An Innovative Approach for Highly Selective Direct Conversion of CO2into Propanol using C2H4and H2

Article abstract of DOI:10.1002/cssc.201402212

Multifunctional catalysts are developed for converting CO₂with C₂H₄and H₂into propanol. Au nanoparticles (NP) supported on TiO₂are found to facilitate this reaction. The activity and selectivity strongly depend on NP size, which can be tuned by the method of Au deposition and by promoting with K. The promoter improves the selectivity to propanol. Under optimized reaction conditions (2 MPa, 473 K, and CO₂/H₂/C₂H₄=1:1:1), CO₂is continuously converted into propanol with a near-to-100 % selectivity. Catalytic tests as well as mechanistic studies by in situ FTIR and temporal analysis of products with isotopic tracers allow the overall reaction scheme to be determined. Propanol is formed through a sequence of reactions starting with reverse water–gas shift to reduce CO₂to CO, which is further consumed in the hydroformylation of ethylene to propanal. The latter is finally hydrogenated to propanol, while propanol hydrogenation to propane is suppressed.

Full text of DOI:10.1002/cssc.201402212

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201402212
An Innovative Approach for Highly Selective Direct
Conversion of CO2 into Propanol using C2H4 and H2
Stefan J. Ahlers, Ursula Bentrup, David Linke, and Evgenii V. Kondratenko*[a]
Multifunctional catalysts are developed for converting CO2
with C2H4 and H2 into propanol. Au nanoparticles (NP) support-
ed on TiO2 are found to facilitate this reaction. The activity and
selectivity strongly depend on NP size, which can be tuned by
the method of Au deposition and by promoting with K. The
promoter improves the selectivity to propanol. Under opti-
mized reaction conditions (2 MPa, 473 K, and CO2/H2/C2H4 =
1:1:1), CO2 is continuously converted into propanol with
a near-to-100% selectivity. Catalytic tests as well as mechanistic
studies by in situ FTIR and temporal analysis of products with
isotopic tracers allow the overall reaction scheme to be deter-
mined. Propanol is formed through a sequence of reactions
starting with reverse water–gas shift to reduce CO2 to CO,
which is further consumed in the hydroformylation of ethylene
to propanal. The latter is finally hydrogenated to propanol,
while propanol hydrogenation to propane is suppressed.
Introduction
Carbon-based compounds play a vital role in everyday life by
fulfilling our needs for a wide range of household and com-
mercial products. For their production, the chemical industry
uses natural gas and oil. With rising energy costs caused by
limited natural resources, the efficiency of current processes
and the development of alternative routes have become a pri-
ority for modern and, particularly, future industry. In addition,
the emission of CO2 contributing to global warming has
strongly increased since the industrial revolution.[1] One way to
circumvent both energy and ecological problems is to use
plant-derived materials and CO2 as feedstock.[2] Biomass is al-
ready applied world-wide to produce fuels and allows for the
generation of bulk chemicals.[3] CO2 also finds an application
for the production of various chemicals, with technologies for
CH3OH and CH4 synthesis being now commercially available.[4]
However, when these products are used as fuels they cannot
contribute to long-term CO2 fixation. In order to ensure the re-
duction of CO2 emissions, it is highly desirable to develop pro-
cesses/catalysts for converting CO2 into products having a long
lifecycle, for example, polymers, dyes, and resins. Such prod-
ucts are typically produced from ethylene and propylene.
Today, propene is mostly produced through the steam crack-
ing of naphtha and liquid petroleum gas, or fluid catalytic
cracking (FCC) of heavy fractions of crude oils. Since the exist-
ing production capacities cannot completely fulfill the continu-
ously growing propene demand, two purpose-built processes
were commercialized: non-oxidative propane dehydrogenation
(DH)[5] and metathesis of ethylene and 2-butenes.[6] Alternative-
ly, propene can be formed through dehydration of propanol.
The latter is commercially produced in a two-step process in-
volving the high-pressure homogeneous hydroformylation of
ethylene over a Co- or Rh-containing catalyst and consecutive
hydrogenation of the resulting propanal to propanol.[7] Various
heterogeneous catalysts were also applied for gas-phase hy-
droformylation of ethylene, resulting in propanal and propa-
nol.[8] However, all hydroformylation processes suffer from the
usage of toxic CO. This drawback can be overcome when
using CO2 instead of CO. The idea behind our concept is to de-
velop a multifunctional heterogeneous catalyst for CO2 conver-
sion with H2 initially to CO via the reverse water–gas shift
(RWGS) reaction, followed by C2H4 hydroformylation with CO
to propanal and hydrogenation of the latter to propanol
(Scheme 1). A similar concept was recently introduced to pro-
Scheme 1. Schematic representation of concept for 1-propanol (propanol)
formation from CO2, H2, and C2H4.
duce carboxylic acids from CO2, liquid olefins, and H2 over a ho-
mogeneous [RhCl(CO)2]2 catalyst with iodine promoters.[9] Im-
mobilized Ru complexes in ionic liquids are also able to cata-
lyze the hydroformylation of 1-hexene with CO2 and H2.[10]
Developing multi-functional catalysts for production of prop-
anol from CO2, C2H4, and H2 is a big challenge because the
RWGS reaction runs above 573 K[11] due to the thermodynamic
constrains, while the hydroformylation requires lower tempera-
tures.[8a,12] Possible side reactions like Fischer–Tropsch (FT) syn-
thesis, methanation, and C2H4 hydrogenation to C2H6 should
[a] S. J. Ahlers, Dr. U. Bentrup, Dr. D. Linke, Dr. E. V. Kondratenko
Leibniz-Institut fꢀr Katalyse e.V. an der Universitꢁt Rostock
Albert-Einstein-Str. 29A, 18059 Rostock (Germany)
Fax: (+49)381-1281-51290
E-mail: evgenii.kondratenko@catalysis.de
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201402212.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
1
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
also be suppressed. According to literature data,[8a,b,11a,c,12,13]
Cu, Co, Rh, and Au are known to support either RWGS and/or
hydroformylation reactions. However, supported Cu-containing
materials do not catalyze the hydroformylation step,[13d] while
Co-based catalysts are active for the formation of hydrocar-
bons through FT synthesis[14] and catalysts with supported Rh
species tend to completely hydrogenate olefins,[15] in our case
C2H4 to C2H6. Supported Au-containing catalysts appear to be
a good alternative as they catalyze the RWGS reaction at mod-
erate temperature up to chemical equilibrium[16] and are active
and selective in the heterogeneous hydroformylation of olefins
with CO.[8a]
potassium content corresponded to the nominal loading
(Table 1). The bimetallic catalysts are abbreviated as follows:
XK-Au/TiO2_a, where “X” stands for the K weight concentration.
The size of supported Au NP and their distribution were ana-
lyzed by high-angle annular dark field scanning transmission
electron microscopy (HAADF-STEM). Representative images
and the corresponding size distributions are shown in Figure 1
and Figure S1 respectively. The smallest Au NP of 2.5 nm in di-
ameter were identified on the surface of Au/TiO2_c, while the
surface of Au/TiO2_b is populated by significantly larger NP of
approximately 22.5 nm. Au/TiO2_a possesses NP of around
7.5 nm. It is worth mentioning that promoting Au/TiO2_a with
K influences the size of the Au NP. An average diameter de-
creases from 7.5 to 4.5 nm after adding 1 wt% K. The size dis-
tribution is also narrow compared to the unpromoted counter-
part. Upon further increase in K loading, the NP become larger
and their distribution broader: in case of 3K-Au/TiO2_a, differ-
ently sized Au NP with two size maxima at 7.5 nm and at
10 nm were identified (Table 1). Due to the low contrast be-
tween K and Ti, elemental energy dispersive X-ray analysis of
selected regions of the images did not reveal where the pro-
moter is located. However, considering the effect of K on the
size of the Au NP, we can conclude that both these compo-
nents are located close to each other.
Based on this background, the main intention of the present
study was to explore the potential of Au-containing catalysts
for the direct conversion of CO2 into propanol in presence of
C2H4 and H2. For this purpose, we applied different methods
for the deposition of Au nanoparticles (NP) on TiO2 with the
aim to tune their size. The monometallic supported catalysts
were additionally doped with K to check if the promoter influ-
ences NP size and catalyst performance. To test their activity
and selectivity in the target reaction, continuous-flow tests
were performed between 473 and 523 K using reaction feeds
with different ratios of CO2/H2/C2H4. We demonstrated that
promoting Au/TiO2 with K increases the selectivity for CO2 con-
version to propanol up to 100%. This positive effect was thor-
oughly elucidated by in situ FTIR analysis and transient experi-
ments with isotopic tracers.
Steady-state catalytic results
The activity and oxo-selectivity (propanol plus propanal) of
monometallic Au-containing catalysts were elucidated under
continuous-flow conditions at 2 MPa using reaction feeds with
different ratios of CO2/H2/C2H4. Figure 2 shows the conversion
of CO2 and the selectivity to CO and propanol obtained over
monometallic Au/TiO2 catalysts using differently composed
CO2/H2/C2H4 feeds.
Results
Catalyst characterization by BET, ICP, and TEM
The Brunauer–Emmett–Teller specific surface areas (SBET) of
bare support and catalytic materials are listed in Table 1. These
values are in the range of 52 to 67 m2 gꢀ1. There is only a slight
decrease in SBET of the support after Au and/or K deposition,
that is, the support morphology has persisted. Our inductively
coupled plasma optical emission spectrometry (ICP-OES) analy-
sis revealed that Au loading differs from the nominal one of
2 wt%. The exact loading was around 1.3 wt% for all catalysts
with the exception for Au/TiO2_c possessing only 0.5 wt%. The
The selectivity was calculated on CO2 basis, that is, products
originated from this feed component were taken into account.
Irrespective of the feed composition, the highest CO2 conver-
sion was obtained over Au/TiO2_a with Au NP of ca. 7.5 nm,
while Au/TiO2_b and Au/TiO2_c with respectively larger or
smaller NP were less active. As expected from the thermody-
namics of the RWGS reaction, CO2 conversion is positively influ-
enced when operating at an excess of H2 (Figure 2).
Propanol formation was observed over Au/TiO2_
a and Au/TiO2_c, whereas Au/TiO2_b did not produce
Table 1. Catalysts and their selected physicochemical and catalytic properties.[a]
any oxygenates. It is worth mentioning that propanol
[b]
[d]
was only formed when CO2, C2H4, and H2 were pres-
ent in the feed, while CO and H2O were the only
products in the absence of C2H4 (Figure 2). Thus,
propanol is formed from CO2 and C2H4 and not
through CO2 hydrogenation. The presence of hydro-
gen is also highly important because propanol forma-
tion only from CO2 and C2H4 is thermodynamically
Catalyst
Loading [wt%] dAu–NP
SBET
X(CO2) S(CO) S(oxo)[c] rAu
Au
K
[nm]
[m2 gꢀ1
]
[%]
[%]
[%]
[hꢀ1
]
TiO2
–
–
–
58
52
67
50
64
65
–
–
–
–
Au/TiO2_a
Au/TiO2_b
Au/TiO2_c
1K-Au/TiO2_a 1.3
2K-Au/TiO2_a 1.4
3K-Au/TiO2_a 1.3
1.3
1.3
0.5
–
–
–
1.0
2.1
2.9
7.5
22.5
2.5
4.5
7.0
3.6
0.5
0.2
3.3
1.2
0.8
83
100
66
25
0
18
0
34
75
100
100
0.39
0
0.63
8.03
2.76
1.79
7.5 & 10.0 59
0
not possible.
A reaction feed with equimolar
[a] Reaction conditions: 2 MPa, 473 K, contact time of 45 gminLꢀ1, CO2/H2/C2H4 =
1:1:1. [b] dAu–NP was determined from HAADF-STEM (Figure S1). [c] S(oxo) stands for
overall selectivity to propanol and propanal. [d] rAu was calculated according to Equa-
tion (4) at close degrees of CO2 conversion.
amounts of CO2, H2, and C2H4 was identified as opti-
mal to produce propanol. When we used a feed with
an excess of hydrogen as required from the reaction
stoichiometry, side ethylene hydrogenation to ethane
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
2
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
proximately 3.3%, while the con-
version of C2H4 to mainly C2H6
increased from 46 to 62%. The
latter non-desired reaction was
suppressed to 35% upon in-
creasing
with a coincident decrease in
CO2 conversion to 0.8%
(Figure 3). Simultaneously, the
oxo-selectivity increased to
K loading to 3 wt%
100% on 2K- and 3K-Au/TiO2_a.
This result is very promising
from a practical viewpoint be-
cause propanol and propanal
can be easily separated from the
non-reacted feed components,
which are further recycled.
Doping with K also strongly in-
fluenced the rate of propanol
formation. The highest Au-relat-
ed rate of 8.03 hꢀ1 was obtained
over 1K-Au/TiO2_a. Upon a fur-
ther increase in K loading, the
rate decreased but was higher
than over the non-promoted
counterparts (Table 1). It is also
worth mentioning that our cata-
lysts showed high on-stream sta-
bility under different reaction
conditions over 250 h on stream
(Figure S2). Taking this informa-
tion into account together with
the fact that the carbon balance
was not worse than 99%, we
propose that carbon deposition
does not play any significant
role in our approach.
To elucidate mechanistic as-
pects of CO2 transformation to
oxygenates, we performed tests
at different degrees of CO2 con-
Figure 1. HAADF-STEM images of (a) Au/TiO2_a, (b) Au/TiO2_b, (c) Au/TiO2_c, (d) 1K-Au/TiO2_a, (e) 2K-Au/TiO2_a,
and (f) 3K-Au/TiO2_a.
strongly dominated. The highest oxo-selectivity of 34% with
the optimized feed was obtained over Au/TiO2_c followed by
Au/TiO2_a with 18%. However, the former catalyst showed sig-
nificantly lower CO2 conversion. Unfortunately, all catalysts also
hydrogenated C2H4 to C2H6 with the selectivity being always
above 95%. Au/TiO2_a, Au/TiO2_b, and Au/TiO2_c revealed an
ethylene conversion of 46, 18, and 30%, respectively. However,
ethane is not a waste product and can be converted to ethyl-
ene in a cracker.
version by varying contact time (t) from 5 to 45 gminLꢀ1.
Figure 4 demonstrates the effect of t on the selectivity to CO,
propanal, propanol, and propane over the K-Au/TiO2_a cata-
lysts. The corresponding data for CO2 conversion are shown in
Figure S3. As expected, the degree of conversion increased
with an increase in t. CO selectivity over Au/TiO2_a, however,
decreased from 97 to 80% in the investigate rage of t, while
propanol selectivity accordingly increased from 2 to 17%.
Traces of propane were observed at t of 30 and 45 gminLꢀ1.
Propanal was not detected at any contact time. Doping of Au/
TiO2_a with K strongly influenced CO selectivity in the whole t
range. Similarly to Au/TiO2_a, this selectivity decreased over
1K-Au/TiO2_a from 80 to 29% upon increase in t from 5 to
45 gminLꢀ1. The propanol selectivity increased from 20 to
70%. However, propane was not detected, while the selectivity
to propanal passed over a maximum with rising t (Figure 4b).
Having generated the potentially propanol-forming
Au/TiO2_a catalyst, we turned our attention to optimizing its
selectivity. To this end, it was doped with different amounts of
K, which is an important promoter for CO2-FT catalysts.[17]
Doping Au/TiO2_a with 1 wt% K resulted in an increase in oxo-
selectivity from 18 to 78% with the related decrease in CO se-
lectivity (Figure 3). CO2 conversion did not change and was ap-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
3
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
*
*
*
Figure 4. Effect of t on selectivity to CO ( ), propanal ( ), propanol ( ), and
*
propane ( ) over (a) Au/TiO2_a, (b) 1K-Au/TiO2_a, (c) 2K-Au/TiO2_a, and
(d) 3K-Au/TiO2_a at 2 MPa and 473 K using a CO2/H2/C2H4 =1:1:1 feed.
CO2, with the former being the main oxygenate in the whole
range of t tested. In agreement with Au/TiO2_a and 1K-Au/
TiO2_a, propanol selectivity increased but started from around
80% at the shortest t of 5 gminLꢀ1. The selectivity to propanal
decreased correspondingly. The relationship between selectivi-
ty and contact time shown in Figure 4 is further analyzed in
the discussion section together with the results presented in
the following section to develop a mechanistic scheme of CO2
conversion into propanal and propanol in the presence of C2H4
and H2.
Effect of doping of Au/TiO2_a with K on CO2 and H2 adsorp-
tion
Figure 2. Conversion of CO2 and selectivity (S) to CO and propanol on Au/
TiO2_a (dark grey bar), Au/TiO2_b (grey bar), and Au/TiO2_c (light grey bar)
against various reaction feeds. x:x:x represents the ratio of CO2/H2/C2H4 in
To clarify the effect of K doping on CO2 conversion, we ana-
lyzed the interaction of CO2 with K-Au/TiO2_a by means of
in situ FTIR spectroscopy and CO2-pulse experiments in the
TAP (temporal analysis of products) reactor. Figure 5 shows the
FTIR spectra recorded after CO2 adsorption on Au/TiO2_a, 1K-
Au/TiO2_a, and 3K-Au/TiO2_a at 473 K. Carboxylate species
(1550 and 1446 cmꢀ1) were identified on the surface of Au/
TiO2_a. Doping this catalyst with 1 wt% K influenced CO2 ad-
sorption in such a way that bidentate carbonates (1581 and
1326 cmꢀ1) were additionally formed. The strength of CO2 in-
teraction with K-Au/TiO2_a further increases with K loading as
concluded from the fact that, besides the bidentate carbonate
species, ionic carbonates (1390 cmꢀ1) were also identified on
the surface of 3K-Au/TiO2_a after its reaction with CO2. To
study the stability of surface CO2-containing species, we ana-
lyzed their reduction by H2 at 473 K. For this purpose, H2 was
fed to the CO2-treated catalyst at the same temperature
(473 K), followed by purging with He. The results obtained are
summarized in Figure 6. Carboxylate species on the surface of
Au/TiO2_a were easily destroyed after reaction with H2 but
were again formed after exposure to a CO2/H2/C2H4 =1:1:1
feed (Figure 6a). In the case of the K-doped catalysts, biden-
tate and ionic carbonate species were only partly decomposed
the feeds. Reaction conditions: 2 MPa, 473 K, contact time of 45 gminLꢀ1
.
Figure 3. Selectivity to CO (blue bar), propanal (yellow bar), and propanol
(green bar) against K content on Au/TiO2_a. Reaction conditions: 2 MPa,
473 K, contact time of 45 gminLꢀ1, CO2/H2/C2H4/N2 =1:1:1:1.
For catalysts with K loading of 2 and 3 wt%, CO and pro-
pane were not detected at any contact time (Figure 4c–d).
Propanol and propanal were the only products formed from
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
4
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
by CO2-pulse experiments in the TAP reactor. In the latter ex-
periments, a CO2/Ar=1:1 mixture was pulsed over the K-Au/
TiO2_a catalysts at 473 K in vacuum. The height-normalized re-
sponses of CO2 are shown in Figure S4. Normalization was per-
formed for accurate comparison of their shape. One can clearly
see that these responses exhibit a broad tailing (non-zero con-
centration at high times). According to the theory of the TAP
reactor,[18] this tailing means that CO2 adsorbs and desorbs.
However, these processes occur with different rates depending
on K loading. Since the tailing increases with rising concentra-
tion of K in K-Au/TiO2_a, we can safely conclude that the ratio
of the rates of adsorption to desorption increases correspond-
ingly. From a chemical viewpoint, the K effect is related to re-
ducing the acidity of the TiO2 surface while simultaneously in-
creasing its basicity. Owing to the acidic properties of CO2, its
adsorption is favored in presence of basic sites.
Figure 5. FTIR spectra of CO2 adsorbed on (a) Au/TiO2_a, (b) 1K-Au/TiO2_a,
and (c) 3K-Au/TiO2_a at 473 K.
Since H2 is required for the conversion of CO2 and C2H4 into
C3-oxygenates, we also investigated mechanistic aspects of ac-
tivation of gas-phase H2 on the surface of K-Au/TiO2_a catalysts.
Two types of experiments were performed in the TAP reactor:
i) single pulsing of an H2/Ne=1:1 mixture and ii) simultaneous
pulsing of H2/Ne=1:1 and D2/Ar=1:1 mixtures. We shall start
with the description of the first experiments. The type of H2 in-
teractions with supported Au NP was elucidated as suggested
by Gleaves et al.[18] To this end, the experimental responses of
H2 and Ne were transformed into a dimensionless form. The
modified response of Ne characterizes the solely diffusion pro-
cess and is called the standard diffusion curve. Such a transfor-
mation of experimental data helps to discriminate between re-
versible and irreversible adsorption of H2 pulsed together with
Ne. In Figure 7(a–b) the dimensionless responses of Ne and H2
Figure 6. FTIR spectra after reacting (i) CO2, (ii) H2, and (iii) CO2/H2/C2H4 =1:1:1
with (a) Au/TiO2_a, (b) 1K-Au/TiO2_a, and (c) 3K-Au/TiO2_a at 473 K.
after reaction with H2. These species also dominate the surface
of 1K-Au/TiO2_a and 3K-Au/TiO2_a under the CO2/H2/C2H4 =
1:1:1 reaction conditions.
Figure 7. Dimensionless responses after pulsing of (a,b) H2/Ne=1:1 and
(c,d) H2/Ne=1:1 and D2/Ar=1:1 over (a,c) Au/TiO2_a and (b,d) 1K-Au/TiO2_
a at 473 K. HD signal in (d) is multiplied by a factor of 10 for better illustra-
tion.
Thus, promoting of Au/TiO2_a with K increases the strength
of CO2 adsorption. This conclusion was additionally supported
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
5
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
after pulsing of an H2/Ne=1:1 mixture over Au/TiO2_a and
1K-Au/TiO2_a at 473 K are compared. The corresponding re-
sults for 2K-Au/TiO2_a are shown in Figure S5. Irrespective of K
loading, the H2 response crosses the Ne response and is locat-
ed below it. This means that H2 reacts reversibly. To derive fur-
ther insights into the H2 adsorption, that is, whether it is disso-
ciative or non-dissociative, we also investigated H/D exchange
by means of H2/Ne=1:1 and D2/Ar=1:1 pulse experiments.
HD was detected only in traces over Au/TiO2_a, while signifi-
cant amounts of HD were observed over the K-containing
counterparts (Figure 7c–d and Figure S5b). It should be espe-
cially highlighted that Au-free 2K/TiO2 did not show any activi-
ty for H/D exchange (Figure S6). Therefore, the presence of HD
on K-Au/TiO2_a indicates that K promotes dissociation of H2
and D2 on the surface of Au NP to yield adsorbed H species,
which can recombine to HD.
partial CO pressure. The corresponding values are given in
Table 1. One can clearly see that the presence of K strongly in-
creases this reaction rate. Importantly, propanol hydrogenation
to propane is also suppressed over K-containing catalysts.
Based on our results and previous literature on the RWGS re-
action[11c] and C2H4 hydroformylation with CO,[8a,b] we suggest
the following mechanistic scheme of propanol formation from
CO2 and C2H4 under consideration of TiO2 surface free of Au
NP. CO2 adsorbs reversibly over the support, while H2 and C2H4
are activated on Au NP.[19] The oxidation state of Au under re-
action conditions is supposed to be zero as concluded from
our X-ray diffraction analysis of used catalysts and previous lit-
erature studies.[11c,20] For thermodynamic reasons, we do not
expect formation of metallic potassium. Surface hydrogen spe-
cies formed on Au NP from gas-phase H2 reduce CO2 adsorbed
in the vicinity of Au NP to yield CO and H2O. From a reactivity
viewpoint, carboxylate species, which are only present on the
surface of K-free TiO2, are more easily converted to CO than bi-
dentate and ionic carbonates, which are formed on the cata-
lysts promoted with high (>1 wt%) K loading. This is a reason
for the low CO2 conversion over K-Au/TiO2_a catalysts with 2
and 3 wt% K (Table 1). Since the rate of CO formation was not
influenced by C2H4 presence, we propose that the RWGS reac-
tion and ethylene activation run independently on each other.
The former reaction probably takes place on the edge of the
NP while ethylene hydroformylation to propanal and hydroge-
nation to ethane compete for the same active sites located on
the surface of NP. Both of these ethylene transformations are
initiated by the formation of a surface ethyl species through
addition of one H atom to adsorbed ethylene.[8a] This inter-
mediate species can react with either another H species to
yield undesired ethane or with gas-phase CO to result in an
adsorbed CH3CH2CO.[8a,b] The latter species was observed in
our FTIR experiments and is characterized by a typical carbonyl
band at approximately 1700 cmꢀ1 (Figure S7). In heterogene-
ous C2H4 hydroformylation with CO,[8a,b] an adsorbed ethyl spe-
cies is formed very rapidly, while reaction of CH3CH2CO with
surface H species limits the formation of propanal.[8a] Bearing
this in mind, we suggest that promoting Au/TiO2_a with K ac-
celerates the latter reaction through an increased generation
of adsorbed H species as proven by H/D exchange (Figure 7
and Figure S5). Importantly, doping with K simultaneously in-
creases the strength of adsorbed CO2 species, thus lowering
the rate of formation of CO required for the hydroformylation
step. Therefore, an optimal loading is required to achieve high
activity to the target product.
Discussion
The results of the steady-state catalytic tests (above) are dis-
cussed to elaborate a mechanistic scheme of CO2 conversion
with C2H4 to propanal/propanol in the presence of H2. The rela-
tionships between products selectivity and CO2 conversion
over Au/TiO2_a and 1K-Au/TiO2_a in Figure 4 show that CO2 is
initially reduced to CO followed by consecutive CO reaction
with C2H4 to yield propanal with the latter being further hydro-
genated to propanol (Scheme 2). The absence of propanal is
Scheme 2. Reaction pathways of CO2 conversion to propanal/propanol in
presence of C2H4 and H2.
related to its rapid hydrogenation to propanol as a result of
the high hydrogenation activity of Au NP as supported by the
formation of propane, which is the end product of such a hy-
drogenation process. Since CO was the only product when
using a CO2/H2/N2/=1:1:2 feed (Figure S2), the absence of CO
over 2K-Au/TiO2_a and 3K-Au/TiO2_a in presence of C2H4 and
H2 (Figure 4c–d) is strong evidence for its reaction with C2H4 to
propanal, which was observed over these catalysts with high
selectivity at low contact times. From a kinetic viewpoint, the
hydroformylation reaction appears to be faster than the RWGS
reaction over the K-promoted catalysts but not over the K-free
counterpart. It is worth mentioning that the selectivity to
propanal decreases with rising t, while the selectivity to propa-
In addition to the effect of K, the size of Au NP has a mani-
fold effect on the catalyst activity for both CO2 conversion to
propanal/propanol and ethylene hydrogenation to ethane.
First of all, an appropriate size of NP was necessary to make
catalysts active for the target reaction. Au/TiO2_b with the
20 nm NP was not able to support the propanal synthesis,
while Au/TiO2_c with the smallest NP of 2.5 nm showed signifi-
cantly lower activity for the formation of oxygenates than the
catalyst possessing NP of 7.5 nm. Furthermore, the size of Au
NP controls the C2H4 conversion to C2H6 as shown in Figure 8;
the conversion passes over a maximum for Au NP of ca. 5 nm.
nol increases and approaches almost 100% at
t of
45 gminLꢀ1, implying that propanol is formed through consec-
utive propanal hydrogenation. Thus, promoting Au/TiO2_a with
K does not change the overall scheme of CO2 conversion into
propanol shown in Scheme 2. We put forward that the pres-
ence of K strongly increases the kinetics of hydroformylation of
C2H4 with CO formed initially from CO2. This statement is sup-
ported by the reaction rate of oxygenate formation deter-
mined at a close degree of CO2 conversion, that is, at a similar
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
6
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Experimental Section
Monometallic Au-containing catalysts with the nominal metal load-
ing of 2 wt% were synthesized according to the methods previous-
ly described in literature using ammonia,[22] urea,[22] and ammonia/
magnesium citrate[23] as reducing agents. Briefly, Au/TiO2_a was
prepared by deposition-precipitation of gold hydroxide from
HAuCl4 (41.1 wt% Au, Chempur) on TiO2 (Anatase, BASF) calcined
at 773 K for 8 h. Ammonia solution (25%, Roth) was added to the
suspension of TiO2 in water solution of HAuCl4 until a pH value of
11 was reached. For preparation of Au/TiO2_b, the deposition–pre-
cipitation of gold hydroxide was done with a urea solution (ꢁ
99%, Roth) instead of ammonia. Au/TiO2_c was synthesized similar-
ly to Au/TiO2_a additionally using a saturated magnesium citrate
solution (ꢁ95.0%, Fluka). All catalysts were filtrated, washed, and
finally dried at 373 K. K-Au/TiO2_a catalysts were prepared by incip-
ient wetness impregnation of dried Au/TiO2_a with a KNO3 solution
(>99%, Merck). The K loading was amounted to 1, 2, and 3 wt%.
All catalysts were finally calcined in air at 473 K for 4 h and then
pressed and sieved to yield a fraction of 250–450 mm.
Figure 8. Effect of size of Au NP on C2H4 conversion to mainly C2H6 over dif-
ferent catalysts at 2 MPa and 473 K using a CO2/H2/C2H4 =1:1:1 feed.
The lower activity of smaller NP is in a good agreement with
a previous study on selective hydrogenation of acetylene in
excess of ethylene over Au/Al2O3.[21] In that paper, Au NP of
3 nm were found to be unable to hydrogenate C2H4.
The weight concentration of Au and K in calcined catalysts was de-
termined by ICP-OES using a Varian 715 emission spectrometer.
Nitrogen adsorption-desorption isotherm was collected at 77 K on
BELSORP-mini II (BEL Japan, Inc.). The specific surface area was cal-
culated from the adsorption applying the Brunauer, Emmett and
Teller equation for the N2 relative pressure range of 0.05<P/P0 <
0.30.
Conclusions
Heterogeneous C2H4 hydroformylation with CO2 to propanal/
propanol was for the first time demonstrated over TiO2-sup-
ported Au-containing catalysts in a single continuous-flow re-
actor. Since propanol can be easily dehydrated to propene, our
concept opens a new way for production of this important
building block of chemical industry. Another important feature
of this concept is its sustainability because it functionalizes
CO2 with C2H4, which can be produced from bioethanol.
Our thorough kinetic and mechanistic studies enabled us to
develop an overall reaction scheme of products formation. CO2
adsorption takes place on bare and K-modified TiO2, while H2
and C2H4 are activated on Au NP. In a first step, CO2 adsorbed
in vicinity of Au undergoes reduction to CO through reverse
water–gas shift reaction. So in situ generated CO is further
used for classical hydroformylation of ethylene to propanal
with the latter being finally hydrogenated to propanol.
HAADF-STEM investigations were performed using an aberration-
corrected JEM-ARM200F microscope operated at 200 kV and
equipped with a JED-2300 energy-dispersive X-ray spectrometer
(EDXS, JEOL) for elemental analysis. The Cs-corrector CESCOR
(CEOS) was used for scanning transmission (STEM) applications.
Samples were deposed on a holey carbon-supported copper grid
(mesh 300). To estimate size distribution of supported Au NP, at
least 100 particles were counted.
The in situ FTIR measurements were carried out in transmission
mode using
a Thermo Scientific Nicolet 6700 spectrometer
equipped with an in-house-developed reaction cell with CaF2 win-
dows connected to a gas-dosing system. The sample powders
were pressed into self-supporting wafers with a diameter of
20 mm and a weight of 50 mg. All spectra were recorded with
a resolution of 4 cmꢀ1 and 64 scans according to the following pro-
tocol. Before FTIR measurements, each sample was heated in a He
flow (50 mLminꢀ1) up to 473 K for 30 min. After this treatment, the
cell was flushed in CO2 (10 mLminꢀ1 CO2 plus 60 mLminꢀ1 He) for
5 min at the same temperature and then closed for 60 min. Subse-
quently, He was fed to the cell for 10 min to remove gaseous spe-
cies followed by feeding H2 (10 mLminꢀ1 H2 plus 60 mLminꢀ1 He)
for 5 min. Again, the cell was closed for 60 min. After this reductive
treatment, the catalyst was flushed in a He flow for 10 min and
then exposed to a CO2/H2/C2H4/He=1:1:1:4 mixture with a total
flow of 70 mLminꢀ1 for 5 min. The catalyst was treated in this mix-
ture under static conditions for 60 min followed by flushing with
He for 10 min. FTIR spectra were recorded after each specific treat-
ment and removal reactive gas-phase species by He. In separate
experiments with fresh samples, propanal (97%, Sigma Aldrich)
was adsorbed at 473 K after pretreatment of the catalysts in a He
flow (50 mLminꢀ1) at 473 K for 30 min followed by feeding H2 for
60 min and flushing with He for 10 min. Propanal was dosed using
Doping of Au/TiO2 with K resulted in a 100% oxo-selectivity
with respect to CO2 conversion at 2 MPa and 473 K using
a CO2/H2/C2H4 =1:1:1 feed. The promoter was established to
influence the size of supported Au NP and the strength of CO2
adsorption. The former catalyst property, which can also be
controlled by the method of gold deposition on TiO2, deter-
mines the activity of catalyst for non-selective ethylene hydro-
genation to ethane: too small (<3 nm) and too large (>
15 nm) NP show low activity to activate C2H4. The positive
effect of K on oxo-selectivity is of a kinetic nature as doping
strongly increases the rate of ethylene hydroformylation owing
to an accelerated formation of adsorbed H species required for
propanol formation from adsorbed CH3CH2CO. Besides the im-
provement of oxo-selectivity, K doping lowers CO2 conversion
because CO2 is strongly adsorbed on basic sites to yield ionic
carbonates, which only slowly decompose to the CO required
for the hydroformylation step.
a
saturator (ambient temperature) and a gas flow of CO2
(10 mLminꢀ1) and He (60 mLminꢀ1). The adsorption procedure was
the same as described above. Background-subtracted adsorbate
spectra were obtained by subtraction of the spectrum after pre-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
7
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
treatment from the spectrum after propanal adsorption and He
flushing.
(Knudsen diffusion regime), any collisions between gas-phase spe-
cies are minimized and therefore, pure heterogeneous reaction
steps were analyzed.
Catalytic tests were performed at 2 MPa between 473 and 623 K in
an in-house developed setup consisting of 51 continuous-flow
fixed-bed reactors operating in parallel. Reactants are CO2 (4.5, Air
Liquide), H2 (5.0, Air Liquide), C2H4 (3.0, Linde), and N2 (5.0, Air Liq-
uide). 300 mg of catalyst (fraction of 250–450 mm) were placed in
stainless-steel tube reactors (i.d. 4 mm) with threefold amount of
SiC for dilution and heated up to 523 K in a N2 flow (6.67 mLminꢀ1
per reactor). Subsequently, the N2 flow was replaced by a flow
(6.67 mLminꢀ1 per reactor) of CO2/H2/C2H4/N2 =4:2:1:1 reaction
feed for 24 h. Then, various CO2/H2/C2H4/N2 feeds were sequentially
fed to the reactors at 473 K and 523 K. In order to derive insights
into the effect of CO2 conversion on product selectivity using
a CO2/H2/C2H4/N2 =1:1:1:1 feed, the modified contact time was
varied from 5 to 45 gminLꢀ1. Steady-state data were collected
after at least 10 h stabilization under each condition. The feed
components and the reaction products were analyzed by an on-
line GC (Varian CP-3800) equipped with FID (HP-Plot Q) and TCD
(HP-Plot Q&Molsieve 5A). The conversion of C2H4 and CO2 was cal-
culated from their inlet and outlet concentrations [Eq. (1)]. The se-
lectivity and yield to propanol, propanal, propane, and CO were
calculated on CO2 basis using Equations (2) and (3), respectively,
taking into account that one C3-hydrocarbon/oxygenate molecule
is formed from one CO2 and one C2H4. Calculation of C2H6 yield
was performed on C2H4 basis.
The reaction mixtures were prepared using Ar (5.0), Ne (5.0), CO2
(4.5), H2 (5.0), and D2 (2.7, Messer Griesheim) without additional pu-
rification. A quadrupole mass spectrometer (HAL RC 301 Hiden An-
alytical) was applied for quantitative analysis of feed components
and reaction products. The following AMUs (atomic mass units)
were used for mass-spectrometric identification of different com-
pounds: 44 (CO2), 4 (D2), 3 (HD), 2 (H2), 20 (Ne, Ar), and 40 (Ar).
Pulses were repeated 10 times for each AMU and averaged to im-
prove the signal-to-noise ratio. The concentration of feed compo-
nents and reaction products were determined from the respective
AMUs using standard fragmentation patterns and sensitivity fac-
tors.
Acknowledgements
The authors thank Dr. Marga-Martina Pohl for carrying out cata-
lyst characterization by HAADF-STEM, Dr. Vita Kondratenko for
H2 adsorption and H/D exchange tests in the TAP reactor and
M.Sc. Christine Rautenberg for in situ FTIR experiments. This work
was supported by the Leibniz-Gemeinschaft under the research
grant SAW-2011-LIKAT-2.
_
_
nf;0 ꢀ nf
Keywords: carbon dioxide · gold · hydroformylation · reverse
water–gas shift · tap reactor
Xf ¼
Yi ¼
Si ¼
ð1Þ
ð2Þ
ð3Þ
_
nf;0
_
_
ni ꢀ ni;0
[1] C. Federsel, R. Jackstell, M. Beller, Angew. Chem. 2010, 122, 6392–6395;
Angew. Chem. Int. Ed. 2010, 49, 6254–6257.
[2] a) M. He, Y. Sun, B. Han, Angew. Chem. 2013, 125, 9798–9812; Angew.
Chem. Int. Ed. 2013, 52, 9620–9633; b) W. Wang, S. Wang, X. Ma, J.
Gong, Chem. Soc. Rev. 2011, 40, 3703–3727.
_
nf;0
Yi
Xf
[3] P. Y. Dapsens, C. Mondelli, J. Pꢁrez-Ramꢂrez, ACS Catal. 2012, 2, 1487–
1499.
_
_
where nf and ni stand for mole flows of feed components and re-
action products, respectively. Subscript 0 is used for inlet mole
flows.
[4] a) G. A. Olah, Angew. Chem. Int. Ed. 2013, 52, 104–107; Angew. Chem.
2013, 125, 112–116; b) G. Centi, S. Perathoner, ChemSusChem 2010, 3,
195–208; c) M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann, F. E.
Kuhn, Angew. Chem. Int. Ed. 2011, 50, 8510; Angew. Chem. 2011, 123,
8662; d) E. V. Kondratenko, G. Mul, J. Baltrusaitis, G. O. Larrazꢃbal, J.
Pꢁrez-Ramꢂrez, Energy Environ. Sci. 2013, 6, 3112–3135.
[5] K. J. Caspary, H. Gehrke, M. Heinritz-Adrian, M. Schwefer in Handbook of
Heterogeneous Catalysis, Vol. 7 (Eds.: G. Ertl, H. Knçzinger, F. Schꢄth, J.
Weitkamp), Wiley-VCH, Weinheim, 2008, pp. 3206–3229.
[6] J. C. Mol, P. W. N. M. van Leeuwen in Handbook of Heterogeneous Catal-
ysis, Vol. 7 (Eds.: G. Ertl, H. Knçzinger, F. Schꢄth, J. Weitkamp), Wiley-
VCH, Weinheim, 2008, pp. 3240–3256.
The Au-related reaction rate was calculated using Equation (4)
xPropanol ꢂ F ꢂ MAu
ð4Þ
rAu
¼
Vm ꢂ mAu
where xPropanol is the mole fraction of propanol in the outlet feed
and Vm the molar volume. mAu represents the amount of gold on
the corresponding catalyst and F is the total feed flow.
Mechanistic aspects of CO2 and H2 interaction with TiO2, Au/TiO2_a,
1K-Au/TiO2_a and 2K-Au/TiO2_a were investigated in the temporal
analysis of products (TAP-2) reactor, a time-resolved technique with
a resolution of approximately 100 ms. The TAP reactor system has
been described in detail elsewhere.[18,24] The catalysts (70 mg) were
packed into the quartz tube reactor (internal diameter=6 mm,
length=40 mm) as a thin layer between two layers of quartz parti-
cles (sieve fraction of 250–350 mm) within the isothermal zone of
the reactor. Before starting pulse experiments, the catalysts were
pretreated at ambient pressure in a flow of 5 vol% H2 in Ar
(20 mLminꢀ1) at 523 K with a ramp of 10 Kminꢀ1. Hereafter, the re-
actor was evacuated to ca. 10ꢀ5 Pa and cooled down to 473 K in
vacuum. Tree types of experiments were performed: (i) single puls-
ing of a H2/Ar=1:1 mixture, (ii) single pulsing of a CO2/Ar=1:1 mix-
ture, and (iii) simultaneous pulsing of H2/Ne=1:1 and D2/Ar=1:1
mixtures. A fresh sample was used for each experiment. The pulse
size of Ar or Ne was around 1015 molecules. Under such conditions
[7] A. J. Papa in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH,
Weinheim, 2000.
[8] a) X. Liu, B. Hu, K. Fujimoto, M. Haruta, M. Tokunaga, Appl. Catal. B
2009, 92, 411–421; b) N. Sivasankar, H. Frei, J. Phys. Chem. C 2011, 115,
7545–7553; c) N. Navidi, J. W. Thybaut, G. B. Marin, Appl. Catal. A 2014,
469, 357–366.
[9] T. G. Ostapowicz, M. Schmitz, M. Krystof, J. Klankermayer, W. Leitner,
Angew. Chem. Int. Ed. 2013, 52, 12119–12123; Angew. Chem. 2013, 125,
12341–12345.
[10] K.-i. Tominaga, Catal. Today 2006, 115, 70–72.
[11] a) C.-S. Chen, W.-H. Cheng, S.-S. Lin, Appl. Catal. A 2004, 257, 97–106;
b) L. Wang, S. Zhang, Y. Liu, J. Rare Earths 2008, 26, 66–70; c) R. Burch,
Phys. Chem. Chem. Phys. 2006, 8, 5483–5500.
[12] T. Hanaoka, H. Arakawa, T. Matsuzaki, Y. Sugi, K. Kanno, Y. Abe, Catal.
Today 2000, 58, 271–280.
[13] a) T. A. Kainulainen, M. K. Niemelꢅ, A. O. I. Krause, Catal. Lett. 1998, 53,
97–101; b) D. Y. Murzin, A. Bernas, T. Salmi, AIChE J. 2012, 58, 2192–
2201; c) T. Kim, F. Celik, D. Hanna, S. Shylesh, S. Werner, A. Bell, Top.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
8
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Catal. 2011, 54, 299–307; d) M. Beller, B. Cornils, C. D. Frohning, C. W.
Kohlpaintner, J. Mol. Catal. A 1995, 104, 17–85.
[19] F. Boccuzzi, A. Chiorino, M. Manzoli, D. Andreeva, T. Tabakova, J. Catal.
1999, 188, 176–185.
[20] A. Goguet, R. Burch, Y. Chen, C. Hardacre, P. Hu, R. W. Joyner, F. C. Meu-
nier, B. S. Mun, D. Thompsett, D. Tibiletti, J. Phys. Chem. C 2007, 111,
16927–16933.
[21] J. Jia, K. Haraki, J. N. Kondo, K. Domen, K. Tamaru, J. Phys. Chem. B 2000,
104, 11153–11156.
[22] N. D. Tran, M. Besson, C. Descorme, New J. Chem. 2011, 35, 2095–2104.
[23] D. A. H. Cunningham, W. Vogel, H. Kageyama, S. Tsubota, M. Haruta, J.
Catal. 1998, 177, 1–10.
[14] Y. Zhang, G. Jacobs, D. E. Sparks, M. E. Dry, B. H. Davis, Catal. Today
2002, 71, 411–418.
[15] M. Guerrero, N. T. Than Chau, S. Noel, A. Denicourt-Nowicki, F. Hapiot, A.
Roucoux, E. Monflier, K. Philippot, Curr. Org. Chem. 2013, 17, 364–399.
[16] H. Sakurai, A. Ueda, T. Kobayashi, M. Haruta, Chem. Commun. 1997,
271–272.
[17] a) R. W. Dorner, D. R. Hardy, F. W. Williams, H. D. Willauer, Appl. Catal. A
2010, 373, 112–121; b) R. W. Dorner, D. R. Hardy, F. W. Williams, H. D.
Willauer, Catal. Commun. 2011, 15, 88–92; c) M. Niemelꢅ, M. Nokkosmꢅ-
ki, Catal. Today 2005, 100, 269–274; d) Y. Yang, H.-W. Xiang, Y.-Y. Xu, L.
Bai, Y.-W. Li, Appl. Catal. A 2004, 266, 181–194.
[24] J. Pꢁrez-Ramꢂrez, E. V. Kondratenko, Catal. Today 2007, 121, 160–169.
Received: March 21, 2014
Revised: April 17, 2014
Published online on && &&, 0000
[18] J. T. Gleaves, G. S. Yablonskii, P. Phanawadee, Y. Schuurman, Appl. Catal.
A 1997, 160, 55–88.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&
9
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
S. J. Ahlers, U. Bentrup, D. Linke,
E. V. Kondratenko*
A flowing conversion: CO2 can be con-
tinuously converted into propanol with
a near 100% selectivity through H2-as-
sisted coupling with C2H4 over K-pro-
moted Au/TiO2 in a single continuous-
flow reactor. Thorough kinetic and
mechanistic studies reveal the overall
reaction scheme of products formation.
The activity and selectivity strongly
depend on nanoparticle size, which can
be tuned by the method of Au deposi-
tion and by doping with K.
&& – &&
An Innovative Approach for Highly
Selective Direct Conversion of CO2
into Propanol using C2H4 and H2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 10
&10
&
ÞÞ
These are not the final page numbers!