Au@TiO2 Core–Shell Composites for the Photocatalytic Reduction of CO2

Article abstract of DOI:10.1002/chem.201801796

Au/TiO₂ catalysts in different geometrical arrangements were designed to explore the role of morphology and structural properties for the photocatalytic reduction of CO₂ with H₂O in the gas-phase. The most active sample was a Au@TiO₂ core–shell catalyst with additional Au nanoparticles (NPs) deposited on the outer surface of the TiO₂ shell. CH₄ and CO are the primary carbon-containing products. Large amounts of H₂ are additionally formed by photocatalytic H₂O splitting. Shell thickness plays a critical role. The highest yields were observed with the thickest layer of TiO₂, stressing the importance of the semiconductor for the reaction. Commercial TiO₂ with and without Au NPs was less active in the production of CH₄ and CO. The enhanced activation of CO₂ on the core–shell system is concluded to result from electronic interaction between the gold core, the titania shell, and the Au NPs on the outer surface. The improved exposure of Au?TiO₂ interface contributes to the beneficial effect.

Full text of DOI:10.1002/chem.201801796

A Journal of
Accepted Article
Title: Au@TiO₂ Core-Shell Composites for the Photocatalytic
Reduction of CO₂
Authors: Anna Pougin, Georgios Dodekatos, Martin Dilla, Harun
Tüysüz, and Jennifer Strunk
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201801796
Link to VoR: http://dx.doi.org/10.1002/chem.201801796
Supported by

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
Au@TiO2 Core-Shell Composites for the Photocatalytic
Reduction of CO2
Anna Pougin[a],†, Georgios Dodekatos[b], Martin Dilla[c], Harun Tüysüz[b]*, Jennifer Strunk[d]*
Abstract: Au/TiO2 catalysts in different geometrical arrangements
were designed to explore the role of morphology and structural
properties for the photocatalytic reduction of CO2 with H2O in the
gas-phase. The most active sample was a Au@TiO2 core-shell
catalyst with additional Au nanoparticles (NPs) deposited on the
outer surface of the TiO2 shell. CH4 and CO are the primary carbon-
containing products. Large amounts of H2 are additionally formed by
photocatalytic H2O splitting. Shell thickness plays a critical role. The
highest yields were observed with the thickest layer of TiO2,
stressing the importance of the semiconductor for the reaction.
Commercial TiO2 with and without Au NPs was less active in the
production of CH4 and CO. The enhanced activation of CO2 on the
core-shell system is concluded to result from electronic interaction
between the gold core, the titania shell, and the Au NPs on the outer
surface. The improved exposure of Au-TiO2 interface contributes to
the beneficial effect.
oxidation reactions. In classical heterogeneous catalysis,
parameters such as the interface between Au and TiO2,[4] the
size and shape of Au,[3b,5] and the oxidation state of TiO2[6] play
a crucial role in the catalytic performance of the prepared
catalyst.
In photocatalysis, modification of TiO2 by deposition of noble
metal nanoparticles on the surface is generally applied in order
to enhance charge carrier separation over the Schottky barrier,
and to better catalyze the reaction on the surface by lowering
the activation energy.[7] This results in improved photocatalytic
performances under UV light exposure. In a recent model study,
some authors of the present work could confirm through infrared
spectroscopy the generally accepted concept of photocatalytic
activity enhancement by improved charge separation: For the
photocatalytic 2-propanol oxidation in the gas phase, the role of
gold in Au/TiO2 was attributed to the removal of excess
electrons otherwise accumulating in TiO2, and their further
transfer onto the co-reactant oxygen.[ 8 ] In a photocatalyst
system in which Au nanoparticles (NPs) had been
photodeposited on TiOx/SBA-15, i.e. on a mesoporous silica
(SBA-15) grafted with titanate (TiOx) species, the facilitation of
an electron transfer reaction from the titania shell around the
gold nanoparticles to gold and from there to the reactants was
Introduction
The photocatalytic reduction of CO2 holds the potential to solve
future shortages of mobile energy sources, whilst reducing the
concentration of atmospheric CO2. Since the first reports on this
reaction forty years have passed,[1] but a practical process and
fundamental understanding is still in development.[2] In 2013, the
yields of the main product, methane, obtained with most TiO2-
based catalysts ranged between 0.1 and 5 µmol gCat-1 h-1, which
is far away from an industrial application.[2] One of the promising
approaches to improve the activity of photocatalysts, and to
possibly also gain understanding of the essential mechanistic
steps, is the addition of noble metals, such as Au, to TiO2. From
the field of classical heterogeneous catalysis, since the
9
]
demonstrated to be
a
likely process.[
Moreover, Au
nanoparticle deposition proved to be a promising method in
order to enable visible light absorption of TiO2 due to the
distinguishing properties of this noble metal. For instance,
Dodekatos et al.[ 10 ] and Schünemann et al.[ 11 ] showed that
selective oxidation of glycerol in the aqueous phase over TiO2-
or SiO2-supported Au or AuCu nanoparticles can be improved
by the aid of visible light irradiation during the thermally
catalyzed reaction. The visible light was absorbed by the Au
nanoparticles to enhance the catalytic performance. The origin
of the light absorbing properties of the Au nanoparticles can be
rationalized by the localized surface plasmon resonance (LSPR)
effect.[ 12] In a simple mechanistic scheme, the LSPR can be
explained by the collective oscillation of the conduction electrons
on the surface of the metal nanoparticle induced by the
interaction of the electrons with the electromagnetic field of the
impinging light and the restoring Coulomb force of the positive
nuclei of the metal.[13]
Apart from attributing enhanced photocatalytic performance to
improved charge carrier separation and visible light absorption,
different energy transfer mechanisms have been proposed for
metal/semiconductor composites.[13b, 14 ] LSPR induces strong
electromagnetic fields around the nanoparticle with higher
intensities than the incident electric field for the initialization of
the LSPR. The intensity of the field depends on the geometry
and size of the metal nanoparticle and the extinction coefficient
of its surroundings.[12-15] A resonantly enhanced near field in the
surrounding medium arises in direct vicinity to the metal
nanoparticle surface, which can be understood in terms of a
concentrated light flux.[13a,14] This may play a key role in
photocatalysis,[13a] potentially also by local heating effects.[13c] In
3
]
pioneering work of Haruta et al.[
TiO2-supported Au
nanoparticles are instead known for their properties in catalytic
[a]
[b]
Dr. A. Pougin
Laboratory of Industrial Chemistry
Ruhr-University Bochum
44780 Bochum, Germany
Dr. G. Dodekatos, Dr. H. Tüysüz
Max-Planck-Institut für Kohlenforschung
45470 Mülheim/Ruhr, Germany
Email: tueysuez@kofo.mpg.de
M. Dilla
Max-Planck-Institute for Chemical Energy Conversion
45470 Mülheim/Ruhr, Germany
Prof. Dr. J. Strunk
[c]
[d]
Leibniz-Institute for Catalysis (LIKAT) at the University of Rostock
18059 Rostock, Germany
Email: jennifer.strunk@catalysis.de
Present address: Evonik Industries, Paul-Baumann-Str. 1, 45772
Marl, Germany
†
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
another mechanism, sometimes referred to as “hot electron
transfer”, the charge carriers are assumed to be directly injected
from the metal to the semiconductor, which is similar to the
mechanism described for dye sensitized photocatalysts.[13,14,16] It
is, however, also possible that the surface chemistry of the noble
applications and encouraged us to investigate Au/TiO2
photocatalysts with core-shell morphology for the CO2 reduction.
In this contribution, TiO2-encapsulated Au nanoparticles are
tested in the photocatalytic CO2 reduction in the gas-phase
under conditions of highest possible purity. Our reaction
conditions exclude additional product formation from impurities.
The geometry of the Au@TiO2-Au nanospheres consists of
three levels: the Au core (~20 nm), surrounded by a TiO2 shell
(Au@TiO2), and additional Au nanoparticles (~3 nm) on the
outer TiO2 shell (Au@TiO2-Au). The thickness of the TiO2 shell
is varied (20, 11, and 8 nm) to study potential influences of a
shielding of the metal core, and to identify the role of a well-
established semiconductor structure for the photocatalytic
process. All catalysts are prepared in two versions, with and
without the additional Au nanoparticles on the outer surface of
the TiO2 shell, to judge upon the influence of exposed and
encapsulated Au nanoparticles. In order to put all findings into
perspective, the best catalysts were tested in an advanced gas
chromatography (GC) system to detect all potential reaction
products. A commercial Au/TiO2 catalyst and two home-made
Au/TiO2 catalysts without core-shell structure were tested as
references.
metals exert
a major influence on the catalytic surface
processes, so that the enhancement is brought about
predominantly by catalytic effects, and not by alterations in light
absorption or charge carrier mobility. In any case, it is the nature
of the semiconductor TiO2 that determines major properties
concerning the lifetime of the charge carriers. The dependency
of charge carrier diffusion dynamics and TiO2 particle size in
17 ]
Au/TiO2 has been studied by Luchao Du et al.[
They
discovered the trend that with bigger TiO2 particles the chances
of electron-hole-recombination are decreased due to the longer
diffusion length of electrons, and the catalyst is more efficient in
a size range between 8 and 50 nm.
Many efforts have been devoted to utilizing the plasmonic
properties for photocatalytic CO2 reduction.[13b,18] For instance,
19
]
Hou et al.[
reported an enhancement in activity in
photocatalytic CO2 reduction over Au/TiO2 in comparison to the
bare support and attributed it to a positive influence from the
surface plasmon. Two different mechanisms were assumed to
be in operation, depending on whether the plasmon was excited
with green light or whether the composite was irradiated with UV
light.[19] Recently, a study conducted by some of the authors of
the present work, performed under reaction conditions of highest
purity, supports a possible influence of plasmonic effects in
noble metal-modified TiO2. However, such effects were clearly
noticeable only for Ag/TiO2, in line with the better overlap of the
absorption spectra of TiO2 with the plasmon resonance
absorption of silver compared to that of gold.[20] Another study
shows that when neither the wavelength of the plasmon
resonance nor the absorption edge of the metal oxide is hit
exactly, but instead a value in between is chosen, the activity is
maximized.[21]
Results and Discussion
Structural characterization of the Au/TiO2 core-shell samples
Representative TEM images of the samples Au@TiO2-08 and
Au@TiO2-08-Au are displayed in Figure 1. In the image of
sample Au@TiO2-08 (Figure 1 a and b) the TiO2 shell thickness
was measured in certain areas, with an average thickness of ~ 8
nm. The gold cores vary slightly in shape. Pure spheres and
slightly elongated structures are observed. The rough diameter
of the gold cores is ~20 nm, and they are clearly encapsulated
by the TiO2 shell. From Figure 1a and b, it is not possible to
distinguish whether or not any part of the gold surface is still
accessible to the surrounding atmosphere, but the predominant
core-shell structure is unambiguously visible. In the TEM image
Although many studies have been reported where
nanostructured core-shell materials with (mostly) noble metal
cores and semiconductor shells were used to perform
photocatalytic hydrogen generation from water[ 22] or pollutant
23
]
degradation,[
such kinds of materials were scarcely
investigated for photocatalytic CO2 reduction.[24] It is expected
that the special morphology can influence the photocatalytic
activity of the material. For example, Seh et al. showed that by
changing the morphology of Au/TiO2 nanostructures the
photocatalytic performance for hydrogen evolution can be
altered.[22f] Janus Au/TiO2 nanostructures exhibited an improved
activity compared to analogous core-shell nanostructures.
Nonetheless, the core-shell structured materials were superior to
simple Au/TiO2 nanocomposites, where Au NPs were deposited
on TiO2 without any specified morphology. The high activities
were attributed to the plasmonic electric fields caused by the Au
NPs, which were more pronounced for large Au NPs (Janus and
core-shell morphology). Furthermore, a stronger localization of
the electric fields was calculated for the Janus nanostructure,
which was consistent with the improved activity. Hence, the
morphology plays a crucial role for the presented photocatalytic
Figure 1. Representative TEM images of samples Au@TiO2-08 (a,b) and
Au@TiO2-08-Au (c,d); the average shell thickness of about 8 nm and the
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
successful deposition of additional gold nanoparticles on the outside of the
TiO2 shell can be clearly identified.
Figure 2. XRD patterns of the Au@TiO2 samples with different shell thickness
(without extra Au deposition on the TiO2 shell).
of sample Au@TiO2-08-Au (Figure 1 c and d), the additional
gold nanoparticles on the outer surface are detected. They are
very uniform in size with an approximate diameter of 3 nm, being
considerably smaller than the gold cores. In Figure 1c it can also
be seen that some of the gold cores are not fully encapsulated
but exhibit some Au surface possibly free of titania (e.g. particle
in the center of the image). Representative TEM images of the
other samples can be found in the supporting information
(Figure S1 and S2). It appears that the gold nanoparticles on the
outer shell cannot be distinguished anymore once the TiO2 shell
becomes too thick.
More information on the surface and adsorption properties of the
samples is gathered with diffuse reflectance infrared Fourier
transform spectroscopy (DRIFTS). Figure 4 shows the full range
spectra of the as made samples after a mild drying procedure in
inert gas. Pretreatment was kept minimal, in order to learn as
much as possible on the initial state of the samples. In the
region between 3800 and 3600 cm-1, three different hydroxyl
groups can be distinguished, located at 3736 cm-1, 3675 cm-1
and 3630 cm-1, which are assigned to two different terminal TiIV-
OH groups on anatase (3736, 3675 cm-1),[26] and a bridging OH
27
]
group between two titanium cations (3630 cm-1),[26,
The XRD patterns of the Au@TiO2 samples are shown in Figure
2. Three reflexes of the gold nanoparticles are observed, which
are located at 38.1° (111), 44.3° (200) and 64.5° (220) 2θ. The
only crystalline TiO2 phase detected in the samples is anatase;
no indication for the presence of rutile or other metastable
phases is evident. Of course, based on XRD data, the presence
or absence of amorphous phases cannot be judged upon.
Relative to the intensity of the most prominent reflection of gold,
the reflections of titania become more pronounced the thicker
the titania shell in the samples is. This is in accordance with the
expectations. The increased reflections may be due to the
increase in the amount of titania in the samples and/or due to an
improved crystallinity of the titania shell. The analogous samples
with additional Au nanoparticles on the TiO2 surface show
identical diffraction patterns (not shown).
In the UV-vis measurements shown in Figure 3, a certain
“support effect” of the TiO2 shell on the Au core becomes
obvious. Upon increasing the shell thickness the plasmon
resonance absorption of the Au core experiences a red shift.[25]
While the maximum is observed at 604 nm for the sample
Au@TiO2-08 with the thinnest TiO2 shell, it is shifted to 630 and
648 nm for TiO2 shells of 11 and 20 nm, respectively. The red
shift in combination with noble metal particle enclosure has
already been studied by Kelly et al.[25]
respectively. When additional gold nanoparticles are deposited
on the outer surface, the amount of terminal OH groups seems
to be slightly higher relative to the bridging OH groups. This
appears reasonable, considering that some nearest neighbor Ti
cations may be covered by the deposited gold nanoparticles
instead of by an OH group. Slight changes in the surface
coverage with hydroxyl groups in response to gold deposition
are also a piece of evidence that the outer gold nanoparticles
are actually attached to the titania shell.
In the region of ~1700 to ~1300 cm-1 (Figure 4, left) a variety of
(bi)carbonate species is detected on all samples, together with
adsorbed water (~1622 cm-1). The presence of (bi)carbonate
species is expected, because the samples have been
synthesized from organic precursors. Upon calcination, the
precursors are oxidized to CO2, which may then readsorb and
may even be incorporated into the structure. Furthermore,
samples have not been dried or cleaned prior to the IR
measurements, so it is likely that physisorption of moisture and
CO2 from ambient air had taken place when samples were
introduced into the DRIFTS sample cell. The CHx impurities
detected on some samples, indicated by bands in the region
between 3000 and 2700 cm-1, presumably also originate from
residual precursors. Additional to the expected surface
impurities, linearly bound CO2 is observed with a band at ~2346
cm-1.[28] The observation of this band is surprising, because such
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
surface of the materials, but are, instead, integrated (trapped)
into their bulk structure. This is supported by the lack of any
changes in these bands after the photocatalytic reaction (Figure
S4), which also proves that these species are not consumed
during the photocatalytic process. For a further discussion of the
peculiar bands of “trapped” CO2 and CO, please see the
Supporting Information. It is interesting to note that at least some
of the gold present in the samples must be slightly negatively
charged, as evidenced by the position of the stretching vibration
of adsorbed CO (2052 cm-1).
Figure 3. UV-vis spectra of Au@TiO2 samples (without Au decoration on the
outside of the TiO2 shell).
Activity tests in photocatalytic CO2 reduction
Preliminary photocatalytic tests of all samples in photocatalytic
CO2 reduction are displayed in Figure 5. All samples have been
measured under identical batch reaction conditions and have
been thoroughly cleaned photocatalytically prior to the reaction
with CO2. The gas composition in all cases consisted of 1.5%
CO2 and ~6000 ppm H2O in ultrapure Helium. In this
measurement series, the used GC system generally allows to
detect methane, methanol and higher hydrocarbons from
concentrations as low as ~1 ppm. H2 and CO cannot be
detected with this device.
From all detectable components, methane was the main product
formed on all samples (Figure 5). While traces of ethane were
detected, methanol was never observed. It is very obvious that
sample Au@TiO2-20-Au is much more active than all other
samples used in this study. Considering the activity of all six
samples, it seems to be
a general trend that methane
productivity increases with a thicker TiO2 shell, and also with
additional gold nanoparticle deposition on the outer TiO2 surface.
It is thus straightforward to expect sample Au@TiO2-20-Au to be
the best one, because it features both a fairly thick TiO2 shell
and gold nanoparticles on the outer surface.
To put the results into perspective, in-depth studies were
performed with the two most active samples with the thickest
TiO2 shell, Au@TiO2-20 and Au@TiO2-20-Au. Measurements
were conducted in an identical set-up, but the gas analysis was
performed with a different GC. This advanced GC analyzes the
gas atmosphere with a barrier discharge ionization detector
Figure 4. DRIFT spectra of as made Au@TiO2 and Au@TiO2-Au samples
after purging in inert gas; top: full range spectra; bottomt: zoom into hydroxyl
group region.
species are usually only observed during low-temperature
adsorption. The same is true for the very weak band observed at
2052 cm-1, which is assigned to CO bound to slightly negatively
charged Au. The band shift to such low wavenumbers may be
explained by the Stark effect or by metal back-donation into
antibonding orbitals of the adsorbed CO.[29] The adsorption of
CO and/or CO2 onto TiO2 and Au/TiO2 is usually known to be
very unstable and it is thus usually performed at low
30
]
temperatures of ~60 to 110 K,[
and rarely at room
temperature.[29] The observations are different in the present
work. Figure S3 (Supporting Information) shows that both
species cannot be removed from Au@TiO2 samples even with a
heat treatment up to 200 °C. In fact, the band intensity is not
even decreased to the slightest extent. It is thus concluded that
the adsorbed CO and CO2 species are not present on the
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
Figure 5. Activity screening for methane formation on all Au@TiO2 and
Au@TiO2-Au samples in photocatalytic CO2 reduction (1.5%CO2, ~6000 ppm
H2O) over 6 h.
Even though all tested catalysts have a different structural
combination of Au in and/or on TiO2 with differing nanoparticle
geometries, the product profiles during CO2 reduction are
largely similar. H2 is formed as overall main product, and CH4
and CO are the main carbon-containing products.
There are two significant differences between the Au@TiO2
core-shell samples and the commercial Au/TiO2 samples. Firstly,
the overall amount of products (related to catalyst mass) is
larger for the core-shell samples, in particular with respect to H2.
Secondly, the core-shell samples also produce significant
amounts of CO, while this product is almost completely absent
for the Au/TiO2 catalysts with conventional geometry. We assign
this to the high CO oxidation activity of the conventional or
commercial Au/TiO2 catalysts already at low temperatures. This
is supported by the temporal evolution of CO: While some CO is
formed initially, it disappears again over the course of time; likely
because it is reoxidized to CO2.
Although a different GC is used for the detailed measurements,
the relative trends in methane production of the two different
Au@TiO2 core-shell systems are fairly well reproduced. Sample
Au@TiO2-20-Au, with additional gold nanoparticles on the
outside of the shell, is roughly three times as active for the
formation of methane as the sample Au@TiO2-20 without the
additional external gold nanoparticles. With the other GC (Figure
4) this ratio was about 4. There are certain deviations in
absolute numbers. In the first measurement with sample
Au@TiO2-20-Au (Figure 4), the amount of methane formed over
the course of six hours was roughly 6.0 µmol/gcat. In the
measurement with the GC containing the BID, methane
production within six hours amounted to 5.05 µmol/gcat. For
sample Au@TiO2-20, methane production was ~1.5 and 1.8
µmol/gcat in the two measurements, respectively. Rather than
being measurement errors, the deviations may be due to a
differing sample orientation towards the incoming light. This has
been confirmed recently for pure TiO2.[ 31 ] In summary, the
relative activity of the samples can be considered to be well
reproducible.
Figure 6. Results of activity testing in CO2 reduction with 6,000 ppm of H2O
and 1.5 % CO2 over a course of 6 h, 200 W Hg/Xe lamp with IR filter; (A)
Au@TiO2-20-Au, (B) Au@TiO2-20, (C) AUROliteTM, (D) Au/P25.
(BID). With this GC it was possible to quantify not only the
formed amounts of methane and ethane, but also of CO and H2.
One GC sampling every 45 min is possible. The produced
amounts were recorded both during the cleaning step and in the
photocatalytic reaction. The results of the initial batch cleaning
procedures are displayed and reported in the supporting
information (Figure S5). After the batch cleaning procedures, the
samples were continuously exposed to cleaning treatment in the
flow mode. Monitoring of desorbing species is not possible
under flow conditions, but after a few hours of flow cleaning, the
reactor is closed off (batch cleaning) and desorbing
contaminants are accumulated for 1 hour before a GC sampling
is performed. This process is repeated until the desorbing
Discussion
The structural characterization of the samples by means of TEM
and XRD confirms the core-shell structure of the materials and
the presence of TiO2 in the anatase phase. The thickness of the
titania shell has an influence on the plasmon absorption of the
central gold core, which shifts to higher wavelength (lower
energy) with thicker TiO2 shell. Although the additional gold
nanoparticles on the outer surface are not detectable by either
XRD or UV-Vis, they are clearly visible in the TEM images for
the Au-TiO2-08-Au sample (Figure 1B). For all samples, the
changes in the hydroxyl group coverage observed by DRIFTS
are an indication for their anchoring to the outer TiO2 shell. At
least some of the gold in the samples must be partially
negatively charged, as evidenced by the position of the weak
vibrational band of adsorbed CO, clearly visible for samples
Au@TiO2-11 and Au@TiO2-11-Au.
amounts of CO, CH4, and H2 were <0.06 μmol/gCat
.
Subsequently, the reactor was filled with the CO2-containing gas
mixture (1.5% CO2, 6000 ppm H2O in He), and the product
formation under batch CO2 reduction conditions was recorded
with the advanced GC for a period of six hours. Apart from the
two core-shell systems, an industrial AUROliteTM catalyst and
the two homemade Au/TiO2 catalysts with conventional
arrangement of gold nanoparticles were also tested (cleaning
was performed identically for all samples; results of the
AUROliteTM catalyst are shown in Figure S5 as an example).
The results of the two Au@TiO2 core-shell systems, the
commercial AUROliteTM and the Au/P25 sample are displayed in
Figure 6. The results of the activity test with the home-made
Au/anatase sample and a reference measurement with pure P25
is displayed in the supporting information (Figure S6). Using the
advanced GC for product analysis, it becomes obvious that the
main product from photocatalytic reactions is not derived from
CO2. Instead, H2 is formed in major quantities with all samples.
All samples are active in photocatalytic CO2 reduction, with the
yields of CH4 increasing both with a thicker TiO2 shell and with
the addition of the small nanoparticles on the outer shell. In this
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
respect, the sample with the thickest shell and the additional
outer nanoparticles (Au@TiO2-20-Au) is by far the most active
sample for the formation of CH4. This was confirmed by two
separate measurements using two different gas chromatographs
for analysis. In-depth studies of the two most active samples
(Au@TiO2-20 and Au@TiO2-20-Au), conducted using an
advanced GC, reveal that CH4 is not the overall main product.
Instead, hydrogen is the predominant product on all samples
studied here, with the yields of H2 exceeding those of CH4 by
roughly an order of magnitude. Hydrogen is also the main
product with the conventional Au/TiO2 reference samples, but
for all conventional Au/TiO2 samples the yields of CH4 are
smaller than those observed for the most active Au@TiO2-Au
sample. Concerning the product distribution, the most striking
difference between the conventional Au/TiO2 photocatalysts and
the core-shell photocatalysts is that the latter also form CO,
while this product is hardly observed on the conventional
catalysts. This is attributed to the high activity of the
conventional Au/TiO2 catalysts in low-temperature CO oxidation,
the actual target reaction for which this photocatalyst system has
been developed. The new core-shell system is thus beneficial in
two respects: Total yields of valuable products are larger, and
CO is observed as an additional product.
UV-vis measurements show that light absorption is governed
both by the Au nanoparticle in the core and the semiconducting
titania shell (Figure 3). The additional small gold nanoparticles
on the outer surface do not exert a significant influence on light
absorption, because they did not cause a visible change in the
UV-Vis spectra (not shown). Considering only the Au@TiO2
samples without additional gold nanoparticles on the outer shell
in a first step, the well-developed structure of the semiconductor
TiO2 in the shell is an important parameter for high activity of the
core-shell systems under UV+vis illumination. One reason may
be a better developed crystal structure (less defects) in the
thicker shell. A possible enhancing influence of “hot electron
transfer” from the gold core to the titania shell seems less
relevant, since such a process is expected to occur most likely in
the samples with the thinnest titania shell (8 nm).
A further decoration with Au nanoparticles on the outer TiO2
surface results in a remarkable three- to fourfold increase in
activity for methane formation in the case of the thickest TiO2
shell. This is clear evidence that the nanoparticles on the outer
surface are much more relevant for photocatalytic CO2 reduction
than the gold core. Both a plasmonic effect of the small external
nanoparticles, as well as improved catalytic properties due to the
exposed gold surface on the small nanoparticles may cause this
significant increase in activity.
The results of the activity tests with the conventional Au/TiO2
reference samples show that a pure catalytic effect alone cannot
explain the higher activity. The conventional catalysts feature
exposed gold nanoparticles of similar size, but total
photocatalytic activity is lower and product distributions are
altered, in particular with respect to the lack of CO production.
The same observation also rules out plasmonic or structural
effects related to the small size of the nanoparticles on the outer
shell as sole reason for the observed differences. If those were
the decisive factor, the reference samples should behave similar
to the core-shell samples. A simple effect of a larger amount of
gold in the core-shell samples is also unlikely: The Au/P25
sample with a much higher gold loading compared to the
Au/anatase samples showed the same trends as the other
Au/TiO2 reference samples, in which predominantly H2 and
almost no CO was formed. Instead, it seems that three aspects
have to work together for the significantly improved
photocatalytic activity for the formation of CH4 and CO, namely
(i) a structurally well-developed anatase TiO2 shell, (ii) a gold
core and (iii) exposed gold nanoparticles on the outer surface. In
such a system, the combination of Au and TiO2 shows
fundamental differences in the (photo)catalytic properties,
evidenced particularly by the inability of the core-shell system to
(re)oxidize CO. Whereas in all conventional Au/TiO2 catalysts
the initially formed CO disappears again within the first few
hours of the experiment, the amounts of CO formed with the
core-shell samples roughly equal the produced amounts of CH4.
The different behavior may originate from structural and
electronic factors. Electromagnetic fields caused by light
irradiation of the core-shell metal-semiconductor-metal structure
may be very different from those caused in conventional
Au/TiO2 catalysts. Furthermore, the interface structure may be
rather different in the core-shell system compared to samples
prepared via the colloidal deposition route, in which reduction
and rutilization of the interface has recently been observed.[32]
No mechanistic studies were conducted here, but the roughly
equal amounts of CO and CH4 formed photocatalytically may be
indicative for the decomposition of a common intermediate in the
mechanism. One option may be acetaldehyde, suggested
previously as intermediate on bare TiO2.[33,37] It must be noted
that the mechanism of CH4 formation on TiO2 is still under
debate, particularly because some of the proposed elementary
steps are thermodynamically not possible on bare TiO2, such as
the formation of the CO2 anion radical located at -1.9 eV on the
electrochemical energy scale. In thermal catalysis with
a
Au/TiO2 system it was found that acetaldehyde strongly
interacts with the perimeter site of Au nanoparticles on TiO2.[34]
If such an interaction also takes place during the photoreaction,
acetaldehyde could be stabilized better with the Au@TiO2-20-Au
catalyst compared to the Au@TiO2-20 catalyst, because the
former has more Au nanoparticle-TiO2 interface. According to
Shkrob et al.,[33] under reaction conditions acetaldehyde breaks
apart and forms CO and CH4. This would explain the product
distribution observed for Au@TiO2-20-Au and Au@TiO2-20,
where CO and CH4 appear in a ratio close to 1 (Figure 6). The
decomposition of an intermediate such as acetaldehyde might
occur similarly on the conventional Au/TiO2 catalysts, such as
AUROlite™ and Au/P25 (Figure 6), but CO is quickly oxidized to
CO2 and does not show the same time profile as for the
Au@TiO2-Au catalysts. Further insight into this issue may be
obtained in future studies.
The results obtained here also indicate that H2, instead, can be
formed photocatalytically under UV+Vis illumination on any
accessible gold surface, regardless whether this surface belongs
to the gold core or to the small nanoparticles on the outer
surface. Since the amounts of H2 formed on samples Au@TiO2-
20-Au and Au@TiO2-20 are almost equal, the Au core seems to
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
have the dominant effect in both cases. In the reference
samples, both a higher gold loading (as in Au/P25) and a larger
size of the gold nanoparticles (as in Au/anatase) seem to
influence the yields of H2 favorably, because in both cases more
H2 is obtained compared to the commercial AUROliteTM catalyst.
Within the first 45 min the largest increase in H2 formation can
be detected, whereby after that period the formation is slowing
down. Since the formation of all products observed in the
present study occurs by processes consuming only electrons,
this decreasing production rate of H2 for all tested samples
(Figure 6) can be explained by the lack of an efficient hole
consuming reaction, that is, the lack of formation of gaseous
dioxygen. As has been described previously for bare TiO2,[31,37]
instead of the formation of O2 in the gas phase, oxygen-derived
species remain on or in the photocatalyst. The presence of
those species eventually favors the back reaction, that is,
oxidation of the formed hydrocarbons and hydrogen to CO2 and
H2O. Further studies on the fate of the holes are in progress and
will be reported at a later time.
additionally formed. The sample with the thickest TiO2 shell was
most active, indicating a major role of the well-developed
semiconductor structure in the reaction. Plasmonic field effects
or a “hot electron transfer” from the gold core seem to be of
minor importance, but may affect the reactivity of the small gold
nanoparticles on the outer TiO2 shell. The improved formation of
carbon-containing products is attributed to an electronic
interaction and a synergetic effect of all three structural features:
(i) the thick TiO2 shell, (ii) the gold core, and (iii) the Au NPs on
the outside. In this arrangement, the latter lose their ability to
rapidly oxidize CO at low temperature, so this product is stable
on the core-shell system. CO and CH4 formation in a ~1:1 ratio
may indicate that both products originate from a common
intermediate. H2 formation is instead observed to a similar
extent on all samples, indicating that this reaction can proceed
on any accessible gold surface, and no particular geometry is
needed.
Our observations are summed up in Figure 7, which provides a
schematic representation of the mode of action of the whole
core-shell system. Charge carriers are generated by
photoexcitation of the semiconductor, that is, the TiO2 shell.
While the hole presumably remains and reacts on or in the TiO2
shell (and is not further considered here), the electron migrates
either to the gold core or to the Au NPs on the outer TiO2 shell.
The reduction of CO2 to CO and CH4 likely occurs on the
interface of the Au NPs exposed on the outer TiO2 shell. The
formation of hydrogen, on the other hand, can proceed on any
exposed gold surface, both on the outer nanoparticles and the
gold core.
Experimental Section
Preparation of Au@TiO2
Au@TiO2 core-shell systems were synthesized based on a method
reported elsewhere[22f]
,
which was modified by our group. Au
nanoparticles used as the core material were prepared by the well-known
Turkevich method via citrate reduction. HAuCl4 • 3 H2O (108 mg,
0.274 mmol, Sigma-Aldrich, > 99.9 %) was dissolved in H2O (180 mL,
millipore Q), heated to 115 °C and kept boiling for 15 min. Subsequently,
sodium citrate dihydrate (180 mg in 5 mL H2O, Sigma-Aldrich) was
quickly added into the vortex of the stirred solution. The mixture is boiled
under reflux for 30 min and, subsequently, cooled down. Aqueous
hydroxypropyl cellulose (300 mg dissolved in 15 mL H2O, Sigma-Aldrich)
is added and the solution is stirred overnight. Gold nanoparticles with
mean particle diameters of 20 nm can be obtained by this procedure. The
aqueous gold solution is concentrated via centrifugation (final volume
25 mL) and mixed with isopropanol (90 mL, Sigma-Aldrich, 99.9 %) in a
250 mL round bottom flask. Aqueous NH3 (3 mL, Sigma-Aldrich, 28 –
30 wt.%) is added to the dispersion and stirred vigorously for 15 min.
Titanium diisopropoxide bis(acetylacetonate) (144 µL in 10 mL
isopropanol for 8 nm shell, Aldrich, 75 wt.% dissolved in isopropanol) is
slowly added into the vortex by a syringe pump and the mixture is further
stirred overnight at RT. The material is washed twice with deionized
water. Subsequently, the purple powder is dried at 90 °C for 1.5 h and
calcined for 1 h at 500 °C (heating ramp: 4 °C min-1). Samples are
denoted as Au@TiO2-X, where X indicates the thickness of the TiO2
shell in nanometers. For the variation of the shell thickness, only the
amount of the titania precursor was varied, while all other synthesis
parameters were kept constant.
Figure 7. Schematic representation of the mode of action of the core-shell
Preparation of Au@TiO2 decorated with Au nanoparticles
The decoration with Au nanoparticles on top of the Au@TiO2 core-shell
materials was achieved by a facile deposition-precipitation method with
urea reported by Zanella et al.[35] Au@TiO2 (20 mg) is dispersed in water
(10 mL, millipore Q) together with urea (0.252 g, Fluka) and the desired
amount of HAuCl4 3 H2O for a total metal loading of 1 wt% at the outer
TiO2 surface. The mixture is then heated up to 90 °C and stirred for 4 h.
The material is washed three times with deionized water and dried for 2 h
at 50 °C under vacuum. The dried samples are then further calcined at
300 °C for 4 h (heating ramp: 2 °C min-1). Resulting samples are denoted
as Au@TiO2-X-Au where X indicates the thickness of the TiO2 shell in
nanometers. An overview can be found in Table S1.
Au@TiO2-Au system.
Conclusions
In this work Au@TiO2-Au catalysts with core-shell morphology
were found to be superior to conventional Au/TiO2 catalysts in
the photocatalytic reduction of CO2. CH4 and CO were the main
carbon-containing products, and significant amounts of H2 are
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
Preparation of Au/TiO2 reference samples
We kindly acknowledge Bernd Spliethoff for the assistance in
HR-TEM analysis. This work is partly supported by MAXNET
Energy research consortium of the Max Planck Society and the
Cluster of Excellence RESOLV (EXC 1069) funded by the
Deutsche Forschungsgemeinschaft. Part of this work was
funded by the German Ministry of Education and Research
(BMBF) within the scope of the funding program “Technologies
for Sustainability and Climate Protection – Chemical Processes
and Use of CO2” (033RC1007A, PhotoKat).
Gold nanoparticles with an average diameter of 4 nm were deposited on
different commercial TiO2 samples using the colloidal deposition method.
The synthesis using the commercial anatase material (~300 m2/g,
Huntsman P&A Germany) is reported in detail in a recent contribution.[32]
The synthesis of the Au/P25 catalyst was performed similarly, using
commercial TiO2 P25 (~50 m2/g, Evonik Industries). The gold loading on
the anatase support was adjusted to 1 wt%. On the P25 support, the
loading of gold was 7.5 wt%. A commercial catalyst (AUROlite™, 1.5
wt% Au on TiO2, STREM Chemicals) was tested as reference. An
overview over all samples studied in this contribution can be found in the
supporting information (Table S1 and S2).
Keywords: High-purity photocatalysis • Gold catalysis •
Activity Testing in Photocatalytic CO2 Reduction
Titanium dioxide • Artificial photosynthesis • Core-shell materials
The activity tests were performed in
a home-made, metal-sealed,
vacuum-tight photoreactor. detailed description can be found
A
elsewhere.[ 36] The full spectrum of a 200 W Hg/Xe lamp is used to
irradiate the photocatalyst (UV+Vis), but the IR part is removed by an IR
water filter attached in front of the lamp output lens to prevent heating of
the sample through the beam. The IR filter is cooled by a cryostat
providing a circulation of cold water (5 °C). Water is dosed by saturation
of the gas phase using stainless-steel saturators cooled to 5°C. The
concentration of water vapor in the gas phase at 5°C and ~1500 mbar
pressure is estimated to be ~6000 ppm. For each test 25 mg of the
respective photocatalyst were used. Details concerning the photocatalytic
cleaning procedure and the actual photocatalytic CO2 reduction can be
found in Ref. [ 37 ]. In brief, each catalyst sample was cleaned
photocatalytically in an atmosphere containing only gaseous water and
helium. This removed hydrocarbon impurities and minimized product
formation from carbon sources other than CO2.[36,37] For testing the
activity in photocatalytic CO2 reduction, each catalyst was brought in
contact with a reactant gas mixture of 1.5 % of CO2 and 0.6 % of H2O
balanced by He. The addition of H2O to the feed gas is enabled by a
[1] M. Halmann, Nature 1978, 275, 115.
[2] E. V. Kondratenko, G. Mul, J. Baltrusaitis, G. O. Larrazábal, J. Pérez-
Ramírez, Energy Environ. Sci 2013, 6, 3112.
[3] a) M. Haruta, Catalysis Today 1997, 36, 153–166; b) T. Hayashi, K.
Tanaka, M. Haruta, J. Catal. 1998, 178, 566–575.
[4] a) M. C. Holz, K. Kähler, K. Tölle, A. C. van Veen, M. Muhler, Phys. Status
Solidi B 2013, 250, 1094–1106; b) D. Widmann, R. J. Behm, Acc. Chem. Res
2014, 47, 740–749.
[5] a) M. Chen, D. Goodman, Catalysis Today 2006, 111, 22–33; b) M. Chen,
K. Luo, D. Kumar, W. Wallace, C.-W. Yi, K. Gath, D. Goodman, Surface
Science 2007, 601, 632–637.
[6] a) M. Comotti, W.-C. Li, B. Spliethoff, F. Schüth, J. Am. Chem. Soc 2006,
128, 917–924; b) A. S. Wörz, U. Heiz, F. Cinquini, G. Pacchioni, J. Phys.
Chem. B 2005, 109, 18418–18426.
[7] a) A. L. Linsebigler, G. Q. Lu, J. T. Yates, Chem. Rev. 1995, 95, 735-758;
b) P. V. Kamat, J. Phys. Chem. C 2007, 111, 2834-2860; c) A. Kudo, Y.
Miseki, Chem. Soc. Rev. 2009, 38, 253-278; d) F. E. Osterloh, Chem. Mater.
2008, 20, 35-54; e) X. Li, J. G. Yu, J. X. Low, Y. P. Fang, J. Xiao, X. B. Chen,
J. Mater. Chem. A 2015, 3, 2485-2534; f) J. Yang, D. Wang, H. Han, C. Li,
Acc. Chem. Res. 2013, 46, 1900-1909; g) Y. Ma, X. Wang, Y. Jia, X. Chen, H.
Han, C. Li, Chem. Rev. 2014, 114, 9987-10043
metal-sealed double saturator which is kept at 5 °C to yield
a
[8] A. Lüken, M. Muhler, J. Strunk, Phys. Chem. Chem. Phys 2015, 17,
10391–10397.
[9] B. Mei, C. Wiktor, S. Turner, A. Pougin, G. van Tendeloo, R. A. Fischer, M.
Muhler, J. Strunk, ACS Catal 2013, 3, 3041–3049.
[10] G. Dodekatos, H. Tüysüz, Catal. Sci. Technol. 2016, 6, 7307-7315.
[11] S. Schünemann, G. Dodekatos, H. Tüysüz, Chem. Mater. 2015, 27, 7743-
7750.
concentration of about 0.6 % of H2O (estimated with the Antoine
equation). The reactor was run in the batch mode. The illumination time
was 6 h for each experimental run. Initially, the photoreactor was filled
with the gas mixture to an overall inside pressure of 1.5 bar. Since the
pressure drops by 100 mbar upon each GC sampling, the pressure drops
to ~1.1 bar within the course of the measurement.
[12] P. Christopher, H. Xin, A. Marimuthu, S. Linic, Nat Mater 2012, 11, 1044-
1050.
[13] a) S. Sarina, E. R. Waclawik, H. Y. Zhu, Green Chem. 2013, 15, 1814-
1833; b) G. Dodekatos, S. Schünemann, H. Tüysüz, Top. Curr. Chem. 2016,
371, 215-252; c) M.A. El-Sayed, Acc. Chem Res. 2001, 34, 257-264.
[14] S. Linic, P. Christopher, D. B. Ingram, Nat. Mater. 2011, 10, 911-921.
[15] K. L. Kelly, E. Coronado, L. L. Zhao, G. C. Schatz, J. Phys. Chem. B
2003, 107, 668–677.
[16] M. Gratzel, Nature 2001, 414, 338-344.
[17] L. Du, A. Furube, K. Yamamoto, K. Hara, R. Katoh, M. Tachiya, J. Phys.
Chem. C 2009, 113, 6454–6462.
[18] K. Li, B. S. Peng, T. Y. Peng, ACS Catal. 2016, 6, 7485-7527.
[19] W. Hou, W. H. Hung, P. Pavaskar, A. Goeppert, M. Aykol, S. B. Cronin,
ACS Catal 2011, 1, 929–936.
[20] M. Dilla, A. Pougin, J. Strunk, J. Energy Chem. 2017, 26, 277-283.
[21] R. Amrollahi, M. S. Hamdy, G. Mul, Journal of Catalysis 2014, 319, 194–
199.
Structural Characterization
Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS)
was measured using a Nicolet Protégé 460 spectrometer equipped with
praying mantis mirrors and an in situ cell (Harrick HVC-DRP-2). The
sample was purged with inert gas throughout the measurement and the
temperature was kept at 35 °C. Since TiO2 is predominantly hydrophilic,
inert gas purge is carried out for 1 h before the first spectrum is recorded
to remove the majority of physisorbed H2O from the catalyst surface.
UV-vis spectroscopy was performed on
a Varian Cary 5 G. The
Au@TiO2 materials were suspended in isopropanol prior to
measurement. The suspension was loaded into quartz cuvettes and
measured in transmission mode.
[22] a) E. Ha, L. Y. S. Lee, H. W. Man, S. C. E. Tsang, K. Y. Wong, ACS Appl.
Mater. Interfaces 2015, 7, 9072-9077; b) S. K. Cushing, J. Li, J. Bright, B. T.
Yost, P. Zheng, A. D. Bristow, N. Wu, J. Phys. Chem. C 2015, 119, 16239-
16244; c) T. Takata, C. S. Pan, M. Nakabayashi, N. Shibata, K. Domen, J.
Am. Chem. Soc. 2015, 137, 9627-9634; d) B.-H. Wu, W.-T. Liu, T.-Y. Chen,
T.-P. Perng, J.-H. Huang, L.-J. Chen, Nano Energy 2016, 27, 412-419; e) C.
H. Li, M. C. Li, S. P. Liu, A. C. Jamison, D. Lee, T. R. Lee, T. C. Lee, ACS
Appl. Mater. Interfaces 2016, 8, 9152-9161; f) Z. W. Seh, S. Liu, M. Low, S.-Y.
Zhang, Z. Liu, A. Mlayah, M.-Y. Han, Adv. Mater. 2012, 24, 2310-2314.
[23] a) X. Zhang, Y. Zhu, X. Yang, S. Wang, J. Shen, B. Lin, C. Li, Nanoscale
2013, 5, 3359-3366; b) S. M. Yoo, S. B. Rawal, J. E. Lee, J. Kim, H.-Y. Ryu,
D.-W. Park, W. I. Lee, Appl. Catal., A 2015, 499, 47-54; c) Y. Q. Sun, Y. G.
Sun, T. Zhang, G. Z. Chen, F. S. Zhang, D. L. Liu, W. P. Cai, Y. Li, X. F. Yang,
C. C. Li, Nanoscale 2016, 8, 10774-10782; d) R. Lee, Y. Kumaresan, S. Y.
Transmission electron microscopy (TEM) studies were performed on a
Hitachi H- 7100 with 100 kV acceleration voltage. Wide angle XRD
patterns collected at room temperature were recorded on
a Stoe
theta/theta diffractometer in Bragg-Brentano geometry (Cu Kα1/2
radiation). The measured patterns were evaluated qualitatively by
comparison with entries from the ICDD PDF-2 powder pattern database.
Acknowledgements
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
Yoon, S. H. Um, I. K. Kwon, G. Y. Jung, RSC Adv. 2017, 7, 7469-7475; e) Y.
Li, J. Zhao, W.-L. You, D. Cheng, W. Ni, Nanoscale 2017, 9, 3925-3933.
[24] a) W. G. Tu, Y. Zhou, H. J. Li, P. Li, Z. G. Zou, Nanoscale 2015, 7, 14232-
14236; b) S. Bera, J. E. Lee, S. B. Rawal, W. I. Lee, Appl. Catal., B 2016, 199,
55-63.
[30] a) L. Mino, A. M. Ferrari, V. Lacivita, G. Spoto, S. Bordiga, A. Zecchina, J.
Phys. Chem. C 2011, 115, 7694–7700; b) C. Deiana, M. Minella, G. Tabacchi,
V. Maurino, E. Fois, G. Martra, Phys. Chem. Chem. Phys 2013, 15, 307-315;
c) H. Noei, L. Jin, H. Qiu, M. Xu, Y. Gao, J. Zhao, M. Kauer, C. Wöll, M.
Muhler, Y. Wang, Phys. Status Solidi B, 2013, 250, 1204-1221; d) W. Grünert,
D. Großmann, H. Noei, M.-M. Pohl, I. Sinev, A. De Toni, Y. Wang, M. Muhler,
Angew. Chem. Int. Ed. 2014, 53, 3245-3249.
[25] K. L. Kelly, E. Coronado, L. L. Zhao, G. C. Schatz, J. Phys. Chem. B
2003, 107, 668–677.
[26] a) A. Šurca Vuk, R. Ješe, M. Gaberšček, B. Orel, G. Dražič, Solar Energy
Materials and Solar Cells 2006, 90, 452–468; b) J. Soria, J. Sanz, I. Sobrados,
J.M. Coronado, A.J. Maira, M.D. Hernández-Alonso, F. Fresno, J. Phys.
Chem. C 2007, 111, 10590-10596; c) S. H. Szczepankiewicz, J. A. Moss, M.
R. Hoffmann, J. Phys. Chem. B 2002, 106, 7654–7658.
[31] M. Dilla, R. Schlögl, J. Strunk, ChemCatChem 2017, 9, 696-703.
[32] A. Pougin, A. Lüken, C. Klinkhammer, D. Hiltrop, M. Kauer, K. Tölle, M.
Havenith, K. Morgenstern, W. Grünert, M. Muhler, J. Strunk, Top. Catal. 2017,
60, 1744-1753.
[33] I. A. Shkrob, T. W. Marin, H. He, P. Zapol, J. Phys. Chem. C 2012, 116,
[27] a) J. T. Carneiro, T. J. Savenije, J. A. Moulijn, G. Mul, J. Phys. Chem. C
2011, 115, 2211–2217; b) P. Du, A. Buenolopez, M. Verbaas, A. Almeida, M.
Makkee, J. Moulijn, G. Mul, Journal of Catalysis 2008, 260, 75–80.
[28] G. Ramis, G. Busca, V. Lorenzelli, Materials Chemistry and Physics 1991,
29, 425–435.
9450–9460.
[34] M. C. Holz, K. Tölle, M. Muhler, Catal. Sci. Technol 2014, 4, 3495–3504.
[35] R. Zanella, S. Giorgio, C. R. Henry, C. Louis, J. Phys. Chem. B 2002, 106,
7634–7642.
[36] B. Mei, A. Pougin, J. Strunk, Journal of Catalysis 2013, 306, 184–189.
[37] A. Pougin, M. Dilla, J. Strunk, Phys. Chem. Chem. Phys. 2016, 18, 10809-
10817.
[29] F. Boccuzzi, A. Chiorino, M. Manzoli, Surface Science 2000, 454-456,
942–946.
This article is protected by copyright. All rights reserved.

10.1002/chem.201801796
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
Layout 1:
FULL PAPER
Wrapped up and decorated:
With Au@TiO2 core-shell
composites featuring additional
Au nanoparticles on the outer
surface, higher yields of CH4 and
CO can be obtained in
Anna Pougin[a],†, Georgios
Dodekatos[b], Martin Dilla[c], Harun
Tüysüz[b]*, Jennifer Strunk[d]*
Page No. – Page No.
photocatalytic CO2 reduction.
Large amounts of H2 from H2O
splitting are simultaneously
produced. Electronic effects and
a catalytic role of gold dominate
the performance of the system.
Au@TiO2 Core-Shell Composites
for the Photocatalytic Reduction
of CO2
This article is protected by copyright. All rights reserved.