You are on page 1of 11

Electrochimica Acta 412 (2022) 140129

Contents lists available at ScienceDirect

Electrochimica Acta
journal homepage: www.journals.elsevier.com/electrochimica-acta

A facile “dark”-deposition approach for Pt single-atom trapping on facetted


anatase TiO2 nanoflakes and use in photocatalytic H2 generation
Gihoon Cha a, 1, Anca Mazare a, b, Imgon Hwang a, Nikita Denisov a, Johannes Will c,
Tadahiro Yokosawa c, Zdeněk Badura d, Giorgio Zoppellaro d, Alexander B. Tesler a,
Erdmann Spiecker c, Patrik Schmuki a, d, e, *
a
Department of Materials Science WW4-LKO, Institute for Surface Science and Corrosion (LKO), Friedrich-Alexander-University of Erlangen-Nuremberg, Martensstrasse
7, D-91058, Erlangen 91058, Germany
b
Advanced Institute for Materials Research (AIMR), National University Corporation Tohoku University (TU), Sendai 980-8577, Japan
c
Institute of Micro- and Nanostructure Research (IMN) and Center for Nanoanalysis and Electron Microscopy (CENEM), Friedrich-Alexander-Universität Erlangen-
Nürnberg (FAU), IZNF, Cauerstraße 3, Erlangen 91058, Germany
d
Regional Center of Advanced Technologies and Materials, Šlechtitelů 27, Olomouc 78371, Czech Republic
e
Department of Chemistry, Faculty of Science, King Abdulaziz University, P.O. Box 80203, Jeddah 21569, Saudi Arabia

A R T I C L E I N F O A B S T R A C T

Keywords: In recent years, single-atom (SA) catalysts have become strongly investigated in electrocatalysis and increasingly
Platinum single-atom also in photocatalysis. Here we demonstrate that single-crystalline anatase TiO2 nanoflake (NF) layers after a
Photocatalysis facile thermal air annealing process provide suitable anchoring sites for Pt SAs. Remarkably, such air annealed
Hydrogen generation
(001) faceted single-crystalline TiO2 nanoflakes can trap and stabilize Pt SAs and assemblies when immersed into
Single-crystal anatase TiO2
a Pt precursor solution for sufficient time in absence of any illumination, i.e., under "dark deposition” conditions.
Single-atom catalysis
“Dark”-deposition These decorated species act as more efficient co-catalysts for the H2 evolution reaction, compared to the classical
Pt nanoparticles (NPs). Here we show that such single-crystal-based TiO2 photocatalyst, decorated with Pt SAs,
can display a 4-fold increase in photocatalytic H2 production at a lower Pt loading than observed for a classical Pt
nanoparticle Pt/TiO2 arrangement.

1. Introduction transfer reactions, typically noble metal co-catalysts are required for
photocatalytic reactions and, in particular, for the photocatalytic H2
Over the past few years, single atom (SA) catalysts have not only evolution reaction (HER). Noble metals such as Au, Pt, Pd, etc. are
been of high interest in thermal heterogeneous catalysis but also entered usually used in the form of nanoparticles (NPs) [15,17-21].
strongly into the field of electrocatalysis where it is used for organic In contrary to the common methods, used to deposit noble metal co-
electrosynthesis, fuel cells, batteries, electrolyzers, and particularly catalyst as NPs, SA co-catalysts have several substantial advantages such
focusing on oxygen reduction or evolution, hydrogen evolution or as a minimal noble metal loading, i.e., reduced costs, and, at the same
oxidation, carbon dioxide reduction or nitrogen reduction reactions) [1, time, a maximal surface-to-volume ratio. One of the most crucial issues
2]. The reason to utilize SAs is mainly due to the maximized for using SA co-catalyst is its surface immobilization on the desired
surface-to-volume ratio of such catalysts and novel reaction pathways substrate. Typical approaches for SA loading include methods based on
that become available at the SA level [3-11]. Most recently, SAs became chemical trapping or molecular-beam-soft-landing [22-24].In general
also attractive as co-catalysts in photocatalytic reactions [3,6,12-16]. In SA traps are based on specific electronic or geometric features of the
photocatalysis, the mobile charge carriers (electron-hole pairs) that are substrates [25].
generated by illuminating a semiconductor, experience charge transfer In the current state of research of the application of SA in photo­
reactions with the environment. To enhance the kinetics of the slow catalysis, two semiconductors are primarily investigated, i.e.,

* Corresponding author at: Department of Materials Science WW4-LKO, Institute for Surface Science and Corrosion (LKO), Friedrich-Alexander-University of
Erlangen-Nuremberg, Martensstrasse 7, D-91058, Erlangen 91058, Germany.
E-mail address: schmuki@ww.uni-erlangen.de (P. Schmuki).
1
Present address: Bavarian Center for Battery Technology (BayBatt), University of Bayreuth, 95447 Bayreuth, Germany.

https://doi.org/10.1016/j.electacta.2022.140129
Received 3 December 2021; Received in revised form 25 February 2022; Accepted 27 February 2022
Available online 1 March 2022
0013-4686/© 2022 Elsevier Ltd. All rights reserved.
G. Cha et al. Electrochimica Acta 412 (2022) 140129

nanostructured C3N4 and anatase TiO2 [6]. The former has good visible photoelectrochemistry in order to gain additional information on rele­
light absorption characteristics. On the other hand, nanostructured vant charge transfer processes.
anatase TiO2 is a benchmark photocatalytic material owing to its
excellent photocorrosion stability and wide range of applications that 2. Experimental
rely on photocatalytic reactions [19,26]. It is worth pointing out that
among the polymorphs of TiO2, anatase shows a higher reactivity to the 2.1. TiO2 nanoflakes fabrication
red-ox potential of water and thus has more optimized energetics than
rutile [26-29]. Before the TiO2 nanoflakes fabrication on FTO substrates (7 Ω•m− 2,
In the field of electrochemistry and photoelectrochemistry, the for­ Pilkington, UK), the FTO glasses were cleaned by ultrasonication in
mation of the anchoring points on the support materials’ surface, which acetone, ethanol, and DI water for 15 min each followed by drying under
can set up and sustain single atom (SA) states, is vital. Thus, the for­ a N2 stream. To fabricate TiO2 nanoflakes (NFs) with preferential (001)
mation of suitable trapping sites is indispensable for granting stability to faceting directly on the FTO substrates, a hydrothermal synthesis was
the SAs. So far, many approaches were investigated for the SA deposition carried out at 150 ◦ C for 15 h using a Teflon-lined stainless steel auto­
including methods such as coprecipitation, atomic layer deposition, clave (capacity 250 mL). The aqueous hydrogen chloride (HCl) solution
photochemical reduction, etc., which are based on metal-containing was prepared by adding 30 mL HCl (37 %) (Sigma-Aldrich, Germany)
precursors that are placed, reduced, and anchored on the support by into 30 mL DI water in a Teflon cell, followed by stirring for 15 min.
defects or vacancies [11]. Two points are further of significant impor­ Afterward, 1.5 mL titanium isopropoxide (Sigma-Aldrich, Germany) was
tance for SACs in electrocatalysis, i.e., selectivity (since alternative re­ added into the diluted HCl solution and vigorously stirred for 15 min at
action pathways compared to classic nanoparticles might be involved), room temperature. As a F− precursor that influences the surface energy
and activity/stability (the strong interaction between SA and the sup­ and fundamental structure unit in the TiO2 crystal, 0.5 g ammonium
port can significantly influence their electrochemical activity and sta­ hexafluorotitanate ((NH4)2TiF6, Sigma-Aldrich, Germany) was added to
bility) [11,30]. Yang et al. reported on a series of impregnation and the prepared solution, and then stirred at room temperature until
annealing approaches to decorate TiO2 nanoparticles with Pt-O units completely dissolved (for 15 min). The cleaned FTO substrates were
that provide a remarkable reactivity [31]. then fixed on the ceramic holder obliquely (the fluorine-doped Tin oxide
In contrast to these approaches, concerning TiO2 and single-atom co- surface was facing downwards). Subsequently, the hydrothermal syn­
catalysts, we showed recently that surface defects (Ti3+-Ov) induced by thesis was carried out followed by cooling down to room temperature in
reducing in Ar/H2 atmosphere (e.g., Pt SA deposition on anatase TiO2 the oven naturally. FTO/TiO2 NF samples were then cleaned with DI
sputtered layers [25], or Pd, Pt or Au on (001) TiO2 nanosheets [32]) or water and dried under a N2 stream. Before the Pt deposition, the TiO2
native defects in anodic TiO2 nanotubes (e.g., Ir or Pt SA deposition on NFs were annealed at different temperatures for 1 h in air, to remove the
anatase TiO2 nanotubes [33]), or on TiO2 nanoflakes (e.g. Pt SA [34]) F-terminated remnant contamination layer from the deposition process
allow a direct, simple immersion reaction of the TiO2 surface with a Pt4+ [44].
precursor to form Pt SAs with a nominal charge (d) of ≈2.
In view of single-crystalline (SC) anatase, facetted SC layers have 2.2. Pt deposition on TiO2 samples (dark / photodeposition)
attracted wide attention. The preferential faceting of anatase was shown
to significantly influence the electron and hole transfer in various pho­ The Pt “dark”- (or photo-) deposition on the prepared TiO2 NFs was
tocatalytic reactions. In the case of single-crystalline anatase with (101) performed by the facile dip-coating method with common Pt precursor
and (001) faceting, the different surface energy of these two facets leads as H2PtCl6 (Metakem GmbH, Germany) in aqueous methanol solution
to the formation of an intrinsic electron junction, which under illumi­ (50 vol %). Normally, a Pt stock solution was prepared initially with a
nation (in an aqueous environment) results in electrons exiting from the certain concentration of H2PtCl6 in DI water according to specifications
(101) and holes from the (001) planes [25,26]. Yang et al [35]. reported before the “dark”- (or photo-) deposition. For instance, 2 × 10− 4 M, 1 ×
the hydrothermal formation of single-crystal anatase TiO2 nanosheets 10− 3 M, and 5 × 10− 3 M of H2PtCl6 stock solutions were prepared for the
(NSs) with a high percentage of (001) facets. The key parameter that was concentrations of 1 × 10− 6 M and 5 × 10− 6 M, and 5 × 10− 5 M,
identified in this process was the presence of fluoride ions, acting as a respectively. The annealed TiO2 nanoflakes samples were immersed into
capping agent that enables the preferential growth of (001) faceted 10 mL of aqueous methanol solution in a quartz finger cell (reactor)
crystals (the surface energy of the (001) facets is reduced via the fluoride followed by the addition of 50 μL of the stock solution into it per each
termination) [35-38]. concentration. Afterward, Ar purging (15 min) was performed to remove
To note is that such (001) faceted powder nanosheets have recently the oxygen present in the solution. (і) “Dark”-deposition - to build up the
gained attention in photocatalytic and also in other TiO2 applications dark conditions, the quartz finger cell was wrapped with aluminum foil
[39-41]. However, the application of such crystallite powder nanosheets after Ar purging, and then kept under stirring at 500 RPM for 24 h
in photoelectrochemistry requires their further coating on a conductive (alternatively for 10 min or 1 h). After the “dark”-deposition, the sam­
substrate, thus making the production process more complicated and ples were rinsed with DI water and then dried under a N2 stream. (ii)
less efficient. In this sense, direct growth of faceted nanosheets on an Pre-photodeposition - after Ar purging, the quartz finger cell that con­
FTO substrate by hydrothermal synthesis is a straightforward approach, taining 50 μL of 1 × 10− 3 M H2PtCl6 stock solution was exposed to the
leading to nanoflake photoanodes [42-44]. light source of λ = 365 nm LED (100 mW) for 3 h leading to the for­
Here, we explore such single-crystalline (001) faceted TiO2 nano­ mation of Pt nanoparticles (NPs) on the surface of TiO2 NFs and,
sheets (or nanoflakes), grown directly on FTO substrates by a hydro­ simultaneously, accompanying by H2 evolution reaction. The illumi­
thermal method, that we further use for Pt single-atom decoration and nated samples were then rinsed with DI water followed by drying under
photocatalytic H2 evolution considering the feasibility to anchor Pt a N2 stream.
single-atoms. We show that a simple air annealing treatment is sufficient
to create suitable trapping sites on the single crystal surface. These traps 2.3. Characterization methods
then react in an immersion treatment of the TiO2 nanosheets in a
chloroplatinic acid solution without further UV illumination – we denote A field-emission scanning electron microscope (FE-SEM, Hitachi,
this process thus as “dark-deposition”. We then explore the SA decorated S4800) equipped with an Energy Dispersive X-Ray (EDAX Genesis, fitted
TiO2 nanoflakes for photocatalytic H2 evolution and find a remarkable to SEM chamber) were used for characterization of the morphology and
reactivity compared to the classic anatase/Pt NP system. Such SAs chemical composition of the samples. HAADF-STEM (High-angle
decorated TiO2 nanoflakes were then also characterized by annular dark-field scanning transmission electron microscopy) and EDS

2
G. Cha et al. Electrochimica Acta 412 (2022) 140129

(Energy-dispersive X-ray spectroscopy) mapping were performed with a Diffusive reflectance spectra were measured by a fiber-based UV-vis-
double-corrected FEI Titan Themis3 300 using an acceleration voltage of IR spectrophotometer (Avantes, ULS2048) equipped with an integrating
300 kV. The microscope is further equipped with a monochromator and sphere AvaSphere-30 using AvaLight-DH-S-BAL balanced power light
a Super-X detector system for STEM-EDS analysis. High-resolution source. The optical bandgap of the TiO2 films was obtained by Kubelka-
STEM-HAADF (HRSTEM-HAADF) images were taken with a collection Munk equation using UV-Vis diffuse reflectance spectra for the indirect
angle from 61 to 200 mrad, a beam current of 120 pA, and a dwell time allowed transitions.
of 50 µs. The crystal phase structure of the samples was obtained by X-
ray diffraction (XRD, X’pert Philips PMD diffractometer) operating with 2.4. Photocatalytic H2 evolution measurements
graphite monochromatized Cu irradiation (wavelength: λ = 1.54056
Å.). The chemical composition of the samples was evaluated by X-ray The H2 evolution measurements, “dark”- or photo- deposited TiO2
Photoelectron Spectroscopy (XPS, PHI5600), and spectra were cali­ NF samples, were performed by immersion the samples into 10 mL of
brated to the Ti 2p peak to 458.0 eV. Fitting of the peaks was performed aqueous methanol solution followed by Ar purging for 15 min. After­
in Multipak software, using an asymmetric peak shape for the metallic ward, the samples were exposed to the LED light source of λ = 365 nm
platinum and a Gaussian-Lorentzian peak shape for the platinum peaks (100 mW). The illuminated area was approximately 1 cm2. For the
attributed to Ptδ+ or Pt4+. EPR spectra were recorded on JEOL-X-320 “normal” photodeposition, the annealed TiO2 nanoflakes were placed in
spectrometer with a variable He temperature set-up ES-CT470 appa­ a quartz reactor containing 5 × 10− 6 M Pt precursor with 10 mL aqueous
ratus. All the samples were measured with similar weight (≈15 mg). In methanol solution and were subjected to illumination with the same LED
all measurements, EPR tubes were kept at the same position in the cavity light source, while at the same time, measuring the H2 evolution ac­
for comparison. Spectra were collected at T = 80 K with following tivity. The photocatalytic H2 evolution was determined by a gas chro­
experimental parameters: modulation amplitude 0.7 mT, modulation matograph (GCMS-QO2010SE, SHIMADZU) with a thermal
frequency 100 kHz, time constant 0.03 s, microwave power 1 mW and conductivity detector (TCD).
sweep time 2 min with 3 accumulations to improve signal to noise ratio.

Fig. 1. (a) SEM image of the as-formed NFs (low magnification), (b) schematic representation of the NF facets, (c,d) high-resolution SEM images of as-formed and
anatase TiO2 NFs (annealed at 450 ◦ C in air). (e) H2 evolution rate at λ = 365 nm of the “dark”-deposited (D-D) as-formed and annealed TiO2 NFs. (f) XPS F 1s high-
resolution peak for the “dark”-deposition as-formed and anatase NFs (F/Ti atomic ratio).

3
G. Cha et al. Electrochimica Acta 412 (2022) 140129

2.5. Photoelectrochemical characterization oxide (FTO) substrates as described in the experimental section. The
TiO2 NF surface, as synthesized, is very smooth and the sheets show F-
Incident photon-to-electron conversion efficiency (IPCE) measure­ termination [43,47] resulting from the used growth process. Previously
ments were carried out in the 300–600 nm range with a setup we demonstrated that annealing in air reduces the F-termination on the
comprising an Oriel 6356 150 W Xe arc lamp as the light source and an TiO2 nanostructures [44]. We annealed the as-formed NFs for 1 h at 150,
Oriel cornerstone 7400 1/8 monochromator as described previously 300, 450, 600, and 800 ◦ C in air (Figs. 2 and S1,2), and the NFs annealed
[45,46]. The measurements were carried out in aqueous 0.1 M Na2SO4 at 450 ◦ C were shown to be optimal for H2 evolution after SA decoration,
solutions, and at an applied potential of +500 mV, in an electrochemical as will be further discussed. The SEM of the surface after partial
cell equipped with a quartz window and operated in a three-electrode F-removal (Fig. 1d) shows clearly roughness – i.e. the presence of surface
configuration with the nanoflake photoanode as a working electrode, defects after this annealing process. After annealing, the samples display
an Ag/AgClSat electrode as the reference electrode, and a Pt foil as the a similar faceted structure but show a reduction in the F-termination by
counter electrode. To measure J-V spectra, the photoelectrode was approx. 50%, as evident from XPS measurements (see F1s peak in
irradiated using a 365 nm wavelength LED light source providing an Fig. 1f).
intensity of 20 mW•cm− 2 and in 0.1 M Na2SO4 electrolyte. The exper­ This is in line with literature reporting that the defluorination in­
iment was carried out in a three-electrode electrochemical cell with Pt troduces Ti3+ surface defects [48,49]. To test the feasibility of using
foil as a counter electrode and Ag/AgClSat as a reference electrode. these surface defects as anchoring points for a Pt co-catalyst on the single
crystal surface, we immersed the annealed surface in an H2PtCl6 solution
3. Results and discussion for up to 24 h. For reference, we also immersed a non-annealed sample.
In neither case in SEM, any surface modification after the H2PtCl6
Fig. 1a, c shows high-resolution SEM images of as-formed (001) exposure can be seen. Then we tested the samples (non-annealed and
facetted TiO2 nanoflakes (NFs) at various magnifications, together with air-annealed after exposure to H2PtCl6) and compared them to the ac­
a schematic representation of the different NF facets (Fig. 1b). The TiO2 tivity of a sample not exposed to H2PtCl6 (as shown in Fig. 1e). As seen,
NF layers were formed by hydrothermal synthesis on fluorine-doped tin the anatase (001) faceted TiO2 NFs result in a negligible H2 generation

Fig. 2. (a) High-resolution SEM images of the as-formed and annealed TiO2 NFs at 300, 450, and 800 ◦ C for 1h showing the (101) and (001) facets. (b) High-
resolution XPS spectra of the as-formed and annealed NFs for Ti 2p, O 1s, and F 1s with (c) corresponding F/Ti ratio obtained from the atomic percentage. (d)
XRD spectra of the as-formed and annealed TiO2 NFs with a comparison of (e) anatase peak (101) and (f) anatase (105) and (211) peaks.

4
G. Cha et al. Electrochimica Acta 412 (2022) 140129

rate at λ = 365 nm (Fig. 1e, black line). When “dark”-deposited, i.e., with
a Pt SA co-catalyst decoration, the annealed TiO2 NFs display superior
H2 evolution rate as well as stability over the experimental duration, i.e.,
24 h (Fig. 1e, blue line), compared to the “dark”-deposited as-formed
faceted TiO2 NFs (Fig. 1e, red line).
We evaluated in more detail the surface morphology, chemistry, and
crystallinity of the as-formed and the NFs annealed at 150, 300, 450,
600, and 800 ◦ C in air – see Fig. 2 and the high-resolution SEM images of
the different facets in Figs. S1,2, and additional XPS and XRD data in
Fig. S3 and Table S1. The formation of holes on the (001) surface and a
roughening (surface defects) on the (101) surface is observed with
increasing annealing temperature, the latter disappears when annealed
at 800 ◦ C (Figs. 2a and S1,2). With increasing the annealing temperature
there is no significant difference in the Ti 2p XPS peaks, whereas, in the
case of F 1s, there is a substantial decrease, which leads to a F/Ti ratio of
almost zero for temperatures higher than 600 ◦ C. There is however a
difference in the O1s peak for NFs annealed at 800 ◦ C where a pro­
nounced shoulder is clearly visible at ≈532–536 eV – this is due to the
presence of SnO2 due to the diffusion of Sn4+ ions into the lattice of TiO2
from the substrate (the fluorine-doped tin oxide glass). This is clearly
visible in the XPS survey spectra where typical Sn peaks appear only
after annealing at 800 ◦ C (Fig. S4).
The XRD patterns are shown in Fig. 2c, and the overlaid anatase
(105) and (211) peaks in Fig. 2d,e, together with the peak profile fitting
in Table S1. The latter yields a preferred alignment of (101) and (200)
lattice planes as compared to (004). This texture is stable upon both Pt
loading and annealing at elevated temperatures. In addition, XRD
analysis yields an initially tensile strained crystal structure of the as-
formed NFs, which relaxes towards anatase bulk equilibrium after
annealing at 450 ◦ C for 1 h. Interestingly, this lattice strain relaxation is
accompanied by hole formation at the surface of the NFs as revealed by Fig. 3. X-band CW EPR spectra of as-formed TiO2 NFs powder (red line) and
after thermal annealing at different temperatures: 1 h at 300 ◦ C (green line),
HR-SEM imaging (see further Figs. 2a and S1,2) and could, thus, be
450 ◦ C (blue line) or 800 ◦ C (magenta line). The EPR spectra were recorded on
related to the out-diffusion of vacant lattice sites. Whereas Pt loading
powder samples, at T = 80 K and with 1.0 mW of applied microwave power.
does not affect the average measured lattice parameters, annealing at
800 ◦ C for 1 h leads again to a tensile strain of the anatase lattice, which
becomes highly dominant, in comparison to the g⊥=1.999 signal (a),
indicates defect formation in the anatase NF bulk at 800 ◦ C.
and an additional contribution from another type of Ti3+ sites can be
Additional information on the defect structure of the NFs (collected
observed, which is characterized by lower g-values (g⊥=1.967 and g||=
from scraped-off powder) was obtained by electron paramagnetic
1.936, marked as c in Fig. 3). It is speculated that these Ti3+ centers (c)
resonance (EPR) spectroscopy Fig. 3. shows the EPR signals for as-
are subjected to smaller orthorhombic distortions compared to the other
formed NFs and annealed NFs at different temperatures. The as-
lattice embedded Ti3+ sites in anatase, and thus such resonances are
prepared NFs (Fig. 3, red trace), characterized by the highest content
similar to those found in lattice embedded Ti3+ centers observed in rutile
of fluorine within the series, shows very weak signals, just above
polymorphs (g⊥=1.97 and g||= 1.94). The presence of surface exposed
background noise, with g⊥=1.999 (marked as a) and at g⊥=1.984
Ti3+ centers that usually express broad and unresolved signals around
(marked as b). These g-values values are consistent with Ti3+ sites, lat­
g~1.93 [53] is not evident in the sample treated at 450 ◦ C, but a minor
tice embedded, located in anatase phase that are experiencing small
fraction of such sites can be seen in the sample annealed at 800 ◦ C, from
differences in lattice distortions [50]. The parallel g-components of the
the uneven background line (zero level signal) at magnetic field higher
Ti3+ fingerprints are known to be much weaker, and especially in
than 337 mT.
as-prepared NFs, remain unresolved. In addition, hyperfine interactions
Another type of defects usually encountered in literature for TiO2
with fluorine nuclei (19F, I=1/2) do not emerge in the EPR spectrum.
nanotubes subjected to air annealing conditions, is voids, that originate
Thermal annealing of NFs leads to an increase of the recorded EPR signal
from protonated titanium vacancies present in the amorphous material
intensities. In the thermally treated samples (Fig. 3, green spectrum is
[54] (a strongly containing hydroxyl material), but are not directly
the sample annealed at 300 ◦ C; blue spectrum corresponds to annealed
evident in our single crystalline material.
at 450 ◦ C and magenta spectrum to annealed at 800 ◦ C), a very weak
Evaluating the photocatalytic H2 evolution efficiency for “dark”-
cluster of resonances develops at low magnetic fields, < 323 mT (with
deposited TiO2 NFs, i.e., Pt SA decorated, for as-formed and NFs
g>ge), which corresponds to resonances addressable to oxygen-based
annealed at a different temperature, we observed that annealing treat­
spin containing sites (e.g. Ti-O•). Significant enhancements of the sig­
ments at 450 ◦ C and 600 ◦ C led to similar H2 evolution rates (Fig. S5).
nals associated to the lattice embedded Ti3+ species are especially
From these, we selected the partially defluorinated samples at 450 ◦ C for
observed after thermal annealing at 450 ◦ C and 800 ◦ C. Similarly to the
further investigations – this considering that previous photo­
as-prepared NFs and in agreement with their XRD analysis, in the
electrochemical investigations showed 450 ◦ C annealed NFs to have
thermally treated NFs become well observable the diverse impacts of
better electron transport time than those annealed at 600 ◦ C [45], and
lattice distortions acting onto the reduced Ti3+ sites, which express two
that annealing at 800 ◦ C had the disadvantage of Sn diffusion. A key
sets of g⊥ values, at g⊥=1.999 (a), g⊥=1.984 (b) and g||= 1.956 in both
parameter for the activity of the samples is the exposure time in the
450 ◦ C and 800 ◦ C samples, and these results are in agreement with
H2PtCl6 solution, which was varied from 10 min to 24 h (see Fig. 4).
previous observations from literature reports [51,52]. However, in the
Results from high-resolution SEM and TEM images demonstrating the
EPR spectrum associated to the sample annealed at the highest tem­
morphology of either “dark”-deposited by immersion in 5 × 10− 6 M
perature (Fig. 3, 800 ◦ C, magenta line), the g⊥=1.984 (b) component

5
G. Cha et al. Electrochimica Acta 412 (2022) 140129

Fig. 4. (a) Photocatalytic H2 evolution rate (μL h− 1) for pre- or normal photodeposition of Pt NPs (NO.6 and NO.7), and different Pt SAs decoration – 5 × 10− 6 M
precursor for 10 min, 1 h and 24 h (NO.1, NO.2 and NO.3, respectively), and 24 h “dark”-deposition using different precursor concentration of 1 × 10− 6 mM (NO.4)
and 5 × 10− 5 mM (NO.5). (b) HR-SEM images of the annealed nanoflakes (NFs) (c–f) various “dark”-deposited treatments leading to Pt SAs and (g) pre-
photodeposited treatment leading to Pt NPs (NO.6). (h) HR-STEM image of the 24 h “dark”-deposited Pt NFs (NO.3) and (i,l) corresponding HAADF-STEM im­
ages of the O/Ti and Pt distribution.

PtCl62− solution for 10 min, 1 h, and 24 h are shown in Fig. 4c–d or for formation of 2–5 nm in diameter Pt NPs (Fig. 4h).
pre-photodeposited Pt NPs in Fig. 4h. For the “dark”-deposition, SEM Regarding the influence of the concentration, we found that among
characterization does not reveal the presence of any NPs on the surface the tested precursor concentrations, i.e., 1 × 10− 6 M, 5 × 10− 6 M, and 5
of the TiO2 NFs. Evidently, with times up to 24, the photocatalytic H2 × 10− 5 M, the highest H2 evolution rate was achieved for the 5 × 10− 6 M
evolution increases and the SEM does not show any trace of particle concentration. For the lowest precursor concentration (1 × 10− 6 M) no
formation. Therefore, in all further experiments, this optimal “dark”-­ Pt NPs were observable. Interestingly, for the highest concentration, i.e.,
deposition immersion time was used to produce SA-loaded NFs. In 5 × 10− 5 M, distinct Pt NPs with a diameter of 4–8 nm were obtained in
contrast, conventional Pt deposition, i.e., samples that were prepared much higher numbers than attained for the pre-photodeposited sample
using the pre-photodeposited approach, as expected, show the (Fig. 4f,g). In addition, both photodeposition and “dark”-deposition with

6
G. Cha et al. Electrochimica Acta 412 (2022) 140129

higher concentration lead to preferential deposition of NPs on the (101) marked area). TEM-EDX images further demonstrate the uniformity of
facets. the Pt SAs trapping on the TiO2 NF surface (Fig. 4j–l). Interestingly, SAs
The presence of Pt SAs and agglomerates of a few SAs on the “dark”- and their agglomerates locate preferably on (101) facets of the NFs
deposited samples is evident from HR-STEM-HAADF, considering also (Fig. S6).
the usual SA literature [3,6,12-16,55] (Fig. 4i), and STEM-EDX images The chemical composition and the coordination states of the anatase
as taken for a sample deposited for 24 h (Fig. 4j–l), see also the FTT and Pt decorated NFs, i.e., Pt NPs obtained by pre-photodeposition and
analysis in Fig. S6 confirming the Pt SAs deposition on the (101) facet of Pt SAs formed by the 24 h “dark”-deposition, were further evaluated by
the NF. The results distinctly point out that facetted anatase layers after the survey and high-resolution XPS analysis (Figs. 5b–f and S7). As
annealing can effectively trap Pt SAs directly by “dark-deposition” (a shown, a clear shift to 72.06 eV is observed in the Pt 4f7/2 peak (Fig. S7a)
simple 24 h immersion in the dark in a H2PtCl6 solution) (Fig. 4h, for the “dark”-deposited Pt SAs, compared to 70.18 eV for the Pt NPs

Fig. 5. (a) XRD spectra of the pre-photodeposited, “dark”-deposited, as-formed, and annealed TiO2 NFs. (b) XPS survey spectra of the annealed, pre-photodeposited,
and “dark”-deposited NFs. High-resolution XPS spectra and peak fitting of the Pt 4f of (c) the pre-photodeposited (Pt NPs) and (d) “dark”-deposited (Pt SAs) on TiO2
NFs. (e) High-resolution XPS spectra of Cl 2p of the samples shown in (b). (f) The atomic percentage (at.%) was obtained from XPS measurements.

7
G. Cha et al. Electrochimica Acta 412 (2022) 140129

decorated NFs. Further fitting of the Pt 4f peaks results as follows. (i) For cannot be detected by XRD (Fig. 5a), while only the anatase crystalline
the pre-photodeposited Pt NPs (NO.6), the typical asymmetric metallic structure together with the SnO2 from the FTO was evident (with no
Pt peaks, i.e., Pt0 4f7/2 and 4f5/2 peaks appear at 70.18 eV and 73.51 eV, change in the measured lattice parameters, compared to the NFs
respectively. (ii) For the “dark”-deposited NFs (NO.3) Pt is present in a annealed at 450 ◦ C for 1 h, see Table S1).
SA state, i.e., Ptδ+ 4f7/2 and 4f5/2 peaks appear at 72.06 eV and 75.43 eV, Finally, we performed a long-term UV illumination experiment for
respectively, having also a small contribution of 11.3 % from Pt2+ at the annealed, “dark”-deposited Pt SAs (NO.3), and pre-photodeposited
74.0 and 77.4 eV. The Ptδ+ peak locations of the Pt SA state (δ ≈ 2+) are Pt NPs (NO.6) for the estimation of photogenerated H2 evaluation
in line with literature on Pt SA [25,32,56]. A very low amount of rates. It is noteworthy that almost a 4-fold increase of photogenerated H2
chlorine is also detected by XPS after the “dark”-deposition (Fig. 5e), and was measured for these optimized Pt SA “dark”-deposited NFs (NO.3) for
considering the Pt/Cl ratio of 1:0.67, the surface attached Pt SAs are only illumination times of up to 24 h, as compared to the pre-photodeposited
mildly coordinated with chloride. The presence of Cl in low amounts and Pt NPs on TiO2 NFs (NO.6) (Fig. 6a). The photocatalytic H2 evolution
the δ ≈ 2+ suggests that the surface trapping of the Pt SAs takes place via rate of the Pt SAs decorated NFs was measured as ~44 µL h− 1 and by
2+
a galvanic displacement reaction Ti3+ →Ti4+ ; Pt 4+ →Ptsurface trapped [25],
longer illumination times from 2 h up to 24 h, it demonstrates only a
as reported for after defective material [2,4,6]. slight decrease to 41.3 µL h− 1, which is about 6 % (Fig. 6b, blue line). At
The Pt SAs prepared by “dark”-deposition on TiO2 NFs demonstrate a the same time, the pre-photodeposited Pt NPs on the NFs demonstrate a
photocatalytic H2 evolution efficiency much higher than that obtained more pronounced reduction in the photogenerated H2 production rate
by the conventional crystalline Pt NPs decoration, even though the pre- by prolonging illumination from 25.27 µL h− 1 to 12.14 µL h− 1, which is
photodeposited sample demonstrates either 3-fold higher Pt loading as 52 % (Fig. 6b, red line). To note that this difference in H2 production rate
measured by XPS (Fig. 5f) or 1.5-fold higher loading as measured by EDX occurs despite the higher Pt at.% loading as calculated from EDX mea­
(Table S2). It should be noted that the higher amount of Pt loading surements (0.1 at.% for pre-photodeposited and 0.07 at.% for “dark”-
obtained by the XPS analysis is due to the high surface sensitivity of the deposited decorated NFs (Table S2). Moreover, while destabilization of
analytical technique. The presence of Pt SAs on the surface of the NFs the SA occurs for the “dark”-deposited NFs (as evident from the SEM

Fig. 6. (a) Amount of H2 evolution photogenerated over 24 h of irradiation at λ = 365 nm for the annealed, the “dark”-deposited (NO.3), and the pre-photodeposited
(NO.6) TiO2 NFs. (b) Photocatalytic H2 evolution rate over 24 h illumination for the samples shown in (a) with additional data at 4, 8, 16, and 20 h of irradiation for
the dark”-deposited NTs (NO.3). (c) Photoelectrochemical J-V curves measured under irradiation at λ = 365 nm for the annealed, the “dark”-deposited (NO.3), and
the pre-photodeposited (NO.6) TiO2 NFs. (d) IPCE and corresponding bandgap (inset plot) spectra of the annealed, the “dark”-deposited (NO.3), and the pre-
photodeposited (NO.6) TiO2 NFs.

8
G. Cha et al. Electrochimica Acta 412 (2022) 140129

images in Fig. S8, Pt particles in the 1–4 nm diameter range), it does not the same NFs. The “dark”-deposition process is facile, low-cost, and does
have a significant influence on the H2 evolution rate. Hence, the reac­ not require special instrumentation, yet forms an efficient single-atom
tivity of the entire photocatalyst is virtually maintained throughout the layer on TiO2 NFs. We have optimized the deposition procedure to
illumination time (on average 41.3 µL h− 1). obtain maximal H2 generation rates by adjacent deposition time and Pt
In order to elucidate more on the effect of Pt on the charge transfer precursor concentration. An optimal “dark”-deposition timeframe is 24
reaction, we acquired the photocurrent-voltage characteristics of the h, and using a minimal Pt precursor concentration of 5 × 10− 6 M out­
TiO2 NFs with and without Pt loading, this for SA and NP decorated performs the classical Pt photodeposited or pre-photodeposited nano­
substrates (Fig. 6c). In general, the voltage dependance of the photo­ particles decoration, showing a stable rate over extended illumination
current can be described by the Gärtner-Butler-Johnson approach as times. We believe that this work introduces a fundamentally new tool in
outlined in Eqs. (1) and (2) [57-59], SAs co-catalyst engineering using the surface defects and the (001)
( [ ( )1/2 ]( )− 1 ) facets of anatase NFs, that show potential in their further use, and may
Jph = qφ0 1 − − exp − αdsc U − − Ufb 1 + αLp (1) inspire the alternative green energy generation community to investi­
gate further the SA catalysis phenomenon.
α = A(hν − Eg)n/2 (hν)− 1
(2)
CRediT authorship contribution statement
where φ0 (s− 1 cm− 2) is the photon flux, q (C) is the charge transferred
per ion, α (m− 1) is the optical absorption coefficient, dsc (nm) is the Gihoon Cha: Investigation, Writing – original draft, Data curation,
depletion layer width for a potential of 1 V across it, e0 (1.6 × 10− 19 C) is Formal analysis. Anca Mazare: Formal analysis, Visualization, Writing –
the electron charge, ε is the dielectric constant of TiO2, ε0 (8.85 × 10− 12 original draft, Writing – review & editing. Imgon Hwang: Investigation,
F m− 1) is the permittivity of vacuum, U (V) is the applied potential, Ufb Data curation, Formal analysis. Nikita Denisov: Investigation, Data
(V) is the flat-band potential, Lp (nm) is the hole diffusion length, A is a curation, Formal analysis. Johannes Will: Investigation, Formal anal­
constant, hν (1240 eVnm− 1) is the photon energy of the incident light, ysis. Tadahiro Yokosawa: Investigation, Data curation, Formal anal­
Eg is the bandgap, and n equals 1 (for direct transitions) or 4 (indirect ysis, Writing – review & editing. Zdeněk Badura: Formal analysis,
transitions).[60] Writing – review & editing. Giorgio Zoppellaro: Resources, Writing –
In line with Eq. (1), photocurrent growth is observed essentially with review & editing. Alexander B. Tesler: Formal analysis, Writing –
(U-Ufb)1/2 for all three samples and reaches a comparable magnitude. In original draft, Writing – review & editing. Erdmann Spiecker: Re­
other words the presence of Pt does not affect the anodic photocurrent sources, Visualization. Patrik Schmuki: Resources, Conceptualization,
(hole transfer reaction) over the entire investigated range, neither in the Writing – original draft, Writing – review & editing, Supervision.
form of SAs nor as NPs. This means the beneficial effect of SAs (and NPs)
in photocatalysis must entirely be ascribed to a facilitated electron
Declaration of Competing Interest
transfer reaction.
Various literature ascribes to Pt SAs an effect on the optical band-gap
The authors declare that they have no known competing financial
[6]. Diffuse reflectance measurements of the same samples demonstrate
interests or personal relationships that could have appeared to influence
identical reflectance in the UV spectral range due to light absorption via
the work reported in this paper.
a band-to-band electron excitation in anatase TiO2 (Fig. S9a) [46]. On
the other hand, both plain and Pt-SA decorated TiO2 NF samples
Acknowledgments
demonstrate identical reflectance spectra in the visible spectral range,
while the Pt-NP TiO2 NF sample shows slightly lower reflectance. The
The authors would like to acknowledge the DFG and the Operational
latter is attributed to the photon absorption by Pt NPs under visible light
research program, Development and Education (European Regional
irradiation due to the plasmonic effect, i.e., the collective oscillation of
Development Fund, Project No. CZ.02.1.01/0.0/0.0/15_003/0000416
free electrons confined to the metallic NP [61,62], while in the case of
of the Ministry of Education, Youth and Sports of the Czech Republic) for
Pt-SA decoration this phenomenon does not occur [63]. Correspond­
financial support. Fahimeh Shahvardanfard is acknowledged for her
ingly, comparable optical bandgap energies of 3.27 eV were obtained,
contribution to this work. J.W. and E.S. acknowledge the research
estimated from the Kubelka-Munk transformations (see Fig. S9b).
training group GRK 1896 “In Situ Microscopy with Electrons, X-rays,
To investigate in more detail Pt effects on the electronic properties,
and Scanning Probes”. A.B.T. acknowledges the DFG (grant number
we acquired IPCE measurements and evaluated the bandgap according
442826449; SCHM 1597/38-1 and FA 336/13-1) for financial support.
to Eq. (2) (Fig. 6d). An. The data for the bandgap evaluation are a replot
of the IPCE spectra (according to an indirect transition) yield for all
Supplementary materials
samples a bandgap value of 3.19 eV – this is well in line with the re­
ported bandgap energies of anatase TiO2 (Fig. 6d, inset) [64]. This
Supplementary material associated with this article can be found, in
clearly shows that in the present case any sort of light absorption effect
the online version, at doi:10.1016/j.electacta.2022.140129.
by Pt SAs can be excluded as a reason for the enhancement of photo­
catalytic H2 generation.
References
4. Conclusions [1] Q. Zhang, J. Guan, Single-atom catalysts for electrocatalytic applications, Adv.
Funct. Mater. 30 (2020) 1–53, https://doi.org/10.1002/adfm.202000768.
In conclusion, the current work introduces a strategy to use (001) [2] W. Zang, Z. Kou, S.J. Pennycook, J. Wang, Heterogeneous single atom
electrocatalysis, where “singles” are “married,”, Adv. Energy Mater. 10 (2020),
faceted TiO2 nanosheets/nanoflakes as a substrate to "dark”-deposit Pt 1903181 https://doi.org/10.1002/aenm.201903181.
single-atoms (SAs). Traps are achieved by annealing of the nanoflakes [3] J. Liu, Catalysis by supported single metal atoms, ACS Catal. 7 (2017) 34–59,
and defluorination, whereas usually SA trapping on TiO2 is established https://doi.org/10.1021/acscatal.6b01534.
[4] B.C. Gates, M. Flytzani-Stephanopoulos, D.A. DIxon, A. Katz, Atomically dispersed
on (Ti3+-Ov) TiO2 states (by annealing in Ar/H2 atmosphere) or native
supported metal catalysts: perspectives and suggestions for future research, Catal.
defects formed in anodic TiO2 nanotubes. Here we introduce single- Sci. Technol. 7 (2017) 4259–4275, https://doi.org/10.1039/c7cy00881c.
crystalline TiO2 NF surfaces with enough defect sites to effectively [5] A. Wang, J. Li, T. Zhang, Heterogeneous single-atom catalysis, Nat. Rev. Chem. 2
reduce Pt precursor and trap it as a SA layer. Such an optimal Pt SAs (2018) 65–81, https://doi.org/10.1038/s41570-018-0010-1.
[6] C. Gao, J. Low, R. Long, T. Kong, J. Zhu, Y. Xiong, Heterogeneous single-atom
decoration of the NFs provides a 4-fold higher reactivity for photo­ photocatalysts: fundamentals and applications, Chem. Rev. 120 (2020)
catalytic H2 production compared to conventional Pt NPs decoration on 12175–12216, https://doi.org/10.1021/acs.chemrev.9b00840.

9
G. Cha et al. Electrochimica Acta 412 (2022) 140129

[7] U. Heiz, A. Sanchez, S. Abbet, W.D. Schneider, Catalytic oxidation of carbon H2 generation, Sol. RRL (2022), 2101026, https://doi.org/10.1002/
monoxide on monodispersed platinum clusters: each atom counts, J. Am. Chem. solr.202101026.
Soc. 121 (1999) 3214–3217, https://doi.org/10.1021/ja983616l. [35] H.G. Yang, C.H. Sun, S.Z. Qiao, J. Zou, G. Liu, S.C. Smith, H.M. Cheng, G.Q. Lu,
[8] M. Flytzani-Stephanopoulos, Gold atoms stabilized on various supports catalyze Anatase TiO2 single crystals with a large percentage of reactive facets, Nature 453
the water-gas shift reaction, Acc. Chem. Res. 47 (2014) 783–792, https://doi.org/ (2008) 638–641, https://doi.org/10.1038/nature06964.
10.1021/ar4001845. [36] W.Q. Fang, J.Z. Zhou, J. Liu, Z.G. Chen, C. Yang, C.H. Sun, G.R. Qian, J. Zou, S.
[9] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Active nonmetallic Au and Pt Z. Qiao, H.G. Yang, Hierarchical structures of single-crystalline anatase TiO2
species on ceria-based water-gas shift catalysts, Science 301 (2003) 935–938, nanosheets dominated by {001} facets, Chem. A Eur. J. 17 (2011) 1423–1427,
https://doi.org/10.1126/science.1085721 (80-.). https://doi.org/10.1002/chem.201002582.
[10] J. Kim, H.E. Kim, H. Lee, Single-atom catalysts of precious metals for [37] M. Dozzi, E. Selli, Specific facets-dominated anatase TiO2: fluorine-mediated
electrochemical reactions, ChemSusChem. 11 (2018) 104–113, https://doi.org/ synthesis and photoactivity, Catalysts 3 (2013) 455–485, https://doi.org/10.3390/
10.1002/cssc.201701306. catal3020455.
[11] C. Zhu, S. Fu, Q. Shi, D. Du, Y. Lin, Single-atom electrocatalysts, Angew. Chem. Int. [38] Q. Wu, M. Liu, Z. Wu, Y. Li, L. Piao, Is photooxidation activity of {001} facets truly
Ed. 56 (2017) 13944–13960, https://doi.org/10.1002/anie.201703864. lower than that of {101} facets for anatase TiO2 crystals? J. Phys. Chem. C 116
[12] B. Xia, Y. Zhang, J. Ran, M. Jaroniec, S.Z. Qiao, Single-atom photocatalysts for (2012) 26800–26804, https://doi.org/10.1021/jp3087495.
emerging reactions, ACS Cent. Sci. 7 (2021) 39–54, https://doi.org/10.1021/ [39] Z. Xiong, Z. Lei, Y. Li, L. Dong, Y. Zhao, J. Zhang, A review on modification of facet-
acscentsci.0c01466. engineered TiO2 for photocatalytic CO2 reduction, J. Photochem. Photobiol. C
[13] B. Wang, H. Cai, S. Shen, Single metal atom photocatalysis, Small Methods 3 Photochem. Rev. 36 (2018) 24–47, https://doi.org/10.1016/j.
(2019), 1800447, https://doi.org/10.1002/smtd.201800447. jphotochemrev.2018.07.002.
[14] L. Zeng, C. Xue, Single metal atom decorated photocatalysts: progress and [40] A. Meng, J. Zhang, D. Xu, B. Cheng, J. Yu, Enhanced photocatalytic H2 -production
challenges, Nano Res. 14 (2021) 934–944, https://doi.org/10.1007/s12274-020- activity of anatase TiO2 nanosheet by selectively depositing dual-cocatalysts on
3099-8. {101} and {001} facets, Appl. Catal. B Environ. 198 (2016) 286–294, https://doi.
[15] G. Marci, L. Palmisano, Heterogeneous Photocatalysis, Elsevier, 2019, https://doi. org/10.1016/j.apcatb.2016.05.074.
org/10.1016/C2016-0-04769-1. [41] P. Liu, X. Huo, Y. Tang, J. Xu, X. Liu, D.K.Y. Wong, A TiO2 nanosheet-g-C3N4
[16] Z. Kou, W. Zang, P. Wang, X. Li, J. Wang, Single atom catalysts: a surface composite photoelectrochemical enzyme biosensor excitable by visible irradiation,
heterocompound perspective, Nanoscale Horiz. 5 (2020) 757–764, https://doi. Anal. Chim. Acta 984 (2017) 86–95, https://doi.org/10.1016/j.aca.2017.06.043.
org/10.1039/D0NH00088D. [42] D. Zhong, Q. Jiang, B. Huang, W.H. Zhang, C. Li, Synthesis and characterization of
[17] C. Fuerstner, K. Faul, K. Burkhard, Photocatalysis in Organic Synthesis, Georg anatase TiO2 nanosheet arrays on FTO substrate, J. Energy Chem. 24 (2015)
Thieme Verlag, Stuttgart, 2019, https://doi.org/10.1055/b-006-161273. 626–631, https://doi.org/10.1016/j.jechem.2015.08.002.
[18] H. Kisch, Semiconductor Photocatalysis, ed.,, Wiley-VCH Verlag GmbH & Co. [43] T. Butburee, P. Kotchasarn, P. Hirunsit, Z. Sun, Q. Tang, P. Khemthong,
KGaA, Weinheim, Germany, 2014, https://doi.org/10.1002/9783527673315. W. Sangkhun, W. Thongsuwan, P. Kumnorkaew, H. Wang, K. Faungnawakij, New
[19] A. Fujishima, X. Zhang, D.A. Tryk, TiO2 photocatalysis and related surface understanding of crystal control and facet selectivity of titanium dioxide ruling
phenomena, Surf. Sci. Rep. 63 (2008) 515–582, https://doi.org/10.1016/j. photocatalytic performance, J. Mater. Chem. A 7 (2019) 8156–8166, https://doi.
surfrep.2008.10.001. org/10.1039/C8TA11475G.
[20] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Environmental [44] F. Shahvardanfard, G. Cha, N. Denisov, B. Osuagwu, P. Schmuki,
applications of semiconductor photocatalysis, Chem. Rev. 95 (1995) 69–96, Photoelectrochemical performance of facet-controlled TiO2 nanosheets grown
https://doi.org/10.1021/cr00033a004. hydrothermally on FTO, Nanoscale Adv. 3 (2021) 747–754, https://doi.org/
[21] T.L. Thompson, J.T. Yates, Surface science studies of the photoactivation of TiO2 10.1039/d0na01017k.
new photochemical processes, Chem. Rev. 106 (2006) 4428–4453, https://doi. [45] X. Zhou, N. Denisov, G. Cha, I. Hwang, P. Schmuki, Photoelectrochemical
org/10.1021/cr050172k. performance of TiO2 photoanodes: nanotube versus nanoflake electrodes,
[22] U. Heiz, A. Sanchez, S. Abbet, W.-D. Schneider, Catalytic oxidation of carbon Electrochem. Commun. 124 (2021), 106937, https://doi.org/10.1016/j.
monoxide on monodispersed platinum clusters: each atom counts, J. Am. Chem. elecom.2021.106937.
Soc. 121 (1999) 3214–3217, https://doi.org/10.1021/ja983616l. [46] N. Denisov, S. Qin, G. Cha, J. Yoo, P. Schmuki, Photoelectrochemical properties of
[23] S. Vajda, M.G. White, Catalysis applications of size-selected cluster deposition, ACS “increasingly dark” TiO2 nanotube arrays, J. Electroanal. Chem. 872 (2020),
Catal. 5 (2015) 7152–7176, https://doi.org/10.1021/acscatal.5b01816. 114098, https://doi.org/10.1016/j.jelechem.2020.114098.
[24] R.E. Palmer, R. Cai, J. Vernieres, Synthesis without solvents: the cluster [47] X.H. Yang, Z. Li, C. Sun, H.G. Yang, C. Li, Hydrothermal stability of {001} faceted
(nanoparticle) beam route to catalysts and sensors, Acc. Chem. Res. 51 (2018) anatase TiO2, Chem. Mater 23 (2011) 3486–3494, https://doi.org/10.1021/
2296–2304, https://doi.org/10.1021/acs.accounts.8b00287. cm2008768.
[25] S. Hejazi, S. Mohajernia, B. Osuagwu, G. Zoppellaro, P. Andryskova, O. Tomanec, [48] F. Wang, W. Ge, T. Shen, B. Ye, Z. Fu, Y. Lu, The effect of bulk/surface defects ratio
S. Kment, R. Zbořil, P. Schmuki, On the controlled loading of single platinum atoms change on the photocatalysis of TiO2 nanosheet film, Appl. Surf. Sci. 410 (2017)
as a Co-catalyst on TiO2 anatase for optimized photocatalytic H2 generation, Adv. 513–518, https://doi.org/10.1016/j.apsusc.2017.03.142.
Mater. (2020) 32, https://doi.org/10.1002/adma.201908505. [49] J. Zhang, J. Wang, Z. Zhao, T. Yu, J. Feng, Y. Yuan, Z. Tang, Y. Liu, Z. Li, Z. Zou,
[26] K. Hashimoto, H. Irie, A. Fujishima, TiO2 photocatalysis: a historical overview and Reconstruction of the (001) surface of TiO2 nanosheets induced by the fluorine-
future prospects, Jpn. J. Appl. Phys. 44 (2005) 8269–8285, https://doi.org/ surfactant removal process under UV-irradiation for dye-sensitized solar cells,
10.1143/JJAP.44.8269. Part 1 Regul. Pap. Short Notes Rev. Pap. Phys. Chem. Chem. Phys. 14 (2012) 4763–4769, https://doi.org/10.1039/
[27] T. Luttrell, S. Halpegamage, J. Tao, A. Kramer, E. Sutter, M. Batzill, Why is anatase c2cp24039d.
a better photocatalyst than rutile? - Model studies on epitaxial TiO2 films, Sci. Rep. [50] S. Livraghi, M. Chiesa, M.C. Paganini, E. Giamello, On the nature of reduced states
4 (2015) 4043, https://doi.org/10.1038/srep04043. in titanium dioxide as monitored by electron paramagnetic resonance. I: the
[28] N. Strataki, V. Bekiari, D.I. Kondarides, P. Lianos, Hydrogen production by anatase case, J. Phys. Chem. C 115 (2011) 25413–25421, https://doi.org/
photocatalytic alcohol reforming employing highly efficient nanocrystalline titania 10.1021/jp209075m.
films, Appl. Catal. B Environ. 77 (2007) 184–189, https://doi.org/10.1016/j. [51] A.M. Czoska, S. Livraghi, M. Chiesa, E. Giamello, S. Agnoli, G. Granozzi, E. Finazzi,
apcatb.2007.07.015. C. Di Valentin, G. Pacchioni, The nature of defects in fluorine-doped TiO2, J. Phys.
[29] M. Ni, M.K.H. Leung, D.Y.C. Leung, K. Sumathy, A review and recent developments Chem. C 112 (2008) 8951–8956, https://doi.org/10.1021/jp8004184.
in photocatalytic water-splitting using TiO2 for hydrogen production, Renew. [52] M. Kus, T. Altantzis, S. Vercauteren, I. Caretti, O. Leenaerts, K.J. Batenburg,
Sustain. Energy Rev. 11 (2007) 401–425, https://doi.org/10.1016/j. M. Mertens, V. Meynen, B. Partoens, S. Van Doorslaer, S. Bals, P. Cool, Mechanistic
rser.2005.01.009. insight into the photocatalytic working of fluorinated anatase {001} nanosheets,
[30] W.H. Cheng, M.H. Richter, M.M. May, J. Ohlmann, D. Lackner, F. Dimroth, J. Phys. Chem. C 121 (2017) 26275–26286, https://doi.org/10.1021/acs.
T. Hannappel, H.A. Atwater, H.J. Lewerenz, Monolithic photoelectrochemical jpcc.7b05586.
device for direct water splitting with 19% efficiency, ACS Energy Lett. 3 (2018) [53] A. Naldoni, M. Altomare, G. Zoppellaro, N. Liu, Š. Kment, R. Zbořil, P. Schmuki,
1795–1800, https://doi.org/10.1021/acsenergylett.8b00920. Photocatalysis with reduced TiO2: from black TiO2 to cocatalyst-free hydrogen
[31] J. Xing, J.F. Chen, Y.H. Li, W.T. Yuan, Y. Zhou, L.R. Zheng, H.F. Wang, P. Hu, production, ACS Catal. 9 (2019) 345–364, https://doi.org/10.1021/
Y. Wang, H.J. Zhao, Y. Wang, H.G. Yang, Stable isolated metal atoms as active sites acscatal.8b04068.
for photocatalytic hydrogen evolution, Chem. A Eur. J. 20 (2014) 2138–2144, [54] M.N. Getz, A. Chatzitakis, X. Liu, P.A. Carvalho, T.S. Bjørheim, T. Norby, Voids in
https://doi.org/10.1002/chem.201303366. walls of mesoporous TiO2 anatase nanotubes by controlled formation and
[32] G. Cha, I. Hwang, S. Hejazi, A.S. Dobrota, I.A. Pašti, B. Osuagwu, H. Kim, J. Will, annihilation of protonated titanium vacancies, Mater. Chem. Phys. 239 (2020),
T. Yokosawa, Z. Badura, Š. Kment, S. Mohajernia, A. Mazare, N.V. Skorodumova, 121953, https://doi.org/10.1016/j.matchemphys.2019.121953.
E. Spiecker, P. Schmuki, As a single atom Pd outperforms Pt as the most active co- [55] M. Kottwitz, Y. Li, H. Wang, A.I. Frenkel, R.G. Nuzzo, Single atom catalysts: a
catalyst for photocatalytic H2 evolution, IScience 24 (2021), 102938, https://doi. review of characterization methods, Chem. Methods 1 (2021) 278–294, https://
org/10.1016/j.isci.2021.102938. doi.org/10.1002/cmtd.202100020.
[33] X. Zhou, I. Hwang, O. Tomanec, D. Fehn, A. Mazare, R. Zboril, K. Meyer, [56] J.P. Simonovis, A. Hunt, I. Waluyo, In situ ambient pressure XPS study of Pt/Cu
P. Schmuki, Advanced photocatalysts: pinning single atom Co-catalysts on titania (111) single-atom alloy in catalytically relevant reaction conditions, J. Phys. D
nanotubes, Adv. Funct. Mater. (2021), 2102843, https://doi.org/10.1002/ Appl. Phys. 54 (2021), 194004, https://doi.org/10.1088/1361-6463/abe07f.
adfm.202102843. [57] M.A. Butler, Photoelectrolysis and physical properties of the semiconducting
[34] S. Qin, N. Denisov, J. Will, J. Kolařík, E. Spiecker, P. Schmuki, A few Pt single electrode WO2, J. Appl. Phys. 48 (1977) 1914–1920, https://doi.org/10.1063/
atoms are responsible for the overall Co-catalytic activity in Pt/TiO2 photocatalytic 1.323948.

10
G. Cha et al. Electrochimica Acta 412 (2022) 140129

[58] W.W. Gärtner, Depletion-layer photoeffects in semiconductors, Phys. Rev. 116 [62] S. Kunwar, M. Sui, P. Pandey, Z. Gu, S. Pandit, J. Lee, Improved configuration and
(1959) 84–87, https://doi.org/10.1103/PhysRev.116.84. LSPR response of platinum nanoparticles via enhanced solid state dewetting of in-
[59] E.J. Johnson, Chapter 6 absorption near the fundamental edge, in: 1967: pp. Pt bilayers, Sci. Rep. 9 (2019) 1329, https://doi.org/10.1038/s41598-018-37849-
153–258. 10.1016/S0080-8784(08)60318-X. 0.
[60] N. Denisov, X. Zhou, G. Cha, P. Schmuki, Photocurrent conversion efficiency of [63] A. Campos, N. Troc, E. Cottancin, M. Pellarin, H.C. Weissker, J. Lermé, M. Kociak,
TiO2 nanotube photoanodes in dependence of illumination intensity, Electrochim. M. Hillenkamp, Plasmonic quantum size effects in silver nanoparticles are
Acta 377 (2021), 137988, https://doi.org/10.1016/j.electacta.2021.137988. dominated by interfaces and local environments, Nat. Phys. 15 (2019) 275–280,
[61] N. Zhang, C. Han, Y.J. Xu, J.J. Foley IV, D. Zhang, J. Codrington, S.K. Gray, Y. Sun, https://doi.org/10.1038/s41567-018-0345-z.
Near-field dielectric scattering promotes optical absorption by platinum [64] D.A.H. Hanaor, C.C. Sorrell, Review of the anatase to rutile phase transformation,
nanoparticles, Nat. Photonics 10 (2016) 473–482, https://doi.org/10.1038/ J. Mater. Sci. 46 (2011) 855–874, https://doi.org/10.1007/s10853-010-5113-0.
nphoton.2016.76.

11

You might also like