You are on page 1of 10

www.acsami.

org Research Article

Plasmonic Heterostructure Functionalized with a Carbene-Linked


Molecular Catalyst for Sustainable and Selective Carbon Dioxide
Reduction
Hwiseok Jun, Shinyoung Choi, Joong Bum Lee, and Yoon Sung Nam*
Cite This: ACS Appl. Mater. Interfaces 2020, 12, 33817−33826 Read Online

ACCESS Metrics & More Article Recommendations *


sı Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: Hybridization of homogeneous catalytic sites with a photoelectrode


is an attractive approach to highly selective and tunable photocatalysis using
heterogeneous platforms. However, weak and unclear surface chemistry often leads
Downloaded via UPPSALA UNIV on April 12, 2024 at 10:03:42 (UTC).

to the dissociation and irregular orientation of catalytic centers, restricting long-


term usability with high selectivity. Well-defined and robust ligands that can persist
under harsh photocatalytic conditions are essential for the success of hybrid-type
photocatalysis. Here, we introduce N-heterocyclic carbene as a durable linker for
the immobilization of a Rubpy complex-based CO2 reduction site (cis-dichloro-
(4,4′-diphosphonato-Rubpy)(p-cymene) (RuCY)) on a p-type gallium nitride/
gold nanoparticle (p-GaN/AuNP) heterostructure. The p-GaN/AuNPs/RuCY
photocathode was coupled with a hematite photoanode to drive photoelectrochemical CO2 reduction along with water oxidation.
Highly selective CO2 reduction into formates, up to 98.2%, was achieved utilizing plasmonic hot electrons accumulated on AuNPs.
The turnover frequency was 1.46 min−1 with a faradic efficiency of 96.8% under visible light illumination (243 mW·cm−2). This work
demonstrates that the N-heterocyclic carbene-mediated surface functionalization with homogeneous catalytic sites is a promising
approach to increase the sustainability and usability of hybrid catalysts.
KEYWORDS: CO2 reduction, heterogenization, N-heterocyclic carbenes, photocatalysis, surface plasmon resonance

■ INTRODUCTION
High demand for clean, renewable energy resources has led to
makes it very challenging to optimize the catalytic perform-
ance.19,20 As a means to resolve the limitations, recent studies
an intensive study of efficient and sustainable photocatalysis, in on grafting molecular catalytic sites to a solid photoelectrode,
which sunlight drives the chemical transformation to produce making a hybrid-type photoelectrode, have inspired researchers
value-added products.1 A variety of molecular catalysts (e.g., in the field to harness both the tunable selectivity of molecular
complexes of cobalt,2 iridium,3 palladium, 4 rhenium, 5 catalysts and efficient light-harvesting and charge transfer of
ruthenium,6 etc.) have been studied as highly selective reactive photoelectrodes.21−23 Indeed, introducing a cocatalyst is
species because their catalytic activities are tunable by the expected to be a promising strategy to achieve high selectivity
molecular manipulation of specified single reactive sites.7,8 and catalytic activity because the specific catalytic site plays the
However, homogeneous catalysis possesses demerits of slow roles of (1) improving selectivity, (2) lowering overpotential,
charge transfer rates9,10 and difficulty in the separation of the (3) promoting charge separation, (4) enhancing CO 2
reaction mixtures,11 which hinder their practical applications. interaction, and (5) suppressing undesired reactions.24,25
Consequently, with a conventional experimental setup, their Therefore, the hybrid-type photoelectrode exhibits synergistic
catalytic activity is generally limited to only 15−30 of turnover effects on ameliorations in the catalytic activity by instantly
number for tens of hours.12 Elevated pressure (50−150 bar) or transferring a high flux of photoinduced charge carriers to
excess organic solvents (90%) has been adopted to upgrade the molecular catalysts,26−28 increasing the ease of recovery and
performance.13,14 reusability29 and securing high reaction selectivity toward the
As another class of catalysts, metal−semiconductor-based desired product.30,31 However, there is the Achilles’ heel of the
photoelectrodes have also been used because photoelectrodes
offer strong light absorption throughout the visible and
infrared regimes by band gap engineering15−17 and chemical Received: May 27, 2020
compatibility with a broad spectrum of solutions.18 Despite Accepted: July 8, 2020
such advantages, the inability to control the surface reactivity Published: July 8, 2020
to selective reaction pathways may limit the versatility of
heterogeneous photoelectrodes. The catalytic activity is
determined by the intrinsic properties of materials, which

© 2020 American Chemical Society https://dx.doi.org/10.1021/acsami.0c09517


33817 ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

hybrid-type photoelectrode, which is the labile linkage holes are extracted to p-GaN and electrons are accumulated in
chemistry that cannot persist under harsh photocatalytic AuNPs and then spontaneously transferred to a CO2 reduction
conditions.32 Therefore, robust and sustainable immobilization catalytic site. cis-Dichloro-(4,4′-diphosphonato-Rubpy)(p-cym-
chemistry is essentially required to achieve high performance ene) (RuCY), a molecular catalyst for selective conversion of
of hybrid-type photoelectrodes for solar-to-chemical con- CO2 to formates,42 was directly immobilized onto the surface
version reactions. of AuNPs via NHC. The generated NHC-RuCY monolayer
N-heterocyclic carbene (NHC) ligands can functionalize a could reduce the overpotential43 and surface resistance44 for
metallic gold surface by generating a self-assembled monolayer. efficient hot electron transfer to convert CO2 at a smaller
NHC has almost double binding energy than that of Au ion, reduction potential (−1.0 VRHE) than bare p-GaN/AuNP
∼343 kJ·mol−1, compared to thiol (170 kJ·mol−1), which is photoelectrode (−1.8 VRHE).45 The immobilization of RuCY
widely used as a ligand for metal and semiconductor and passivation of the gold surface via NHC ligands resulted in
nanoparticles.33 The bond strengths on the surface of metallic successful alteration of the major product from CO generated
Au, according to the previous studies,34,35 also imply that the at the surface of AuNPs to the desired product, formates,
NHC moiety (149 kJ·mol−1) is a more plausible species than generated at RuCY. The full PEC reactions exhibited excellent
thiol (126 kJ·mol−1) for stable immobilization of small faradic efficiency (FE) (96.8%) and high selectivity (up to
molecules. The high binding energy between Au and NHC 98.2%).


originates from the chemical properties of the filled sp2 orbital
of CNHC being a strong σ-donor and the empty p orbital of RESULTS AND DISCUSSION
CNHC being an acceptor for π back-donation.36 There have
been examples that manipulate the reactivity of metal catalysts Plasmonic p-GaN/AuNP Heterostructure for Efficient
using NHC as a surface ligand.37−40 However, despite the high Hot Hole Collection. Hot electrons generated from metal−
energy benefits, it is exceptionally rare to use NHC as a linker semiconductor heterostructures have been extensively studied
between two components that play different roles in hybrid- for photocatalytic applications.42,46−50 The removal of hot
type catalysts for intentionally controlled product alteration holes from the metal d-band to p-type semiconductors through
and improved reactivity. The ultrastable and unreactive the heterojunction can affect the relaxation mechanism of hot
properties of NHC against thermal, chemical, and electrical electrons in the metal sp-band, which can lead to the improved
stresses, therefore, led us to consider it as a new linker suitable performances of metal−semiconductor heterostructures.51,52
for the surface functionalization of plasmonic heterostructures On the contrary, only a few studies have been reported on hot
with molecular catalytic sites.40,41 hole collection to drive a catalytic reaction.45,53,54 The driving
Herein, we report the photoelectrochemical (PEC) reduc- force can be greater if hot holes are generated from the d-band
tion of carbon dioxide (CO2) into formates by ruthenium that has a large density of states.52,55,56 The strong oxidation
complexes hybridized with a plasmonic p-GaN/gold nano- power of the holes can drive an oxidation reaction effectively
particle (AuNP) heterojunction photoelectrode (Figure 1). We when transferred to the appropriate catalytic site compared to
the sp-band holes that have a relatively small energy.
Plasmonic gold nanostructures have an interband transition
threshold of ∼1.8 eV, and photons with energies below the
threshold only contribute to intraband transition (excitation at
the sp-band) producing hot electrons to overcome the barrier
height of the Schottky junction for n-type semiconductor.57,58
However, even if photons with energies higher than 1.8 eV are
received, very few hot electrons are transferred to the n-type
semiconductor because the d-band of Au is located more than
2 eV below the Fermi level (Figure S2a).58 However, the
excited d-band holes at the p-GaN/AuNP interface can be
easily transferred to p-GaN due to high energy over the barrier
height (Figure S2b).
As an sp-band-active plasmonic light-harvesting material,
therefore, we prepared a p-GaN/AuNP heterostructure by
depositing atomic Au layers onto an epitaxial p-GaN wafer on
sapphire using electron-beam evaporation (Figure S3a). The
Figure 1. Schematic illustration of the photoelectrochemical α- bare p-GaN substrate is transparent and has no significant
Fe2O3/fluorinated tin oxide (FTO)∥p-GaN/AuNPs/RuCY full absorption in the visible regime (>400 nm) (Figure 2a). The
reaction (a) and immobilization of RuCY on the surface of AuNPs fringes in the absorption spectra originate from Fabry−Pérot
via NHC ligand (b). interference, which is due to the high refractive index of GaN
substrate.59,60 Therefore, interference or undesired effects by
aim to demonstrate that NHC is a promising candidate for the p-GaN substrate are negligible under visible light
surface functionalization to develop sustainable hybrid-type illumination. A broad absorption peak appeared around 700
photocatalysts. The water oxidation reaction is driven at a nm after the deposition of Au layers with a thickness of 1.5 nm.
counter α-Fe2O3 (hematite) electrode consuming hot holes When the p-GaN/Au substrate was annealed at 300 °C, the
that originate from the d-band of plasmonic AuNPs. The p- absorption maximum was blue-shifted to 570 nm, indicating
GaN/AuNP heterojunction generates a Schottky barrier for the formation of plasmonic AuNPs. The scanning electron
hot hole collection. The entire Z-scheme band structure is microscopy (SEM) image of p-GaN/AuNPs exhibits the
described in Figure S1. Due to the shape of the band structure, formation of AuNPs on the surface of p-GaN (Figure 2b), and
33818 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

Photoelectrochemical Analysis of p-GaN/AuNPs/


RuCY Photocathode. We determined the photovoltages of
the prepared heterostructures using a Pt foil as a counter
electrode under AM 1.5 illumination with a long-pass filter
(>400 nm). AuNPs on the p-GaN effectively absorb visible
light over 400 nm and produced hot carriers through the
interband transition. The p-GaN/AuNP heterostructure
exhibited a photovoltage of ∼720 mV, which was much higher
than the photovoltage generated at Au/TiO2 (∼200 mV)42
and over 55-fold higher than bare p-GaN (13 mV) (Figure 3a).
Figure 2. Absorption spectra of photocathodes in the fabrication
process (a) and SEM image of the p-GaN/AuNPs (b).

their average diameter was ∼15.5 nm with 7.84 standard


deviation (Figure S3b).
RuCY Immobilization via N-Heterocyclic Carbene
Chemistry. RuCY is a ruthenium bipyridine complex that
serves as a molecular catalyst to selectively convert CO2 into
formates with high selectivity (∼95%) as reported in our
previous study.42 RuCY follows a proton-involving reaction
mechanism with an active [Ru−H]+ (ruthenium hydride
complex) form. NHC-functionalized RuCY (denoted as Figure 3. Open-circuit voltages of the bare p-GaN (black line (a)), p-
GaN/AuNPs (blue line (a)), and p-GaN/AuNPs/RuCY (red line
“NHC-RuCY”) was prepared as described in the Experimental (b)). Two graphic panels are used because of different relaxation
Section and used for immobilization onto AuNPs via Au− kinetics.
NHC linkages. The terminal groups of bipyridine were
transformed to possess propargyl moieties (−CCH) for The high photovoltage suggests that hot hole collection from
click conjugation with azide groups (−N3) in N,N′-diisopropyl the p-GaN/AuNP heterostructure can more efficiently drive
benzimidazolium carbenes, resulting in a highly stable triazole photocatalytic reactions than hot electron extraction.61,62 After
bridge. Isopropyl groups at carbene-nitrogens are known to be the light was turned off, the photovoltage decreased slowly
an optimal substituent to strengthen Au−CNHC binding with because of efficient carrier separation with a low recombination
no hindrance by bulkiness.43 The prepared NHC-RuCY was rate at the p-GaN/AuNP junction. The photovoltage of p-
self-assembled at the surface of AuNPs on p-GaN, as GaN/AuNPs/RuCY increased linearly in the early stage of
confirmed by X-ray photoelectron spectroscopy (XPS) analysis light illumination (Figure S6). RuCY reduces protons to
for Ru 3d and N 1s scans (Figure S3c). construct a [Ru−H] complex, while the self-consumption of
To obtain evidence for strong and sustainable immobiliza-
hot electrons occurs at the photoelectrode.63 The photovoltage
tion of NHC-RuCY onto AuNPs, we carried out an
was eventually saturated to ∼600 mV (Figure 3b).
electrochemical analysis to examine the surface stabilization
We also examined the redox properties of each photo-
effect of NHC ligand (Figure S4). Cyclic voltammetry was
electrode using cyclic voltammetry. In the CO2-free buffer, p-
performed in the range of −1.2 to +0.45 VRHE, which does not
GaN/AuNPs exhibited a high photocurrent density of around
cause the delamination of AuNPs from the substrate. The
changes of the largest current density at −1.2 VRHE in both 20 μA·cm−2 (Figure S7). However, the immobilization of
bare AuNPs and NHC-RuCY-modified AuNPs were limited to RuCY decreased the photocurrent density to 5 μA·cm−2, which
<10% (from −1.08 to −1.13 μA and from −0.71 to −0.78 μA, might result from the passivation of AuNPs by the self-
respectively) during 30 cycles of analysis. However, the assembled RuCY monolayer in the absence of substances for
cysteamine-modified AuNPs exhibited over 40% of the RuCY. p-GaN/AuNPs and p-GaN/AuNPs/RuCY exhibited
difference (from −0.44 to −0.74 μA). The result indicates different reduction peaks and onset potential values in the
that the NHC-modified surface is much more stable than the CO2-saturated buffer (Figure 4a). A weak reduction peak was
thiol-modified surface. observed at −0.41 VRHE with an onset potential of +0.12 VRHE
Next, we determined whether NHC can replace the thiol for p-GaN/AuNPs/RuCY, while p-GaN/AuNPs exhibited a
ligand of cysteamine prelinked to AuNPs through a conven- peak at −0.87 VRHE with an onset potential of −0.06 VRHE.
tional Au−S bond. The ligand exchange by NHC-RuCY was Interestingly, the photocurrent density of p-GaN/AuNPs at
analyzed using XPS (Figure S5). In the N 1s scan, the −1.2 VRHE significantly decreased from 20 μA·cm−2, in the
population of nitrogen atoms increased because NHC-RuCY CO2-free buffer, to 3 μA·cm−2, in the CO2-saturated buffer,
has seven nitrogen atoms, while cysteamine has only one and p-GaN/AuNPs/RuCY exhibited over 1.5 times larger
nitrogen atom. The total counts of nitrogen atoms did not photocurrent density at −0.4 VRHE than p-GaN/AuNPs
increase seven times because of the larger surface coverage of (Figure 4b). The surface of AuNPs cannot easily convert
benzimidazolium carbene compared to cysteamine. No protons to molecular hydrogen in the presence of CO2 because
significant sulfur 2p peak remained in the NHC-RuCY-treated the molar concentration of protons was only 0.2 mM, while the
sample, indicating the complete elimination of cysteamine. The saturated molar concentration of CO2 was 41 mM in acetate
Ru 3d scan also confirmed that the counts of carbon atoms buffer at pH 3.7. The small reduction peak originated from
increased as the ruthenium signal was generated. The results CO2 conversion into CO at the surface of AuNPs. p-GaN/
support the successful immobilization of NHC-RuCY on the AuNPs produced CO as the major product with two-tenth of
surface of AuNPs. H2 as a side product (Figure 4c). On the other hand, the
33819 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

R is a resistor and CPE is a constant phase element. The entire


Nyquist plots with detailed fitted parameters are described in
Table S1 and Figure S8. Notably, the plots in Figure 5c,d
indicate that both CO2 and NHC-RuCY increased the
interface resistance (Rinter) because CO2 has a higher reduction
potential than hydrogen, and the NHC-RuCY monolayer
blocks the direct contact of electrolyte with the highly
conductive Au surface. Besides, the resistance inside the
electrodes (Rel) increased in the dark when NHC-RuCY was
immobilized (entries 5 and 7 compared to entries 1 and 3,
respectively), which supports the work function-reducing effect
of surface functionalization.43 Moreover, the NHC-RuCY
monolayer also significantly decreased Rinter when p-GaN/
AuNPs/RuCY was exposed to the incident light. In the dark, p-
GaN/AuNPs/RuCY exhibited a very high Rinter of 3225 kΩ
(entry 7) while p-GaN/AuNPs showed only 674 kΩ (entry 3).
In contrast, Rinter of p-GaN/AuNPs/RuCY became even
smaller (237 kΩ, entry 8) than that of p-GaN/AuNPs (288
kΩ, entry 4) under light illumination. Since materials with a
Figure 4. Linear-sweep voltammograms (a) and chronoamperograms small work function generally exhibit a low threshold voltage
at −0.4 VRHE (b) of p-GaN/AuNPs (blue lines) and p-GaN/AuNPs/ and high electron mobility toward the electrolyte solution, the
RuCY (red lines) in CO2-saturated acetate buffer (pH 3.7). The scan NHC-RuCY monolayer on the surface of AuNPs promotes
rate was 50 mV·s−1 in (a). (c) Product analysis of photo- electron transfer through the surface, and, consistently, RuCY
electrochemical CO2 reduction with p-GaN/AuNPs and p-GaN/ immobilized via NHC ligands in this work takes on the role of
AuNPs/RuCY. External bias: −0.4 VRHE; counter electrode: Pt foil;
efficiently delivering electrons to CO2 and suppressing the
light source: solar simulator AM 1.5 with a 400 nm undercut filter.
undesired reaction that occurs directly at the Au surface.
Plasmon-Driven Photoelectrochemical CO2 Reduc-
NHC-RuCY monolayer successfully altered the major product tion and Water Oxidation. The p-GaN/AuNPs/RuCY
from CO to formates even with a higher total amount of photocathode was coupled to a hematite (α-Fe2O3) nanowire
products. The different reactivity might result from the photoanode to conduct PEC full reaction, i.e., CO2 reduction
reduced work function of gold43 and the decreased resistance at the photocathode and water oxidation at the photoanode.
at the electrode/electrolyte interface44 by the self-assembled The hematite nanowires were grown on the surface of FTO-
NHC-RuCY monolayer, promoting electron transfer for the coated glass according to a well-known hydrothermal method
reaction. (Figure S9a),64 and their water oxidation activity was analyzed
To determine whether p-GaN/AuNPs/RuCY is a more before the full reaction. The additional sintering of hematite
effective electron injecting electrode than nonfunctionalized p- nanowires on the FTO substrate for Sn-doping was also
GaN/AuNPs, we employed electrochemical impedance spec- performed at 800 °C for 20 min to promote catalytic activity.64
troscopy (EIS) with an equivalent circuit (Figure 5a,b), where The sintering process increased the current density from 1.7 to
121 μA·cm−2 (Figure S9b) while decreasing the oxidation
potential from +1.7 to +0.9 VRHE (Figure S9c).
The catalytic performance was examined at pH 6 under four
different external biases from −1.45 to −0.4 VRHE (Table 1 and
Figure 6a). The calibration data for product quantification are
presented in Figure S10, and it was confirmed through the
isotope-labeling study using 13CO2 that the source of formate
was carbon dioxide (Figure S11). The formate selectivity
decreased as side reactions, especially, the generation of
methanol that is a six-electron-requiring product increased
when a more negative potential was applied. The PEC full
reaction exhibited the maximum performance at −1.05 VRHE,
which is a faradic efficiency of 96.8% with a formate selectivity
of 91.8%. However, faradic efficiencies significantly decreased
above −1.05 VRHE (62.4% at −0.85 VRHE and 42.1% at −0.4
VRHE) despite very high formate selectivity (92.5 and 98.2%,
respectively). Thus, in the following experiments, the photo-
electrochemical reactions were conducted at −1.05 VRHE to
minimize the loss of electrical energy.
Figure 5. Layouts of p-GaN/AuNPs (a) and p-GaN/AuNPs/RuCY
Chronoamperometry analysis indicates that the photo-
(b) for impedance spectroscopy with their equivalent circuits used to
fit the parameters, where “el” and “inter” indicate the electrode and current depends on the externally added gaseous CO2. The
the electrode/electrolyte interface, respectively. Nyquist plots of p- α-Fe2O3/FTO∥p-GaN/AuNPs/RuCY full reaction exhibited a
GaN/AuNPs and p-GaN/AuNPs/RuCY electrodes in a CO2- photocurrent density of ∼40 μA·cm−2 in the CO2-saturated
saturated solution in the dark or under light illumination (c) and buffer though the photocurrent gradually increased (Figure
under light illumination with or without the dissolved CO2 (d). 6b). In contrast to Figure 4b, where a Pt foil was used as a
33820 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

Table 1. Photoelectrochemical CO2 Reduction with Water Oxidation at Various External Biasesa
entry ext. bias/V (vs RHE) current/μA CO/nmol MeOH/nmol HCOO−/nmol formate selectivity/% FE/%b
1 −0.4 2.36 <10c
<1c
223 98.2 42.3
2 −0.85 8.83 <10c 33 1141 92.5 62.4
3 −1.05 15.11 60 71 3010 91.8 96.8
4 −1.45 19.74 119 181 3633 84.6 97.0
a
Results based on the average values from three independent reaction batches. Reaction conditions: electrolyte = 0.2 M phosphate buffer at pH 6
with CO2 saturation; reaction time = 12 h; and light intensity = 243 mW·cm−2 (Xe lamp, >400 nm). bFaradic efficiency was calculated by taking
the number of electrons required for each product. cConcentrations indistinguishable from noises in the spectrum.

Figure 6. Product selectivity with faradic efficiency at four different external biases for 12 h reactions (a) and photocurrent density (b) of α-Fe2O3/
FTO∥p-GaN/AuNPs/RuCY full reaction. The number of Ru atoms on the surface of p-GaN/AuNP substrates monitored during the reaction (c)
and stable i−t graph during 20 h of the reaction (d) at −1.05 VRHE.

counter electrode, the hematite photoanode exhibited slower on Au surface, simultaneous CO2 reduction and water
kinetics of water oxidation. After 5 min of standing by, the oxidation reaction were successfully conducted for at least 20
steady-state photocurrent density was 61 μA·cm−2 (Figure h without a decrease of the catalytic activity (Figure 6d).
S12). Once the light was turned off, interestingly, the We also confirmed the long-term stable catalytic activity by
photocurrent decayed in multistep kinetics exhibiting a 5−6 quantitatively analyzing CO2-derived products and O2 evolved
min delayed complete recovery to zero since shallow traps and at the photocathode and photoanode, respectively. The
deep traps of the photoelectrode can hold the hot carriers for product generated at the photocathode was mainly formates,
up to thousands of seconds.65 The delayed current decay alongside small proportions (<3%) of methanol and CO
occurred only in the p-GaN/AuNPs/RuCY electrode (Figure (Figure 7a). According to the calculation based on the number
S12). of RuCY (∼3.36 nmol per a photocathode), the hybrid catalyst
Next, we examined the robustness of the NHC ligand and exhibited a high turnover number of 1757 for 20 h, which
the long-term stability of the hybrid photocathode. The corresponds to a turnover frequency of 1.46 min−1 to produce
number of RuCY on the electrode was determined using formates, while a similar photoelectrochemical configuration in
inductively coupled plasma mass spectrometry (ICP-MS)
(Figure 6c). The average amount of the immobilized NHC-
RuCY was 3.36 nmol, and no significant changes were
observed while thiol-linked RuCY molecules were detached
gradually, and physisorbed RuCY molecules were eliminated
within 2 h (Figure S13). The hollow point of HS-RuCY at 20 h
indicates the detached amount of Ru in the reaction solution.
The stable binding property of the NHC ligand on gold or
transition metal surfaces comes from the unordinary electronic
structure that allows the carbon atom to be divalent with a
highly directional sp2-hybridization,66 while thiol-based mono-
layers are prone to degrade because of their air and thermal
Figure 7. Photoelectrochemical CO2 reduction at p-GaN/AuNPs/
instability. The compared XPS survey scans also support the RuCY (a) and water oxidation at α-Fe2O3/FTO (b) in the CO2-
chemical consistency before and after the reaction with no saturated buffer (0.2 M phosphate, pH 6) under visible light
significant difference in the ratio of Au, C, N, and Ru signals illumination (>400 nm). External bias was −1.05 VRHE, and the
(Figure S14). Due to the strong affinity and stability of NHC loaded amount of RuCY was 3.36 nmol.

33821 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

the most recent study using polymeric molecular catalytic site 4,4′-dicarboxyl-2,2′-bipyridyl (2 mmol, 1 equiv) were dissolved in 10
exhibited an overall turnover number of 25 in 5 h with a faradic mL of dichloromethane, and three drops of dimethylformamide
efficiency of 85%.67 As the counter electrode, the α-Fe2O3/ (DMF) was added to the solution. After 30 min, the solvent and
FTO photoanode produced gaseous O2 by half the amount of volatile components were removed by a rotary evaporator. Then, the
crude solid was redissolved in dichloromethane, and propargylamine
CO2 conversion products (Figure 7b). Because the faradic (2.2 equiv) and diisopropylethylamine (DIPEA, 3 equiv) were added
efficiency of this reaction condition was 96.8%, the 2:1 ratio of to the solution. The reaction mixture was stirred for over 12 h. After
formate to oxygen was reasonable.


the designated time of reaction, the mixture solution was quickly
filtered through a short silica gel column by negative pressure to
CONCLUSIONS eliminate unreacted propargylamine, DIPEA, and DIPEA−H+ salt.
The residual solution was concentrated under reduced pressure and
We presented the first example of the hybrid photocathode
dissolved in acetonitrile. Dichloro-(p-cymene)ruthenium(II) dimer
design using carbene-based functionalization of a metal− (2.1 equiv) was suspended in the solution. The heterogeneous
semiconductor heterostructure with molecular catalytic sites. suspension was bubbled by Ar for 20 min, and then, the mixture was
This work demonstrated that the p-GaN/AuNPs/RuCY refluxed for 4 h. After the designated time, the pot was placed in a
hybrid photocathode can efficiently collect plasmon-induced refrigerator at 4 °C for over 6−12 h to precipitate enough of the
hot holes via the Schottky barrier junction of p-GaN and suspended brown product. The precipitate was filtered and washed
AuNPs and drive highly selective CO2 reduction for formate three times with cold acetonitrile. The obtained solid was dried in a
production with simultaneous water oxidation by a hematite vacuum oven at 40 °C for over 6 h to give cis-dichloro-(4,4′-bis-
photoanode under visible light illumination. The successful propargylamido-Rubpy)(p-cymene).
immobilization of RuCY to the plasmonic photoelectrode was
1
H NMR (CDCl3, 300 MHz): δ 10.54 (d, 2H), 9.87 (t, 2H), 9.24
(d, 2H), 8.29 (m, 2H), 5.82 (d, 2H), 5.67 (d, 2H), 4.30 (s, 4H), 2.62
achieved using strong and stable NHC chemistry. Due to the
(m, 1H), 2.32 (s, 3H), 2.23 (s, 2H), 1.05 (d, 6H) (Figure S16).
work function lowering effect of AuNPs by the NHC Synthesis of cis-Dichloro-(4,4′-bis-(diisopropyl benzimida-
monolayer and resistance-reducing effect of the interface by zolium-triazolyl)amido-Rubpy)(p-cymene) (NHC-RuCY Precur-
redox-active surface functionalization, ∼97% of the overall sor) (Scheme S3). The azide-terminated carbene precursor (100 mg,
faradic efficiency was obtained at the external bias of −1.05 V 0.27 mmol) and propargyl RuCY (84 mg, 0.14 mmol) were dissolved
(vs reversible hydrogen electrode (RHE)). The NHC ligand in tetrahydrofuran (THF) (6 mL) and water (6 mL). CuSO4 (4% w/v
provided a long-term stable surface layer to maintain the aqueous solution, 1.84 mL, 0.294 mmol) was added followed by the
catalytic activity over 20 h. We believe that the high selectivity dropwise addition of a freshly prepared sodium ascorbate solution (1
M aqueous solution, 0.59 mL). The reaction mixture was stirred at
and stability of the carbene-linked hybrid electrode can lead to room temperature for 2 h. After removal of THF under vacuum,
advances in the field of solar-to-chemical conversion and chloroform (10 mL) and conc. NH3 (3 mL) were added to the
artificial photosynthesis.


solution, which was allowed to stir for an additional 30 min at room
temperature to quench all of the Cu(I) by complexation of
EXPERIMENTAL SECTION [Cu(NH3)6]. The aqueous phase was discarded and the organic
Synthesis of 5-Azido-1,3-diisopropylbenzimidazolium Io- phase was washed twice with water and once with brine and then
dide (Scheme S1). A mixture of 6-bromo-1H-benzimidazole (2.54 dried with anhydrous MgSO4. The product was obtained after the
mmol), sodium azide (5.08 mmol), sodium ascorbate (0.13 mmol), filtration of MgSO4 and the following solvent evaporation. The
copper iodide (0.25 mmol), and ethylenediamine ligand (0.38 mmol) residual solid was redissolved in DMSO (30 mg·mL−1), and 50 mL of
was placed in a 50 mL round-bottomed flask containing 10 mL of acetonitrile was slowly added to the solution using a dropping funnel
ethanol−water (7:3) solution. The reaction mixture was degassed and for over 1 h. The precipitate was washed with acetonitrile two more
purged with argon gas, and then stirred under a reflux condition for 3 times and dried in a vacuum oven at 40 °C for over 6 h.
h. After the designated time, the reaction mixture was allowed to cool
1
H NMR (DMSO-d6, 300 MHz): δ 9.38 (s, 2H), 8.32 (s, 2H), 7.88
to room temperature, and the crude mixture was purified by (d, 2H), 7.74 (d, 2H), 7.27 (s, 2H), 7.00 (d, 2H), 6.95 (s, 2H), 6.94
extraction with ethyl acetate and short flash column chromatography (d, 2H), 6.17 (m, 2H), 6.05 (m, 2H), 4.88 (m, 4H), 4.20 (d, 4H),
to remove unreacted residue and impurities. Then, 6-azido-1H- 2.73 (m, 1H), 2.27 (s, 3H), 1.58 (m, 12 H), 1.22 (d, 6H), 1.15 (m,
benzimidazole (1.57 mmol) and cesium carbonate (2.35 mmol) were 12H). High-resolution mass spectrometry (HR-MS) (electrospray
added to a 100 mL round-bottomed flask. Acetonitrile (25 mL) was ionization (ESI), m/z, [M − 2I]2+): calcd molecular weight (M.W.)
added to the flask followed by the addition of 2-iodopropane (4.71 556.1873, found M.W. 556.1844 (Figure S17).
mmol). The reaction mixture was heated to a reflux condition at 85− After the preparation of NHC-RuCY precursors, the active NHC-
90 °C and stirred overnight. The volatile chemicals were removed in RuCY was obtained by dissolving in THF, followed by slow addition
vacuo, and the residual solid was then extracted three times from the of 1.1 equiv potassium tert-butoxide (t-BuOK) in an ice bath. The
ethyl acetate/water mixed solution. The combined organic layer was reaction was stirred for 30 min, and then, THF was evaporated under
then dried with anhydrous magnesium sulfate and then filtered over a reduced pressure. The resulting solid was redissolved in DMSO and
glass filter. The filtrate was concentrated under reduced pressure, and used for surface modification. The solution (5 mg·mL−1 in DMSO) of
the resultant viscous compound was added to a thick-walled sealed activated NHC-RuCY was placed in a 24-well plate carrying the p-
tube. Neat 2-iodopropane (2 mL, 20 mmol) was then added to the GaN/AuNPs substrates. Then, the molecular catalyst was immobi-
flask. The flask was tightly sealed with a Teflon screw cap and stirred lized onto the surface of AuNPs during shaking incubation at room
at 90−100 °C for 96 h. After the designated reaction time, the flask temperature for over 12 h.
was cooled to room temperature and the volatile 2-iodopropane was Synthesis of 4,4′-Dithiazolidine-2,2′-bipyridine (Scheme
removed in vacuo. Acetonitrile was added to dissolve the solid S4). A suspension of 4,4′-dicarboxaldehyde-2,2′-bipyridine (500 mg,
completely. Around 30 mL of diethyl ether was then added to the 2.356 mmol) in 30 mL of absolute ethanol was slightly heated to
solution and stirred until precipitation ceased. The precipitate was obtain a clear solution. Another 10 mL of ethanol solution of
triturated by diethyl ether to give the product for the next step. cysteamine hydrochloride (802 mg, 7.068 mmol) was neutralized by
1
H NMR (dimethyl sulfoxide (DMSO)-d6, 300 MHz): δ 9.40 (s, addition of trimethylamine (1.03 mL, 7.421 mmol) and was then
1H), 7.77 (d, 1H), 6.99 (d, 1H), 6.93 (s, 1H), 4.90 (m, 2H), 1.58 (d, added to the bipyridine solution. The mixture solution was refluxed
6H), 1.17 (d, 6H) (Figure S15). for 6 h and cooled to room temperature. After the cooled mixture was
Synthesis of cis-Dichloro-(4,4′-bis-propargylamido-Rubpy)- filtered to remove undissolved substances, the solvent was removed by
(p-cymene) (Scheme S2). Oxalyl chloride (6 mmol, 3 equiv) and a rotary evaporator. The residual solid was redissolved in 50 mL of a

33822 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

1:1 mixed solution of water and dichloromethane, and the organic 2 h. During the heat treatment, the FeOOH nanowires were
layer was extracted three times. The collected organic solution was converted into hematite nanowires without deterioration of the
dried by anhydrous magnesium sulfate, filtered, and concentrated by a nanowire structures. To further reduce the defects of the hematite
rotary evaporator. Fifty milliliters of a 1:1 mixed solution of ethanol nanowires and Sn-doping, the second sintering process64 was carried
and diethyl ether was added to the residue, and the mixture was out at 800 °C for an additional 20 min.
heated until the solution become clear. Standing still for 1 h at room Photoelectrochemical Analysis of the Half-Reaction and
temperature for slow cooling gave a pale-yellow precipitate, and the the Full Reaction at p-GaN/AuNPs/RuCY Photocathode. The
solid was washed with cold diethyl ether three times. half-reaction and the full reaction were conducted to analyze
1
H NMR (CDCl3, 300 MHz): δ 8.64 (d, 2H), 8.50 (s, 2H), 7.41 electrochemical behaviors of p-GaN/AuNPs/RuCY in three-electrode
(d, 2H), 5.66 (s, 2H), 3.48 (m, 2H), 3.28 (m, 2H), 3.11 (m, 4H) configurations with Ag/AgCl electrode as the reference electrode in
(Figure S18). 0.2 M acetate buffer (pH 3.7) and 0.2 M phosphate buffer (pH 6),
Synthesis of 4,4′-Bis(methylaminoethanethiol)Rubpy(p- respectively. Pt foil was used as a counter electrode for the half-
cymene) (Thiol-RuCY) (Scheme S5). The resulting powder in reaction, and α-Fe2O3/FTO photoanode was used for the PEC full
Scheme S4 (100 mg, 0.303 mmol) and dichloro(p-cymene)- reaction. The photocurrent was obtained under 1 sun AM 1.5
ruthenium(II) dimer (92.6 mg, 0.151 mmol) were suspended in 30 illumination (>400 nm) using a solar simulator (Asahi Spectra, HAL-
mL of pure acetonitrile, and the heterogeneous suspension was 320). The electronic potentials were converted into the reversible
bubbled by Ar gas for 20 min. The mixture was refluxed for 4 h, and hydrogen electrode (RHE) scale using the following equation: E(RHE)
the suspension became dark during the reaction. After the designated = E(Ag/AgCl) + (0.059 × pH) + E(Ag/AgCl)0, where E(Ag/AgCl)0 = 0.197 V
time, the reaction pot was placed in a refrigerator overnight to at 25 °C.
precipitate the brown solid. The resulting solid was washed with cold Formate Production by the Plasmon-Driven Photoelec-
diethyl ether three times and dried in a vacuum oven at 40 °C for 6 h trochemical CO 2 Reduction with Simultaneous Water
to give dithiazolidine-Rubpy(p-cymene). Then, the thiazolidine- Oxidation. The photoelectrochemical CO2 reduction and water
terminated RuCY was dissolved in 30 mL of 0.5 M K2CO3 solution. oxidation reactions were conducted in a three-electrode system to
Sodium borohydride (90.8 mg, 2.4 mmol) in 3 mL of 0.25 M NaOH elucidate the catalytic activity of p-GaN/AuNPs/RuCY at a certain
solution was then added to the mixture. The solution was stirred for 2 electric potential. The aqueous solution was a 0.2 M KH2PO4/
h while maintaining the temperature under 20 °C and cooled to 0−4 Na2HPO4 buffer, and the pH of the solution was adjusted to 6.0
°C in an ice bath. Glacial acetic acid was added to the reaction because RuCY significantly loses its reactivity with carbon dioxide
mixture until the final pH become 5. After a quick filtration through a above this pH, as described previously.42 The electrolyte solutions
3 cm layer of celite to remove residual salt, the mixture was poured were degassed under reduced pressure and N2 bubbling, and then, the
into 50 mL of tetrahydrofuran, and a dark brown solid was solution was saturated with CO2 by bubbling at a slightly positive
precipitated. The supernatant solution was decanted, and the pressure for 1 h. All three electrodes, p-GaN/AuNPs/RuCY, α-
precipitate was resuspended and triturated using diethyl ether to Fe2O3/FTO, and Ag/AgCl, were put in a 10 mL vial containing 4 mL
give the thiol-terminated RuCY. of the electrolyte solution, and the gas phase was purged with Ar,
1
H NMR (DMSO-d6, 300 MHz): δ 8.62 (d, 2H), 8.44 (s, 2H), while the vial was tightly sealed with a rubber septum and paraffin
7.46 (d, 2H), 6.31 (d, 2H), 6.07 (d, 2H), 5.70 (d, 4H), 3.58 (m, 2H), seal-tape. The vial was placed in front of a xenon lamp (XLS300-OFR,
3.20 (m, 4H), 2.98 (m, 4H), 2.59 (m, 1H), 2.16 (s, 3H), 0.95 (d, SPECTRO), and the light was irradiated for designated periods of
6H). HR-MS (ESI, m/z, [M − Cl]+): calcd M.W. 605.1113, found time with a 400 nm long-pass filter. Formate ions were protonated by
M.W. 605.1091 (Figure S19). addition of sulfuric acid (10 vol % 1.0 M H2SO4) and extracted as
Preparation of the p-GaN/AuNP Photoelectrodes. The formic acid by ethyl acetate three times.69 Chemical species in the
known procedure in the literature45 was followed with slight aqueous and gas phases were determined, respectively, by gas
modification for the actual experimental environment. The plasmonic chromatography (GC) (YL6500 GC) (aqueous phase: detector =
p-GaN/Au photoelectrode was prepared via electron-beam evapo- pulsed discharge ionization detector (PDD); carrier gas = helium;
ration of the atomic Au layer onto commercial p-type gallium nitride flow rate = 1.0 mL·min−1; temperature conditions = 40 °C for 3 min,
substrates epitaxially grown on sapphire (Pam-Xiamen, c-axis 0001). then elevated to 200 °C at a rate of 30 °C·min−1; column for PDD =
First, the p-GaN substrates were thoroughly cleaned with 0.02% (v/v) Agilent Technologies, HP-FFAP, 30 m × 0.53 mm, 1.00 μm. Gas
of NH4OH solution and copious amounts of deionized water. Then, phase: detector = thermal conductivity detector (TCD); carrier gas =
the substrates were dried in an oven at 60 °C for 1 h. A film of Au argon; flow rate = 10.0 mL·min−1; temperature condition = isocratic
with 1.5 nm thickness was deposited on the p-GaN surface by physical 90 °C; column for TCD = SUPELCO Analytical, Mol Sieve 5A, 6 ft ×
vapor deposition under a pressure of 1 × 10−7 Torr. The Au film was 1/8 in.).
then annealed in an ambient condition at 300 °C (heating rate: 10 Isotope-Labeled Study for Photoelectrochemical Formate
°C·min−1) for 1 h to construct the spherical geometry of gold Production. Two phosphate buffer solutions were bubbled with
12
nanoparticles, as well as to ensure the stable adhesion between gold CO2 and 13CO2, respectively. The 13CO2 gas was generated by
and p-GaN interface. addition of 1.0 M sulfuric acid into a Ba13CO3-containing flask.14
Hematite (α-Fe2O3) Nanowires on FTO Substrate as a After 20 h of light illumination, the formate ions were protonated and
Photoanode. The α-Fe2O3/FTO photoanode was prepared extracted by ethyl acetate. Then, gas chromatography-mass spectrom-
according to a hydrothermal method as previously reported68 with etry (GC-MS) analysis (EI mode, ISQ QD300, Thermo Fisher
slight modification. Fluorinated tin oxide (FTO) substrates were Scientific) was conducted to determine whether the formic acid was
washed three times in an ultrasonicator with a 1:1:1 mixture solution labeled with 13C.
of deionized water, 2-propanol, and acetone. After rinsing the Calculation of Faradic Efficiency. The faradic efficiency (FE) of
substrates with copious amounts of deionized water, they are dried in the photoelectrochemical CO2 reduction for each product was
air at 100 °C for 1 h. A 50 mL poly(tetrafluoroethylene) (PTFE) liner calculated as follows
was filled with 25 mL of an aqueous solution containing ferric
chloride (FeCl3, 0.15 M), sodium nitrate (NaNO3, 1 M), and 131 μL FE (%) = 100 × (Ni × F × Pm)/(I × t )
of hydrochloric acid (37 wt % HCl). The washed FTO substrates
were put into the PTFE liner leaning against the wall with the where Ni is the number of electrons required to produce a particular
conducting face down. The autoclave was placed in an oven at a product, F is the Faraday constant (96 485 C·mol−1, 1 C = 1 A·s), Pm
temperature of 100 °C and taken out after 4 h. After the autoclave was is the determined molarity of the product, I is the average current
cooled to room temperature, the substrates covered with uniform iron during the reaction time, and t is the reaction time (12 h = 43 200 s).
oxyhydroxide (FeOOH) nanowires were rinsed with deionized water The number of electrons involved with the reduction reactions is
and subsequently annealed at 550 °C (heating rate: 10 °C·min−1) for two for H2, CO, and formate and six for methanol. The molarity of

33823 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

the products was determined by calibrating the GC using a serial Rhenium Bipyridine Complex. J. Am. Chem. Soc. 2014, 136, 6021−
dilution of each standard reagent in known concentrations. 6030.

■ ASSOCIATED CONTENT
* Supporting Information

(6) Kuramochi, Y.; Fukaya, K.; Yoshida, M.; Ishida, H. trans-(Cl)-
[Ru(5,5′-diamide-2,2′-bipyridine)(CO)2Cl2]: Synthesis, Structure,
and Photocatalytic CO2 Reduction Activity. Chem. − Eur. J. 2015,
21, 10049−10060.
The Supporting Information is available free of charge at (7) Zecchina, A.; Bordiga, S.; Groppo, E. The Structure and
https://pubs.acs.org/doi/10.1021/acsami.0c09517. Reactivity of Single and Multiple Sites on Heterogeneous and
Detailed methods, characterizations, and additional data Homogeneous Catalysts: Analogies, Differences, and Challenges for
of solar-to-chemical conversion (PDF) Characterization Methods. Selective Nanocatalysts and Nanoscience;


John Wiley & Sons, Ltd., 2011; pp 1−27.
(8) Roy, N.; Suzuki, N.; Terashima, C.; Fujishima, A. Recent
AUTHOR INFORMATION Improvements in the Production of Solar Fuels: From CO2 Reduction
Corresponding Author to Water Splitting and Artificial Photosynthesis. Bull. Chem. Soc. Jpn.
Yoon Sung Nam − Department of Materials Science and 2019, 92, 178−192.
Engineering and KAIST Institute for NanoCentury, Korea (9) Zhao, G.; Huang, X.; Wang, X.; Wang, X. Progress in Catalyst
Advanced Institute of Science and Technology, Daejeon 34141, Exploration for Heterogeneous CO2 Reduction and Utilization: A
Republic of Korea; orcid.org/0000-0002-7302-6928; Critical Review. J. Mater. Chem. A 2017, 5, 21625−21649.
(10) Francke, R.; Schille, B.; Roemelt, M. Homogeneously Catalyzed
Email: yoonsung@kaist.ac.kr Electroreduction of Carbon DioxideMethods, Mechanisms, and
Authors Catalysts. Chem. Rev. 2018, 118, 4631−4701.
(11) Cole-Hamilton, D. J. Homogeneous Catalysis–New Ap-
Hwiseok Jun − Department of Materials Science and
proaches to Catalyst Separation, Recovery, and Recycling. Science
Engineering, Korea Advanced Institute of Science and 2003, 299, 1702−1706.
Technology, Daejeon 34141, Republic of Korea; orcid.org/ (12) Takeda, H.; Koike, K.; Inoue, H.; Ishitani, O. Development of
0000-0002-1266-001X an Efficient Photocatalytic System for CO2 Reduction Using
Shinyoung Choi − Department of Materials Science and Rhenium(I) Complexes Based on Mechanistic Studies. J. Am. Chem.
Engineering, Korea Advanced Institute of Science and Soc. 2008, 130, 2023−2031.
Technology, Daejeon 34141, Republic of Korea; orcid.org/ (13) Méndez, M. A.; Voyame, P.; Girault, H. H. Interfacial
0000-0001-6220-1180 Photoreduction of Supercritical CO2 by an Aqueous Catalyst.
Joong Bum Lee − Department of Materials Science and Angew. Chem., Int. Ed. 2011, 50, 7391−7394.
Engineering, Korea Advanced Institute of Science and (14) Kuramochi, Y.; Kamiya, M.; Ishida, H. Photocatalytic CO2
Technology, Daejeon 34141, Republic of Korea; orcid.org/ Reduction in N,N-Dimethylacetamide/Water as an Alternative
0000-0001-7751-1720 Solvent System. Inorg. Chem. 2014, 53, 3326−3332.
(15) Swearer, D. F.; Zhao, H.; Zhou, L.; Zhang, C.; Robatjazi, H.;
Complete contact information is available at: Martirez, J. M. P.; Krauter, C. M.; Yazdi, S.; McClain, M. J.; Ringe, E.;
https://pubs.acs.org/10.1021/acsami.0c09517 Carter, E. A.; Nordlander, P.; Halas, N. J. Heterometallic Antenna−
Reactor Complexes for Photocatalysis. Proc. Nat. Acad. Sci. U.S.A.
Notes 2016, 113, 8916−8920.
The authors declare no competing financial interest. (16) Shi, X.; Ueno, K.; Oshikiri, T.; Sun, Q.; Sasaki, K.; Misawa, H.

■ ACKNOWLEDGMENTS
This work was supported by the Basic Science Research
Enhanced Water Splitting Under Modal Strong Coupling Conditions.
Nat. Nanotechnol. 2018, 13, 953−958.
(17) Park, J.; Jun Kim, H.; Nam, S.; Kim, H.; Choi, H.-J.; Jang, Y. J.;
Sung Lee, J.; Shin, J.; Lee, H.; Baik, J. M. Two-Dimensional Metal-
Program and Nano·Material Technology Development Pro- Dielectric Hybrid-Structured Film With Titanium Oxide for
gram through the National Research Foundation of Korea Enhanced Visible Light Absorption and Photo-catalytic Application.
(NRF) funded by the Ministry of Science and ICT (NRF- Nano Energy 2016, 21, 115−122.
2020R1A2C2004168, NRF-2017M3A7B4052797, and NRF- (18) Hara, K.; Kobayashi, H.; Komanoya, T.; Huang, S. J.; Pruski,
2017M3A7B4042235). M.; Fukuoka, A. Supported Metal Catalysts for Green Reactions.

■ REFERENCES
(1) Enthaler, S.; von Langermann, J.; Schmidt, T. Carbon Dioxide
Heterogeneous Catalysis for Today’s Challenges: Synthesis, Character-
ization and Applications; Royal Society of Chemistry, 2015; Chapter 3,
pp 61−76.
(19) Samsudin, E. M.; Hamid, S. B. A.; Juan, J. C.; Basirun, W. J.;
and Formic acidThe Couple for Environmental-friendly Hydrogen
Storage? Energy Environ. Sci. 2010, 3, 1207−1217. Centi, G. Enhancement of the Intrinsic Photocatalytic Activity of
(2) Lee, G. Y.; Kim, I.; Lim, J.; Yang, M. Y.; Choi, D. S.; Gu, Y.; Oh, TiO2 in the Degradation of 1,3,5-triazine Herbicides by Doping With
Y.; Kang, S. H.; Nam, Y. S.; Kim, S. O. Spontaneous Linker-free N,F. Chem. Eng. J. 2015, 280, 330−343.
Binding of Polyoxometalates on Nitrogen-doped Carbon Nanotubes (20) Hoque, M. A.; Guzman, M. I. Photocatalytic Activity:
for Efficient Water Oxidation. J. Mater. Chem. A 2017, 5, 1941−1947. Experimental Features to Report in Heterogeneous Photocatalysis.
(3) Kang, P.; Meyer, T. J.; Brookhart, M. Selective Electrocatalytic Materials 2018, 11, No. 1990.
Reduction of Carbon Dioxide to Formate by a Water-soluble Iridium (21) Sekizawa, K.; Maeda, K.; Domen, K.; Koike, K.; Ishitani, O.
Pincer Catalyst. Chem. Sci. 2013, 4, 3497−3502. Artificial Z-Scheme Constructed with a Supramolecular Metal
(4) DuBois, D. L.; Miedaner, A.; Haltiwanger, R. C. Electrochemical Complex and Semiconductor for the Photocatalytic Reduction of
Reduction of Carbon Dioxide Catalyzed by [Pd(triphosphine)- CO2. J. Am. Chem. Soc. 2013, 135, 4596−4599.
(solvent)](BF4)2 Complexes: Synthetic and Mechanistic Studies. J. (22) Kuriki, R.; Sekizawa, K.; Ishitani, O.; Maeda, K. Visible-Light-
Am. Chem. Soc. 1991, 113, 8753−8764. Driven CO2 Reduction with Carbon Nitride: Enhancing the Activity
(5) Kou, Y.; Nabetani, Y.; Masui, D.; Shimada, T.; Takagi, S.; of Ruthenium Catalysts. Angew. Chem., Int. Ed. 2015, 54, 2406−2409.
Tachibana, H.; Inoue, H. Direct Detection of Key Reaction (23) Sekizawa, K.; Sato, S.; Arai, T.; Morikawa, T. Solar-Driven
Intermediates in Photochemical CO2 Reduction Sensitized by a Photocatalytic CO2 Reduction in Water Utilizing a Ruthenium

33824 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

Complex Catalyst on p-Type Fe2O3 with a Multiheterojunction. ACS (40) Crudden, C. M.; Horton, J. H.; Ebralidze, I. I.; Zenkina, O. V.;
Catal. 2018, 8, 1405−1416. McLean, A. B.; Drevniok, B.; She, Z.; Kraatz, H.-B.; Mosey, N. J.;
(24) Li, X.; Yu, J.; Jaroniec, M.; Chen, X. Cocatalytsts for Selective Seki, T.; Keske, E. C.; Leake, J. D.; Rousina-Webb, A.; Wu, G. Ultra
Photoreduction of CO2 into Solar Fuels. Chem. Rev. 2019, 119, Stable Self-assembled Monolayers of N-Heterocyclic Carbenes on
3962−4179. Gold. Nat. Chem. 2014, 6, 409−414.
(25) Chang, X.; Wang, T.; Yang, P.; Zhang, G.; Gong, J. The (41) Lee, J. B.; Choi, S.; Kim, J.; Nam, Y. S. Plasmonically-Assisted
Development of Cocatalysts for Photoelectrochemical CO2 Reduc- Nanoarchitectures for Solar Water Splitting: Obstacles and Break-
tion. Adv. Mater. 2018, 31, No. 1804710. throughs. Nano Today 2017, 16, 61−81.
(26) Dubois, K. D.; Li, G. Innovative Photocatalysts for Solar Fuel (42) Jun, H.; Choi, S.; Yang, M. Y.; Nam, Y. S. A Ruthenium-Based
Generation by CO2 Reduction. In New and Future Developments in Plasmonic Hybrid Photocatalyst for Aqueous Carbon Dioxide
Catalysis; Suib, S. L., Ed.; Elsevier: Amsterdam, 2013; Chapter 9, pp Conversion With a High Reaction Rate and Selectivity. J. Mater.
219−241. Chem. A 2019, 7, 17254−17260.
(27) De Vos, A.; Lejaeghere, K.; Muniz Miranda, F.; Stevens, C. V.; (43) Kim, H. K.; Hyla, A. S.; Winget, P.; Li, H.; Wyss, C. M.; Jordan,
Van Der Voort, P.; Van Speybroeck, V. Electronic Properties of A. J.; Larrain, F. A.; Sadighi, J. P.; Fuentes-Hernandez, C.; Kippelen,
Heterogenized Ru(ii) Polypyridyl Photoredox Complexes on B.; Brédas, J.-L.; Barlow, S.; Marder, S. R. Reduction of the Work
Covalent Triazine Frameworks. J. Mater. Chem. A 2019, 7, 8433− Function of Gold by N-Heterocyclic Carbenes. Chem. Mater. 2017,
8442. 29, 3403−3411.
(28) Luo, Y.-C.; Chu, K.-L.; Shi, J.-Y.; Wu, D.-J.; Wang, X.-D.; (44) Chae, S. Y.; Choi, J. Y.; Kim, Y.; Le Tri Nguyen, D.; Joo, O.-S.
Mayor, M.; Su, C.-Y. Heterogenization of Photochemical Molecular Photoelectrochemical CO2 Reduction with a Rhenium Organo-
Devices: Embedding a Metal−Organic Cage into a ZIF-8-Derived metallic Redox Mediator at Semiconductor/Aqueous Liquid Junction
Matrix To Promote Proton and Electron Transfer. J. Am. Chem. Soc. Interfaces. Angew. Chem., Int. Ed. 2019, 58, 16395−16399.
2019, 141, 13057−13065. (45) DuChene, J. S.; Tagliabue, G.; Welch, A. J.; Cheng, W.-H.;
(29) Collis, A. E. C.; Horváth, I. T. Heterogenization of Atwater, H. A. Hot Hole Collection and Photoelectrochemical CO2
Homogeneous Catalytic Systems. Catal. Sci. Technol. 2011, 1, 912− Reduction with Plasmonic Au/p-GaN Photocathodes. Nano Lett.
919. 2018, 18, 2545−2550.
(30) Zhang, B.; Sun, L. Artificial Photosynthesis: Opportunities and (46) Li, W.; Valentine, J. Metamaterial Perfect Absorber Based Hot
Challenges of Molecular Catalysts. Chem. Soc. Rev. 2019, 48, 2216− Electron Photodetection. Nano Lett. 2014, 14, 3510−3514.
2264. (47) Li, W.; Coppens, Z. J.; Besteiro, L. V.; Wang, W.; Govorov, A.
(31) Sheehan, S. W.; Thomsen, J. M.; Hintermair, U.; Crabtree, R.
O.; Valentine, J. Circularly Polarized Light Detection With Hot
H.; Brudvig, G. W.; Schmuttenmaer, C. A. A Molecular Catalyst for
Electrons in Chiral Plasmonic Metamaterials. Nat. Commun. 2015, 6,
Water Oxidation That Binds to Metal Oxide Surfaces. Nat. Commun.
No. 8379.
2015, 6, No. 6469.
(48) Choi, S.; Nam, Y. S. Gold−Titanium Dioxide Half-Dome
(32) Youngblood, W. J.; Lee, S.-H. A.; Kobayashi, Y.; Hernandez-
Heterostructures for Plasmonic Hydrogen Evolution. ACS Appl.
Pagan, E. A.; Hoertz, P. G.; Moore, T. A.; Moore, A. L.; Gust, D.;
Energy Mater. 2018, 1, 5169−5175.
Mallouk, T. E. Photoassisted Overall Water Splitting in a Visible
(49) Kim, J.; Son, H. Y.; Nam, Y. S. Multilayered Plasmonic
Light-Absorbing Dye-Sensitized Photoelectrochemical Cell. J. Am.
Heterostructure of Gold and Titania Nanoparticles for Solar Fuel
Chem. Soc. 2009, 131, 926−927.
(33) Pyykkö, P.; Runeberg, N. Comparative Theoretical Study of N- Production. Sci. Rep. 2018, 8, No. 10464.
Heterocyclic Carbenes and Other Ligands Bound to AuI. Chem. − (50) Son, H. Y.; Jun, H.; Kim, K. R.; Hong, C. A.; Nam, Y. S.
Asian J. 2006, 1, 623−628. Tannin-mediated Assembly of Gold−Titanium Oxide Hybrid Nano-
(34) Crudden, C. M.; Horton, J. H.; Narouz, M. R.; Li, Z.; Smith, C. particles for Plasmonic Photochemical Applications. J. Ind. Eng. Chem.
A.; Munro, K.; Baddeley, C. J.; Larrea, C. R.; Drevniok, B.; 2018, 63, 420−425.
Thanabalasingam, B.; McLean, A. B.; Zenkina, O. V.; Ebralidze, I. (51) Sundararaman, R.; Narang, P.; Jermyn, A. S.; Goddard, W. A.,
I.; She, Z.; Lraatz, H. B.; Mosey, N. J.; Saunders, L. N.; Yagi, A. III; Atwater, H. A. Theoretical Predictions for Hot-Carrier Generation
Simple Direct Formation of Self-Assembled N-Heterocyclic Carbene From Surface Plasmon Decay. Nat. Commun. 2014, 5, No. 5788.
Monolayers on Gold and Their Application in Biosensing. Nat. (52) Brown, A. M.; Sundararaman, R.; Narang, P.; Goddard, W. A.;
Commun. 2016, 7, No. 12654. Atwater, H. A. Nonradiative Plasmon Decay and Hot Carrier
(35) Lavrich, D. J.; Wetterer, S. M.; Bernasek, S. L.; Scoles, G. Dynamics: Effects of Phonons, Surfaces, and Geometry. ACS Nano
Physisorption and Chemisorption of Alkanethiols and Alkyl Sulfides 2016, 10, 957−966.
on Au(111). J. Phys. Chem. B 1998, 102, 3456−3465. (53) Schlather, A. E.; Manjavacas, A.; Lauchner, A.; Marangoni, V.
(36) Zhukhovitskiy, A. V.; MacLeod, M. J.; Johnson, J. A. Carbene S.; DeSantis, C. J.; Nordlander, P.; Halas, N. J. Hot Hole
Ligands in Surface Chemistry: From Stabilization of Discrete Photoelectrochemistry on Au@SiO2@Au Nanoparticles. J. Phys.
Elemental Allotropes to Modification of Nanoscale and Bulk Chem. Lett. 2017, 8, 2060−2067.
Substrates. Chem. Rev. 2015, 115, 11503−11532. (54) Fu, A.; Chen, X.; Tong, L.; Wang, D.; Liu, L.; Ye, J. Remarkable
(37) Mercs, L.; Albrecht, M. Beyond catalysis: N-Heterocyclic Visible-Light Photocatalytic Activity Enhancement over Au/p-type
Carbene Complexes as Components for Medicinal, Luminescent, and TiO2 Promoted by Efficient Interfacial Charge Transfer. ACS Appl.
Functional Materials Applications. Chem. Soc. Rev. 2010, 39, 1903− Mater. Interfaces 2019, 11, 24154−24163.
1912. (55) Govorov, A. O.; Zhang, H.; Gun’ko, Y. K. Theory of
(38) Cao, Z.; Kim, D.; Hong, D.; Yu, Y.; Xu, J.; Lin, S.; Wen, X.; Photoinjection of Hot Plasmonic Carriers from Metal Nanostructures
Nichols, E. M.; Jeong, K.; Reimer, J. A.; Yang, P.; Chang, C. J. A Into Semiconductors and Surface Molecules. J. Phys. Chem. C 2013,
Molecular Surface Functionalization Approach to Tuning Nano- 117, 16616−16631.
particle Electrocatalysts for Carbon Dioxide Reduction. J. Am. Chem. (56) Liu, J. G.; Zhang, H.; Link, S.; Nordlander, P. Relaxation of
Soc. 2016, 138, 8120−8125. Plasmon-Induced Hot Carriers. ACS Photonics 2018, 5, 2584−2595.
(39) Cao, Z.; Derrick, J. S.; Xu, J.; Gao, R.; Gong, M.; Nichols, E. (57) Derkachova, A.; Kolwas, K.; Demchenko, I. Dielectric Function
M.; Smith, P. T.; Liu, X.; Wen, X.; Copéret, C.; Chang, C. J. Chelating for Gold in Plasmonics Applications: Size Dependence of Plasmon
N-Heterocyclic Carbene Ligands Enable Tuning of Electrocatalytic Resonance Frequencies and Damping Rates for Nanospheres.
CO2 Reduction to Formate and Carbon Monoxide: Surface Plasmonics 2016, 11, 941−951.
Organometallic Chemistry. Angew. Chem., Int. Ed. 2018, 57, 4981− (58) Tagliabue, G.; Jermyn, A. S.; Sundararaman, R.; Welch, A. J.;
4985. DuChene, J. S.; Pala, R.; Davoyan, A. R.; Narang, P.; Atwater, H. A.

33825 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826
ACS Applied Materials & Interfaces www.acsami.org Research Article

Quantifying the Role of Surface Plasmon Excitation and Hot Carrier


Transport in Plasmonic Devices. Nat. Commun. 2018, 9, No. 3394.
(59) Hums, C.; Finger, T.; Christen, J.; Dadgar, A.; et al. Fabry-
Perot Effects in InGaN/GaN Heterostructures on Si-Substrate. J.
Appl. Phys. 2007, 101, No. 033113.
(60) Kumari, S.; Khare, A.; Gupta, R.; Tomar, M.; Gupta, V. Fabry-
perot Modes Enhanced Pump-Probe Coupling in Gold Micro-Disk
Patterned Ruby Thin Film. Opt. Mater. 2017, 72, 375−379.
(61) Moskovits, M. The Case for Plasmon-Derived Hot Carrier
Devices. Nat. Nanotechnol. 2015, 10, 6−8.
(62) Kim, Y.; Dumett Torres, D.; Jain, P. K. Activation Energies of
Plasmonic Catalysts. Nano Lett. 2016, 16, 3399−3407.
(63) Pugh, J. R.; Bruce, M. R. M.; Sullivan, B. P.; Meyer, T. J.
Formation of a Metal-Hydride Bond and the Insertion of Carbon
Dioxide. Key Steps in the Electrocatalytic Reduction of Carbon
Dioxide to Formate Anion. Inorg. Chem. 1991, 30, 86−91.
(64) Ling, Y.; Wang, G.; Wheeler, D. A.; Zhang, J. Z.; Li, Y. Sn-
Doped Hematite Nanostructures for Photoelectrochemical Water
Splitting. Nano Lett. 2011, 11, 2119−2125.
(65) Jiang, J.; Ling, C.; Xu, T.; Wang, W.; Niu, X.; Zafar, A.; Yan, Z.;
Wang, X.; You, Y.; Sun, L.; Lu, J.; Wang, J.; Ni, Z. Defect Engineering
for Modulating the Trap States in 2D Photoconductors. Adv. Mater.
2018, 30, No. 1804332.
(66) Gaggioli, C. A.; Bistoni, G.; Ciancaleoni, G.; Tarantelli, F.;
Belpassi, L.; Belanzoni, P. Modulating the Bonding Properties of N-
Heterocyclic Carbenes (NHCs): A Systematic Charge-Displacement
Analysis. Chem. − Eur. J. 2017, 23, 7558−7569.
(67) Kamata, R.; Kumagai, H.; Yamazaki, Y.; Sahara, G.; Ishitani, O.
Photoelectrochemical CO2 Reduction Using a Ru(II)-Re(I) Supra-
molecular Photocatalyst Connected to a Vinyl Polymer on a NiO
Electrode. ACS Appl. Mater. Interfaces 2019, 11, 5632−5641.
(68) Tang, P.; Xie, H.; Ros, C.; Han, L.; Biset-Peiró, M.; He, Y.;
Kramer, W.; Rodríguez, A. P.; Saucedo, E.; Galán-Mascarós, J. R.;
Andreu, T.; Morante, J. R.; Arbiol, J. Enhanced Photoelectrochemical
Water Splitting of Hematite Multilayer Nanowire Photoanodes by
Tuning the Surface State via Bottom-Up Interfacial Engineering.
Energy Environ. Sci. 2017, 10, 2124−2136.
(69) Cicchillo, R. M.; Zhang, H.; Blodgett, J. A. V.; Whitteck, J. T.;
Li, G.; Nair, S. K.; van der Donk, W. A.; Metcalf, W. W. An Unusual
Carbon−Carbon Bond Cleavage Reaction During Phosphinothricin
Biosynthesis. Nature 2009, 459, 871−874.

33826 https://dx.doi.org/10.1021/acsami.0c09517
ACS Appl. Mater. Interfaces 2020, 12, 33817−33826

You might also like