Propanol formation from CO2 and C2H4 with H2 over Au/TiO2: Effect of support and K doping

Article abstract of DOI:10.1016/j.cattod.2015.04.006

The direct conversion of CO₂ with C₂H₄ and H₂ into propanol was investigated over K-promoted Au/TiO₂-r (rutile) and Au/TiO₂-a (anatase) catalysts at 473 K and 523 K. For both catalytic systems, the propanol selectivity increased with rising K loading, while the conversion of CO₂ decreased. Both effects of the promoter diminished at 523 K. However, irrespective of K loading and reaction temperature, K-Au/TiO₂-r materials showed higher propanol selectivity than K-Au/TiO₂-a. In addition, the non-desired conversion of ethylene into ethane over the rutile-based catalysts decreased with rising temperature from 473 K to 523 K, while an increase was observed for the anatase-based materials. Taking into account the results of HAADF-STEM analysis, the metal-support interaction influencing both size and distribution of Au species was assumed to be a reason for the support effect on the target and side reactions.

Full text of DOI:10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Propanol formation from CO and C H with H over Au/TiO :
2
2
4
2
2
Effect of support and K doping
a
b
a
a
Stefan J. Ahlers , Ralph Kraehnert , Carsten Kreyenschulte , Marga-Martina Pohl ,
a
a,∗
David Linke , Evgenii V. Kondratenko
Leibniz-Institut für Katalyse e.V. an der Universität Rostock, Albert-Einstein-Str. 29a, 18059 Rostock, Germany
Technische Universität Berlin, Straße des 17. Juni 124, 10623 Berlin, Germany
a
b
a r t i c l e i n f o
a b s t r a c t
Article history:
2
The direct conversion of CO with C H and H into propanol was investigated over K-promoted Au/TiO r
(
2
2
4
2
Received 30 October 2014
Received in revised form 31 March 2015
Accepted 2 April 2015
rutile) and Au/TiO2 a (anatase) catalysts at 473 K and 523 K. For both catalytic systems, the propanol
selectivity increased with rising K loading, while the conversion of CO2 decreased. Both effects of the
promoter diminished at 523 K. However, irrespective of K loading and reaction temperature, K–Au/TiO2 r
materials showed higher propanol selectivity than K–Au/TiO2 a. In addition, the non-desired conversion
of ethylene into ethane over the rutile-based catalysts decreased with rising temperature from 473 K to
Available online xxx
Keywords:
Carbon dioxide
Gold
5
23 K, while an increase was observed for the anatase-based materials. Taking into account the results of
HAADF-STEM analysis, the metal-support interaction influencing both size and distribution of Au species
was assumed to be a reason for the support effect on the target and side reactions.
© 2015 Elsevier B.V. All rights reserved.
TiO2
Hydroformylation
Ethylene
Propanol
1
. Introduction
Over the last years, the economic and particularly environ-
exothermic with the enthalpy value being approximately 3 times
higher than for the methanol synthesis from CO . It is also worth
2
mentioning that ethylene required for propanol synthesis can be
easily obtained from bio-ethanol through its dehydration [17,18],
while hydrogen can be produced from water [19–21]. Thus, this
new approach does not depend on fossil non-renewable carbon
sources and can be considered as sustainable.
mental incentives motivated the search for alternative carbon
sources for industrial production of fuels and commodity chemi-
cals. Biomass [1–3] and carbon dioxide [4–7] are considered to be
such alternatives. Up to date, there are commercial processes for
converting biomass into biofuels [1], or in more particular, into bio-
ethanol [8,9]. CO2 can be hydrogenated to methane or methanol
K = −167 kJ mol−1
CO + C H + 3 H → C H7OH + H O(ꢀrH
)
2
2
4
2
3
2
573
[
7,10,11]. Methanol can be used for the production of ethylene and
propylene through a methanol to olefins technology [12,13]. These
olefins are two most important building blocks in the chemical
(1)
industry. However, the conversion of CO into these olefins consists
2
of several process steps with their own advantages and drawbacks
From a mechanistic viewpoint, propanol is formed from CO2
according to a scheme consisting of three consecutive reactions
[
14].
Recently, we have introduced an approach [15] for the direct
[
15]. CO2 is primarily converted into CO via the reverse water-gas
conversion of CO2 into 1-propanol (propanol), which is one of the
basic chemicals for the daily life [16]. Moreover, this alcohol can
be easily and selectively dehydrated to propylene thus opening a
new way for the production of this important olefin. Besides CO2
as a reactant, C H and H are also needed (Eq. (1)). This reaction is
shift (RWGS) reaction and then the in situ produced CO partici-
pates in the hydroformylation of C H4 to propanal in the presence
2
of H . In the last step, the intermediate propanal is hydrogenated to
2
propanol. Au nanoparticles (NP) supported on TiO anatase (TiO2 a)
2
2
4
2
were established to catalyze all these steps finally yielding propanol
with near-to 100% selectivity on the basis of CO when the catalysts
2
were additionally promoted by potassium oxide. However, such
materials suffer from low propanol productivity and high activity
for non-desired ethylene hydrogenation into ethane.
∗ Corresponding author. Tel.: +49 3811281290; fax: +49 381128151290.
E-mail address: Evgenii.Kondratenko@catalysis.de (E.V. Kondratenko).
http://dx.doi.org/10.1016/j.cattod.2015.04.006
0
920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
S.J. Ahlers et al. / Catalysis Today xxx (2015) xxx–xxx
2
The purpose of the present contribution was to further explore
the potential of CO2 conversion with ethylene and hydrogen into
propanol. The focus was set (i) to increase the conversion of CO2
into propanol without significant loss of high selectivity, i.e. to
improve propanol yield and (ii) to analyze factors governing non-
selective ethane formation from ethylene. To this end, we used TiO2
Scanning electron microscopy on a SEM, JEOL 7401F operated at
10.0 kV and equipped with a Bruker-AXS XFlash® 4030 was applied
for analyzing surface morphology of the powder particles and for
estimating size and distribution of Au NP. The latter information
was additionally obtained from STEM measurements of samples
at 200 kV with an aberration-corrected JEM-ARM200F (JEOL, Cor-
rector: CEOS). The microscope is equipped with a JED-2300 (JEOL)
energy-dispersive X-ray-spectrometer (EDXS) for chemical anal-
ysis. The aberration corrected STEM imaging (high-angle annular
dark field (HAADF) and annular bright field (ABF)) were per-
formed under the following conditions. HAADF and ABF both were
done with a spot size of approximately 0.13 nm, a convergence
(
rutile or anatase) supported catalysts with a fixed amount of Au
and different K loading for elucidating the effect of support mor-
phology on Au NP’s size and distribution as well as on their activity
and selectivity in the target reaction. The fresh and used catalysts
were characterized by HAADF-STEM (high-angle annular dark field
scanning transmission electron microscopy), SEM (scanning elec-
tron microscopy), and XRD (X-ray diffraction). Catalytic tests were
performed at 473 K as well as at 523 K and 2 MPa using various
CO /H /C H /N reaction feeds.
◦
angle of 30–36 and collection semi-angles for HAADF and ABF
of 90–170 mrad and 11–22 mrad respectively. The sample was
deposed without any pretreatment on a holey carbon supported
Cu-grid (mesh 300) and transferred to the microscope. To estimate
size distribution of supported Au NP from HAADF-STEM and SEM
images, at least 100 particles were counted using ImageJ program
[22].
2
2
2
4
2
2
. Experimental
2.1. Catalyst preparation
X-ray diffraction (XRD) of samples in powder form was carried
out using a X’Pert Pro (Panalytical, Almelo) operated as a reflection
setup in the theta/theta mode. The alignment was checked against a
Au/TiO2 r was prepared by deposition-precipitation (DP) of gold
hydroxide from HAuCl4 (41.1 wt% Au, Chempur) on TiO2 in the
rutile structure (Sachtleben) as described elsewhere for the prepa-
ration of Au/TiO2 a [15]. 2 g of TiO2 r powder were added at room
temperature under continuous stirring to a 1.57 mM HAuCl4 solu-
tion (200 ml) and further stirred for 30 min. DP of gold hydroxide
was done with 0.5 ml ammonia solution (25%, Roth). The final solu-
◦
silicon standard. The data were collected in the 2ꢀ range of 5–80 .
The program suite WINXPow by STOE&CIE with inclusion of the
Powder Diffraction File PDF2 of the ICDD (International Centre of
Diffraction Data) was used to identify the phase composition of the
samples. The size of Au crystallites were obtained from XRD pattern
◦
tion was further stirred for 10 min before filtration. The K–Au/TiO2
r
at 2ꢀ of 38.2 and applying the Scherrer equation after fitting with
catalysts were prepared by incipient wetness impregnation of dried
Au/TiO2 r with KNO3 solution (>99%, Merck) to achieve K concen-
tration of 1, 2, 3, and 4 wt%. The impregnated catalyst precursors
were finally calcined in air at 473 K for 4 h. Hereafter, they were
pressed and sieved to obtain particles of a fraction of 250–450 ␮m.
The catalysts were denoted XK-Au/TiO2 r and XK-Au/TiO2 a from
a pseudo-Voigt-function.
2.3. Catalytic testing
Catalytic tests were carried out at 2 MPa and 473 K and 523 K
using an in-house developed setup equipped with 50 continuous-
[
15], where “X” stands for the particular K weight concentration
flow fixed-bed reactors operating in parallel. CO (4.5, Air Liquide),
2
(Table 1).
H (5.0, Air Liquide), C H (3.0, Linde), and N (5.0, Air Liquide) were
2
2
4
2
the feed components. 300 mg of catalyst (fraction of 250–450 ␮m)
were loaded into stainless-steel tube reactors (i.d. 4 mm) with a
threefold amount of SiC for dilution. Before testing, the catalysts
2
.2. Catalyst characterization
−
1
The weight concentration of Au and K in calcined catalysts
were initially heated to 523 K in a N2 flow (6.67 ml min
per
(Table 1) was determined by inductively coupled plasma optical
reactor) followed by feeding a CO /H /C H /N = 4:2:1:1 mixture
2
2
2
4
2
emission spectrometry (ICP-OES) using a Varian 715 emission spec-
trometer.
at the same flow rate for 24 h. Hereafter, the reaction temper-
ature was reduced to 473 K, at which the catalysts were again
tested for 24 h. In the next step, the feed has been changed to
CO /H /C H /N = 1:1:1:1 for further tests at 473 K and 523 K.
Steady-state data were obtained after at least 10 h on stream under
each condition. The reaction products and feed components were
analyzed by an on-line GC (Varian CP-3800) equipped with both
Nitrogen adsorption–desorption isotherms for each cata-
lyst were collected at 77 K on a BELSORP-mini II (BEL Japan,
Inc.) setup. The specific surface area was calculated from the
adsorption isotherm applying the Brunauer, Emmett, and Teller
2
2
2
4
2
0
equation for the N2 relative pressure range of 0.05 < P/P < 0.30.
Table 1
Catalysts and their selected physicochemical and catalytic properties.
BET/m2
g
−1
(used)
Y(propanol)a
% (at 473 K)
Y(propanol)a
/% (at 523 K)
Catalyst
Loading/wt%
dAu NP/nm (STEM)
dAu NP/nm (SEM)
dAu/nm (XRD)
/
Au
K
Rutile supported catalysts
b
n/ac
5.5
Au/TiO2 r
2.0
1.9
1.8
1.8
1.7
–
7.0
7.5
6.5
8.5
6.0
5.0–5.5
5.0–5.5
5.5
51
60
59
68
58
1.0
1.9
1.6
1.4
1.1
0.4
1.8
2.2
3.1
2.7
b
b
1K–Au/TiO2 r
2K–Au/TiO2 r
3K–Au/TiO2 r
4K–Au/TiO2 r
0.9
1.9
2.8
3.6
5.5
n/a
b
c
6.0 and 7.5
7.0 and 9.0b
5.5
Anatase supported catalysts [15]
Au/TiO2 a
1.3
1.3
1.3
–
1.0
2.9
7.5
4.5
n/a
n/a
7.5
n/ac
55
62
60
0.7
2.4
0.8
0.0
1.0
1.4
c
1K–Au/TiO2 a
K–Au/TiO2 a
n/a
c
3
7.5 and 10.0
n/a
a
b
c
−1
Reaction conditions: 2 MPa, contact time of 45 g min l , CO2/H2/C2H4/N2 = 1:1:1:1.
Sub-nanometer (<2 nm) Au particles were also found.
Au reflexes overlaid with those of TiO2.
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
S.J. Ahlers et al. / Catalysis Today xxx (2015) xxx–xxx
3
flame ionization and thermal conductivity detectors with HP-Plot
Q and HP-Plot Q and Molsieve 5A columns respectively. The yields
to propanol, propanal, and CO were calculated on CO2 basis using
(approximately 20 nm). The same distribution was also valid for
K-free Au/TiO2 r (Fig. S4).
As proven by XRD, the rutile structure of TiO2 did not change
after deposition of Au and K as well as after performing catalytic
tests at 473 K and 523 K for at least 250 h on stream (Fig. S5). The
crystallite size of TiO2 r increased from 15.1 nm to 18.2 nm after cal-
cination at 523 K. Deposition of Au and K led to a further increase
to 19.3 nm. Changes in the structure of anatase to rutile were not
investigated because this usually occurs at temperatures above
Eq. (2). It was considered that one C -oxygenate molecule is formed
3
from one CO2 and one C H . The conversion of CO and C H4 was
2
4
2
2
calculated from the sum of the corresponding yields (Eq. (3)). Selec-
tivity was calculated according to Eq. (4).
n˙ i − n˙ i,0
Yi =
× 100%
(2)
(3)
n˙
f,0
8
23 K [23].
ꢀ
SBET of used K–Au/TiO a and K–Au/TiO2 r catalysts was around
2
Xf =
Yi
2
−1
i
50–60 m g . For the latter catalysts, a reduction in S
BET
compared
2
−1
to the initial SBET of 107 m g occurred upon catalysts calcination
due to sintering of bundled TiO2 r crystallites [24].
and
Yi
Si = Xf × 100%
(4)
3
.1.2. Surface morphology
In order to check if preparation method, i.e. Au deposition
where n˙ f and n˙ i stand for mole flows of feed components and reac-
tion products respectively. Subscript 0 is used for inlet mole flows.
and/or subsequent K impregnation and reaction conditions influ-
ence surface support morphology, we analyzed bare supports and
supported catalysts by scanning electron microscopy (SEM). Rep-
resentative SEM images of unloaded TiO2 a and TiO2 r are shown
in Fig. 2. Additional detailed images are summarized in Figs. S6
and S7. The fresh TiO2 r support consisted of needle-like crystal-
lites with a thickness of around 15–20 nm (Fig. 2(b)), while mainly
squares with rounded edges in the size of 20–30 nm were identified
for TiO2 a (Fig. 2(a)). The morphology of both supports did not sig-
nificantly change both after Au and K deposition as well as after
performing catalytic tests. However, the size of TiO2 a particles
3
3
3
. Results and discussion
.1. Catalysts characterization
.1.1. Surface and bulk characteristics of K–Au/TiO2
The size distribution of supported Au NP was determined by
HAADF-STEM. SEM and XRD were additionally applied. Fig. 1(a,
b) shows representative HAADF-STEM images of Au/TiO2 r and
4
K–Au/TiO2 r after catalytic testing. The images of all used and
selected fresh catalysts together with the particle size histograms
are summarized in Figs. S1 and S2. Irrespective of K loading, Au NP
with an average diameter between 5 nm and 9 nm were identified
on the surface of K–Au/TiO2 r catalysts (Table 1). Any correlation
between the loading and the size could not be established. Their
anatase-based counterparts possessed similarly sized NP. The size
of Au NP determined by HAADF-STEM is in a good agreement with
the values obtained from XRD and SEM despite the fact that XRD
measures only average crystallite sizes and makes no real state-
ment on the Au NP size distribution and SEM is not able to detect
Au NP smaller than 4 nm (Table 1).
However, it should be particularly highlighted that the surface
of K–Au/TiO2 r is additionally populated by Au NP smaller than
2
S3). Even single Au atoms were detected (Fig. S3). Their presence
does not depend of K loading and catalyst history, i.e. fresh or used
(20–30 nm) slightly decreased to 11–23 nm for fresh 3K–Au/TiO2 a
and to 13–22 nm for used 3K–Au/TiO2 a respectively (Fig. 2(c), Fig.
S7). The slight decrease might be due to the alkali treatment of the
support during the precipitation of Au.
Similarly to the catalysts based on TiO2 a, small changes were
also identified in the surface morphology of those based on TiO2 r.
From the comparison of the images of fresh TiO2 r (Fig. 2(b)) and
used Au/TiO2 r (Fig. 2(d)) one can conclude for the latter mate-
rial that the endings of the needles were blunted and the titania
particles were now alike to small rods. This came along with an
increase in the crystallite sizes from 15.1 nm to 17.5 nm as deter-
mined by XRD. Similar morphological changes were also found for
3
K–Au/TiO2 r but with a lower extent. Such modifications were
nm, which we call sub-nanometer Au particles (Fig. 1(a, b) and Fig.
probably a reason for the decrease in SBET after calcining fresh mate-
rials and their exposure to reaction feeds. Taking into account the
fact that such decrease and morphological changes were less pro-
nounced in the catalysts containing K, one may suggest that the
promoter stabilized surface morphology against restructuring dur-
ing catalytic tests, similar to the less decreasing SBET for K doped
catalysts after the experiment (Table 1).
(
Fig. S3). Importantly, such sub-nanometer Au particles were not
detected on the surface of K–Au/TiO2 a irrespective of K loading
Fig. 1(c, d)). Thus, the support morphology but not the presence of
(
promoter determines the formation of sub-nanometer Au particles
on catalysts surface. Such species did not change their size under
reaction conditions, while larger Au NP grew as concluded from the
HAADF-STEM images of fresh and spent catalysts in Fig. S3.
3.2. Catalytic performance of K–Au/TiO2
3.2.1. Factors determining catalysts activity
Composition SEM analysis shown in Fig. 2(e, f) of anatase- and
rutile-based catalysts further supports that Au NP larger than 4 nm
did not significantly change their size after catalytic tests over 250 h
on stream. This technique enabled us to conclude that the support
morphology influences surface density and distribution of Au NP.
As seen from Fig. 2(e) for used 3K–Au/TiO2 a, Au NP were spread
all over the support. A few TiO2 a particles with a higher surface
density of Au NP but almost the same size were also detected. Yet,
some bigger NP with the size of approximately 40 nm were addi-
tionally identified. In contrast to TiO2 a-based catalysts, there was
no agglomeration of Au NP on the surface of TiO2 r-based cata-
lysts. Fundamental aspects of this effect were not investigated in
the present contribution. The overview images of 3K–Au/TiO2 r in
Fig. 3 shows the conversion of CO and C H over all K–Au/TiO r
2
2
4
2
and K–Au/TiO2 a catalysts and the selectivity to propanal, propanol,
and carbon monoxide based on CO2 conversion at 473 K and 523 K.
The only other main product was ethane formed from ethyl-
ene. Further minor side products were propane, C -hydrocarbons
6
(propanal condensation [25] and further dehydration), and pen-
tanal/pentanol (C H4 dimerization and further hydroformylation).
2
We shall start with the discussion of catalysts activity. Conversion
values of CO2 and C H4 over the unpromoted Au/TiO2 r catalyst at
2
473 K were around 2.8% and 39% respectively (Fig. 3(a)), which are
lower than 3.6% and 44% obtained over Au/TiO2 a (Fig. 3(b)). One
can see that the conversion of ethylene was significantly higher
than the conversion of CO . This is due to the fact that C H was
Fig. 2(f) shows a homogeneous distribution of Au NP over TiO2
with Au NP in the range of 5.0–5.5 nm and some bigger Au NP
r
2
2
4
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
4
S.J. Ahlers et al. / Catalysis Today xxx (2015) xxx–xxx
Fig. 1. HAADF-STEM images of used (a) Au/TiO2 r, (b) 4K–Au/TiO2 r, (c) Au/TiO2 a, and (d) 3K–Au/TiO2 a.
not only converted to propanal/propanol but mainly hydrogenated
to C H .
3 wt%. The stronger effect of temperature on the conversion over
K-containing catalysts compared to their K-free counterparts may
be explained by the fact that the activation energy of reduction of
ionic carbonates formed in the presence of K is higher than that
of carboxylate species prevailing on the surface of K-free Au/TiO2
catalysts [15].
2
6
In agreement with K–Au/TiO2 a [15], doping of Au/TiO2 r with K
led to a decrease in CO conversion. This is due to the fact that pres-
2
ence of K favors the formation of ionic carbonates, which are stable
against their reduction by H to generate gas-phase CO. Irrespective
2
of the support morphology, ethylene conversion increased after
promoting Au/TiO2 with 1 wt% K. However, it started to decrease
with a further increase in the K loading (Fig. 3(a)). The increase in
the conversion can be explained by the fact that the activation of
It is worth mentioning that the effect of temperature on the eth-
ylene conversion mainly to ethane is more complex and depends
on the kind of supporting material, i.e. anatase vs. rutile. For
the K–Au/TiO2 a catalysts, the conversion increased with rising
temperature from 473 K to 523 K (Fig. 3(b)). For Au/TiO2 r, the con-
version increased only slightly from 39% to 44% but decreased over
all K-containing Au/TiO2 r catalysts. The higher the K loading, the
stronger was the decrease (Fig. 3(a)). Such reversed temperature
effect can be explained when the coverage by adsorbed ethyl-
ene or/and hydrogen species participating in ethane formation is
strongly reduced with rising temperature. This depends on the ratio
of the reaction constants of adsorption to desorption. No further
fundamental investigations were yet carried out to deeper analyze
this abnormal behavior.
gas-phase H on the surface of Au NP is improved in the presence of
2
K [15]. This catalyst property was not found to depend on rising K
content. However, the concentration of basic sites should increase
with promoter loading and negatively influence the adsorption and
consequently conversion of C H over the catalysts with high K
2
4
content [26].
As expected, CO2 conversion over all tested catalysts rose when
the reaction temperature was increased from 473 K to 523 K. How-
ever, the abovementioned negative effect of K doping on the
conversion became less pronounced and depended on the kind of
support. For K–Au/TiO2 r at 473 K, it decreased from 2.8% to 1.1%,
i.e. by a factor of 2.5 upon increasing K loading from 0 wt% to 4 wt%
while the corresponding factor at 523 K was 1.7 (Fig. 3(a)). Such
factors for the K–Au/TiO2 a system were 4.5 and 1.6 at 473 K and
3.2.2. Origins of catalyst selectivity
Propanal and propanol were the major products formed from
CO2 and C H . CO was additionally converted to CO, while ethyl-
2
4
2
5
23 K respectively, when K loading was increased from 0 wt% to
ene was hydrogenated to ethane. The undesired hydrogenation to
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
S.J. Ahlers et al. / Catalysis Today xxx (2015) xxx–xxx
5
Fig. 2. SEM images of bare (a) TiO2 a and (b) TiO2 r as well as different catalysts after catalytic tests: (c) Au/TiO2 a, (d) Au/TiO2 r, (e) 3K–Au/TiO2 a, and (f) 3K–Au/TiO2 r.
Images (a)–(d) were recorded in SE (morphology) and (e)–(f) in BSE mode (z contrast).
ethane dominated. The selectivity to ethane over K–Au/TiO2 r cat-
alysts at 473 K slightly increased with K loading from 97% to 98%,
while it did not practically depend on the loading and was around
K–Au/TiO2 a catalytic system showed very similar results with
respect to doping with K (Fig. 3(d)). The positive effect of K on
the performance of K–Au/TiO2 a was previously explained by the
fact that K accelerated the desorption step to propanal proba-
bly through an increased generation of adsorbed H species [15].
Comparing propanol selectivity at 473 K with the one at 523 K, it
is obvious that the positive effect of K loading became less pro-
9
8% over their anatase-based counterparts. When the reaction tem-
perature was increased to 523 K, the selectivity decreased from 99%
to 93% and from 99% to 96% with rising K loading over K–Au/TiO2
and K–Au/TiO2 a respectively.
r
r
Concerning the propanol selectivity, the unpromoted Au/TiO2
nounced at higher temperature (Fig. 3(d)). However, K–Au/TiO2 r
catalyst produced propanol with a selectivity of 37% at 473 K,
while such value over Au/TiO2 a was 20% (Fig. 3). Notably, the lat-
ter catalyst did not form propanol at 523 K but this alcohol was
observed over Au/TiO2 r with a selectivity of 6% thus highlight
the importance of support morphology for catalytic performance
of supported Au species. Doping Au/TiO2 r with K led to a strong
decrease in the selectivity to CO with a simultaneous increase
in the selectivity to propanol (Fig. 3(c)). CO was not observed
over 3K–Au/TiO2 r and 4K–Au/TiO2 r, i.e. a combined selectivity
catalysts performed more selectively than K–Au/TiO2 a catalysts.
For example, CO was the only product formed from CO2 over
Au/TiO2 a, while Au/TiO2 r converted CO to propanol with a selec-
2
tivity of 6% which increased to 64% with an increase in K loading
from 0 wt% to 4 wt%. The higher K loading also improved the
propanol selectivity of K–Au/TiO2 a materials, which however did
not exceed 45%. However, owing to the fact that TiO2 r-based
catalysts are more selective for propanol production than their
TiO2 a-based counterparts, they showed an average higher yield
of propanol (Table 1). The highest yield of propanol of 3.1% was
to C -oxygenates (propanal + propanol) of 100% was achieved. The
3
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
6
S.J. Ahlers et al. / Catalysis Today xxx (2015) xxx–xxx
Fig. 3. Conversion of CO2 (
,
) and C2H4 (
,
) over (a) K–Au/TiO2 r and (b) K–Au/TiO2 a catalysts at 473 K (closed symbols) and 523 K (open symbols) as well as
selectivity to CO (dark gray bar), propanol (light gray bar), and propanal (black bar) over (c) K–Au/TiO2 r and (d) K–Au/TiO2 a catalysts at 473 K (closed bars) and 523 K
−
1
(
dashed bars). Reaction conditions: 2 MPa, contact time of 45 g min l , CO2/H2/C2H4/N2 = 1:1:1:1.
achieved over 3K–Au/TiO r, which corresponds to the space time
in their physical properties from larger Au NP. For example, sub-
nanometer Au particles tend to have a partial charge [27,28] and
were also reported to be active for the water gas-shift reaction
[29]. Based on this, we assume that the catalysts ability for the
CO adsorption, or like in our specific case, the insertion of CO in
the Au–Ethyl bond, can also be accelerated. The latter reaction was
suggested to be the rate-determining step in the hydroformyla-
tion of low olefins [15,30]. Considering the above discussion, it is
assumed that the presence of sub-nanometer Au particles might
be a reason for higher propanol selectivity over the rutile-based
catalysts.
2
−1
−1
yield of 34 gpropanol
h
kgcatalyst .
Importantly, the effect of CO conversion on the propanol selec-
2
tivity can be excluded because similarly composed catalysts are
compared at close degrees of the conversion. Taking this experi-
mental condition into account, the selectivity difference between
the anatase- and rutile-based catalysts can be explained as follows.
From a mechanistic viewpoint [15], CO2 is primarily converted
into CO, which further reacts with C H4 and H2 to propanal. This
2
aldehyde is quickly hydrogenated to propanol. Based on this reac-
tion scheme, the selectivity to CO is determined by the rates of
(
i) CO generation and (ii) its consecutive reaction with ethylene
In summary, two important effects of support morphology on
catalytic performance of K–Au/TiO2 should be especially empha-
sized: (i) a higher propanol selectivity was achieved and (ii) the
non-desired C H4 hydrogenation to C H6 was slightly suppressed
with an increase in reaction temperature from 473 K to 523 K over
rutile-based catalysts compared with their anatase-based coun-
terparts. Operating at 523 K favors a higher CO2 conversion. In
combination with a higher selectivity to propanol, this ends up
in an increase in the yield to propanol. The inhibiting effect of
to propanal. The lower CO selectivity over K–Au/TiO2 r than over
K–Au/TiO2 a at close CO2 conversion (similar activity for CO gen-
eration) can only be explained when propanol is formed faster
over rutile-based catalysts. Since propanal/propanol are exclu-
sively formed on the surface of Au species but not on the bare
support, metal-support interaction, which influence either the
properties or the kind (size/form) of metal particles, should be
responsible for the effect of support morphology on propanol for-
mation. Our thorough STEM analysis revealed that anatase-based
catalysts possess Au NP between 4.5 nm and 10 nm. Importantly,
sub-nanometer Au particles were exclusively observed on the sur-
face of rutile-supported catalysts irrespective of K loading. Larger
Au NP with an average size of around 7 nm were also identified on
these materials. In other words, they possess two different types
of Au species strongly differing in their size. These species can
show different activity for CO formation and its consecutive con-
version. It is well known that sub-nanometer Au particles differ
2
2
temperature on C H6 formation also opens the possibility to oper-
2
ate at high partial pressures of H2 for increasing CO2 conversion
into CO, which is required for C H4 hydroformylation. However,
2
additional tests are further required to explore the potential of
operation at high partial pressures of H and reaction temperatures.
2
From a viewpoint of catalyst design, in addition to stabilization
of sub-nanometer Au particles, the choice of selectivity-governing
promoter should be focused on suppressing non-selective eth-
ylene hydrogenation to ethane without negative effect on CO2
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006

G Model
CATTOD-9567; No. of Pages7
ARTICLE IN PRESS
S.J. Ahlers et al. / Catalysis Today xxx (2015) xxx–xxx
7
conversion. The promoter should not favor the formation of ionic
carbonates but weaken surface acidity for a lower ethylene conver-
sion.
References
[
[
1] H.B. Goyal, D. Seal, R.C. Saxena, Renew. Sustain. Energy Rev. 12 (2008)
04–517.
2] P.Y. Dapsens, C. Mondelli, J. Pérez-Ramírez, ACS Catal.
1499.
5
2 (2012) 1487–
4
. Conclusions
[
[
[
3] A.V. Bridgwater, Biomass Bioenergy 38 (2012) 68–94.
4] W. Wang, S. Wang, X. Ma, J. Gong, Chem. Soc. Rev. 40 (2011) 3703–3727.
5] M. He, Y. Sun, B. Han, Angew. Chem. Int. Ed. 52 (2013) 9620–9633.
The present study including catalytic runs and thorough char-
acterizations by HAADF-STEM, SEM, and XRD highlighted the
importance of support morphology (anatase vs. rutile) for stabi-
[6] G. Centi, S. Perathoner, ChemSusChem 3 (2010) 195–208.
[
7] E.V. Kondratenko, G. Mul, J. Baltrusaitis, G.O. Larrazábal, J. Pérez-Ramírez,
Energy Environ. Sci. 6 (2013) 3112–3135.
lizing Au NP of a certain type on the surface of K–Au/TiO catalysts.
2
[
8] Y. Sun, J. Cheng, Bioresour. Technol. 83 (2002) 1–11.
In addition to the nanoparticles between 5 nm and 10 nm present
on the surface of K–Au/TiO2 r and K–Au/TiO2 a, sub-nanometer
[9] C.E. Wyman, Annu. Rev. Energy Environ. 24 (1999) 189–226.
10] G.A. Olah, Angew. Chem. Int. Ed. 52 (2013) 104–107.
11] J. Ott, V. Gronemann, F. Pontzen, E. Fiedler, G. Grossmann, D.B. Kersebohm,
G. Weiss, C. Witte, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH
Verlag GmbH & Co. KGaA, 2000.
[
[
(
<2 nm) Au particles were observed on the surface of the former
materials only. Moreover, the application of rutile support ensured
homogeneous distribution of both kinds of NP. As a consequence,
catalytic performance of K–Au/TiO2 r and K–Au/TiO2 a differed.
Irrespective of K loading, rutile-supported catalysts performed
superior to their anatase-based counterparts in terms of propanol
selectivity. This might be partially related to the presence of sub-
[12] M. Stöcker, Microporous Mesoporous Mater. 29 (1999) 3–48.
[
[
[
13] J.F. Haw, W. Song, D.M. Marcus, J.B. Nicholas, Acc. Chem. Res. 36 (2003)
17–326.
3
14] H. Zimmermann, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH
Verlag GmbH & Co. KGaA, 2000.
15] S.J. Ahlers, U. Bentrup, D. Linke, E.V. Kondratenko, ChemSusChem 7 (2014)
2631–2639.
nanometer Au particles. Promoting both Au/TiO2 a and Au/TiO2
r
[16] A.J. Papa, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag
GmbH & Co. KGaA, 1993.
with K strongly increased propanol selectivity but decreased CO2
conversion. Non-desired ethylene hydrogenation to ethane was
found to decrease over K–Au/TiO2 r with an increase in reaction
temperature from 473 K to 523 K.
[
17] I. Takahara, M. Saito, M. Inaba, K. Murata, Catal. Lett. 105 (2005)
49–252.
2
[18] H. Zimmermann, R. Walzl, Ullmann’s Encyclopedia of Industrial Chemistry,
Wiley-VCH Verlag GmbH & Co. KGaA, 2000.
[
19] M. Ni, M.K.H. Leung, D.Y.C. Leung, K. Sumathy, Renew. Sustain. Energy Rev. 11
2007) 401–425.
(
Acknowledgements
[
[
[
[
[
20] O. Khaselev, J.A. Turner, Science 280 (1998) 425–427.
21] D.B. Levin, L. Pitt, M. Love, Int. J. Hydrogen Energy 29 (2004) 173–185.
22] T.J. Collins, Biotechniques 43 (2007) S25–S30.
23] R. Ren, Z. Yang, L.L. Shaw, J. Mater. Sci. 35 (2000) 6015–6026.
24] X. Bokhimi, R. Zanella, A. Morales, J. Phys. Chem.
15210–15216.
The authors thank Benjamin Paul for carrying out catalyst char-
acterization by SEM, Anja Simmula for the ICP-OES analysis and
Dr. Matthias Schneider and Claudia Heber for XRD analysis of
K–Au/TiO2 r and K–Au/TiO2 a catalysts. This work was supported
by the Leibniz-Gemeinschaft under the research grant SAW-2011-
LIKAT-2.
C 111 (2007)
[
[
25] T.A. Kainulainen, M.K. Niemelä, A.O.I. Krause, Catal. Lett. 53 (1998)
7–101.
26] W.D. Mross, Catal. Rev. 25 (1983) 591–637.
9
[27] Y. Denkwitz, Z. Zhao, U. Hörmann, U. Kaiser, V. Plzak, R.J. Behm, J. Catal. 251
2007) 363–373.
(
[
[
28] P. Kast, G. Ku cˇ erová, R.J. Behm, Catal. Today 244 (2015) 146–160.
29] Q. Fu, W. Deng, H. Saltsburg, M. Flytzani-Stephanopoulos, Appl. Catal. B 56
Appendix A. Supplementary data
(
2005) 57–68.
[30] X. Liu, B. Hu, K. Fujimoto, M. Haruta, M. Tokunaga, Appl. Catal. B 92 (2009)
11–421.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2015.04.
4
006
Please cite this article in press as: S.J. Ahlers, et al., Propanol formation from CO2 and C H4 with H2 over Au/TiO : Effect of support and
2
2
K doping, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.04.006