectivity in the total products was 34.2 C-mol%. In addition,
the yield of ethanol based on methanol was 32%. To get
higher reaction efficiency at lower temperature, exploration of
new catalytic systems is highly desirable. Herein, we report the
efficient synthesis of ethanol from methanol, CO2 and H2
under milder conditions over [RuCl2(CO)3]2, Co4(CO)12 bi-
metallic catalysts with LiI as a promoter and N-ethyl-2-pyrroli-
done (NEP) as the solvent (Scheme 1).
Scheme 1 Synthesis of ethanol via methanol homologation with CO2
and H2.

This journal is © The Royal Society of Chemistry 2019 Green Chem.

 View Article Online

Paper Green Chemistry

Results and discussion
Catalytic system
To obtain a better reaction result at lower temperature, we
tested various catalytic systems on methanol homologation
using CO2, as shown in Table 1.
The catalytic system consisting of [Ru(CO)3Cl2]2, Co4(CO)12,
LiI, and N-ethyl-2-pyrrolidone (NEP) could efficiently promote
the reaction at 160 °C (entry 1), which is much lower than that
reported previously.7 The representative graphs of GC and
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.
the subsequent well known Ru–Co catalyzed methanol homo-
logation. When single [Ru(CO)3Cl2]2 was used as the catalyst,
ethanol could be generated but the activity and selectivity of
ethanol were much lower (entry 2). In addition, CO was the
major product, indicating that the Ru catalyst was effective for
the RWGS reaction, while no ethanol was formed and the reac-
tion activity was very low when single Co4(CO)12 was utilized to
catalyze the reaction (entry 3). These results revealed that the
Ru complex was the major catalyst and the Co complex acted
as a cocatalyst. Obviously, evident synergy existed between the

GC-MS are shown in Fig. S1 and S2.† The TOF of ethanol
based on the Ru catalyst was as high as 7.5 h−1 and the selecti-
vity of ethanol in total products reached 65.0 C-mol%. Both
the activity and selectivity of ethanol were much higher than
those in the previous report.7 Besides CH4 (13.8 C-mol%),
evident CO (21.2 C-mol%) was observed in the product. This
suggested that the reaction occurred via a RWGS reaction and
Ru and Co catalysts (entries 1–3). We combined other
Ru halides, RuCl3 and RuI3, with Co4(CO)12, respectively, and
the results demonstrated that Cl was a better halide ligand
(entries 4 and 5). Further experiments indicated that the co-
existence of carbonyl and Cl in the Ru precursor gave the best
catalytic result (entries 1, 6 and 7). Using Co4(CO)12 as a co-
catalyst, we also tried catalysts other than the Ru complex

Table 1 Different catalytic systems for ethanol synthesis from methanol, CO2 and H2 a

Selectivity (C-mol%)
b −1
Entry Catalyst precursors Promoter TOF (h ) Ethanol CO CH4

1c [Ru(CO)3Cl2]2, Co4(CO)12 LiI 7.5 65.0 21.2 13.8

2 [Ru(CO)3Cl2]2 LiI 0.5 5.3 49.6 45.2
3d Co4(CO)12 LiI 0 0 5.8 94.2
4 RuCl3, Co4(CO)12 LiI 6.1 62.4 25.9 11.7
5d RuI3, Co4(CO)12 LiI 5.1 56.4 18.1 25.5
6 Ru(PPh3)3Cl2, Co4(CO)12 LiI 5.6 60.4 24.1 15.6
7 Ru(2,2′-bipyridyl)3Cl2·6H2O, Co4(CO)12 LiI 0.1 4.8 0 95.2
8d Mn2(CO)10, Co4(CO)12 LiI 0.1 10.9 7.5 81.5
9d Fe3(CO)12, Co4(CO)12 LiI <0.1 7.8 26.1 66.2
10d FeCl2, Co4(CO)12 LiI 0.1 4.7 17.7 77.5
11d Rh6(CO)16, Co4(CO)12 LiI 0.3 10.4 0 89.6
12d RhCl3, Co4(CO)12 LiI 0.2 9.0 0 91.0
13d Ir4(CO)12, Co4(CO)12 LiI 0.1 1.9 0.4 97.8
14d IrCl3, Co4(CO)12 LiI 0.2 7.3 0 92.7
15 [Ru(CO)3Cl2]2, Co2(CO)8 LiI 6.4 62.5 25.1 12.5
16 [Ru(CO)3Cl2]2, CoCl2 LiI 5.6 46.6 25.7 27.7
17 [Ru(CO)3Cl2]2, CoBr2 LiI 4.6 45.0 25.6 29.4
18 [Ru(CO)3Cl2]2, CoI2 LiI 5.3 50.8 19.6 29.6
19 [Ru(CO)3Cl2]2, CoSO4·7H2O LiI 5.7 46.9 21.9 31.2
20 [Ru(CO)3Cl2]2, Ir4(CO)12 LiI 1.3 14.7 47.5 37.8
21d [Ru(CO)3Cl2]2, Rh6(CO)16 LiI 0.5 6.8 39.5 53.7
22d [Ru(CO)3Cl2]2, Fe3(CO)12 LiI 0.3 4.4 69.1 26.5
23d [Ru(CO)3Cl2]2, NiCl2 LiI 0 0 7.7 92.3
24d [Ru(CO)3Cl2]2, PdCl2 LiI <0.1 0.2 14.6 85.2
25 [Ru(CO)3Cl2]2, Co4(CO)12 LiBr 2.3 34.1 54.1 11.8
26 [Ru(CO)3Cl2]2, Co4(CO)12 LiCl 0.9 15.9 79.1 4.9
27d [Ru(CO)3Cl2]2, Co4(CO)12 LiBF4 <0.1 4.8 0 95.2
28d [Ru(CO)3Cl2]2, Co4(CO)12 LiOAc 0.1 14.6 31.7 53.8
29d [Ru(CO)3Cl2]2, Co4(CO)12 LiOH·H2O <0.1 1.4 0.1 98.6
30 [Ru(CO)3Cl2]2, Co4(CO)12 I2 0.1 0.7 0 99.3
31 [Ru(CO)3Cl2]2, Co4(CO)12 NaI 5.3 57.1 27.9 14.9
32 [Ru(CO)3Cl2]2, Co4(CO)12 KI 4.1 50.3 33.7 15.9
33d [Ru(CO)3Cl2]2, Co4(CO)12 RbI 2.9 50.3 32.7 16.9
34d [Ru(CO)3Cl2]2, Co4(CO)12 CsI 1.1 23.9 57.2 19.0
35d [Ru(CO)3Cl2]2, Co4(CO)12 MgI2 2.9 38.8 28.4 32.8
36d [Ru(CO)3Cl2]2, Co4(CO)12 CaI2 1.9 23.0 25.1 51.9
37 [Ru(CO)3Cl2]2, Co4(CO)12 AlI3 0 0 0 100.0
38 [Ru(CO)3Cl2]2, Co4(CO)12 SnI4 <0.1 0.4 2.3 97.3
39d [Ru(CO)3Cl2]2, Co4(CO)12 ZnI2 0.1 2.2 72.1 25.7
a
Reaction conditions: 10 μmol Ru catalyst and 60 μmol Co catalyst (based on the metal), 2.5 mmol promoter, 2 mL solvent, 2.6 mmol methanol,
2 MPa CO2 and 6 MPa H2 (at room temperature), 160 °C, 15 h. b TOF denotes moles of ethanol produced per mole of the Ru catalyst per hour in
the steady state. c Yield of ethanol based on the methanol feedstock was 43.5%. d A black precipitate was observed after the reaction.

Green Chem. This journal is © The Royal Society of Chemistry 2019

 View Article Online

Green Chemistry Paper

(Mn2(CO)10, Fe3(CO)12, FeCl2, Rh6(CO)16, RhCl3, Ir4(CO)12, and Table 2 Effect of different solvents on methanol homologation with
IrCl3), but all the catalytic results were very poor (entries 8–14). CO2 and H2 a
Hence [Ru(CO)3Cl2]2 was a suitable catalyst for the reaction.
Products (μmol)
With [Ru(CO)3Cl2]2 as the major catalyst, we tested other Co
cocatalysts (Co2(CO)8, CoCl2, CoBr2, CoI2, and CoSO4·7H2O), Entry Solvent Ethanol CO CH4
where Co4(CO)12 was proved to be the best one (entries 1 and b
1 141 29 3228
15–19). We also conducted the experiments with cocatalysts
b
other than the Co complex (Ir4(CO)12, Rh6(CO)16, Fe3(CO)12, 2 0 0 3389
b
NiCl2, and PdCl2), but the catalytic performances were all poor 3 0 1407 96
(entries 20–24). Therefore, Co4(CO)12 was an appropriate coca- 4b
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.

0 261 17
talyst for the reaction. b
5 0 776 101
Based on [Ru(CO)3Cl2]2 and Co4(CO)12, we used other Li
halides instead of the LiI promoter (LiBr and LiCl), and the 6 1053 890 474
results showed that the promoting effect of the Li halides fol- 7 0 102 2072
lowed the sequence: LiI > LiBr > LiCl (entries 1, 25 and 26). We
also tried other Li salts (LiBF4, LiOAc, and LiOH·H2O), but the 8 0 48 6
ethanol yield was very low (entries 27–29). The control test
9 18 854 286
using iodine (I2) as the promoter demonstrated that the iodide
anion was necessary for this reaction (entry 30). The peculiar 10b 0 1617 260
iodide effects in transition metal catalyzed reactions have been
systematically discussed elsewhere.8 On the basis of the iodide 11b 0 1268 275
effect, we tested other alkali metal iodides, i.e., NaI, KI, RbI,
and CsI (entries 31–34). It was observed that the sequence of 12b 0 915 627
the promoting effect is as follows: LiI > NaI > KI > RbI > CsI. 13 0 1615 608
We also tried alkaline metal iodides (MgI2 and CaI2), but their
performance was not as good as that of an alkali metal iodide 14 1128 737 480
and they followed the order: MgI2 > CaI2 (entries 35 and 36). 15 289 564 1526
After testing some commonly used transition metal iodides
(AlI3, SnI4, and ZnI2), we found that their performances were 16b 7 153 2061
all very poor (entries 37–39). Consequently, LiI was the best
17 15 353 1933
promoter for the reaction. As discussed above, the eminent
promoting effect of LiI may lie in two aspects: the stronger 18 381 601 723
Lewis acidity and smaller size of Li+, and the better nucleo-
philicity of I−. In a word, the catalytic system consisting of 19 241 872 329
[Ru(CO)3Cl2]2, Co4(CO)12 and LiI was fit for the target reaction.
20 0 25 176
Solvent effect
21 33 62 14
The role of a solvent in homogeneous transition metal cata-
lyzed reactions is very crucial. Even a small modification of the
22 1071 791 503
solvent structure may cause a large change in the rate and in
the pattern of the processes in solution.9 But the role of the
solvent in hydrogenating CO2 to C2+ products has not been
a
Reaction conditions: 10 μmol [RuCl2(CO)3]2, 60 μmol Co4(CO)12
(based on the metal), 2.5 mmol LiI, 2 mL solvent, 2.6 mmol methanol,
adequately discussed in the literature. In this report, we con-
2 MPa CO2 and 6 MPa H2 (at room temperature), 160 °C, 15 h. b A
ducted systematic investigation of the solvent effect in the black precipitate was observed after the reaction.
target reaction (Table 2).
It is known that dioxane and tetrahydrofuran (THF) are
commonly used solvents in conventional methanol homologa-
tion with syngas (CO/H2).10 When dioxane was used in this These facts suggested that the RWGS reaction could hardly
work, little ethanol was observed due to decomposition of take place in dioxane or THF, which is necessary for ethanol
dioxane, and methanol was mostly converted to methane formation in the target reaction. When we used diethyl-
(entry 1). We also tested THF as a reaction solvent, no ethanol methylamine as the solvent, no ethanol was detected and both
was detected and methanol was totally converted to methane catalyst and solvent decomposed significantly (entry 3).
(entry 2). In the above two tests, CH3I could form easily but Nevertheless, a considerable amount of CO was produced in
little or no CO was observed. For comparison, when the reac- the reaction, indicating that the amine group could effectively
tion was conducted in THF using 0.5 MPa CO, instead of CO2, catalyze the RWGS reaction but not the ethanol formation. The
a considerable amount of ethanol was produced (Fig. S3†). basicity of the amine group may explain its performance in

This journal is © The Royal Society of Chemistry 2019 Green Chem.


Paper
activating acidic CO2. We also tested pyridine as the solvent;
the activity of the RWGS reaction was lower than that in di-
ethylmethylamine and the catalyst was also unstable (entry 4).
We further tried 1-methylpyrrolidine, a cyclic saturated amine,
as the solvent, and the RWGS reaction was stably and efficien-
tly accelerated (entry 5). When we turned to NMP, a cyclic
amide, the catalyst was stable and a remarkable amount of
ethanol was generated (entry 6). The coexistence of the amine
group and carbonyl group was necessary for ethanol formation
because no ethanol was observed when cyclopentanone was
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.
used as the solvent and the catalyst was not stable (entry 7).
The carbonyl group helped to dissolve LiI, which is a key com-
ponent for the C–C bond formation step in ethanol synthesis.
But the cyclopentanone was also unstable and could be readily
hydrogenated to cyclopentanol under reaction conditions.
When the O atom in NMP was changed to an S atom, no
ethanol was generated and the catalyst was unstable; moreover,
the RWGS reaction was also inhibited (entry 8). These facts
indicated that there should be some synergy between the
amine and carbonyl groups in the cyclic amide solvent.
Compared with the carbonyl group, the sulfinyl group is less
polar, which may not be eligible to trigger the synergy during
the target reaction. The substituent on the amine group was
crucial to ethanol formation. When the methyl group on NMP
was replaced by a H atom, the RWGS process took place but no
ethanol was detected in the reaction (entry 9). The cyclic
nature of the solvent structure was also required for producing
ethanol. When the reaction was carried out in a non-cyclic
amide like solvent, such as N,N-dimethylformamide (DMF),
N,N-dimethylacetamide (DMA) and 1,1,3,3-tetramethylurea
(TMU), no ethanol was formed but the RWGS reaction pro-
ceeded smoothly (entries 10–12). We further substituted the C
atom, with an O atom, near the carbonyl group in the cyclic
amide, but ethanol generation was totally prohibited (entries 6
and 13). In short, the cyclic amide with an alkyl substituent on
an N atom was the elementary structure of a suitable solvent
for the target reaction. The amine-based group was responsible
for the RWGS reaction, while the alkyl group on the N atom,
the neighboring carbonyl group, and the cyclic structure were
necessary for ethanol production.
We further investigated the impact of the substituent of the
N atom on the catalytic performance. Based on the above dis-
cussion, we learned that NMP is an effective solvent for the
target reaction. When the methyl group on the N atom was
replaced by an ethyl group, the activity and selectivity of
ethanol formation were better than those using NMP (entries 6
and 14). The superiority of the ethyl group over the methyl
group has a sound experimental basis under different con-
ditions, as shown in Table S1.† The ethyl group has a stronger
electron donating ability than the methyl group, which could
strengthen the basicity of the N atom and accordingly acceler-
ate the reaction. When the ethyl group was further changed
into larger ones, both ethanol and CO production decreased
(entries 15–18). This phenomenon could be ascribed to the
steric effect of these substituents. The final impact of the sub-
stituent on the reaction relied on the balance between the elec-
Green Chem.
View Article Online
Green Chemistry
tronic and steric effects. The ring size of the cyclic amide
solvent was also important for the catalytic performance. NMP
is a five membered ring cyclic amide. When 1-methyl-2-piperi-
done, a six membered ring cyclic amide, was used as the
solvent, both the catalytic activity and selectivity became worse
(entry 19). The smaller ring size often possesses a stronger
ring tension, which helps to expose the lone pair electrons on
the N atom as an active site, and accelerate the reaction. The
relative position of the amine group and carbonyl group also
affected the catalytic performance. When the N-methyl group
of the 1-methyl-2-piperidone solvent was transferred to the
para position, no ethanol was detected and the catalyst was
unstable (entry 20). Similarly, when 1-benzyl-3-pyrrolidinone,
instead of 1-benzyl-2-pyrrolidinone, was used as the solvent,
ethanol production was markedly lower (entries 18 and 21). In
addition, a remarkable amount of a LiI precipitate was
observed. Therefore, only when the amine group and carbonyl
group were located at the ortho position, the target reaction
could proceed effectively. There is a strong possibility that the
conjugation between the lone pair electrons of the N atom and
π electrons of the carbonyl double bond could make the two
functional groups cooperate well. Moreover, the p–π conju-
gation between the two groups helps to stabilize the carbonyl
group under reaction conditions. The number of amine
groups also influenced the reaction. 1,3-Dimethyl-2-imidazoli-
dinone (DMI) was a slightly better solvent than NMP in pro-
moting ethanol synthesis (entries 6 and 22), while it was not
as good as NEP (entries 14 and 22). The isotope labeling test
using 13CH3OH ruled out the possibility that ethanol was
formed from the ethyl group on the NEP molecule (Fig. S4†).
In a word, NEP was an appropriate solvent for the target reac-
tion. In addition, it has low volatility and is widely used as a
solvent in different processes such as those in the pharma-
ceutical industry.
Impact of reaction parameters
Fig. 1 depicts the formation rate and selectivity of ethanol at
different temperatures. The reaction could hardly take place at
120 °C, while obvious ethanol formation was observed at
Fig. 1 The TOF and selectivity of ethanol at different temperatures.
Reaction conditions: 10 μmol [RuCl2(CO)3]2, 60 μmol Co4(CO)12 (based
on the metal), 2.5 mmol LiI, 2 mL NEP, 2.6 mmol methanol, 2 MPa CO2
and 6 MPa H2 (at room temperature), 15 h.
This journal is © The Royal Society of Chemistry 2019
 View Article Online

Green Chemistry Paper

Table 3 Effect of reaction parameters on the reactiona

Selectivity [C-mol%]
Ru/Co LiI CO2/H2 TOF
Entry (μmol) (mmol) (MPa) (h−1) Ethanol CO CH4

1b 10/60 2.5 2/6 0.1 2.4 96.2 1.4

2 5/65 2.5 2/6 11.2 56.4 18.1 25.5
3 10/60 2.5 2/6 7.5 65.0 21.2 13.8
4 15/55 2.5 2/6 4.8 62.1 23.9 14.0
5 20/50 2.5 2/6 4.2 66.2 23.6 10.1
6 60/10 2.5 2/6 0.4 18.0 32.4 49.7
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.

7 5/30 2.5 2/6 10.3 59.0 21.9 19.1

8 15/90 2.5 2/6 5.5 66.6 21.6 11.8
9 10/60 1 2/6 4.3 48.8 34.7 16.5
Fig. 2 Arrhenius plots for ln TOF (h−1) versus the inversed temperature 10 10/60 2 2/6 6.6 61.6 25.0 13.5
(from 403 to 433 K) at low conversion. Reaction conditions: 10 μmol 11 10/60 3 2/6 7.0 65.1 20.3 14.5
[RuCl2(CO)3]2, 60 μmol Co4(CO)12 (based on the metal), 2.5 mmol LiI, 12 10/60 3.5 2/6 5.4 46.7 24.3 28.9
2 mL NEP, 2.6 mmol methanol, 2 MPa CO2 and 6 MPa H2 (at room
13 10/60 4 2/6 4.6 39.5 22.8 37.7
14c 10/60 2.5 0/0 0 — — —
temperature), 6 h.
15c 10/60 2.5 0/6 0 — — —
16 10/60 2.5 2/0 0 — — —
17 10/60 2.5 0.5/1.5 1.6 50.6 15.1 34.3
18 10/60 2.5 1/3 4.7 65.9 18.0 16.2
130 °C. The yield of ethanol increased rapidly with the elevat- 19 10/60 2.5 3/9 8.7 60.5 28.4 11.2
ing temperature. At 160 °C, the TOF of ethanol reached 7.5 h−1 20 10/60 2.5 1/7 6.7 64.9 18.5 16.5
21 10/60 2.5 3/5 7.2 57.9 27.3 14.8
and the selectivity of ethanol was as high as 65%. When the 22 10/60 2.5 4/4 6.8 53.5 34.0 12.5
temperature was further increased, the ethanol yield had a minor 23 10/60 2.5 5/3 6.2 47.8 35.4 16.9
change but the ethanol selectivity decreased sharply due to the 24 10/60 2.5 6/2 5.6 45.8 34.6 19.6
rapid yield of CH4. Therefore, the suitable temperature for this a
Reaction conditions: 10 μmol [RuCl2(CO)3]2, 60 μmol Co4(CO)12
reaction was 160 °C. We calculated the apparent activation (based on the metal), 2.5 mmol LiI, 2 mL NEP, 2.6 mmol methanol, 2
energy from the Arrhenius plot using the data in the temperature MPa CO2 and 6 MPa H2 (at room temperature), 160 °C, 15 h. b No
methanol was added before the reaction. c A black precipitate was
range from 403 K to 433 K, as displayed in Fig. 2. The result indi- observed after the reaction.
cated that the apparent activation energy of the catalytic system
was 116 kJ mol−1, which was much lower than that in ethanol
synthesis via CO2 hydrogenation reported previously.4d This is set the total pressure at 8 MPa and changed the ratio of
consistent with the result that the reaction promoted by our cata- CO2/H2, and the results showed that 1/3 was a suitable ratio
lytic system could proceed at lower temperature. (entries 3 and 20–24).
At 160 °C, we studied other reaction parameters (Table 3).
The methanol substrate was necessary for constructing Recyclability
ethanol molecules. When CO2 hydrogenation was catalyzed by
Generally speaking, recycling of a homogeneous catalyst is very
the optimized catalytic system, little alcohol product was pro-
difficult. Due to good stability, our catalytic system can be recycled
duced (entry 1). In addition, little methane was observed in
and reused for at least five cycles, as shown in Fig. 3. The result
CO2 hydrogenation, which indicated that in the target reaction
showed that the catalytic activity did not change considerably,
the methane byproduct was mostly formed from the methanol
and the TON of ethanol reached 555.5 after the five cycles.
substrate (Fig. S1 and S5†). The dosage ratio of the Ru/Co cata-
lysts was crucial to the catalytic performance. We fixed the
total molar amount of the catalysts at 70 μmol and altered the
molar ratio of the Ru/Co catalysts, and found that 10 μmol Ru
and 60 μmol Co gave the best result (entries 2–6). When the
molar ratio of Ru/Co was fixed at 1/6, the appropriate usage of
the catalyst was also 10 μmol Ru and 60 μmol Co (entries 3, 7
and 8). We also tested the usage of the LiI dosage and
observed that 2.5 mmol of LiI was fit for the reaction (entries 3
and 9–13). Too much I− from LiI would occupy the active sites
of the catalyst, inhibiting the reaction. The gas pressure also
affected the catalytic performance. Without CO2 and/or H2, no
product was formed, further indicating that ethanol generated
in the reaction was not from NEP (entries 14–16). We fixed the
Fig. 3 The results of the recycling tests. Reaction conditions: 10 μmol
ratio of CO2/H2 at 1/3, and changed the total pressure of the [RuCl2(CO)3]2 and 60 μmol Co4(CO)12 (based on the metal), 2.5 mmol LiI,
gases. The results demonstrated that 2 MPa CO2 and 6 MPa H2 2 mL NEP, 2.6 mmol methanol, 2 MPa CO2 and 6 MPa H2 (at room
were suitable for the reaction (entries 3 and 17–19). We also temperature), 160 °C, 15 h.

This journal is © The Royal Society of Chemistry 2019 Green Chem.


Paper
Time course of the reaction
Fig. 4 illustrates the time course of the reaction. The amount
of methanol decreased continuously during the reaction.
Ethanol emerged at 2 h, and the amount of ethanol grew stea-
dily until 15 h. Then the reaction rate became slower because
the methanol substrate was gradually used up. During the
whole course, the amount of methane increased slowly and
steadily. The content of CO increased rapidly at the beginning
(0–3 h), and then it remained at a constant level with the reac-
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.
tion continuing. This confirmed that CO generated in situ
rapidly participated in ethanol production.
Mechanistic discussion
As discussed in the former paragraphs, the target reaction
involved the RWGS reaction to produce CO and subsequent
methanol homologation with CO/H2 (Fig. S6†). The mecha-
nisms of the Ru catalyzed RWGS reaction11 and Ru–Co acceler-
ated methanol homologation12 have been reported elsewhere.
Based on the mechanistic study in the literature and the
experimental data of our catalytic system, we rendered the
possible mechanism of this work (Scheme 2). Firstly, methanol
reacted with LiI to form CH3I and LiOH (Step 1). CH3I was
known as the intermediate of methanol homologation and
could form spontaneously from methanol and iodide at elev-
ated temperature.13 The formation of CH3I could be acceler-
ated by the Lewis acidic cation (Li+).7 The tracer experiments
Fig. 4 Time course of the reaction. Reaction conditions: 10 μmol
[RuCl2(CO)3]2 and 60 μmol Co4(CO)12 (based on the metal), 2.5 mmol LiI,
2 mL NEP, 2.6 mmol methanol, 2 MPa CO2 and 6 MPa H2 (at room temp-
erature), 160 °C.
Scheme 2 Proposed mechanism of the reaction.
Green Chem.
View Article Online
Green Chemistry
using CH3OD and CH318OH affirmed that the OH broke away
from CH3OH during the reaction, supporting the formation of
CH3I (Fig. S7 and S8†). Secondly, CH3Co*I was generated by
oxidative addition of CH3I to the active Co species (Co*)
(Step 2). The oxidative addition of CH3I to the Co complex has
also been reported in other Co catalyzed methanol homologa-
tion reactions.12 Thirdly, CO, generated by the RWGS reaction
(Step 3), inserted into the CH3–Co* bond and resulted in the
CH3COCo* complex (Step 4). The insertion of CO into the
metal–carbon bond is a basic step in organic chemistry.14
Fourthly, acetaldehyde was reductively eliminated from
CH3COCo*I in the presence of H2, during which Co* was
regenerated and HI was formed (Step 5, Fig. S9†).
Acetaldehyde was in situ reduced to ethanol by H2 catalyzed by
the Ru catalyst (Step 6, Fig. S10†). The tracer experiment using
13
CH3OH also verified that the methyl group of methanol was
transferred into the methyl group of the 13CH3CH2OH mole-
cule, which is consistent with the proposed mechanism
(Fig. S4†). The tracer experiment using D2 instead of H2
revealed that at most six D atoms could enter one ethanol
molecule, indicating an obvious H–D exchange during the
reaction (Fig. S11†). The detailed study of such exchange reac-
tions has been reported elsewhere.15 Finally, LiOH and HI gen-
erated in situ neutralized spontaneously to form LiI and H2O
(Step 7). After these steps, all the catalytic species were regener-
ated for the next cycle. The active centers of the Ru catalyst in
the RWGS reaction and the Co catalyst in methanol homologa-
tion with CO/H2 have been studied elsewhere.11,12 The basic
structure of the Ru and Co active centers in this work should
be similar to those reported in the literature.
Conclusions
In summary, we have studied Ru–Co bimetallic catalysts for
methanol homologation with CO2 and H2 to produce ethanol
under mild conditions. The catalytic system consisting of
[RuCl2(CO)3]2/Co4(CO)12, LiI and NEP shows excellent perform-
ance for this reaction. The reaction can be carried out at
160 °C. The turnover frequency (TOF) based on Ru and the
selectivity of ethanol in total products can reach 7.5 h−1 and
65.0 C-mol%, respectively. In addition, the yield of ethanol
reaches 43.5%. A detailed study indicates that the reaction pro-
ceeds via the RWGS reaction and subsequent methanol homo-
logation using CO/H2. The outstanding activity and selectivity
of the catalytic system results from the synergy of the catalytic
components. The solvent is very important for the reaction.
NEP could effectively accelerate the RWGS reaction and metha-
nol homologation, rendering the best catalytic result. The cata-
lyst can be recycled at least five times without any obvious
change of the catalytic performance, and the TON of ethanol
reached 555.5 in five cycles. The transformation of CO2 into
useful chemicals and fuels is one of the major areas in green
chemistry, and this work provides an interesting route to
produce ethanol via methanol homologation using CO2 as a
C1 feedstock.
This journal is © The Royal Society of Chemistry 2019

Green Chemistry
Experimental
Chemicals
Tris(2,2′-bipyridine) dichlororuthenium(II) hexahydrate
(Ru(2,2′-bipyridyl)3Cl2·6H2O, 98+%), iridium carbonyl
(Ir4(Co)12, 98+%), iron dodecacarbonyl (Fe3(CO)12, 99%) and
acetaldehyde (CH3CHO, 99.5%) were provided by Acros
Organics. Dicobaltoctacarbonyl (Co2(CO)8), 1-methyl-
pyrrolidine-2-thione (97%) and diethylmethylamine (98%)
were purchased from TCI Shanghai Co., Ltd.
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.
Dodecacarbonyltetracobalt (Co4(CO)12, 98%), hexarhodium
hexadecacarbonyl (Rh6((CO)16, 98%), rhodium trichloride
anhydrous (RhCl3, 99.9%), sodium iodide (NaI, 99.5%), potass-
ium iodide (KI, 99.9%), tin(IV) iodide (SnI4, 99.998%), iridium
(III) chloride, anhydrous (IrCl3, 99.99%), N,N-dimethyl-
acetamide (DMA, 99%), lithium chloride (LiCl, 99%), ferrous
chloride (FeCl2, 99.5%), magnesium iodide (MgI2, 99.996%),
calcium iodide (CaI2, 99.99%), lithium iodide (LiI, 99.95%),
rubidium iodide (RbI, 99.8%), cesium iodide (CsI, 99.9%), pal-
ladium chloride (PdCl2, 99.9%), ruthenium(III) iodide anhy-
drous (RuI3, Ru 20.5%), cobalt(II) iodide anhydrous (CoI2,
99.5%), manganese carbonyl (Mn2(CO)10, 30.6%), lithium
tetrafluoroborate (LiBF4, 98%), lithium hydroxide mono-
hydrate (LiOH·H2O, LiOH 56.5%), 1-methyl-2-pyrrolidinone
(NMP, 99%), N,N-dimethylformamide (DMF, 99%) and 1,1,3,3-
tetramethylurea (TMU, 99%) were offered by Alfa Aesar China
Co, Ltd. Methanol-D1 (CH3OD, 99.5 atom% D) and 1-benzyl-3-
pyrrolidinone (98%) were provided by Beijing InnoChem
Science & Technology Co., Ltd. Cobalt(II) chloride (CoCl2,
99%), sulfolane (99%), tetrahydrofuran (THF, 99.5%), iodine
(I2, 99.8%), 1,4-dioxane (99.5%), 1-methylpyrrolidine (98%),
1-methyl-4-piperidone (98%), cyclopentanone (99.5%), zinc
iodide (ZnI2, 98%), lithium acetate (LiOAc, 99%), 2-pyrrolidi-
none (98%) and pyridine (99.5%) were obtained from J&K
Scientific Ltd. Lithium bromide (LiBr, 99.5%), tris(triphenyl-
phosphine)ruthenium(II) chloride (Ru(PPh3)3Cl2, 99%), and
1,3-dimethyl-2-imidazolidinone (DMI, 98+%) were provided by
Adamas Reagent Co., Ltd. Toluene (99.5+%) was purchased
from Beijing Chemical Works. Methanol-18O (CH318OH, 95%)
and 1-methyl-2-piperidone (99%) were offered by Sigma-
Aldrich Co, LLC. Dichlorotricarbonylruthenium dimer
([Ru(CO)3Cl2]2, 98%), nickel(II) chloride anhydrous (NiCl2,
98%) and aluminum iodide (AlI3, 95%) were purchased from
Strem Chemicals, Inc. 1-Ethyl-2-pyrrolidone (NEP, 99%),
1-benzyl-2-pyrrolidinone (98%), 3-methyl-2-oxazolidone (98%),
1-octyl-2-pyrrolidone (98%), 1-cyclohexyl-2-pyrrolidone (99%)
and 1-butyl-2-pyrrolidone (98+%) were obtained from Aladdin.
Cobalt(II) bromide (CoBr2, 98+%) was purchased from
Fluorochem Ltd. Cobalt sulfate heptahydrate (CoSO4·7H2O,
99.5%) was offered by Amethyst Chemicals. CO2 (99.99%), H2
(99.99%) and CO (99.99%) were provided by Beijing Analytical
Instrument Company.
Catalytic reaction
The apparatus and procedures were similar to those reported
previously.6 All the reactions were carried out in a 16 mL
This journal is © The Royal Society of Chemistry 2019
View Article Online
Paper
Teflon-lined stainless steel batch reactor equipped with a mag-
netic stirrer. The inner diameter of the reactor was 18 mm. In
a typical experiment, desired amounts of Ru or/and Co or
another metal, LiI or another promoter, and methanol and
2 mL NEP or another solvent were added into the reactor. After
the air in the reactor was replaced with 1 MPa CO2 three times,
CO2 and H2 were charged into the reactor for the desired
pressure at room temperature, respectively. The reactor was
placed in an air bath of constant temperature, and the stirrer
was set at 800 rpm. After the reaction, the reactor was cooled
in an ice-water bath and the residual gas was released carefully
in a gas bag. The liquid mixture was analyzed by using a gas
chromatograph (GC, Agilent 7890 B) equipped with a flame
ionization detector (FID) and a DB-5 capillary column using
toluene as the internal standard. The liquid products were
identified using GC-MS (Agilent-7890B-5977A) and by compar-
ing the retention times with the respective standards in the GC
traces. The amounts of the products were calculated from the
relevant GC data. The gaseous samples were detected by a GC
(Agilent 4890D) equipped with a TCD detector and a packed
column (carbon molecular sieve TDX-01, 3 mm in diameter
and 1 m in length) using argon as the carrier gas.
Recycling test
After the reaction, the reactor was cooled down in an ice-water
bath and the residual gas was released carefully. The amounts
of the products were analyzed as discussed above. Then the
alcohols generated and unreacted methanol were removed
under vacuum at 80 °C for 10 h. Then the catalytic system was
used directly for the next cycle.
Conflicts of interest
There are no conflicts of interest to declare.
Acknowledgements
The authors thank the National Natural Science Foundation of
China (21875262, 21733011, and 21533011), National Key
Research and Development Program of China
(2017YFA0403102), Beijing Municipal Science & Technology
Commission (Z181100004218004), and the Chinese Academy
of Sciences (QYZDY-SSW-SLH013).
References
1 (a) M. Y. He, Y. H. Sun and B. X. Han, Angew. Chem., Int.
Ed., 2013, 52, 9620–9633; (b) Q. Liu, L. P. Wu, R. Jackstell
and M. Beller, Nat. Commun., 2015, 6, 5933; (c) M. Aresta,
A. Dibenedetto and A. Angelini, Chem. Rev., 2014, 114,
1709–1742; (d) P. G. Jessop, F. Joó and C. C. Tai, Coord.
Chem. Rev., 2004, 248, 2425–2442; (e) J. Klankermayer,
Angew. Chem., Int. Ed., 2016, 55, 7296–7343; (f ) H. Zhou
and X. B. Lu, Sci. China: Chem., 2017, 60, 904–911;
Green Chem.

Paper
(g) Q. L. Qian, J. J. Zhang, M. Cui and B. X. Han, Nat.
Commun., 2016, 7, 11481; (h) Y. Liu, H. Zhou, J. Z. Guo,
W. M. Ren and X. B. Lu, Angew. Chem., Int. Ed., 2017, 56,
4862–4866; (i) J. Kothandaraman, A. Goeppert, M. Czaun,
G. A. Olah and G. K. S. Prakash, Green Chem., 2016, 18,
5831–5838; ( j) X. Y. Ren, Z. Y. Zheng, L. Zhang, Z. Wang,
C. G. Xia and K. L. Ding, Angew. Chem., Int. Ed., 2017, 56,
310–313; (k) Y. N. Gao, S. Z. Liu, Z. Q. Zhao, H. C. Tao and
Z. Y. Sun, Acta Phys.-Chim. Sin., 2018, 34, 858–872;
(l) M. Cui, Q. L. Qian, Z. H. He, Z. F. Zhang, J. Ma,
Published on 03 January 2019. Downloaded by University of Kansas on 1/21/2019 9:22:21 AM.
T. B. Wu, G. Y. Yang and B. X. Han, Chem. Sci., 2016, 7,
5200–5205; (m) M. Cui, Q. L. Qian, J. J. Zhang, C. J. Chen
and B. X. Han, Green Chem., 2017, 19, 3558–3565.
2 (a) F. L. Liao, X. P. Wu, J. W. Zheng, M. M. J. Li,
A. Kroner, Z. Y. Zeng, X. L. Hong, Y. Z. Yuan, X. Q. Gong
and S. C. E. Tsang, Green Chem., 2017, 19, 270–280;
(b) S. Kattel, P. J. Ramírez, J. G. Chen, J. A. Rodriguez and
P. Liu, Science, 2017, 355, 1296–1299; (c) S. Wesselbaum,
T. vom Stein, J. Klankermayer and W. Leitner, Angew.
Chem., Int. Ed., 2012, 51, 7499–7502; (d) N. M. Rezayee,
C. A. Huff and M. S. Sanford, J. Am. Chem. Soc., 2015,
137, 1028–1031; (e) S. Kattel, B. H. Yan, Y. X. Yang,
J. G. Chen and P. Liu, J. Am. Chem. Soc., 2016, 138,
12440–12450.
3 (a) M. Beller and U. T. Bornscheuer, Angew. Chem., Int. Ed.,
2014, 53, 4527–4528; (b) G. Prieto, ChemSusChem, 2017, 10,
1056–1070.
4 (a) Z. H. He, Q. L. Qian, J. Ma, Q. L. Meng, H. C. Zhou,
J. L. Song, Z. M. Liu and B. X. Han, Angew. Chem., Int. Ed.,
2016, 55, 737–741; (b) Q. L. Qian, M. Cui, Z. H. He,
C. Y. Wu, Q. G. Zhu, Z. F. Zhang, J. Ma, G. Y. Yang,
J. J. Zhang and B. X. Han, Chem. Sci., 2015, 6, 5685–5689;
(c) M. Cui, Q. L. Qian, Z. H. He, Z. F. Zhang, J. Ma,
T. B. Wu, G. Y. Yang and B. X. Han, Chem. Sci., 2016, 7,
5200–5205; (d) S. X. Bai, Q. Shao, P. T. Wang, Q. G. Dai,
Green Chem.
View Article Online
Green Chemistry
X. Y. Wang and X. Q. Huang, J. Am. Chem. Soc., 2017, 139,
6827–6830; (e) H. Kusama, K. Okabe, K. Sayama and
H. Arakawa, Energy, 1997, 22, 343–348; (f ) K. K. Bando,
K. Soga, K. Kunimori and H. Arakawa, Appl. Catal., A, 1998,
175, 67–81; (g) H. Kurakata, Y. Izumi and K. Aika, Chem.
Commun., 1996, 389–390; (h) T. Inui and T. Yamamoto,
Catal. Today, 1998, 45, 209–214; (i) T. Inui, T. Yamamoto,
M. Inoue, H. Hara, T. Takeguchi and J. B. Kim, Appl. Catal.,
A, 1999, 186, 395–406; ( j) L. S. Davy, D. Nieskens, Y. Ferrari,
R. Liu and J. Kolonko, Catal. Commun., 2011, 14, 111–113;
(k) S. G. Li, H. J. Guo, C. R. Luo, H. R. Zhang, L. Xiong,
X. D. Chen and L. L. Ma, Chem. Lett., 2013, 143, 345–355.
5 J. J. Zhang, Q. L. Qian, M. Cui, C. J. Chen, S. S. Liu and
B. X. Han, Green Chem., 2017, 19, 4396–4401.
6 Q. L. Qian, M. Cui, J. J. Zhang, J. F. Xiang, J. L. Song,
G. Y. Yang and B. X. Han, Green Chem., 2018, 20, 206–213.
7 K. Tominaga, Y. Sasaki, T. Watanabe and M. Saito, Stud.
Surf. Sci. Catal., 1998, 114, 495–498.
8 P. M. Maitlis, A. Haynes, B. R. James, M. Catellani and
G. P. Chiusoli, Dalton Trans., 2004, 3409–3419.
9 V. Gutmann, Coord. Chem. Rev., 1976, 18, 225–255.
10 Texaco Inc., US Pat, US4409405-A, 1983.
11 (a) K. Tsuchiya, J. D. Huang and K. Tominaga, ACS Catal.,
2013, 3, 2865–2868; (b) W. Li, S. Guo and L. Guo, Catal.
Surv. Asia, 2017, 21, 185–197.
12 M. Röper, H. Loevenich and J. Korff, J. Mol. Catal., 1982,
17, 315–322.
13 M. Röper and H. Loevenich, J. Organomet. Chem., 1983,
255, 95–102.
14 (a) G. Braca, G. Sbrana, G. Valentini, G. Andrich and
G. Gregorio, J. Am. Chem. Soc., 1978, 100, 6238–6240;
(b) G. Braca, L. Paladini, G. Sbrana, G. Valentini,
G. Andrich and G. Gregorio, Ind. Eng. Chem. Prod. Res. Dev.,
1981, 20, 115–122.
15 T. Junk and W. J. Catallo, Chem. Soc. Rev., 1997, 26, 401–406.
This journal is © The Royal Society of Chemistry 2019