Covalent Immobilization of a Molecular Catalyst on Cu2O Photocathodes for CO2 Reduction

Article abstract of DOI:10.1021/jacs.5b12157

Sunlight-driven CO₂ reduction is a promising way to close the anthropogenic carbon cycle. Integrating light harvester and electrocatalyst functions into a single photoelectrode, which converts solar energy and CO₂ directly into reduced carbon species, is under extensive investigation. The immobilization of rhenium-containing CO₂ reduction catalysts on the surface of a protected Cu₂O-based photocathode allows for the design of a photofunctional unit combining the advantages of molecular catalysts with inorganic photoabsorbers. To achieve large current densities, a nanostructured TiO₂ scaffold, processed at low temperature, was deposited on the surface of protected Cu₂O photocathodes. This led to a 40-fold enhancement of the catalytic photocurrent as compared to planar devices, resulting in the sunlight-driven evolution of CO at large current densities and with high selectivity. Potentiodynamic and spectroelectrochemical measurements point toward a similar mechanism for the catalyst in the bound and unbound form, whereas no significant production of CO was observed from the scaffold in the absence of a molecular catalyst.

Full text of DOI:10.1021/jacs.5b12157

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Covalent Immobilization of a Molecular Catalyst
2
2
on CuO Photocathodes for CO Reduction
Marcel Schreier, Jingshan Luo, Peng Gao, Thomas Moehl, Matthew T. Mayer, and Michael Grätzel
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b12157 • Publication Date (Web): 23 Jan 2016
Downloaded from http://pubs.acs.org on January 24, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Covalent Immobilization of a Molecular Catalyst on Cu O
2
Photocathodes for CO Reduction
2
Marcel Schreier, Jingshan Luo, Peng Gao, Thomas Moehl, Matthew T. Mayer*, Michael Grätzel*
Institut des Sciences et Ingénierie Chimiques, Ecole Polytechnique Fédérale de Lausanne, Station 6, CH-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
015 Lausanne, Switzerland
ABSTRACT
Sunlight-driven CO reduction is a promising way to close the anthropogenic carbon cycle. Integrating
2
light harvester and electrocatalyst functions into a single photoelectrode, which converts solar energy and
CO directly into reduced carbon species, is under extensive investigation. The immobilization of
rhenium-containing CO reduction catalysts on the surface of a protected Cu O-based photocathode allows
for the design of a photofunctional unit combining the advantages of molecular catalysts with inorganic
photoabsorbers. To achieve large current densities, a nanostructured TiO scaffold, processed at low
temperature, was deposited on the surface of protected Cu O photocathodes. This led to a 40-fold
2
2
2
2
2
enhancement of the catalytic photocurrent as compared to planar devices, resulting in the sunlight-driven
evolution of CO at large current densities and with high selectivity. Potentiodynamic and
spectroelectrochemical measurements point towards a similar mechanism for the catalyst in the bound-
and unbound form whereas no significant production of CO was observed from the scaffold in absence of
a molecular catalyst.
Keywords: CO
2
reduction, Re-catalyst, Cu
2
O, TiO scaffold, photoelectrochemistry
2
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
Ever increasing atmospheric concentrations of carbon dioxide (CO
2
) are having a profound impact on the
1
world’s climate system. The need to control CO
2
levels has led to intense research into alternative energy
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
systems that do not rely on fossil fuels as a source of primary energy. Solar energy is an overwhelmingly
2
abundant resource great enough to fulfill all of the world’s energy demands, but due to its intermittent
3
nature, it must be stored to become a truly useful source of energy. Various approaches for the storage of
solar power have been investigated, including its direct transformation into chemical energy to yield “solar
fuels” which has received substantial interest in recent years.4,5 The most attractive approach entails using
sunlight to reduce CO
2
into carbon-based fuels,6–8 thereby allowing the use of the existing energy
infrastructures while simultaneously closing the anthropogenic carbon cycle.
Achieving direct conversion of sunlight and CO into chemical fuels requires the linking of light
2
absorbing and catalytic components, accomplished in nature by the process of photosynthesis. A
promising artificial system for the production of solar fuels at higher efficiency than the natural
9
counterpart is the photoelectrochemical cell, a device in which semiconductors, functionalized with
catalysts, are illuminated while immersed in an electrolyte containing the reactants. Photo-generated
charges in the semiconductor are separated and directed to the respective catalysts, where they
subsequently drive the fuel generating reactions. This approach has been extended to tandem structures,
achieving unassisted water splitting and CO
reduction.10–16
2
The development of efficient and selective catalysts is crucial to the success of catalytic fuel-forming
reactions. Molecular catalysts for hydrogen evolution17,18 and carbon dioxide reduction19 have attracted
significant attention due to their well-controlled properties and tunable nature, making them an interesting
20
subject of study. Significant progress has been made with respect to stability and activity towards the
desired reactions.21
ACS Paragon Plus Environment

Page 3 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
An attractive approach to driving molecular catalysts using the power of sunlight is to incorporate the
catalysts into photoelectrochemical systems. Such an application can be achieved in either homogenous or
heterogeneous approaches, whereas the former employs the catalysts as a dissolved species in the
electrolyte.7,22,23 This, however, requires large amounts of catalyst, most of which is not actively
participating in the reaction, and in addition, causes parasitic light absorption, thereby limiting the
photocurrent density observed from these systems.7,24 Furthermore, many molecular catalysts rely on
precious metal centers, thus providing incentive for minimizing the amount of required catalyst as much
as possible. Both of these challenges can be addressed by binding the molecular catalyst directly to the
surface of the photoelectrode, which comes down to heterogenizing a homogeneous catalyst. Thereby, the
use of catalyst is reduced to an amount which is actively participating in the reaction, and some recent
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
works have demonstrated the ability of this approach to significantly improve catalytic yield in
electrocatalytic and photocatalytic systems.25–28
Recently, the use of metal oxides, particularly TiO
2
, towards the protection of otherwise unstable
photoelectrode materials, such as Cu O, has drawn much attention in the solar fuels research
2
community.8,29–31 These metal oxide overlayers constitute an ideal substrate for the covalent binding of
molecules by the means of molecular linking groups such as carboxylates and phosphonates,26,32–34
approaches which have been crucial towards the realization of efficient dye sensitized solar cells.35
However, despite the attractive prospect of decreasing catalyst loading while retaining efficiency, few
examples of molecular catalysts linked to surfaces of photoelectrodes have been shown.36–38 In the case of
CO
reduction, all reports on immobilized catalysts have so far focused on dye-catalyst complexes39,40 and
2
polymer films.12–14,41–43
Here, we demonstrate for the first time the production of carbon monoxide (CO) as a solar fuel by the
covalent attachment of molecular CO
2
reduction catalysts to the surface of TiO
2
-protected Cu O
2
photocathodes by the means of phosphonate linkers (Figure 1). CO is an attractive target for solar fuel
synthesis due to its application in the Fischer-Tropsch and gas-to-methanol processes, which allow its
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
transformation into virtually any carbon based fuel, the cost of which is largely driven by the preparation
44
of syngas. Importantly, we show that the presence of a low-temperature processed mesoporous scaffold
on the surface of the photoelectrode is a crucial requirement to achieve efficient CO catalysis,
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
demonstrating an innovative and important application of these scaffolds which originally enabled the
breakthrough of dye sensitized solar cells.35
Figure 1. Schematic (not to scale) of protected Cu
Re(bipy) CO reduction catalyst 1.
2
O photoelectrode with covalently bound
2
RESULTS AND DISCUSSION
The virtues of the meso-porous scaffold
Complexes based on Re, particularly Re(bpy)(CO) Cl first introduced by Lehn et al, have shown excellent
3
performance as well as robustness when used as homogeneous electrocatalysts for the selective conversion
to CO in acetonitrile.19,21,45,46 At the same time, TiO
-protected Cu O photoelectrodes hold the
of CO
2
2
2
2
9
record performance as low-cost p-type light harvesters for water splitting. To enable covalent binding to
ACS Paragon Plus Environment

Page 5 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
the TiO
2
surface protecting the Cu O photocathode, the bipyridyl ligand of the original Lehn complex was
2
modified with phosphonate linking groups to form catalyst 1 (Figure 1, Figure S1). A methylene bridge
was incorporated between the phosphonate and bipyridyl moiety to minimize modification of the
electronic structure of the catalyst, a negative consequence of the ligand modification as has been
described elsewhere.45
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cu
2
O photocathodes were prepared by electrodeposition followed by atomic layer deposition of 20 nm Al
29
doped ZnO (AZO) and 100 nm of TiO
2
, as previously described. The TiO surface was subsequently
2
functionalized with catalyst 1 by soaking the device in a 1 mM solution of catalyst in MeCN for 24 hours,
followed by copious rinsing with neat MeCN. However, subsequent photoelectrochemical testing did not
show substantial photocurrents, pointing towards the absence of catalytic activity. The maximum
-2
photocurrent densities under CO were found to be merely 65 A cm , which indicated a negligible rate
2
of catalytic charge transfer under these conditions (Figure S2).7,29 In addition, measurements with varying
light intensities showed a non-linear photocurrent response, indicating an impeded charge extraction
which leads to major charge recombination within the photoelectrode. Therefore, the simple
functionalization of a standard Cu
2
O photocathode with catalyst 1 was found to be inadequate for
achieving significant rates of light-driven CO
2
reduction.
We hypothesized that the catalyst loading on this flat TiO surface was too low to achieve significant
2
currents, introducing a bottleneck to the flux of photogenerated charge from the photocathode. This was
further confirmed by FTIR, showing no signal relative to any bound catalyst. In their early stages of
development, dye sensitized solar cells suffered from analogous limitations where the light absorption of a
planar device was too low to achieve high efficiency, a problem eventually overcome by loading the dyes
onto thick, transparent films of mesoporous TiO
.35 We theorized that a similar approach might find
2
success here, wherein enhancing the roughness factor and catalyst loading could enable better
compatibility with the photogenerated electron flux from the photocathode, and therefore sought to
deposit porous films atop the photocathode surface.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
These mesoporous TiO
2
films are traditionally processed at 500 °C in air, a treatment necessary to oxidize
47
the binder material and to sinter the TiO
2
particles. Unfortunately, this method is not compatible with the
photocathode employed here since the p–n junction between Cu
2
O and AZO has been found to deactivate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
upon heating beyond 250 °C, thereby precluding any high temperature treatment for protected Cu
photocathodes.48 However, the curing of TiO
2
O
2
films under ultraviolet (UV) light has shown promising
results as an alternative to high-temperature processing.49,50 Inspired by these works, mesoporous films of
.5 to 5 um thickness with 18 nm TiO particles were deposited onto fluorine-doped tin oxide (FTO)
substrates and treated by UV light at room temperature for 48 hours followed by TiCl processing at 70
treatment was found to be the key to achieve films, which were stable under the testing
4
2
4
°C. The TiCl
conditions.
4
Catalyst immobilization on low-temperature mesoporous TiO films
2
The successful UV-curing of TiO
disappearance of modes associated with ethyl cellulose for both thermally and UV cured films (Figure
a). From AC-impedance measurements, it was found that films prepared at low temperature showed an
2
films was confirmed by FTIR measurements, showing the
4
7
2
additional RC element which could be attributed to a contact resistance between individual particles.
Interestingly, this element almost completely disappeared at more negative potentials as shown in Figure
2
b, and the films thus behave similar to thermally cured films under these conditions.
After heating the samples to 150 °C under vacuum to remove water from the scaffold, catalyst 1 was
applied onto the surface of mesoporous TiO by soaking the substrate in a 1 mM solution of 1 for 24
2
hours. The samples were then copiously rinsed with MeCN to ensure removal of any unbound species. To
verify for the presence of bound catalysts, FTIR-ATR measurements were carried out on the sample
surface. As shown in Figure 2c, strong peaks corresponding to the carbonyl vibrations on the Re complex
-1
could be observed at 1911 and 2035 cm , which are not present on a bare TiO
2
sample before the
treatment with catalyst. These peaks are blue-shifted in the bound form as compared to their powder
-1
spectrum (1868 and 2012 cm ), suggesting a decreased electron density on the Re center upon binding to
ACS Paragon Plus Environment

Page 7 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
TiO
2
. The bound catalyst introduced a slight yellow coloration of the TiO films as can be seen from the
2
photographs and UV-Vis transmission spectra in Figure 2d. The presence of grafted catalyst was further
confirmed by X-ray photoelectron spectroscopy (XPS) measurements on the photocathodes which show
the appearance of a clearly visible signal for Re (Figure 2e) at 40.4 and 42.7 eV, in agreement with
previously reported values.51 Additionally, nitrogen and phosphorous as well as oxygen due to the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
complex could be confirmed (Figure S3). The amount of catalyst present on the scaffold was determined
by ICP-OES of rhenium after complete dissolution in acid and found to be ~ 85 nmol cm-2.
To prove that binding was mediated by the phosphonate group and indeed covalent, the catalyst binding
procedure was repeated using a complex without binding group (species 2, Figure S1).7,45 After rinsing, no
FTIR signal from bound molecules could be observed in this case (Figure S4), thereby establishing that
the phosphonate groups were crucial for achieving catalyst immobilization on the scaffold.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
c
d
e
f
Figure 2: a) FTIR-ATR measurement of TiO films before curing and after thermal and UV curing. Ethyl
2
cellulose is efficiently eliminated upon UV curing, yielding films which are essentially indistinguishable
from thermal processing. b) EIS Spectra of thermally- and UV cured films. The additional charge transfer
ACS Paragon Plus Environment

Page 9 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
resistance which is observed for UV cured films disappears at more negative potentials. (c) FTIR-ATR
measurements of films before and after binding of catalyst 1, compared to the catalyst powder spectrum,
confirming catalyst binding to the film. (d) Optical absorption spectra of TiO
bound catalyst. The inset shows a photo of a TiO film with and without bound catalyst, as an illustration
of the catalyst color. (e) XPS detail spectrum of the Re4f peak, confirming the presence of Re-catalyst on
the TiO film surface compared to a UV cured TiO film without catalyst. (f) Cyclic voltammetry of
catalyst 1 bound to UV cured TiO on FTO. The catalytic onset can be clearly distinguished in the
presence of CO
2
films with and without
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
2
2
2
.
Electrochemical characterization of immobilized catalyst
To examine the electrochemical behavior of the unbound catalyst itself, cyclic voltammetry scans were
carried out on a MeCN solution of catalyst 1 using a Glassy Carbon electrode (Figure S5). After a first
+
reduction at -1.62 V vs. Fc /Fc, two minor reduction peaks could be observed at -1.77 V and -1.88 V.
2
Under CO , the second peak leads into a catalytic wave, confirming the activity of the complex towards
2
reducing CO . Furthermore, as a positive side effect of the bipyridine modification, its catalytic onset is
shifted 195 mV positive compared to Re(bpy)(CO)
3
Cl which was introduced by Lehn et al. (Figure S5).46
Cyclic voltammetry measurements of catalyst 1 bound to UV cured TiO films show a strong activity
2
-
2
towards the electroreduction of CO
2
(Figure 2f), exhibiting current densities exceeding 7 mA cm at -2.4
+
+
V vs. Fc /Fc in the presence of CO
2
. The onset of current is observed at approximately -1.7 V vs. Fc /Fc in
only one reduction peak, which is interesting since commonly Re(bpy) catalysts are known to be activated
through a 2-step reduction process in systems where homogeneously dissolved catalyst is driven by an
inert electrode.45,52
To investigate the catalytic activation in the present system, UV-Vis spectroelectrochemical experiments
were carried out. The system under study offers a powerful tool for combining electrochemical and
spectroscopic study. The scaffold roughness enables the presence of a high density of grafted species,
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
while the semi-transparency of the TiO
2
scaffold allows the use of transmittance-based measurements.
From the analysis of potential-dependent light absorption (Figure S6), we observed a catalyst transition
+
characterized by two absorption peaks at 507 and 578 nm appearing at –1.9 V vs. Fc /Fc when polarized
in the absence of CO
attribute the spectral change to the formation of a stable anion of catalyst 1. Furthermore, this absorption
was only observed in the presence of Ar, but not under CO , indicating that the observed species is being
2
, a potential coinciding with the catalytic onset under CO as shown in Figure 2f. We
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
2
2
rapidly turned over in the presence of carbon dioxide and suggesting that it corresponds to the catalytically
active species. In addition, the onset of catalytic current from the unbound catalyst at glassy carbon
(
Figure S5) was observed at the same potential. These data and the spectroelectrochemical observation of
a catalytically active intermediate at the potential of the catalytic reduction event at glassy carbon, leads us
to the conclusion that the catalytic mechanism remains unchanged by binding the catalyst to the TiO
scaffold.
2
ACS Paragon Plus Environment

Page 11 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
a
b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PorousTiO2
400 nm
ALD Al:ZnO+ TiO2
c
Cu O
2
Au
400 nm
F:SnO2
400 nm
Figure 3: a) SEM cross section of the modified photoelectrode. Layers have been colored for better
readability. The device has the following structure: FTO (580 nm), gold (550 nm), Cu O (500 nm – 1μm),
ALD Al:ZnO (20 nm), ALD TiO (100 nm), mesoporous TiO (4.5 – 5 μm, full scale shown in Figure S7).
b) Top-down view of the unmodified photoelectrode surface (corresponding to the green layer in (a))
2
2
2
exhibiting low roughness, and c) after adding the mesoporous TiO network which leads to a highly
2
porous structure, exhibiting a large surface area.
CO
2
reduction photoelectrocatalysis
Having established UV-cured TiO
2
as a viable substrate for active grafted catalysts, we applied this low-
O/AZO/TiO photocathodes. The
different layers of the device can be clearly distinguished from the electron micrographs shown in Figure
. Layer thicknesses are indicated in the figure caption. The layers fulfill the following functions: Cu
serves as photoabsorber and forms a p-n junction with AZO, whereas the ALD TiO layer serves as a
temperature process to deposit the mesoporous scaffold onto Cu
2
2
3
2
O
2
protection layer to avoid photocorrosion of Cu
2
O and mesoporous TiO
2
serves as a scaffold to support the
29
53
molecular catalyst. Due to its nature as a buried p-n junction device, the transfer of electrons to the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
catalyst is dictated by the offset of the conduction band of TiO and the catalyst redox potential. The
2
+
54
flatband potential of mesoporous TiO
2
in acetonitrile is known to be in the range of -2.4 V vs. Fc /Fc,
+
whereas the onset of catalytic activity of catalyst 1 at glassy carbon is observed at -1.9 V vs. Fc /Fc. Thus,
a significant driving force can be expected to exist for electron transfer to take place. A linear sweep
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
voltammogram of the mesoporous TiO
4a. Under CO , substantial photocurrents exceeding 2.5 mA cm were observed. Importantly, the presence
of a mesoporous scaffold led to a 40-fold increase in current compared to flat photocathodes. In
comparison to the catalyst activity on a dark electrode, i.e. a mesoporous TiO film deposited directly on
2
modified photoelectrode loaded with catalyst 1 is shown in Figure
-2
2
2
FTO (Figure 2f), the catalytic current begins at potentials about 600 mV more positive on the
photoelectrode, defining the photovoltage produced by the electrode. Both the photocurrent density and
photovoltage achieved here represent an improvement over our previous work wherein a protected Cu O
2
7
photocathode was immersed in a solution of dissolved catalyst, and the enhancement in photocurrent can
be largely attributed to decreased parasitic light absorption by the catalyst when it is selectively grafted
onto the electrode surface. A linear dependence on incident light intensity was observed for this system
(
Figure S8), showing that the catalyst is now capable of supporting the flux of photogenerated electrons
from the photocathode, enabled by loading onto the high area surface of the mesoporous TiO scaffold.
2
a
b
Figure 4: a) Photoelectrochemical performance under chopped simulated sunlight of Cu O photocathodes
2
ACS Paragon Plus Environment

Page 13 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
with (red) and without (blue) a mesoporous TiO
2
layer on the surface, tested under CO with molecular Re
2
-1
catalyst covalently bound to the surface. Linear sweep at 10 mV s . b) Chopped-light polarization test
+
under CO
2
at -1.9 V vs. Fc /Fc of mesoporous TiO
2
layer modified Cu O photocathode with catalyst.
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Our previous observations when using a photocathode to drive a homogenous catalyst in solution led to
the hypothesis that consecutive catalyst reductions were hindered by electrostatic repulsion between the
polarized semiconductor electrode and charged catalyst intermediates, inhibiting subsequent CO
2
7
reduction. In that study, the inhibition could be surmounted by addition of a protic additive to solution. In
this report, efficient catalyst turnover was accomplished without the need for protic additives,
demonstrating how covalently linking the catalyst to the electrode surface is a useful approach toward
enabling sequential charge transfer between semiconductor electrode and CO reduction catalyst, in
2
addition to greatly decreasing the required catalyst amount and minimizing its parasitic absorption.
+
Potentiostatic testing at -1.9 V vs. Fc /Fc under chopped light illumination showed sustained catalytic
activity towards the reduction of CO
2
(Figure 4b), during which a gradual decrease in photocurrent could
be observed as will be discussed below.
In order to verify the catalytic activity towards CO
2
conversion, the evolved gases were quantified by gas
-supported catalyst on FTO. It was found
chromatography during constant polarization of UV cured TiO
2
that the sample reached faradaic yields for CO between 80 and 95%, whereas in absence of catalyst,
negligible CO was observed (Figure 5). Furthermore, when polarized under inert helium saturation, no
production of gas-phase species was observed (Figure S9). Taken together with the high turnover numbers
observed, as estimated below, these experiments indicate that the observed carbon monoxide originates
from a catalytic reaction of CO
2
rather than from decomposition of the catalyst, and that the molecular
to take place.
catalyst is indeed required for turnover of CO
2
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5: Quantification of the faradaic efficiency of carbon monoxide production by catalyst-modified
+
+
TiO
2
films polarized at -2.05 V vs. Fc /Fc and from a bare TiO
2
film at -2.20 V vs. Fc /Fc under CO
2
saturation. The presence of catalyst is necessary for CO
2
to be turned over.
Loss of catalytic activity
During potentiostatic testing of the photoelectrode under chopped light, a gradual decrease in photocurrent
Figure 4b) and in faradaic efficiency for CO (Figure 5) was observed over time. To shed light on these
(
processes, a series of techniques were used to determine the changes to the electrodes resulting from
+
extended operation. After 2 hours of polarization at –1.9 V vs. Fc /Fc under chopped light of a catalyst
coated photoelectrode followed by rinsing in neat MeCN, surface analysis by XPS confirmed the
persistence of the elements attributable to the immobilized catalyst (Re, N, and P), albeit at a slightly
lower intensity (Figure S3). The peaks associated with Re did not experience a significant shift, indicating
no change in oxidation state. At the same time, FTIR-ATR spectra showed the persistence of the C-O
stretches related to the bound catalyst, similarly at a slightly reduced intensity (Figure S10). From the
observation of a decreased signal intensity, one may assume that a part of the catalyst is desorbing which
is the most straightforward explanation for the decrease in photocurrent. However, further experiments
ACS Paragon Plus Environment

Page 15 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
gave evidence that other processes, such as the structural change of catalysts, are likely to play a role. Our
7
previous report rules out the instability of the Cu
catalyst and TiO
experiments were conducted using thermally cured mesoporous TiO
2
O photoabsorber underlayer, so the issue remains on
2
mesoporous scaffold. To eliminate contribution from the scaffold, additional
films on FTO, where a similar
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
degradation was observed, suggesting that the scaffold does not play a role in the deactivation process.
Furthermore, it was found that immersing a completely deactivated sample in a 1 mM solution of catalyst
for 24 h led to almost complete recovery of the catalytic activity (Figure 6a). In contrast, when a
deactivated sample was rinsed, dried and subsequently tested in fresh electrolyte, no regain of activity was
observed. Despite the loss of catalytic activity, the presence of catalyst was consistently observed by FTIR
spectroscopy in between the measurements (Figure 6b). It is notable that even after 3 h of polarization at -
+
2.2 V vs. Fc /Fc, strong signals from the catalyst C-O stretches were visible. Despite being weaker than in
the case of a fresh sample, the observation of these signals indicates that loss of catalyst is not the only
process leading to the observed decrease in current. Support for this conclusion comes from ICP-OES
analysis, which even after 3 h of continuous polarization still detects between 60 and 70 % of the initial
rhenium on the substrates (Table S1). Structural change of the catalyst molecules themselves is therefore
likely to play a significant role as well, as can be seen from the distinct color change of samples after
deactivation (Figure S11). Previously, the formation of a formato-rhenium complex was suggested as
55
being at the origin of the loss of catalytic activity, but in our case no evidence for this effect can be
obtained from FTIR, where no significant shift of the carbonyl stretches is observed. In contrast, we find
-1
that the vibrations in the region from 1400 to 1600 cm , related to the bipyridine group, are bleached after
testing, suggesting that a structural change of the bipyridine may lead to catalyst deactivation. This effect
is likely intensified by the fact that only a very small amount of catalyst is present in the system, which
cannot be replenished by diffusion of pristine catalyst from a bulk solution, as it is the case in a
homogeneous electrocatalytic system. As an example, from integrating the current towards CO during a
+
1
.5 h test at -2.05 V vs. Fc /Fc (Figure 5), the catalyst molecules exceed 70 turnovers, which is larger than
what was achieved in previous reports under similar conditions.26,27,51,55,56
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
b
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6: a) CV curves recorded from a thermally treated TiO sample at different stages of testing. Re-
2
exposing a deactivated sample to fresh catalyst (red trace) led to a performance comparable to the fresh
sample (blue trace). In contrast, the performance of a sample tested for 3 h without recatalyzation (green
trace) is significantly lower than of a fresh sample. b) At each stage of the experiment ATR-FTIR
measurements were carried out. The C-O stretches associated with the catalyst are observed at each stage.
Advantages and challenges
The catalyst binding strategy developed here successfully enabled CO reduction to CO while using orders
2
of magnitude smaller quantities of expensive catalyst as compared to systems of homogeneous catalyst
solutions, and minimizing parasitic absorption due to the catalyst. Since the covalent linking strategy uses
a small amount of catalyst with a high proportion of catalytic participation, in contrast to homogeneous
systems where only a miniscule fraction of catalyst is interacting with the electrode at a given moment,
each grafted catalyst molecule experiences a significantly higher number of turnovers. Therefore, it is
subject to the acceleration of any degradation pathways, as opposed to cases where abundant fresh catalyst
molecules are present in solution. We consider this an advantage of heterogenized homogeneous catalyst
systems, since they can provide a more direct picture of catalyst change than is possible in homogeneous
systems.
ACS Paragon Plus Environment

Page 17 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
While detailed study of the deactivation is a goal of ongoing work, we have examined the system by
multiple techniques before, during, and after operation. The hybrid heterogeneous system demonstrated
here should warrant further analysis by in situ techniques (optical spectroscopy, transient absorption
spectroscopy) to elucidate the mechanisms, kinetics, and deactivation pathways of molecular catalysts
with the added benefit of electrochemical control. Binding the catalyst to an electro-active surface enables
biasing all catalyst molecules at the same potential, reduces the influence of diffusion processes, and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
facilitates the multiple electron transfers necessary for a multi-step reaction like CO
2
reduction. For
instance, it has been reported that surface immobilization of catalyst can lead to increased lifetimes of
26
catalyst intermediates. This strategy of using immobilized catalysts on a nanostructured scaffold may
prove attractive for diverse electrochemical and photoelectrochemical applications, including but not
limited to water splitting, CO
-reduction and bioelectrochemistry.57-59
2
Although there are reports of immobilized catalysts for heterogeneous photocatalytic CO
2
reduction on
particles in solution,26-28,32-34 relatively few studies exist where catalysts are grafted onto an electrode
4
1
surface, and even fewer which report electrochemical characterization. Among these, to the best of our
knowledge, this is the first report of examining the potential and the photoelectrochemistry of
phosphonate-mediated covalently bound catalysts for CO
2
reduction. While the performance observed
with this system is astonishing and even exceeds the performance recently reported with silicon
4
3
nanowires, numerous further improvements can be envisioned. Increasing the robustness of the catalyst
as well as the binding strategy will improve the stability of the system. Furthermore, the replacement of
rhenium with an earth-abundant metal core will lead to a system that is free of precious metals, thereby
decreasing the system cost. This will enable another step towards the realization of scalable solar fuels
devices, which will be key to achieving an energy system based on renewable resources.
Finally, while the study conducted here examined a photocathode in three-electrode configuration, where
the oxidation of the electrolyte serves as the source of electrons, a complete solar-to-fuel conversion
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
process requires a balanced and sustainable counter electrode oxidation process, such as oxygen evolution
reduction reaction.8,14
from water, to supply electrons for the CO
2
CONCLUSION
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
We demonstrated the light-assisted reduction of carbon dioxide by the use of a covalently bound
molecular catalyst on mesoporous TiO modified Cu O photocathodes. The observed photocurrents
2
2
represent a significant improvement over what has previously been demonstrated with molecular catalysts
homogeneously dissolved in the electrolyte solution. Furthermore, in addition to being based on a low-
cost photoelectrode, the observed photocurrents are comparable to or better than previous reports of
immobilized catalysts on photoelectrodes both towards CO reduction and water splitting. These
2
improvements can largely be attributed to the use of a nanostructured photoelectrode surface which we
have shown to be necessary to achieve large photocurrents with covalently bound catalysts. This work
represents the first demonstration of the use of phosphonate linking groups to immobilize a molecular fuel
generation electrocatalysts on the surface of a photoelectrode and opens the path toward further research
into improving the system stability and efficiency, as well as more detailed investigation of the
mechanistic details of molecular electrocatalysts.
EXPERIMENTAL SECTION
Synthesis of catalyst (1)
The anchoring bipyridine synthesized for this study has a CH
2
spacer between the PO
3
2
H group and the
bipyridine core. A direct link between the catalyst and the semiconductor is not a strict requirement for
high electron injection efficiency and the isolated anchoring group has little influence in the electronic
structure of the catalyst itself. Therefore, Re(I) complexes derived from bipyridine 4 were synthesized.
Utilization of a spacer group between the phosphonic acid and the bipyridine ligand can also provide a
ACS Paragon Plus Environment

Page 19 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
chance to synthesize catalyst 1 with the redox center positioned a considerable distance away from the
semiconductor surface.
4
,4‘-Bis(hydroxymethyl)-2,2‘-bipyridine (2). 8.2 g of sodium borohydride was added in one portion to a
suspension of 4,4‘-Bis(methyl carboxylate)-2,2‘-bipyridine (3.0 g, 10.0 mmol) in 200 mL of absolute
ethanol. The mixture was refluxed for 3 h and cooled to room temperature, and then 200 mL of an
ammonium chloride saturated water solution was added to decompose the excess borohydride. The
ethanol was removed under vacuum and the precipitated solid dissolved in a minimal amount of water.
The resulting solution was extracted with ethyl acetate (5 × 200 mL) and dried over sodium sulfate, and
the solvent was removed under vacuum. The desired solid was obtained in 79% yield and was used
without further purification.
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1H NMR (400 MHz, CD
3
2
OD, ):ꢀ 4.75 (4H, s, CH ); 7.43 (2H, d, J = 5.5 Hz, aryl H on C5 and C5‘); 8.25
(
2H, s, aryl H on C3 and C3‘); 9.00 (2H, d, J = 5.5 Hz, aryl H on C6 and C6‘).
4,4‘-Bis(bromomethyl)-2,2‘-bipyridine (3). The bipyridine 2 (1.06 g, 4.9 mmol) was dissolved in a
mixture of 48% HBr (30 mL) and concentrated sulfuric acid (10 mL). The resulting solution was refluxed
135 °C) for 6 h and then allowed to cool to room temperature, and 40 mL of water was added. The pH
was adjusted to neutral with NaOH solution and the resulting precipitate filtered, washed with water (pH =
), and air-dried. The product was dissolved in chloroform (40 mL) and filtered. The solution was dried
over magnesium sulfate and evaporated to dryness, yielding 1.42 g of 3 (85% yield) as a white powder.
(
7
1H NMR (400 MHz, CDCl
, ):ꢀ 4.50 (4H, s, CH2); 7.38 (2H, d, J = 5 Hz, aryl H on C5 and C5‘); 8.45
2H, s, aryl H on C3 and C3‘); 8.68 (2H, d, J = 5 Hz, aryl H on C6 and C6‘).
3
(
4,4‘-Bis(diethylmethylphosphonate)-2,2‘-bipyridine (4). A chloroform (8 mL) solution of 3 (1.1 g, 3.2
mmol) and 12 mL of triethyl phosphite was refluxed (156 °C) for 3 h under nitrogen. The excess
phosphite was removed under high vacuum, and then the crude product was purified by column
chromatography on silica gel (eluent ethyl acetate/methanol 80/20) yielding 1.17 g (80%) of 4.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1H NMR (300 MHz, CDCl
apparent quintet, J = 7 Hz, OCH
C3 and C3‘); 8.62 (2H, d, J = 5 Hz, aryl H on C6 and C6‘).
31P NMR (81 MHz, CDCl
, ):ꢀ 25.37.
, ):ꢀ 1.29 (12H, t, J = 7 Hz, CH
); 3.23 (4H, d, J = 22 Hz, CH P); 4.09 (8H,
3
3
2
2
); 7.35−7.38 (2H, m, aryl H on C5 and C5‘); 8.34−8.37 (2H, m, aryl H on
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
[
Re(4,4’-diethylmethylphosphonate-2,2’-bipyridine)(CO)
3
Cl] (5). Re(CO) Cl (0.5 g, 1.38 mmol) was
5
dissolved in 50 mL of hot toluene. An equimolar amount of 4,4‘-Bis(diethylmethylphosphonate)-2,2‘-
bipyridine (0.630 g, 1.38 mmol) was added to the hot solution, and the reaction mixture was stirred with
reflux for 1 h. Once the solution reached reflux the mixture began to change color from clear, to yellow,
and finally to orange. After 1 h of reflux the reaction mixture was removed from heat and cooled in a
freezer. A solid precipitate was obtained from cold toluene. The yellow solid was dried overnight in a
vacuum oven at 90 °C. Spectroscopically pure product was obtained from the reaction with an overall
1
yield of 90 %. H NMR (400 MHz, Chloroform-d, ): 8.95 (d, J = 5.7 Hz, 2H), 8.22 (t, J = 2.1 Hz, 2H),
7.45 (dt, J = 5.6, 2.0 Hz, 2H), 4.20 – 4.06 (m, 8H), 3.33 (s, 2H), 3.27 (s, 2H), 1.33 (td, J = 7.1, 1.7 Hz,
12H).
[Re(4,4’-dimethylphosphonic acid-2,2’-bipyridine)(CO) Cl] (Catalyst 1). 200 mg (0.252 mmol) of 5
3
was added to 1 M HCl (5.0 mL) and the mixture was kept at 70 °C for 6 h. Upon cooling to room
temperature, the precipitate formed was collected by vacuum filtration and washed with cold methanol
1
and water. Drying first in air and then in vacuum gave 1 as a yellow solid. H NMR (400 MHz,
Acetonitrile-d3, ): 8.93 – 8.80 (m, 2H), 8.41 (s, 2H), 7.52 – 7.41 (m, 2H), 3.44 – 3.27 (m, 4H) (Figure
S12).
Synthesis of catalyst (2)
Catalyst (2) was synthesized as previously described.7
Fabrication of photoelectrodes and reference samples
ACS Paragon Plus Environment

Page 21 of 28
Journal of the American Chemical Society
1
2
3
30 nm of Cr, followed by 450 nm (nominal) of Au were deposited on FTO glass (TEC-15, Solaronix,
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Switzerland) by sputter deposition. The deposition was done using an Alliance Concept DP650 Sputtering
system (AllianceConcept, France). Cu
2
O thin films were deposited onto the gold layer by
electrodeposition from a copper lactate solution as described previously47 for 105 minutes under a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-2
galvanostatic current density of 0.1 mA cm . A Pt mesh was used as counter electrode. The copper lactate
solution was prepared by dissolving 7.58 g of CuSO (Sigma Aldrich), 67.5 g of lactic acid (Fluka), and
1.77 g of K HPO (Sigma Aldrich) in 250 ml H O. 2 M KOH (Reactolab) was slowly added to the
4
2
2
4
2
reaction mixture to reach pH 12. During this step, the color of the solution changed from dark blue to light
blue, and finally to dark blue again. Precise adjustment of the pH was achieved using small amounts of
KOH and lactic acid.
Atomic Layer Deposition of the Al:ZnO and TiO
2
layers was carried out on a Savannah 100 system
7
(
Cambridge Nanotech, USA) as described previously. Al:ZnO was deposited at 200 °C, whereas the
precursors, diethylzinc (ABCR), trimethylaluminum (ABCR), and H
temperature. Under constant vacuum, the precursor was pulsed for 15 ms, followed by a 10 s wait time.
Subsequently, H O was pulsed using the same sequence. After 20 cycles of ZnO, 1 cycle of Al was
deposited using the same pulse sequence. This was repeated for 5 times, which led to the deposition of 20
nm of Al:ZnO. TiO was deposited under constant vacuum at 150 °C using tetrakis (dimethylamino)
titanium, TDMAT (99%, Sigma Aldrich) at 75 °C, and H (50%, Sigma Aldrich) at room temperature,
2
O (18 MΩ) were kept at room
2
O
2 3
2
2
O
2
using a pulse length of 10 ms, followed by 10 s wait time. 1500 cycles were carried out to yield a TiO
film of 100 nm thickness.
2
After ALD deposition of overlayers, the surface of the photoelectrode was subjected to a TiCl
bath treatment. This was achieved by incubating the samples in 150 mL of 40 mM TiCl solution for 30
minutes at 70 °C. The samples were subsequently rinsed with water and dried in an air stream.
Crystalline mesoporous TiO was deposited onto the surface of the cathodes by screen printing.
Commercial 18-NRT TiO paste (Dyesol) was deposited by carrying out 2 printing steps using a 61T (30.4
4
chemical
4
2
2
3
-2
cm m ) screen. The thus-formed films were analyzed by profilometry to exhibit a thickness of 4.5 to 5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
μm.
Curing of the films was carried out at room temperature for 48 hours by keeping the film under UV
irradiation at 1 mm distance from the lamp (GPH 369T5VA/4 18 W, 253.7 nm peak, Heraeus).
Subsequently, the TiCl treatment was repeated as described above.
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Before catalyst immobilization, water was removed from the mesoporous scaffold by treating the
electrode for 30 minutes at 150 °C under vacuum. Subsequently, the catalyst was immobilized by
immersing the electrode for 24 h in a 1 mM solution of catalyst in acetonitrile, followed by copious
rinsing with dry acetonitrile.
Reference samples of mesoporous TiO films were prepared directly on FTO glass. UV treated samples
2
were prepared as follows: FTO glass was pretreated with TiCl
scaffold and subsequent treatment steps as described above.
4
, followed by deposition of the mesoporous
Thermally cured samples were prepared as follows: FTO glass was pretreated with TiCl
mesoporous TiO was deposited as above. In the next step, the films were cured according to a well-
established thermal procedure. TiCl
4
and subsequently
2
47
4
post treatment was carried out as above, followed by heating to 500
°C in a hot air stream for 30 minutes. If catalyst was deposited on the sample, it was immersed in catalyst
solution as above while still in a warm state.
Device characterization
Scanning electron microscopy was carried out using a Zeiss Merlin microscope. FTIR-ATR measurements
were carried out using a Spectrum GX system (Perking Elmer), equipped with an ATR extension
-1
(
MIRacle, Pike Technologies). Each spectrum consists of 50 scans at 4 cm resolution. Catalyst loading
onto the TiO films were determined as follows: After determination of the surface area, each sample was
introduced in a 100 mL Erlenmeyer flask. Subsequently, 2 – 3 mL of Milli-Q water, 7.5 mL HCl (Merck
318, Suprapur) and 2.5 mL HNO (Merck 441, Suprapur) were added. After standing overnight, the
2
3
mixture was refluxed for 2 h and 20 mL of Milli-Q water was added. The solution was subsequently
ACS Paragon Plus Environment

Page 23 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
filtered and the filter washed 5 times with 2.0 M HNO
3
. Subsequently, the volume was completed to 50.0
mL with 2.0 M HNO
3
and analyzed by ICP-OES. Profilometry of TiO films was carried out using an
2
Alpha-Step 500 profilometer. XPS measurements were carried out using a PHI VersaProbe II Scanning
XPS Microprobe.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Electrochemical testing
Solar simulation was carried out using a LOT Oriel light source, equipped with a 450 W Xe arc lamp
(
Ozone Free, Osram). The light intensity was reduced by a metal grid which served as a neutral density
filter, followed by a KG3 filter (Edmund Optics). The system was tuned to 1 sun intensity using a
calibration diode. Light chopping was carried out using an electromechanical chopper at a frequency of 1
Hz. Samples were tested in a custom-built gastight PEC test cell. The test solution was dry acetonitrile
(
Sigma Aldrich) with 0.1 M Bu
4
NPF (tetrabutylammonium hexafluorophosphate, Sigma Aldrich).
6
Electrochemical measurements were performed using a Gamry Interface 1000 and a BioLogic SP-300
potentiostat. A Ag/AgCl wire, separated from the cell by a porous diaphragm, was used as a pseudo
reference electrode and its potential was calibrated to ferrocene. A Pt wire in the same compartment was
used as counter electrode. A 3 mm diameter Glassy Carbon rotating disk electrode (Metrohm) was used
for electrochemical characterization of catalysts in solution. In all cases, the solution was saturated either
with Ar (Carbagas, 99.9999%) or with CO
2
(Carbagas 99.998%) by sparging the solution under stirring
+
-1
for 20 min. CVs were carried out between -0.4 and -2.4 V vs. Fc /Fc at 100 mV s , whereas linear sweep
-1
scans were carried out between the same potentials at 10 mV s . Potential-stepping AC-impedance
measurements were carried out using a BioLogic SP-300 between 1 MHz and 0.1 Hz.
Gas Measurements
-1
Gas Measurements were carried out under 10 mL min flow of CO using the same equipment as
2
described above. Blank tests were carried out, verifying that the counter-electrode in the cell produced
negligible CO. Polarization was carried out using the potentiostats described above and the effluent gas
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
was led into the sampling loop of a gas chromatograph after evaporated acetonitrile was condensed out at -
25 °C. The gas chromatograph (Trace ULTRA, Thermo Scientific) was equipped with a ShinCarbon
column (Restek) and a PDD detector (Vici). Helium was used as carrier gas.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Catalyst turnover numbers (TON) were calculated by integrating the amount of CO produced during the
polarization test and comparing to the measured amount of catalyst.
Spectroelectrochemical measurements
Spectroelectrochemistry was carried out using the UV-Vis equipment and the custom-made test cell
described above. An ocean-optics halogen lamp was used to illuminate transparent TiO films on FTO,
2
prepared as described above. A condenser lens behind the film collected the transmitted photons and fed
them to an OceanOptics photodiode array by the means of a fiber optic cable. The potential was stepped
from -0.4 to -2.4 V vs. Fc+/Fc using a Gamry Framework script. After 20 s equilibration time at each
potential, a custom-made LabView software recorded a spectrum from the photodiode array.
ASSOCIATED CONTENT
Supporting Information. Experimental methods, schemes and supporting results. This material is available
free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGEMENTS
This project was funded by Siemens AG. We would like to express our warmest thanks to T. M. Aufar
Kari for the preparation of many samples and for conducting measurements. Further thanks go to Prof.
Hubert Girault for helpful discussions, Frédéric Gumy for providing LabView code, to Sylvain Coudret
and Karine Vernez Thomas for carrying out ICP-OES analysis and to Pierre Mettreaux for carrying out
XPS measurements.
Author Information
Corresponding Authors
ACS Paragon Plus Environment

Page 25 of 28
Journal of the American Chemical Society
1
2
3
*
matthew.mayer@epfl.ch
* michael.graetzel@epfl.ch
Present Addresses
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
T.M.: Department of Chemistry, University of Zurich, Switzerland
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
REFERENCES
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(1) Climate Change 2013: The Physical Science Basis - Summary for Policy Makers; IPCC, Series Ed.;
IPCC, 2013.
(
(
2) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. 2006, 103 (43), 15729.
3) The Economist. June 13, 2015, http://www.economist.com/news/business/21654082-threaten-fossil-
fuels-solar-power-must-solve-its-intermittency-problem-banishing-clouds.
(
4) Faunce, T.; Styring, S.; Wasielewski, M. R.; Brudvig, G. W.; Rutherford, A. W.; Messinger, J.; Lee,
A. F.; Hill, C. L.; deGroot, H.; Fontecave, M.; MacFarlane, D. R.; Hankamer, B.; Nocera, D. G.;
Tiede, D. M.; Dau, H.; Hillier, W.; Wang, L.; Amal, R. Energy Environ. Sci. 2013, 6 (4), 1074.
5) Luo, J.; Im, J.-H.; Mayer, M. T.; Schreier, M.; Nazeeruddin, M. K.; Park, N.-G.; Tilley, S. D.; Fan,
H. J.; Grätzel, M. Science 2014, 345 (6204), 1593.
(
(
6) Kumar, B.; Llorente, M.; Froehlich, J.; Dang, T.; Sathrum, A.; Kubiak, C. P. Annu. Rev. Phys.
Chem. 2012, 63 (1), 541.
(7) Schreier, M.; Gao, P.; Mayer, M. T.; Luo, J.; Moehl, T.; Nazeeruddin, M. K.; Tilley, S. D.; Grätzel,
M. Energy Environ. Sci. 2015, 8, 855.
(
8) Schreier, M.; Curvat, L.; Giordano, F.; Steier, L.; Abate, A.; Zakeeruddin, S. M.; Luo, J.; Mayer, M.
T.; Grätzel, M. Nat. Commun. 2015, 6, 7326.
(
(
9) Grätzel, M. Nature 2001, 414 (6861), 338.
10) Bornoz, P.; Abdi, F. F.; Tilley, S. D.; Dam, B.; van de Krol, R.; Graetzel, M.; Sivula, K. J. Phys.
Chem. C 2014, 118 (30), 16959.
(
(
11) Brillet, J.; Yum, J.-H.; Cornuz, M.; Hisatomi, T.; Solarska, R.; Augustynski, J.; Graetzel, M.; Sivula,
K. Nat. Photonics 2012, 6 (12), 824.
12) Sato, S.; Arai, T.; Morikawa, T.; Uemura, K.; Suzuki, T. M.; Tanaka, H.; Kajino, T. J. Am. Chem.
Soc. 2011, 133 (39), 15240.
(13) Arai, T.; Sato, S.; Kajino, T.; Morikawa, T. Energy Environ. Sci. 2013, 6 (4), 1274.
(14) Arai, T.; Sato, S.; Morikawa, T. Energy Environ. Sci. 2015, 8 (7), 1998.
(
15) Dias, P.; Schreier, M.; Tilley, S. D.; Luo, J.; Azevedo, J.; Andrade, L.; Bi, D.; Hagfeldt, A.; Mendes,
A.; Grätzel, M.; Mayer, M. T. Adv. Energy Mater. 2015, 5, 1501537.
(
16) Luo, J.; Li, Z.; Nishiwaki, S.; Schreier, M.; Mayer, M. T.; Cendula, P.; Lee, Y. H.; Fu, K.; Cao, A.;
Nazeeruddin, M. K.; Romanyuk, Y. E.; Buecheler, S.; Tilley, S. D.; Wong, L. H.; Tiwari, A. N.;
Grätzel, M. Adv. Energy Mater. 2015, 5, 1501520.
(
(
(
(
17) Dempsey, J. L.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Acc. Chem. Res. 2009, 42 (12), 1995.
18) Savéant, J.-M. Chem. Rev. 2008, 108 (7), 2348.
19) Costentin, C.; Robert, M.; Savéant, J.-M. Chem. Soc. Rev. 2013, 42 (6), 2423.
20) Willkomm, J.; Muresan, N. M.; Reisner, E. Chem. Sci. 2015, 6 (5), 2727.
(21) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. Science 2012, 338 (6103), 90.
(22) Kumar, B.; Smieja, J. M.; Kubiak, C. P. J. Phys. Chem. C 2010, 114 (33), 14220.
(23) Guerrero, A.; Haro, M.; Bellani, S.; Antognazza, M. R.; Meda, L.; Gimenez, S.; Bisquert, J. Energy
Environ. Sci. 2014, 7 (11), 3666.
(
(
(
24) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. J. Am. Chem. Soc. 2012, 134 (27), 11235.
25) Cecchet, F.; Alebbi, M.; Bignozzi, C. A.; Paolucci, F. Inorganica Chim. Acta 2006, 359 (12), 3871.
26) Windle, C. D.; Pastor, E.; Reynal, A.; Whitwood, A. C.; Vaynzof, Y.; Durrant, J. R.; Perutz, R. N.;
Reisner, E. Chem. – Eur. J. 2015, 21 (9), 3746.
(
(
27) Liu, C.; Dubois, K. D.; Louis, M. E.; Vorushilov, A. S.; Li, G. ACS Catal. 2013, 3 (4), 655.
28) Dubois, K. D.; He, H.; Liu, C.; Vorushilov, A. S.; Li, G. J. Mol. Catal. Chem. 2012, 363–364, 208.
ACS Paragon Plus Environment

Page 27 of 28
Journal of the American Chemical Society
1
2
3
(
29) Paracchino, A.; Laporte, V.; Sivula, K.; Grätzel, M.; Thimsen, E. Nat. Mater. 2011, 10 (6), 456.
30) Seger, B.; Pedersen, T.; Laursen, A. B.; Vesborg, P. C. K.; Hansen, O.; Chorkendorff, I. J. Am.
Chem. Soc. 2013, 135 (3), 1057.
31) Seger, B.; Tilley, D. S.; Pedersen, T.; Vesborg, P. C. K.; Hansen, O.; Grätzel, M.; Chorkendorff, I.
RSC Adv. 2013, 3 (48), 25902.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
(
(
(
32) Ha, E.-G.; Chang, J.-A.; Byun, S.-M.; Pac, C.; Jang, D.-M.; Park, J.; Kang, S. O. Chem. Commun.
2014, 50 (34), 4462.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
33) Suzuki, T. M.; Tanaka, H.; Morikawa, T.; Iwaki, M.; Sato, S.; Saeki, S.; Inoue, M.; Kajino, T.;
Motohiro, T. Chem. Commun. 2011, 47 (30), 8673.
(34) Anfuso, C. L.; Snoeberger, R. C.; Ricks, A. M.; Liu, W.; Xiao, D.; Batista, V. S.; Lian, T. J. Am.
Chem. Soc. 2011, 133 (18), 6922.
(
(
(
35) O’Regan, B.; Grätzel, M. Nature 1991, 353 (6346), 737.
36) Krawicz, A.; Cedeno, D.; Moore, G. F. Phys. Chem. Chem. Phys. 2014, 16 (30), 15818.
37) Krawicz, A.; Yang, J.; Anzenberg, E.; Yano, J.; Sharp, I. D.; Moore, G. F. J. Am. Chem. Soc. 2013,
135 (32), 11861.
(
(
38) Seo, J.; Pekarek, R. T.; Rose, M. J. Chem. Commun. 2015, 51 (68), 13264.
39) Kou, Y.; Nakatani, S.; Sunagawa, G.; Tachikawa, Y.; Masui, D.; Shimada, T.; Takagi, S.; Tryk, D.
A.; Nabetani, Y.; Tachibana, H.; Inoue, H. J. Catal. 2014, 310, 57.
40) Sahara, G.; Abe, R.; Higashi, M.; Morikawa, T.; Maeda, K.; Ueda, Y.; Ishitani, O. Chem. Commun.
2015, 51 (53), 10722.
(
(41) Tian, H. ChemSusChem 2015, 8, 3746.
(42) Arai, T.; Sato, S.; Uemura, K.; Morikawa, T.; Kajino, T.; Motohiro, T. Chem. Commun. 2010, 46
(
37), 6944.
43) Torralba-Peñalver, E.; Luo, Y.; Compain, J.-D.; Chardon-Noblat, S.; Fabre, B. ACS Catal. 2015, 5
10), 6138.
(
(
(
(
(
(
44) Dry, M. E. Catal. Today 2002, 71 (3–4), 227.
45) Smieja, J. M.; Kubiak, C. P. Inorg. Chem. 2010, 49 (20), 9283.
46) Hawecker, J.; Lehn, J.-M.; Ziessel, R. J. Chem. Soc. Chem. Commun. 1984, No. 6, 328.
47) Ito, S.; Murakami, T. N.; Comte, P.; Liska, P.; Grätzel, C.; Nazeeruddin, M. K.; Grätzel, M. Thin
Solid Films 2008, 516 (14), 4613.
48) Paracchino, A.; Mathews, N.; Hisatomi, T.; Stefik, M.; Tilley, S. D.; Grätzel, M. Energy Environ.
Sci. 2012, 5 (9), 8673.
(
(49) Zardetto, V.; Di Giacomo, F.; Garcia-Alonso, D.; Keuning, W.; Creatore, M.; Mazzuca, C.; Reale,
A.; Di Carlo, A.; Brown, T. M. Adv. Energy Mater. 2013, 3 (10), 1292.
50) Arakawa, H.; Yamaguchi, T.; Sutou, T.; Koishi, Y.; Tobe, N.; Matsumoto, D.; Nagai, T. Curr. Appl.
Phys. 2010, 10 (2, Supplement), S157.
(
(
(
51) Blakemore, J. D.; Gupta, A.; Warren, J. J.; Brunschwig, B. S.; Gray, H. B. J. Am. Chem. Soc. 2013,
135, 18288.
52) Keith, J. A.; Grice, K. A.; Kubiak, C. P.; Carter, E. A. J. Am. Chem. Soc. 2013, 135 (42), 15823.
53) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi, Q.; Santori, E. A.;
(
Lewis, N. S. Chem. Rev. 2010, 110 (11), 6446.
(54) Redmond, G.; Fitzmaurice, D. The Journal of Physical Chemistry 1993, 97 (7), 1426.
(55) Takeda, H.; Ohashi, M.; Tani, T.; Ishitani, O.; Inagaki, S. Inorg. Chem. 2010, 49 (10), 4554.
(
56) Won, D.-I.; Lee, J.-S.; Ji, J.-M.; Jung, W.-J.; Son, H.-J.; Pac, C.; Kang, S. O. J. Am. Chem. Soc.
015, 137 (42), 13679.
57) Fan, K.; Li, F.; Wang, L.; Daniel, Q.; Chen, H.; Gabrielsson, E.; Sun, J.; Sun, L. ChemSusChem
015, 8 (19), 3242.
2
(
2
(
58) Liu, R.; Stephani, C.; Han, J. J.; Tan, K. L.; Wang, D. Angew. Chem. Int. Ed. 2013, 52 (15),
4225.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
59) White, J. L.; Baruch, M. F.; Pander III, J. E.; Hu, Y.; Fortmeyer, I. C.; Park, J. E.; Zhang, T.;
Liao, K.; Gu, J.; Yan, Y.; Shaw, T. W.; Abelev, E.; Bocarsly, A. B. Chem. Rev. 2015, 115,
12888.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
TOC GRAPHIC
ACS Paragon Plus Environment