Photoelectrochemical CO2 Reduction into Syngas with the Metal/Oxide Interface

Article abstract of DOI:10.1021/jacs.8b03067

Photoelectrochemical (PEC) reduction of CO₂ with H₂O not only provides an opportunity for reducing net CO₂ emissions but also produces value-added chemical feedstocks and fuels. Syngas, a mixture of CO and H₂, i

Full text of DOI:10.1021/jacs.8b03067

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Photoelectrochemical CO2 Reduction into Syngas with the Metal/
Oxide Interface
Sheng Chu,†,⊥ Pengfei Ou,‡,⊥ Pegah Ghamari,†,⊥ Srinivas Vanka,†,§ Baowen Zhou,† Ishiang Shih,†
Jun Song,*,‡ and Zetian Mi*,†,§
†Department of Electrical and Computer Engineering, McGill University, 3480 University Street, Montreal, Quebec H3A 0E9,
Canada
‡Department of Mining and Materials Engineering, McGill University, 3610 University Street, Montreal, Quebec H3A 0C5, Canada
§Department of Electrical Engineering and Computer Science, University of Michigan, 1301 Beal Avenue, Ann Arbor, Michigan
48109, United States
S
* Supporting Information
ABSTRACT: Photoelectrochemical (PEC) reduction of CO2 with H2O not only
provides an opportunity for reducing net CO2 emissions but also produces value-
added chemical feedstocks and fuels. Syngas, a mixture of CO and H2, is a key
feedstock for the production of methanol and other commodity hydrocarbons in
industry. However, it is challenging to achieve efficient and stable PEC CO2 reduction
into syngas with controlled composition owing to the difficulties associated with the
chemical inertness of CO2 and complex reaction network of CO2 conversion. Herein,
by employing a metal/oxide interface to spontaneously activate CO2 molecule and
stabilize the key reaction intermediates, we report a benchmarking solar-to-syngas
efficiency of 0.87% and a high turnover number of 24 800, as well as a desirable high
stability of 10 h. Moreover, the CO/H2 ratios in the composition can be tuned in a
wide range between 4:1 and 1:6 with a total unity Faradaic efficiency. On the basis of experimental measurements and theoretical
calculations, we present that the metal/oxide interface provides multifunctional catalytic sites with complementary chemical
properties for CO2 activation and conversion, leading to a unique pathway that is inaccessible with the individual components.
The present approach opens new opportunities to rationally develop high-performance PEC systems for selective CO2 reduction
into valuable carbon-based chemicals and fuels.
established commercial processes for liquid fuels synthesis.
Providing different CO/H2 ratio in syngas mixtures can also be
used for different downstream products (e.g., 1:3, 1:2 and 1:1
for methane, methanol and oxo-alcohols, respectively).16−18
Therefore, the syngas route would allow a flexible platform for
integration with a wide window of catalytic systems in a broad
CO2-recycling scheme without the strict requirement of
suppression of H2 evolution reaction.19−31
To date, various semiconductor photocathodes, including p-
Si,32−34 ZnTe,35,36 CdTe,37 p-InP,37 Cu2O38,39 and p-NiO,40,41
have been investigated for PEC CO2 reduction into CO, usually
in conjunction with a molecular metal-complex or metal
cocatalyst (e.g., Au, Ag and derivatives) to realize selective CO
production. However, it remains challenging to develop
efficient and stable PEC catalytic system that can activate
inert CO2 molecule at low overpotential or even spontaneously,
and selectively produce syngas with controlled composition in a
wide range to meet different downstream products. It has been
reported that pure metal catalyst with a simple monofunctional
site usually has a weak interaction with CO2 molecule and
INTRODUCTION
■
The development of solar-powered CO2 reduction with H2O
holds the promise to mitigate greenhouse gas (CO2) emission
into the atmosphere, while simultaneously converting renew-
able solar energy into storable value-added chemicals and
fuels.1−5 The photoelectrochemical (PEC) route, which
combines light harvesting photovoltaic and electrochemical
components into a monolithically integrated device, has
received considerable attention for application in CO2
reduction recently.6−9 Among the wide variety of CO2
reduction products, CO is a gaseous product that requires
only two proton−electron transfers, and thus a kinetically
feasible choice compared to CH3OH and CH4, which require
six and eight proton−electron transfers to form one molecule,
respectively.10,11 Moreover, CO is an important bulk chemical
to form methanol and other commodity hydrocarbons via
syngas intermediate, a mixture of CO and H2, by using well-
established standard industrial processes such as Fischer−
Tropsch technology.12−15 These attributes, together with the
almost inevitable H2 evolution in aqueous PEC cell can be
valorized to produce syngas mixtures, rendering the proposed
syngas production from CO2 and H2O conversion a
technologically and economically viable pathway to leverage
Received: March 19, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Characterization of Pt-TiO2/GaN/n+-p Si sample. (a) Schematic illustration of the structure, (b) 45°-tilted SEM image shows GaN
nanowire growth vertically on the Si substrate, (c) TEM image illustrates Pt nanoparticles distributed uniformly on the GaN nanowire surface, (d)
HRTEM image, and EDX analysis of the center (e) and edge region (f) indicates the coating of GaN nanowire and Pt nanoparticles with ultrathin
TiO2 layer. The Cu peaks in EDX arise from the TEM sample grid.
cannot provide multiple sites for stabilizing the key reaction
intermediates with optimal binding strength, which leads to
impractically high overpotential and low catalytic efficiency/
stabilty.42,43 Here we show that efficient and stable syngas
production with tunable composition from PEC CO2 reduction
can be achieved with the introduction of a completely different
type of sites: metal/oxide interface. On the basis of experiments
and first-principles theoretical calculations, it is found that the
metal/oxide interface can spontaneously activate CO2 and
stabilize the key reaction intermediates for facilitating CO
production. The intimate metal/oxide interface provides the
multifunctional combination of metal and oxide catalytic sites
with complementary chemical properties, which opens new
reaction channels that are not possible with the individual
components. The versatility of using metal/oxide interface is
demonstrated by the combination of different metals (Pt and
Pd) and oxides (TiO2 and ZnO). Remarkably, although pristine
metal catalytically favors the proton reduction to evolve H2, the
coverage of metal with oxide to form metal/oxide interface
shows preferential activity for CO2 reduction over H2 evolution.
As an example, by rationally integrating the Pt/TiO2 cocatalyst
with the strong light harvesting of p-n Si junction and efficient
electron extraction effect of GaN nanowire arrays (Pt-TiO2/
GaN/n+-p Si), a record half-cell solar-to-syngas (STS)
efficiency of 0.87% and a benchmark turnover number
(TON) of 24 800 have been achieved in an aqueous PEC
system. The durability of the PEC system for highly stable
syngas production of 10 h has been demonstrated as well.
on p-n Si junction was selected as the platform to load the Pt-
TiO2 system. The schematic design of Pt-TiO2/GaN/n+-p Si is
illustrated in Figure 1a. The sample was prepared in two major
steps. First, GaN nanowire arrays were grown on p-n Si wafer
by plasma-assisted molecular beam epitaxy (see the Supporting
Information).31,44 Such a structure takes advantage of the
strong light absorption capability of Si (bandgap of 1.1 eV) and
efficient electron extraction effect as well as large surface area
provided by GaN nanowires. Moreover, the light absorption
and catalytic reaction sites are decoupled spatially in the
structure, providing an ideal platform to investigate the effect of
cocatalysts on the catalytic performance without affecting the
optical properties. Second, Pt nanoparticles and TiO2 ultrathin
layer were deposited on GaN nanowires surface in sequential
order using photodeposition and atomic-layer deposition
(ALD) process, respectively (see the Supporting Information).
The intimate Pt/TiO2 interface provides multiple sites and
unique channels that facilitate the CO2 activation and reaction
pathways for syngas production.
Characterization of Structure. The morphology and
chemical component of the Pt-TiO2/GaN/n+-p Si hetero-
structures were studied using scanning electron microscopy
(SEM), transmission electron microscopy (TEM), energy-
dispersive X-ray spectroscopy (EDX) and inductively coupled
plasma−atomic emission spectroscopy (ICP-AES) analysis.
The cross-sectional SEM image (Figure 1b) shows that the
GaN nanowires are aligned vertically to the Si substrate with an
average diameter of ∼50 nm ( 15 nm) and height of 250 nm
( 50 nm). TEM image in Figure 1c reveals that Pt
nanoparticles of 2−3 nm size are uniformly deposited on the
GaN nanowires surface. High-resolution TEM (HRTEM)
image (Figure 1d), along with EDX analysis in the center
and edge regions of nanowire (Figures 1e and 1f, respectively),
RESULTS AND DISCUSSION
■
Design and Synthesis of Pt-TiO2/GaN/n+-p Si. We
choose Pt-TiO2 as an example to demonstrate the validity of
metal/oxide interface for PEC CO2 reduction. GaN nanowire
B
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. (a) Faradaic efficiencies for CO (gray bars) and H2 (red bars), and solar-to-syngas efficiency of Pt-TiO2/GaN/n+-p Si photocathode as a
function of potential in CO2-saturated 0.5 M KHCO3 solution (pH 7.5). (b) Chronoamperometry data and FEs for CO and H2 of Pt-TiO2/GaN/
n+-p Si photocathode at +0.27 V vs RHE. The dashed lines denote cleaning of photoelectrode and purging of the PEC cell with CO2. (c) J−V curves
of bare GaN/n+-p Si, GaN/n+-p Si with individual Pt or TiO2 cocatalyst, and Pt-TiO2/GaN/n+-p Si. (d) Faradaic efficiencies for CO at +0.27 V vs
RHE. The FEs for CO of GaN/n+-p Si and TiO2/GaN/n+-p Si photocathodes were measured at −0.33 V vs RHE due to the negligible photocurrent
at an applied positive potential. The light illumination is ∼8 suns.
indicates the coating of GaN nanowire with ultrathin TiO2
layer. The TiO2 layer is amorphous and has a thickness of ∼1
nm, which corresponds to 18 ALD cycles of TiO2 deposition.
The lattice spacings of 0.22 and 0.26 nm corresponds to the
(111) facet of Pt and (002) lattice plane of GaN, indicating the
preferred nanowire growth along ⟨0001⟩ direction (c-axis). The
loading amounts of Pt and Ti in Pt-TiO2/GaN/n+-p Si were
determined to be 4.9 and 48.3 nmol cm−2, respectively, by
using ICP-AES analysis.
decreased CO FE at a more negative potential than +0.37 V is
mainly due to the limited CO2 mass transport in the electrolyte
at high CO generation rate.46,47 The kinetic limitation was
evidenced by the saturated current density for CO generation
in the high applied bias region (Figure S3a, Supporting
Information). In addition, different Tafel slopes for the CO2
reduction and H2 evolution reactions could lead to the above-
mentioned bias-dependent reaction selectivity. To evaluate
their contribution, the Tafel plots for CO and H2 evolution
were drawn by using the corresponding partial current density,
as shown in Figure S3b in the Supporting Information. The
Tafel slopes were calculated by using data points more positive
than +0.37 V, as the slope increases dramatically at more
negative potentials due to the mass-transport limitations.48,49 It
was found that the Tafel slopes for CO and H2 evolution were
386 and 119 mV dec−1, respectively. The different Tafel slopes
result in the bias-dependent reaction selectivity largely in the
low bias region. At all the applied potentials, a total FE of 97
8% was obtained for the cogeneration of CO and H2, with no
appreciable amount of other gas products detected by gas
chromatograph (GC) and liquid products (e.g., HCOOH and
CH3OH) analyzed by nuclear magnetic resonance (NMR)
spectroscopy. To demonstrate that the generated CO from
CO2 reduction, isotopic experiment using 13CO2 was
conducted. The signal at m/z = 29 assigned to 13CO appeared
in the gas chromatography−mass spectrometry analysis (Figure
S4, Supporting Information), indicating the CO product is
formed from the reduction of CO2.
Photoelectrochemical Properties. PEC performance of
the sample was investigated in CO2-saturated 0.5 M KHCO3
solution (pH 7.5) under 300 W xenon lamp irradiation (800
mW cm−2, ∼8 suns) in a conventional three-electrode cell. To
reveal the interaction of photocathode with CO2, the current−
potential (J−V) curves of Pt-TiO2/GaN/n+-p Si in a CO2 or
Ar-saturated electrolyte was compared (Figure S1, Supporting
Information). There is a large enhancement in the photo-
current generation under CO2 atmosphere compared to that of
Ar atmosphere, indicating an interaction between the electrode
surface and CO2 molecule for CO2 reduction. Figure 2a shows
the Faradaic efficiencies (FEs) for CO and H2 on Pt-TiO2/
GaN/n+-p Si at applied potential between +0.47 V and +0.07 V
vs reversible hydrogen electrode (RHE) in CO2-saturated
electrolyte (hereafter, all the potentials are referenced to the
RHE unless otherwise specified). The corresponding chro-
noamperometry data at different applied potentials are shown
in Figure S2 (Supporting Information). At an applied potential
of +0.47 V, the photocathode exhibited a high CO FE of 78%,
indicating the major extracted photogenerated electrons were
used for selectively CO2-to-CO conversion at the catalyst
surface. By tuning the potential from +0.47 V to +0.07 V, the
CO/H2 ratio can be tuned in a large range between 4:1 and 1:6.
At + 0.27 V, a CO/H2 ratio of 1:2 is obtained, which is a
desirable composition of syngas mixtures for methanol
synthesis and Fischer−Tropsch hydrocarbon formation.45 The
A highlight of our system is the highly positive onset
potential of +0.47 V (underpotential of 580 mV to the CO2/
CO equilibrium potential at −0.11 V) for producing high CO
FE of 78% in an aqueous PEC cell. The reported onset
potentials, FEs and solar energy conversion efficiency of
different photocathodes for CO production in an aqueous PEC
cell are compared (Table S1, Supporting Information). Among
C
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
various reported photocathodes, our system features the lowest
onset potential, which is 170 mV positive shifted compared
with the best value reported in the literature.36 The extremely
low onset potential of our photocathode is attributed to
coupling effects including strong light harvesting of p-n Si
junction, efficient electron extraction of GaN nanowire arrays,
and extremely fast syngas production kinetics on Pt-TiO2 dual
cocatalysts. The STS efficiencies of our system at different
applied potentials were calculated according to the measured
photocurrent density and FEs for CO and H2 (eq 1, Supporting
Information). As shown in Figure 2a, at +0.17 V, the STS
efficiency reached 0.87%, which greatly outperforms other
reported photocathodes and sets a new benchmark reported to
date (Table S1, Supporting Information).
carrier recombination at the surface.54,55 It is also supposed that
Pt/TiO2 interface is more resistant to CO poisoning than Pt
alone as shown in thermochemical catalysis,56−58 which could
contribute to the enhanced syngas production on metal/oxide
interface. Figure 2d shows the comparison of FEs of CO for the
four samples. Besides CO product, the remaining balance of
photocurrent drives H2 evolution from proton reduction. It is
shown that CO FEs are very low on bare GaN/n+-p Si, and
with individual Pt or TiO2 cocatalyst (1.7%, 2% and 5.6%,
respectively). In contrast, the CO formation selectivity
increases greatly to 32% by loading Pt-TiO2 dual cocatalyst,
indicating a synergetic effect between Pt and TiO2. We attribute
the synergy to the strong interaction at the intimate metal/
oxide interface, which provides the multifunctional adsorption/
reaction sites for CO2 activation and conversion. It is worth
mentioning that there is an optimized thickness of ∼1 nm TiO2
for maximum catalytic activity and CO selectivity (Figure S6,
Supporting Information). Very thin TiO2 deposition yields less
interfacial reactive sites, while increasing the TiO2 thickness
over 1 nm resulted in limited mass transport of reactants to the
interfacial sites and large tunneling resistance to charge carrier
transport associated with thick TiO2 layer.59,60
CO2 Adsorption and Activation. To elucidate the role of
metal/oxide interface for the conversion of CO2 to CO from
the fundamental atomic level, density functional theory (DFT)
calculations were employed using Ti3O6H6/Pt(111) to describe
the Pt/TiO2 interface (see Supporting Information for
computational details). The hydroxylation of titania cluster
(Ti3O6H6) was considered in the calculations to account for the
effect of PEC CO2 reduction conditions in an aqueous
environment.61,62 The optimized geometry of Ti3O6H6/Pt-
(111) is presented in Figure S7 (Supporting Information). As
CO2 adsorption and activation on catalyst surface is the initial
and often the rate-determining step for the whole CO2
reduction process,63,64 we first investigate the CO2 adsorption
characteristics on Ti3O6H6/Pt(111) surface. The calculation of
CO2 adsorption on pristine Pt(111) was also performed as a
comparison. Figures 3a and 3b show the optimized
configurations of CO2 adsorption on the pristine Pt(111) and
Ti3O6H6/Pt(111) surface, respectively. It was found that CO2
retains the original linear configuration on pristine Pt(111),
similar to its isolated gas-phase state. In contrast, there are
strong interactions between CO2 molecule and the Ti3O6H6/
Pt(111) interface, with C atom strongly binding to the Pt atom
underneath with a bond length of 2.02 Å and one O atom (O2)
attaching to the Ti atom with a shorter bond length of 1.96 Å.
Such a strong bonding between CO2 and Ti3O6H6/Pt(111)
interface results in a significant bending of CO2 molecule from
its originally linear form to an O−C−O angle of 125.02°, thus
forming a tridentate configuration that facilitates its subsequent
transformations.64 In addition, the strong interaction of CO2
with the interface weakens the two C−O bonds of CO2, leading
to elongated C−O bonds (1.22 and 1.32 Å) from the original
bond length of 1.18 Å in the isolated CO2 molecule (Table S2,
Supporting Information). The weakened C−O bonds and the
formed bent CO2 configurations indicate a remarkable
activation of CO2 molecule upon chemisorption at the
interface, which is in contrast with the negligible activation of
CO2 on pristine Pt(111). This result agrees well with the
observations in the field of thermochemical catalysis that CO2
transformation is greatly enhanced with metal/oxide interface
as compared to that with pure metal.65−69 It is worth noting
that the CO2 activation mechanism at metal/oxide interface has
The durability of Pt-TiO2/GaN/n+-p Si photocathode was
investigated at a constant potential of +0.27 V by five
consecutive runs with each run of 2 h (Figure 2b). After each
cycle, the products of CO and H2 were analyzed by GC, the
electrode was thoroughly cleaned by deionized water and the
PEC cell was purged by CO2 for 20 min. During the five runs of
10 h operation, the electrode showed similar behavior in terms
of photocurrent density and product selectivity, indicating the
high stability of the sample during the syngas production
process. The initial decrease of high photocurrent density in
each run is likely due to the limited mass transfer of reactants or
products at high reaction rates, which can be recovered in the
next run after the cleaning of photoelectrode surface. The CO/
H2 ratio in the products was kept nearly 1:2 during the five
cycles of operation, which is a desirable syngas composition for
synthetizing downstream products including methanol and
liquid hydrocarbons.45 In addition, the SEM, TEM, and XPS
analysis of Pt-TiO2/GaN/n+-p Si photocathode after the PEC
reaction were performed, as shown in Figure S5 (Supporting
Information). No appreciable change of GaN nanowires and
Pt-TiO2 catalysts were found. The total turnover number
(TON), defined as the ratio of the total amount of syngas
evolved (264 μmol) to the amount of Pt-TiO2 catalyst (10.64
nmol, calculated from the catalyst loadings and electrode
sample area of 0.2 cm2), reached 24 800, which is at least 1 or 2
orders of magnitude higher than previously reported values for
syngas or CO formation from PEC or photochemical CO2
reduction.31,50−53
Investigation of the Catalytic Mechanisms. To under-
stand the underlying catalytic mechanism and the role of basic
components for the PEC performance of the Pt-TiO2/GaN/n+-
p Si photocathode, we conducted a series of control
experiments. Figure 2c shows the comparison of LSV curves
for bare GaN/n+-p Si, GaN/n+-p Si with individual Pt or TiO2
cocatalyst, and Pt-TiO2/GaN/n+-p Si. The bare GaN/n+-p Si
displays a poor PEC performance with a negligible photo-
current density and highly negative onset potential. The loading
of Pt cocatalyst can greatly improve the PEC performance with
an onset potential of about +0.47 V and photocurrent density
of ∼50 mA cm−2 at −0.33 V, while TiO2 alone shows a small
photocurrent density of 5 mA cm−2 at −0.33 V. Compared to
bare Pt, significantly higher photocurrent density of ∼120 mA
cm−2 at −0.33 V is attained when Pt and TiO2 are loaded
simultaneously. It is proposed that the formation of intimate
Pt/TiO2 interface stabilizes the reaction intermediates and
reduces the activation barrier for syngas production, which are
validated by theoretical calculations discussed below. In
addition, the ultrathin TiO2 overlayer may passivate the
nanowire surface states and reduce the probability of charge
D
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Pt/GaN/n+-p Si and Pt-TiO2/GaN/n+-p Si was tested by CO2
adsorption−desorption measurements (Figure S8, Supporting
Information). The CO2 adsorption amount over Pt-TiO2/
GaN/n+-p Si was 1.91 μmol cm−2, which was 7 times higher
than that of Pt/GaN/n+-p Si (0.27 μmol cm−2). As a
comparison, the CO2 adsorption amount on plain GaN/n+-p
Si was 0.24 μmol cm−2, indicating the low propensity of Pt for
CO2 chemisorption. The combined experimental and theoreti-
cal results explain well the different behaviors in the PEC
studies that pristine Pt does not favor CO2 reduction, while the
construction of Pt/TiO2 interface shows greatly enhanced
activity for CO2 reduction.
To further investigate the detailed bonding interaction
between CO2 and Ti3O6H6/Pt(111) interface, the differential
charge density (DCD) was examined, shown in Figure 3c. The
yellow and blue regions indicate electronic charge accumulation
and depletion, respectively. Strong electronic coupling between
CO2 and the interface was evidenced by the electron charge
density redistribution around the interfacial region. Notable
electron accumulation near the O2 atom in CO2 and electron
depletion around the neighboring Ti nucleus indicates an ionic-
like Ti−O bonding, while the electron accumulation between
Pt and C atoms suggests the formation of covalent Pt−C
bonding. Overall, substantial electrons are transferred from the
interface to CO2 molecule, resulting in the formation of
activated *CO2− anion and eventually the enhanced CO2
reduction activity. Quantitative estimate of the electron transfer
was studied by the Bader charge analysis (Table S2, Supporting
Information).74,75 It was found that CO2 attracted 0.684e from
the substrate for CO2 adsorption at the Ti3O6H6/Pt(111)
interface, as compared to 0.0263e in the case of pristine Pt.
Reaction Energetics. Moreover, to gain insights into the
selective CO evolution from CO2 reduction at molecular level,
DFT calculations were also performed to understand the
reaction energetics of the CO2 → CO pathway. As suggested by
previous studies,76−78 we considered the following reaction
steps:
Figure 3. Side views of optimized configurations of CO2 adsorbed on
the (a) Pt(111) surface and (b) Ti3O6H6/Pt(111) surface. (c)
Differential charge density of CO2 adsorbed at the Ti3O6H6/Pt(111)
interface. Regions of yellow and blue indicate electronic charge gain
and loss, respectively. Isosurface contours of electron density
differences were drawn at 0.002 e/Bohr3. (d) Calculated free energy
diagrams for CO2 reduction to CO on Pt(111) and Ti3O6H6/Pt(111)
surfaces at 0 V vs RHE. The optimized structures for each step are also
shown. To improve legibility, a break region was added from 0.25 to
3.75 on the Y axis due to the large energy barriers for the CO2
reduction on Pt(111) surface. Pt: gray, Ti: blue, O: red, C: brown and
H: white.
CO2(g) + * + H+(aq) + e− → *COOH
(1)
a certain degree of similarity to that reported on individual
metal oxide (e.g., TiO2) with oxygen vacancies, in which one of
the O atoms in CO2 is coordinating to an undercoordinated Ti
atom at the edge of the cluster (i.e., essentially an O
vacancy).70−72 For example, Somorjai and co-workers proposed
a similar mechanism to explain carbonyl bond activation at the
Pt/TiO2 interface via oxygen vacancy on TiO2 surface.70
The energetics associated with CO2 adsorption on Pt(111)
and Ti3O6H6/Pt(111) surfaces were also calculated and
*COOH + H+(aq) + e− → *CO + H2O(l)
(2)
*CO → CO(g) + *
(3)
where a lone asterisk (*) represents a surface adsorption site
and * symbol before a molecule denotes a surface-bound
species. Figure 3d shows the calculated free energy diagram of
CO2 reduction on Pt(111) and Ti3O6H6/Pt(111). On pristine
Pt(111), the first step of CO2 activation to form *COOH
intermediate is highly endergonic with a free energy change
(ΔG) of 5.08 eV, which is the rate-limiting step for the whole
CO2 reduction process. In contrast, on the Ti3O6H6/Pt(111)
interface, *COOH formation is exergonic owing to the strong
binding to the interfacial sites, with C and O atoms in COOH
binding to Pt(111) and Ti of Ti3O6H6, respectively. Similarly,
the strong binding and stabilization of *CO intermediates were
also observed with cooperative interactions with both metal and
oxide in the interface, resulting in the facile formation of *CO.
The rate-limiting step in the Ti3O6H6/Pt(111) system is the
CO desorption, but with a much smaller free energy change of
0.88 eV as compared to 5.08 eV on pristine Pt(111). This result
suggests that there are sites of different nature with
complementary chemical properties in the metal/oxide inter-
face that work in synergy to facilitate the CO2 reduction into
analyzed in terms of the adsorption energy (Ead) and
CO2
def
deformation energy (E ) (Table S2, Supporting Informa-
tion). Here Ead represents the net energy increased upon
CO2
def
adsorption. E denotes the energy change from the distortion
of a linear CO2 molecule into a buckled configuration,
correlating with the degree of CO2 activation.73 The Ead and
CO2
def
E
of CO2 adsorption at Ti3O6H6/Pt(111) interface are
−0.80 and 2.65 eV respectively, as compared with those of 4.44
and 0.01 eV on pristine Pt(111). The negative Ead value implies
the exothermic process of CO2 adsorption at Ti3O6H6/Pt(111)
interface, while positive Ead value indicates the unfavorable CO2
adsorption on pristine Pt(111). In addition, the large positive
CO2
def
value of E
in the case of Ti3O6H6/Pt(111) confirms that
CO2 is activated spontaneously at the interface, in strong
contrast to the marginal value on pristine Pt(111).
Experimentally, the amount of CO2 adsorption capacity over
E
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. (a) XPS of Pt 4f of Pt/GaN/n+-p Si and Pt-TiO2/GaN/n+-p Si. (b) Electron localized function (ELF) of Ti3O6H6/Pt(111). The
probability of finding electron pairs varies from 0 (blue color) to 1 (red color).
Figure 5. (a) Faradaic efficiencies for CO of Pd/GaN/n+-p Si, Pd-TiO2/GaN/n+-p Si, Pt/GaN/n+-p Si and Pt-ZnO/GaN/n+-p Si. The
measurements were performed at +0.27 V vs RHE for 100 min. (b) Calculated free energy diagrams for CO2 reduction to CO on Pd(111), Pt(111),
Ti3O6H6/Pd(111) and Zn6O6H7/Pt(111) surfaces at 0 V vs RHE. The optimized structures for each step are also shown. To improve legibility, a
break region was added from 0.25 to 2.75 on the Y axis due to the large energy barriers for the CO2 reduction on Pd(111) and Pt(111) surface. Pd:
pine green, Pt: gray, Ti: blue, Zn: purple, O: red, C: brown and H: white.
CO. In addition, the effects of the electrolyte and applied
potential were considered in DFT calculations, similar
conclusions were obtained. The calculation results and details
are available in the Supporting Information (Figures S9 and
S10).
between metal and oxide may also contribute to the selective
CO2 reduction into CO on Pt-TiO2/GaN/n+-p Si photo-
cathode. The electronic properties of Pt were evaluated using
the peak energy of Pt 4f by X-ray photoelectron spectroscopy
(XPS) analysis (Figure 4a). Compared to Pt/GaN/n+-p Si, a
notable shift of ca. 0.5 eV to higher binding energy position was
observed for Pt 4f in Pt-TiO2/GaN/n+-p Si. This shift is less
pronounced than the binding energy difference between Pt0
and Pt2+ in PtO (ca. 1.5 eV),83 indicating the presence of
electron deficient Pt species (Ptn+) in Pt-TiO2/GaN/n+-p Si. A
significant electronic modification by strong metal/oxide
interaction is likely responsible for this change of Pt oxidation
state.84−87 To confirm the strong interaction between the metal
and oxide, we computed the electron localized function (ELF)
for Ti3O6H6/Pt(111) system, as shown in Figure 4b. Topology
analysis of ELF can effectively characterize the nature of
different chemical bonding schemes,88 and has been used to
estimate the degree of metal−support interactions.89 The ELF
map of Ti3O6H6/Pt(111) shows that there is a significant
electron redistribution in the regions between Pt and Ti3O6H6,
indicating strong interactions between them. The strong
interactions can modify the electronic property of Pt and
hence enhance CO2 reduction.
Considering that H2 product from proton reduction is the
other important component in the syngas mixture besides CO,
we also calculated free energy diagrams for H2 evolution on
pristine Pt(111) and Ti3O6H6/Pt(111), as shown in Figure
S11a (Supporting Information). Ti3O6H6/Pt(111) shows a
slightly lowered energy barrier than that on pristine Pt(111) by
0.06 eV. Considering that the uncertainty associated with DFT
energy calculations is on the same order,79 the calculated
energy barriers for hydrogen evolution reaction are comparable
in the two cases. Recent studies have shown that the CO2
reduction selectivity in competition with H2 evolution is related
to the difference between their two thermodynamic limiting
potentials (denoted as UL(CO2) − UL(H2)).80−82 Therefore,
we have calculated the difference between limiting potentials
for CO and H2 evolution. As seen in Figure S11b (Supporting
Information), Ti3O6H6/Pt(111) displays a significantly more
positive value for UL(CO2) − UL(H2) than that on pristine
Pt(111), indicating higher selectivity for CO2 reduction to CO.
Strong Metal/Oxide Interaction. In addition to the
important role of the metal/oxide interface in activating CO2
and stabilizing the key reaction intermediates, the electronic
modification of the Pt catalyst owing to the strong interaction
Generalization to Other Metal/Oxide Systems. By
understanding the CO2 activation and conversion at the Pt/
TiO2 interface on an atomic level, we propose that the findings
can be extended to other metal/oxide systems. To show the
F
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
generality, Pd-TiO2/GaN/n+-p Si and Pt-ZnO/GaN/n+-p Si
were synthesized by varying either metal or oxide components
(see the Supporting Information). The chemical components
and structures were confirmed by TEM and EDX analysis
(Figures S12−S13, Supporting Information). By using ICP-
AES analysis, the loading amounts of Pd and Ti in Pd-TiO2/
GaN/n+-p Si, Pt and Zn in Pt-ZnO/GaN/n+-p Si were
determined to be 5.4 and 46.1, 4.7, and 39.1 nmol cm−2,
respectively. The FEs of CO for Pd-TiO2/GaN/n+-p Si and Pt-
ZnO/GaN/n+-p Si were measured and compared with Pd/
GaN/n+-p Si and Pt/GaN/n+-p Si, respectively (Figure 5a).
The CO FEs of Pd-TiO2/GaN/n+-p Si and Pt-ZnO/GaN/n+-p
Si are four and 11 times higher than that with individual metal
cocatalysts, similar to the trend observed in Pt-TiO2/GaN/n+-p
Si system. In addition, the free energy diagram of CO2
reduction into CO were calculated to validate the experimental
observations. Ti3O6H6/Pd(111) and Zn6O6H7/Pt(111) were
used in the DFT calculations to describe the Pd/TiO2 and Pt/
ZnO interface, respectively. As seen in Figure 5b, Ti3O6H6/
Pd(111) and Zn6O6H7/Pt(111) shows a significantly lowered
energy barrier than those on pristine Pd(111) and Pt(111).
Similarly, it was found that the formation of *CO from CO2
reduction via *COOH intermediate is a facile downhill process
in the presence of metal/oxide interface, while the first step of
CO2 activation to form *COOH is highly endergonic on pure
metal surface. Although quantitative differences exist between
different systems, similar qualitative trend indicates the critical
role of metal/oxide interfaces in activating CO2, and stabilizing
the key reaction intermediates for facilitating CO production.
The present study could therefore provide a promising set of
principles to enhance the CO2 reduction performance by
tuning the compositions and structures of metal/oxide
interface.
ORCID
Sheng Chu: 0000-0001-9850-5012
Jun Song: 0000-0003-3675-574X
Zetian Mi: 0000-0001-9494-7390
Author Contributions
⊥S.C., P. O., and P.G. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by Emissions Reduction Alberta
(ERA). Part of the work was conducted in the Microfabrication
Facility at McGill University and the Michigan Center for
Materials Characterization at the University of Michigan (NSF
#DMR-0723032). We greatly acknowledge the financial
support from McGill Engineering Doctoral Award and National
Sciences and Engineering Research Council (NSERC)
Discovery grant (grant # RGPIN-2017-05187). We also
thank Supercomputer Consortium Laval UQAM McGill and
Eastern Quebec for providing computing power.
REFERENCES
■
(1) Kumar, B.; Llorente, M.; Froehlich, J.; Dang, T.; Sathrum, A.;
Kubiak, C. P. Annu. Rev. Phys. Chem. 2012, 63, 541.
(2) White, J. L.; Baruch, M. F.; Pander Iii, J. E.; Hu, Y.; Fortmeyer, I.
C.; Park, J. E.; Zhang, T.; Liao, K.; Gu, J.; Yan, Y.; Shaw, T. W.; Abelev,
E.; Bocarsly, A. B. Chem. Rev. 2015, 115, 12888.
(3) Habisreutinger, S. N.; Schmidt-Mende, L.; Stolarczyk, J. K.
Angew. Chem., Int. Ed. 2013, 52, 7372.
(4) Tu, W. G.; Zhou, Y.; Zou, Z. G. Adv. Mater. 2014, 26, 4607.
(5) Xie, S. J.; Zhang, Q. H.; Liu, G. D.; Wang, Y. Chem. Commun.
2016, 52, 35.
(6) Wang, P.; Wang, S.; Wang, H.; Wu, Z.; Wang, L. Part. Part. Syst.
Char. 2018, 35, 1700371.
CONCLUSIONS
■
(7) Pang, H.; Masuda, T.; Ye, J. H. Chem. - Asian J. 2018, 13, 127.
(8) Zhang, N.; Long, R.; Gao, C.; Xiong, Y. Sci. China Mater. 2018,
61, 771.
(9) Kalamaras, E.; Maroto-Valer, M.; Shao, M.; Xuan, J.; Wang, H.
Catal. Today 2018, DOI: 10.1016/j.cattod.2018.02.045.
(10) Vesborg, P. C. K.; Seger, B. Chem. Mater. 2016, 28, 8844.
(11) Karatairi, E.; Miller, J. E. MRS Bull. 2017, 42, 878.
(12) Bell, A. T. Catal. Rev.: Sci. Eng. 1981, 23, 203.
(13) Waugh, K. C. Catal. Today 1992, 15, 51.
(14) Dry, M. E. Catal. Today 2002, 71, 227.
(15) Khodakov, A. Y.; Chu, W.; Fongarland, P. Chem. Rev. 2007, 107,
1692.
(16) Wenzel, M.; Rihko-Struckmann, L.; Sundmacher, K. AIChE J.
2017, 63, 15.
(17) Foit, S. R.; Vinke, I. C.; de Haart, L. G. J.; Eichel, R. A. Angew.
In summary, we have demonstrated an efficient and stable PEC
CO2 reduction system for syngas production with controlled
composition, by employing a metal/oxide interface to activate
inert CO2 molecule and stabilize the key reaction intermediates.
Using Pt/TiO2 as an example, a benchmarking solar-to-syngas
efficiency of 0.87% and a high turnover number of 24 800 were
achieved. Moreover, the PEC system exhibited highly stable
syngas production in the 10 h duration test. On the basis of
experimental measurements and theoretical calculations, it was
found that the synergistic interactions at the metal/oxide
interface provide unique reaction channels that structurally and
electronically facilitate CO2 conversion into CO. This work
may open new opportunities for the design and development of
high-performance photoelectrochemical systems for selective
CO2 reduction.
Chem., Int. Ed. 2017, 56, 5402.
(18) Hernandez, S.; Farkhondehfal, M. A.; Sastre, F.; Makkee, M.;
Saracco, G.; Russo, N. Green Chem. 2017, 19, 2326.
(19) Ross, M. B.; Dinh, C. T.; Li, Y.; Kim, D.; De Luna, P.; Sargent,
E. H.; Yang, P. D. J. Am. Chem. Soc. 2017, 139, 9359.
(20) Sheng, W. C.; Kattel, S.; Yao, S. Y.; Yan, B. H.; Liang, Z. X.;
Hawxhurst, C. J.; Wu, Q. Y.; Chen, J. G. G. Energy Environ. Sci. 2017,
10, 1180.
(21) Xu, J. Q.; Li, X. D.; Liu, W.; Sun, Y. F.; Ju, Z. Y.; Yao, T.; Wang,
C. M.; Ju, H. X.; Zhu, J. F.; Wei, S. Q.; Xie, Y. Angew. Chem., Int. Ed.
2017, 56, 9121.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.8b03067.
Experimental and computational methods, additional
figures and tables (PDF)
(22) Guo, S. J.; Zhao, S. Q.; Wu, X. Q.; Li, H.; Zhou, Y. J.; Zhu, C.;
Yang, N. J.; Jiang, X.; Gao, J.; Bai, L.; Liu, Y.; Lifshitz, Y.; Lee, S. T.;
Kang, Z. H. Nat. Commun. 2017, 8, 1828.
(23) He, R.; Zhang, A.; Ding, Y. L.; Kong, T. Y.; Xiao, Q.; Li, H. L.;
Liu, Y.; Zeng, J. Adv. Mater. 2018, 30, 1705872.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: jun.song2@mcgill.ca. Phone: 1 514 398 4592.
*E-mail: ztmi@umich.edu. Phone: 1 734 764 3963.
■
G
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(24) Furler, P.; Scheffe, J. R.; Steinfeld, A. Energy Environ. Sci. 2012,
5, 6098.
(58) Shi, F.; Baker, L. R.; Hervier, A.; Somorjai, G. A.; Komvopoulos,
K. Nano Lett. 2013, 13, 4469.
(25) Kumar, B.; Smieja, J. M.; Sasayama, A. F.; Kubiak, C. P. Chem.
Commun. 2012, 48, 272.
(26) Kang, P.; Chen, Z. F.; Nayak, A.; Zhang, S.; Meyer, T. J. Energy
Environ. Sci. 2014, 7, 4007.
(27) Lee, J. S.; Won, D. I.; Jung, W. J.; Son, H. J.; Pac, C.; Kang, S. O.
(59) Chen, Y. W.; Prange, J. D.; Duhnen, S.; Park, Y.; Gunji, M.;
Chidsey, C. E. D.; McIntyre, P. C. Nat. Mater. 2011, 10, 539.
(60) Kim, H. J.; Kearney, K. L.; Le, L. H.; Haber, Z. J.; Rockett, A. A.;
Rose, M. J. J. Phys. Chem. C 2016, 120, 25697.
(61) Kattel, S.; Yan, B. H.; Yang, Y. X.; Chen, J. G. G.; Liu, P. J. Am.
Chem. Soc. 2016, 138, 12440.
(62) Gao, D.; Zhang, Y.; Zhou, Z.; Cai, F.; Zhao, X.; Huang, W.; Li,
Y.; Zhu, J.; Liu, P.; Yang, F.; Wang, G.; Bao, X. J. Am. Chem. Soc. 2017,
139, 5652.
Angew. Chem., Int. Ed. 2017, 56, 976.
(28) Li, F. F.; Lau, J.; Licht, S. Adv. Sci. 2015, 2, 1500260.
(29) Li, D. W.; Ouyang, S. X.; Xu, H.; Lu, D.; Zhao, M.; Zhang, X. L.;
Ye, J. H. Chem. Commun. 2016, 52, 5989.
(30) Urbain, F.; Tang, P. Y.; Carretero, N. M.; Andreu, T.; Gerling, L.
G.; Voz, C.; Arbiol, J.; Morante, J. R. Energy Environ. Sci. 2017, 10,
2256.
(31) Chu, S.; Fan, S. Z.; Wang, Y. J.; Rossouw, D.; Wang, Y. C.;
Botton, G. A.; Mi, Z. Angew. Chem., Int. Ed. 2016, 55, 14260.
(32) Hinogami, R.; Nakamura, Y.; Yae, S.; Nakato, Y. J. Phys. Chem. B
1998, 102, 974.
(63) Chang, X. X.; Wang, T.; Gong, J. L. Energy Environ. Sci. 2016, 9,
2177.
(64) Alvarez, A.; Borges, M.; Corral-Perez, J. J.; Olcina, J. G.; Hu, L.
J.; Cornu, D.; Huang, R.; Stoian, D.; Urakawa, A. ChemPhysChem
2017, 18, 3135.
(65) Graciani, J.; Mudiyanselage, K.; Xu, F.; Baber, A. E.; Evans, J.;
Senanayake, S. D.; Stacchiola, D. J.; Liu, P.; Hrbek, J.; Sanz, J. F.;
Rodriguez, J. A. Science 2014, 345, 546.
(66) Rodriguez, J. A.; Liu, P.; Stacchiola, D. J.; Senanayake, S. D.;
White, M. G.; Chen, J. G. G. ACS Catal. 2015, 5, 6696.
(67) Kattel, S.; Yan, B. H.; Chen, J. G. G.; Liu, P. J. Catal. 2016, 343,
115.
(33) Kong, Q.; Kim, D.; Liu, C.; Yu, Y.; Su, Y.; Li, Y.; Yang, P. D.
Nano Lett. 2016, 16, 5675.
(34) Song, J. T.; Ryoo, H.; Cho, M.; Kim, J.; Kim, J. G.; Chung, S. Y.;
Oh, J. Adv. Energy Mater. 2017, 7, 1601103.
(35) Jang, Y. J.; Jang, J. W.; Lee, J.; Kim, J. H.; Kumagai, H.; Lee, J.;
Minegishi, T.; Kubota, J.; Domen, K.; Lee, J. S. Energy Environ. Sci.
2015, 8, 3597.
(36) Jang, Y. J.; Jeong, I.; Lee, J.; Lee, J.; Ko, M. J.; Lee, J. S. ACS
Nano 2016, 10, 6980.
(37) Yoneyama, H.; Sugimura, K.; Kuwabata, S. J. Electroanal. Chem.
Interfacial Electrochem. 1988, 249, 143.
(38) Torralba-Penalver, E.; Luo, Y.; Compain, J. D.; Chardon-Noblat,
(68) Kattel, S.; Liu, P.; Chen, J. G. G. J. Am. Chem. Soc. 2017, 139,
9739.
(69) Kattel, S.; Ramirez, P. J.; Chen, J. G.; Rodriguez, J. A.; Liu, P.
Science 2017, 355, 1296.
(70) Baker, L. R.; Kennedy, G.; Van Spronsen, M.; Hervier, A.; Cai,
X.; Chen, S.; Wang, L.-W.; Somorjai, G. A. J. Am. Chem. Soc. 2012,
134, 14208.
(71) Huygh, S.; Bogaerts, A.; Neyts, E. C. J. Phys. Chem. C 2016, 120,
21659.
(72) Ji, Y. F.; Luo, Y. J. Am. Chem. Soc. 2016, 138, 15896.
(73) AlOtaibi, B.; Kong, X.; Vanka, S.; Woo, S. Y.; Pofelski, A.;
Oudjedi, F.; Fan, S.; Kibria, M. G.; Botton, G. A.; Ji, W.; Guo, H.; Mi,
Z. ACS Energy Lett. 2016, 1, 246.
S.; Fabre, B. ACS Catal. 2015, 5, 6138.
(39) Schreier, M.; Luo, J. S.; Gao, P.; Moehl, T.; Mayer, M. T.;
Gratzel, M. J. Am. Chem. Soc. 2016, 138, 1938.
(40) Sahara, G.; Abe, R.; Higashi, M.; Morikawa, T.; Maeda, K.;
Ueda, Y.; Ishitani, O. Chem. Commun. 2015, 51, 10722.
(41) Sahara, G.; Kumagai, H.; Maeda, K.; Kaeffer, N.; Artero, V.;
Higashi, M.; Abe, R.; Ishitani, O. J. Am. Chem. Soc. 2016, 138, 14152.
(42) Hansen, H. A.; Varley, J. B.; Peterson, A. A.; Norskov, J. K. J.
Phys. Chem. Lett. 2013, 4, 388.
(43) Hansen, H. A.; Shi, C.; Lausche, A. C.; Peterson, A. A.; Norskov,
J. K. Phys. Chem. Chem. Phys. 2016, 18, 9194.
(44) Fan, S. Z.; AlOtaibi, B.; Woo, S. Y.; Wang, Y. J.; Botton, G. A.;
Mi, Z. Nano Lett. 2015, 15, 2721.
(45) Rostrup-Nielsen, J. R. Catal. Today 2000, 63, 159.
(46) Kuhl, K. P.; Hatsukade, T.; Cave, E. R.; Abram, D. N.;
Kibsgaard, J.; Jaramillo, T. F. J. Am. Chem. Soc. 2014, 136, 14107.
(47) Singh, M. R.; Clark, E. L.; Bell, A. T. Phys. Chem. Chem. Phys.
2015, 17, 18924.
(48) Rosen, J.; Hutchings, G. S.; Lu, Q.; Rivera, S.; Zhou, Y.; Vlachos,
D. G.; Jiao, F. ACS Catal. 2015, 5, 4293.
(49) Kim, S. K.; Zhang, Y.-J.; Bergstrom, H.; Michalsky, R.; Peterson,
A. ACS Catal. 2016, 6, 2003.
(50) Takeda, H.; Ohashi, K.; Sekine, A.; Ishitani, O. J. Am. Chem. Soc.
(74) Bader, R. F. Acc. Chem. Res. 1985, 18, 9.
(75) Henkelman, G.; Arnaldsson, A.; Jon
2006, 36, 354.
́
sson, H. Comput. Mater. Sci.
(76) Peterson, A. A.; Abild-Pedersen, F.; Studt, F.; Rossmeisl, J.;
Nørskov, J. K. Energy Environ. Sci. 2010, 3, 1311.
(77) Asadi, M.; Kim, K.; Liu, C.; Addepalli, A. V.; Abbasi, P.; Yasaei,
P.; Phillips, P.; Behranginia, A.; Cerrato, J. M.; Haasch, R. Science 2016,
353, 467.
(78) Li, Q.; Fu, J. J.; Zhu, W. L.; Chen, Z. Z.; Shen, B.; Wu, L. H.; Xi,
Z.; Wang, T. Y.; Lu, G.; Zhu, J. J.; Sun, S. H. J. Am. Chem. Soc. 2017,
139, 4290.
(79) Wellendorff, J.; Lundgaard, K. T.; Møgelhøj, A.; Petzold, V.;
Landis, D. D.; Nørskov, J. K.; Bligaard, T.; Jacobsen, K. W. Phys. Rev.
B: Condens. Matter Mater. Phys. 2012, 85, 235149.
(80) Shi, C.; Hansen, H. A.; Lausche, A. C.; Nørskov, J. K. Phys.
Chem. Chem. Phys. 2014, 16, 4720.
(81) Kim, D.; Xie, C. L.; Becknell, N.; Yu, Y.; Karamad, M.; Chan, K.;
Crumlin, E. J.; Norskov, J. K.; Yang, P. D. J. Am. Chem. Soc. 2017, 139,
8329.
2016, 138, 4354.
(51) Hong, D. C.; Tsukakoshi, Y.; Kotani, H.; Ishizuka, T.; Kojima,
T. J. Am. Chem. Soc. 2017, 139, 6538.
(52) Rao, H.; Bonin, J.; Robert, M. ChemSusChem 2017, 10, 4447.
(53) Lian, S. C.; Kodaimati, M. S.; Weiss, E. A. ACS Nano 2018, 12,
568.
(54) Lin, Y. J.; Kapadia, R.; Yang, J. H.; Zheng, M.; Chen, K.; Hettick,
M.; Yin, X. T.; Battaglia, C.; Sharp, I. D.; Ager, J. W.; Javey, A. J. Phys.
Chem. C 2015, 119, 2308.
(55) Qiu, J.; Zeng, G. T.; Ha, M. A.; Ge, M. Y.; Lin, Y. J.; Hettick, M.;
Hou, B. Y.; Alexandrova, A. N.; Javey, A.; Cronin, S. B. Nano Lett.
2015, 15, 6177.
(82) Li, X. G.; Bi, W. T.; Chen, M. L.; Sun, Y. X.; Ju, H. X.; Yan, W.
S.; Zhu, J. F.; Wu, X. J.; Chu, W. S.; Wu, C. Z.; Xie, Y. J. Am. Chem.
Soc. 2017, 139, 14889.
(83) Li, Y. H.; Xing, J.; Chen, Z. J.; Li, Z.; Tian, F.; Zheng, L. R.;
Wang, H. F.; Hu, P.; Zhao, H. J.; Yang, H. G. Nat. Commun. 2013, 4,
2500.
(84) Tauster, S. J.; Fung, S. C.; Garten, R. L. J. Am. Chem. Soc. 1978,
100, 170.
(85) Li, Q. Y.; Wang, K.; Zhang, S. L.; Zhang, M.; Yang, H. J.; Jin, Z.
S. J. Mol. Catal. A: Chem. 2006, 258, 83.
(56) Hepel, M.; Kumarihamy, I.; Zhong, C. Electrochem. Commun.
2006, 8, 1439.
(86) Park, J. Y.; Baker, L. R.; Somorjai, G. A. Chem. Rev. 2015, 115,
2781.
(57) Guo, X.; Guo, D.; Qiu, X.; Chen, L.; Zhu, W. J. Power Sources
2009, 194, 281.
H
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(87) Chen, P. R.; Khetan, A.; Yang, F. K.; Migunov, V.; Weide, P.;
Sturmer, S. P.; Guo, P. H.; Kahler, K.; Xia, W.; Mayer, J.; Pitsch, H.;
Simon, U.; Muhler, M. ACS Catal. 2017, 7, 1197.
(88) Silvi, B.; Savin, A. Nature 1994, 371, 683.
(89) Vidal, A. B.; Feria, L.; Evans, J.; Takahashi, Y.; Liu, P.;
Nakamura, K.; Illas, F.; Rodriguez, J. A. J. Phys. Chem. Lett. 2012, 3,
2275.
I
DOI: 10.1021/jacs.8b03067
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX