You are on page 1of 28

Review

Fundamentals and Challenges


of Electrochemical CO2 Reduction
Using Two-Dimensional Materials
Zhenyu Sun,1,* Tao Ma,1 Hengcong Tao,1 Qun Fan,1 and Buxing Han2,*

Electrochemical CO2 reduction (ECR) to value-added fuels and chemicals pro- The Bigger Picture
vides a ‘‘clean’’ and efficient way to mitigate energy shortages and to lower Transforming CO2 into fuels and
the global carbon footprint. The unique structures of two-dimensional (2D) chemicals by using
nanosheets and their tunable electronic properties make these nanostructured electrocatalysis is a promising
materials intriguing in catalysis. Various 2D nanosheets are showing promise for strategy for providing a long-term
CO2 reduction, depending on the preferred reaction product (HCOOH, CO, solution to mitigating global
CH4, CH3OH, or CH3COOH). In this review, we focus on recent progress that warming and energy supply
has been achieved in using these 2D materials for ECR. We highlight procedures problems. Two-dimensional (2D)
available for tuning catalytic activities of 2D materials and describe the funda- nanosheets featuring abundant
mentals and future challenges of CO2 catalysis by 2D nanosheets. surface atoms of low coordination
and large specific surface area
appear to be beneficial for fast
INTRODUCTION
interfacial charge transfer and
Energy consumption continues to grow rapidly and is expected to reach a level facile electrochemical catalysis.
approximately two times greater than that of current consumption by 2050.1 Fossil The properties of nanosheet can
fuels will likely continue to be a major source of energy for the forseeable future. It be readily tuned by variations in
has been predicted that in the coming few decades (2010–2060), about 500 giga- their thickness, by modification
tons of CO2 will be generated from combustion of fossil fuels.2 Excessive anthropo- with heteroatoms, or by external
morphic CO2 emission has caused problems associated with resources, environ- stimuli such as electric field, strain,
ment, and climate, known as the ‘‘greenhouse effect.’’ To alleviate these adverse and illumination, providing new
effects, converting CO2 into fuels or commodity chemicals ushers in hope for an en- routes for engineering 2D
ergy transition from today’s ‘‘fossil fuel economy’’ to a sustainable ‘‘CO2 economy.’’ materials for CO2 electrocatalysis.
CO2 can be converted through diverse routes, including biochemical,3 electrochem- We describe the latest advances
ical,4 photochemical,5 radiochemical,6 and thermochemical reactions.7,8 Among in electrocatalytic CO2 reduction
these approaches, direct electroreduction of CO2 into hydrocarbons, oxygenates, using 2D materials and their
or CO is attractive because of (1) environmental compatibility coupling with car- hybrids. Relationships between
bon-free renewable energy resources (solar, tidal, and wind), (2) operating under structure and properties in these
ambient temperature and pressure, (3) control of reactions by adjusting external pa- emerging 2D electrocatalysts are
rameters such as electrolytes and applied voltages, and (4) engineering and eco- discussed. A foundational
nomic feasibility. Nonetheless, the linear molecule CO2 is stable and chemically inert summary for electrocatalytic CO2
with a low electron affinity and a large energy gap (13.7 eV) between its lowest un- conversion and possible reaction
occupied molecular orbital and highest occupied molecular orbital. CO2 transforma- pathways are highlighted.
tion is dominated by nucleophilic attacks at the carbon, which is an uphill process
requiring a substantial input of energy (750 kJ mol1 required for dissociation of
the C=O bond). Many studies of electrochemical CO2 reduction (ECR) have used wa-
ter as electrolyte, but CO2 dissolves poorly in water at acidic or neutral pH values.
Homogeneous9 and heterogeneous catalysts10 have been applied to accelerate
this kinetically slow reduction reaction. Although high selectivity is achieved in ho-
mogeneous catalysis, most homogeneous systems suffer from drawbacks, such as
high cost, toxicity, poor stability, and complex post-separation, limiting practical

560 Chem 3, 560–587, October 12, 2017 ª 2017 Elsevier Inc.


application, but which are less likely to occur in heterogeneous catalysis. Therefore,
major efforts have been made to develop heterogeneous electrocatalysts including
(bi-)metals, metal oxides and sulfides, carbon-based materials, and organic
frameworks.10

Despite advances achieved in the electrocatalytic reduction of CO2, this field still
faces challenges of (1) large overpotential (or low energetic efficiency), (2) slow elec-
tron transfer kinetics resulting in low exchange current densities, (3) unsatisfactory
selectivity, implying costly separation steps, and (4) deactivation of electrodes in
less than 100 hr, restricting practical use and technological commercialization.
Most CO2 electrocatalysts reported so far operate below 20 mA cm2, which is, how-
ever, far less than commercial electrolyzers typically operating with over 70% effi-
ciency at current densities above 200 mA cm2.11 In addition, the scarcity and
high cost of noble metal catalysts (Pt, Au, and Pd) are obstacles to their large-scale
applications. From these scenarios, heightened research efforts have focused on the
design and synthesis of novel, cost-effective, and robust electrocatalysts that can
reduce CO2 at high rates with a minimum amount of overpotential.

Research on two-dimensional (2D) nanosheets has burgeoned in the past decade


after the isolation of graphene. A large range of 2D compounds, including group
IV and V elements, metals, chalcogenides, oxides, halides, nitrides, carbides, hy-
droxides, hydrides, phosphates, phosphonates, and covalent organic frameworks,
have been reported. The family of 2D materials display insulating, semiconducting
(with direct and indirect band gaps ranging from UV to infrared throughout the
visible region), and metallic behavior. Of particular interest is that enhanced and
novel electronic and optical properties emerge when layered materials are thinned
to a single layer or to few-layer flakes (referred to as ‘‘nanosheets’’) as a result of
quantum confinement. Semiconducting 2D materials exhibit strong enhancement
of Coulomb interactions among charge carriers and defects, leading to longer
lived excitons and trions than those in bulk.12 Another important feature is the spe-
cific surface area created with reduced dimensionality. In addition, the physico-
chemical properties of 2D materials can be effectively tuned in diverse ways,
such as doping, heterostructuring, chemical functionalization, and alloying.13
Layered materials with interlayer cohesive energies of 200 meV or less per atom
can be exfoliated down to nanosheets via top-down methods (e.g., mechanical
or chemical exfoliation to overcome interlayer van der Waals forces).14 Single-
and few-layer nanosheets can also be grown through bottom-up approaches
(e.g., chemical vapor deposition, wet-chemical synthesis). These strategies, in
combination with characterization using spectroscopic techniques and theoretical
calculations, open new opportunities to engineer nanosheets for catalysis and en-
ergy applications.

Catalysis with 2D nanosheets has sparked increasing interest because such sheets
have unique structural and electronic properties, albeit this area is still in its in- 1StateKey Laboratory of Organic-Inorganic
fancy.15 Ultrathin nanosheets with open double-sided surfaces possess abundant Composites, Beijing University of Chemical
Technology, Beijing 100029, People’s Republic of
exposed surface atoms that can easily escape from the respective lattice to form va-
China
cancy-type defects. Such vacancy defects, together with the structural disorder that 2BeijingNational Laboratory for Molecular
usually appears in nanosheets, can reduce the coordination number of the surface Sciences, Key Laboratory of Colloid and Interface
atoms, leading to dangling bonds and enhancing catalytic performance. An increase and Thermodynamics, Institute of Chemistry,
Chinese Academy of Sciences, Beijing 100190,
in low-coordinated surface sites promotes chemisorption of reactants. Control of People’s Republic of China
vacancy defects allows one to modify the electronic structure and thereby tailor cor- *Correspondence: sunzy@mail.buct.edu.cn (Z.S.),
responding catalytic activities. Equally importantly, atomic sites at the edges of hanbx@iccas.ac.cn (B.H.)
nanosheets with low coordination can afford interesting catalytic properties. The https://doi.org/10.1016/j.chempr.2017.09.009

Chem 3, 560–587, October 12, 2017 561


well-defined structure of 2D nanosheets also provides an applicable platform for
investigation of catalytic mechanisms at the atomic level.

2D nanosheets with a substantial number of exposed active sites appear to be bene-


ficial for fast interfacial charge transfer and facile electrochemical catalysis. Very
promising results have already been realized in developing nanosheets for clean en-
ergy conversion (involving water, hydrogen, and oxygen) competitive with noble
metal electrocatalysts.16 The properties of nanosheets can be readily tuned by vary-
ing their thickness, and by external stimuli such as electric field, strain, illumination,
providing new routes to engineer 2D materials for CO2 (photo)electrocatalysis.
Nanosheets can be chemically doped with heteroatoms or surface decorated with
catalytic nanoparticles or heterostructures to direct particularly desired CO2 chem-
istries, and such surface modifications can be used to design multiple mechanistic
catalytic steps. Further, the limiting specific surface area imparted by nanosheets of-
fers a promise of limiting catalytic activity. Recently, a diversity of 2D materials
ranging from metals, metal oxides to chalcogenides, and even metal-free catalysts
have been demonstrated to have great potential for electrocatalytic CO2 reduc-
tion.17–22 In the following sections, we review recent advances in electrocatalytic
CO2 reduction using 2D materials. Relationships between structure and properties
in these emerging 2D electrocatalysts are discussed. A foundational summary for
electrocatalytic CO2 conversion and possible reaction pathways are highlighted.
An outlook of future prospects for conversion of CO2 into fuels by nanosheet electro-
catalysis is described.

A PERSPECTIVE OF ELECTROCHEMICAL CO2 REDUCTION


Heterogeneous electrochemical reduction of CO2 occurs at electrode-electrolyte in-
terfaces. Three steps are mainly considered when modeling these heterogeneous
catalytic processes:

(1) Chemical adsorption of CO2 on the surface of a catalyst (cathode).


(2) Electron transfer and/or proton migration to break C–O bonds and/or form
C–H bonds.
(3) Rearrangement of product species followed by desorption from electrode
surface and diffusion into electrolyte.

Aqueous electrocatalytic reduction of CO2 is particularly challenging because of oc-


currences of competitive proton reduction, leading to poor selectivity for carbona-
ceous products. Another issue is the low solubility of CO2 in water (0.034 M),
adversely affecting such diffusion-controlled reactions. There are several ap-
proaches that can be used to enhance CO2 dissolution, involving (1) the use of
nonaqueous solutions; (2) low-temperature working conditions; (3) increasing CO2
partial pressure. The CO2 pressure also influences the thermodynamics of the
ECR. Some metal catalysts that are inactive at atmospheric pressure, however, can
reduce CO2 to CO, formate, or hydrocarbons at elevated pressures. Alternatively,
the rate at which CO2 flows into the catholyte below 50 mL CO2 min1 affects cata-
lytic activity. Higher CO2 flow rates yield larger reaction rates, faradic efficiencies,
and product selectivity.23

Parameters that have been used as a gauge for evaluating the performance of a cata-
lyst include the following: (1) onset potential or overpotential (h, i.e., the difference
between the thermodynamic and actual electrode reduction voltages); (2) current
density (current divided by the geometric surface area of working electrode); (3)

562 Chem 3, 560–587, October 12, 2017


faradic efficiency (EFaradic = anF=Q, where a is the number of electrons transferred, n
is the number of moles for a given product, F is Faraday’s constant [96,485 C mol1],
and Q is all the charge passed throughout the electrolysis process); (4) energetic ef-
ficiency (Eenergetic = ðE 0 =E 0 + hÞ3EFaradic , where E0 is standard potential); (5) turnover
frequency (TOF), a measure of per-site activity of catalysts, such as TOF (s1) for CO
formation,

i0 ðA cm2 Þ 3 EFaradic ðCOÞ


;
½active  site densityðsites cm2 Þ 3 ½1:602 3 1019 C=e  3 2e =CO2 
and (6) Tafel slope. A Tafel plot relates to overpotential verses logarithm of the par-
tial current density of a specific product. The Tafel slope derived is an indicator of
reaction pathways and the rate-determining step (RDS). A slope of 118 mV
dec1 implies that the RDS for CO2 reduction is the initial step of the CO2,– gener-
ation, whereas the RDS is the chemical step following the fast one-electron pre-equi-
librium when the Tafel slope is 59 mV dec1.24 A good catalyst should minimize the
activation barrier for CO2 reduction in relation to proton reduction, driving CO2
reduction selectively at low overpotential (i.e., high energy efficiency) with satisfac-
tory reaction rates (i.e., high turnover number).

The total cell voltage required for ECR includes potentials for both anodic and
cathodic processes (Ecell = Eanode  Ecathode). Real-world ECR and oxygen evolution
reaction (OER) catalysts require overpotentials of several hundred millivolts to attain
satisfactory reaction rates. As a consequence, cell voltages usually exceed the reac-
tion formal potentials. Note that the anode catalyst plays an important role in CO2
conversion because it consumes almost half of the electrical input. Therefore,
improving the anode efficiency and lowering OER overpotentials can reduce the
overall energy requirements for CO2 reduction, making this carbon mitigation strat-
egy more practical.

Thermodynamics and Kinetics of CO2 Reduction


Direct reduction of CO2 involves one-electron transfer to form CO2 anion radical
(CO2,–). However, this step is highly unfavorable, having a negative formal redox po-
tential 1.97 V (versus standard hydrogen electrode [SHE]) in an aprotic solvent
(N,N0 -dimethylformamide [DMF]) and 1.90 V in water (pH 7). Catalytic strategies
have thus been developed to bypass the formation of CO2,– through proton-assis-
ted multiple-electron transfer to reduce CO2 at lower energetic costs. Depending
on the number of electrons and protons transferred, CO2 can be reduced to 16
different products, including CO, oxalic acid (H2C2O4) (C2O42, basic medium), for-
mic acid (HCOOH) (HCOO, basic medium), formaldehyde (HCHO), C, methanol
(CH3OH), methane (CH4), ethylene (C2H4), ethanol (C2H5OH), ethane (C2H6), and
n-propanol (C3H7OH) (Equations 1–20; CO2 reduction potentials versus SHE at pH
7). The competing proton reduction feeding hydrogen evolution is a two-electron
process (Equation 21).4,10,25

CO2 + 2H + + 2e /HCOOH E 0 redox =  0:610 V (Equation 1)


CO2 + 2H2 O + 2e /HCOOH + OH E 0 redox =  1:491 V (Equation 2)
CO2 + 2H + + 2e /CO + H2 O E 0 redox =  0:530 V (Equation 3)
CO2 + 2H2 O + 2e /CO + 2OH E 0 redox =  1:347 V (Equation 4)
2CO2 + 2H + + 2e /H2 C2 O4 E 0 redox =  0:913 V (Equation 5)
2CO2 + 2e /C2 O2
4 E 0 redox =  1:003 V (Equation 6)
+ 
CO2 + 4H + 4e /HCHO + H2 O E 0 redox =  0:480 V (Equation 7)
CO2 + 3H2 O + 4e /HCHO + 4OH E 0 redox =  1:311 V (Equation 8)

Chem 3, 560–587, October 12, 2017 563


CO2 + 4H + + 4e /C + 2H2 O E 0 redox =  0:200 V (Equation 9)
CO2 + 2H2 O + 4e /C + 4OH E 0 redox =  1:040 V (Equation 10)
CO2 + 6H + + 6e /CH3 OH + H2 O E 0 redox =  0:380 V (Equation 11)
CO2 + 5H2 O + 6e /CH3 OH + 6OH E 0 redox =  1:225 V (Equation 12)
CO2 + 8H + + 8e /CH4 + 2H2 O E 0 redox =  0:240 V (Equation 13)
CO2 + 6H2 O + 8e /CH4 + 8OH E 0 redox =  1:072 V (Equation 14)
2CO2 + 12H + + 12e /C2 H4 + 4H2 O E 0 redox =  0:349 V (Equation 15)
2CO2 + 8H2 O + 12e /C2 H4 + 12OH E 0 redox =  1:177 V (Equation 16)
2CO2 + 12H + + 12e /C2 H5 OH + 3H2 O E 0 redox =  0:329 V (Equation 17)
2CO2 + 9H2 O + 12e /C2 H5 OH + 12OH E 0 redox =  1:157 V (Equation 18)
2CO2 + 14H + + 14e /C2 H6 + 4H2 O E 0 redox =  0:270 V (Equation 19)
3CO2 + 18H + + 18e /C3 H7 OH + H2 O E 0 redox =  0:310 V (Equation 20)
2H + + 2e /H2 E 0 redox =  0:42 V (Equation 21)
The effects of conditions such as the types of electrocatalysts (composition, size,
shape, oxidation state, and crystallographic structure), electrolytes (cation, anion,
concentration, and pH), temperature, pressure, and applied potential can be super-
imposed onto the reaction scheme. A reduction reaction with a more positive E0 is
thermodynamically more favorable according to the relationship DG = nFE0,
where n is the number of electrons transferred during the redox reaction and F is
the Faraday constant. On the basis of E0, the reductions of CO2 toward hydrocarbon
or alcohol products should be thermodynamically more favorable than CO,
HCOOH, HCHO, and H2 production. However, that is actually not the case because
in addition to a thermodynamic barrier, CO2 reduction also has a kinetic depen-
dence on the concentration of available protons in solution. This implies that a
preferred catalyst likely entails catalytic sites transferring electrons that are close
to sites providing protons. The hydrogenation of adsorbed C1 intermediates is kinet-
ically easier than the formation of C–C bonds, limiting the rate and selectivity of C2
and higher hydrocarbon production. The maximum faradic efficiency for C2H4, the
simplest C2 product, was reported to be 60%,26 whereas that of C3 product
(C3H7OH) is no more than 30%.27,28

In the multi-electron transfer reduction of CO2, the adsorption energies of reaction


intermediates seem to follow linear scaling relationships, thus limiting catalytic effi-
ciency.29,30 In order to improve catalytic activity, breaking such linear scaling rela-
tions is required, which can be realized by: (1) reducing coordination numbers, (2)
doping with p-block elements, (3) introducing oxophilic sites, and (4) coating the
catalyst surface with active ligands.30

Electrolytes
In CO2 electroreduction, electrolytes provide a medium to transfer coupled elec-
trons and protons (e/H+). The type and concentration of electrolytes affect catalyst
activity and selectivity. It should also be noted that trace impurities (metal and
organic) from electrolytes can cause the deactivation of catalytic sites and affect
catalyst performance.

Aqueous Solutions
Most studies in electrochemical CO2 reduction have focused on weakly acidic or
alkaline CO2-saturated aqueous electrolytes containing inorganic salts with
HCO3, SO42, or Cl anions and alkali metal cations (e.g., Na+ and K+).

Cationic species alter the outer Helmholtz plane potential and can also influence the
hydrogen coverage on the electrode by delivering water molecules from their

564 Chem 3, 560–587, October 12, 2017


solvation shell to the electrode. Meanwhile, cations can affect relative concentra-
tions of charged species (such as anion radical intermediates) close to the electrode,
thus affecting product selectivity and current density.31 The rate of CO2 electrore-
duction to CH3OH was found to increase with the surface charge of electrolyte cat-
ions in the following order: Na+ < Mg2+ < Ca2+ < Ba2+ < Al3+ < Zr4+.32 Cation size
was shown to influence product selectivity. Large cations favor the formation of
HCOOH on a Hg electrode, C2H4 on a Cu electrode,33 and CO on a Ag electrode.31
Large cations were also demonstrated to suppress H2 evolution.31,34 H2 evolution
prevailed over CO2 reduction in Li+ electrolytes, whereas CO2 reduction was favor-
able in Na+, K+, and Cs+ solutions.

Anionic species (i.e., Cl, ClO4, SO42, HCO3, and H2PO4) with different buffer
capacities affect the local pH at the electrode and hence the nature and selectivity
of products formed.35 A high local pH inhibits H2 evolution because of low proton
concentration. C2H4 and alcohols are favored in KCl, K2SO4, KClO4, and dilute
KHCO3 solutions under a non-equilibrium local high pH, whereas CH4 is preferen-
tially generated in KH2PO4 and concentrated KHCO3 solutions.35 The CH4/C2H4
product ratio increases with increasing HCO3 concentration. The effects of halo-
gens on ECR were also studied with a Cu mesh electrode in aqueous electrolytes
of 3 M KCl, KBr, and KI.36 It was found that the bond between adsorbed halides
(e.g., Br, Cl, or I) and Cu facilitated electron transfer from these anions to the
vacant orbital of CO2 and promoted CO2 reduction. A stronger halide adsorption
to the electrode made CO2 more intensely restrained, thus enhancing reduction
current. In addition, specifically adsorbed halides inhibited proton adsorption
and induced a higher hydrogen overvoltage. A recent study from Strasser and col-
leagues revealed that different halogen anions in electrolytes resulted in distinct
product distributions for ECR on a Cu electrode.37 Addition of both Cl and Br
increased CO selectivity. On the contrary, the presence of I decreased the selec-
tivity toward CO, whereas CH4 formation was six times higher than with halide-free
electrolytes.

Organic Electrolytes
Nonaqueous electrolytes were applied for ECR to increase CO2 solubility and also to
suppress hydrogen evolution, thereby improving faradic efficiency. The solubility of
CO2 in DMSO and acetonitrile (AN) is approximately four times that in water,
whereas its solubility in CH3OH and propylene carbonate is about five and eight
times greater, respectively. DMF is a good solvent for CO2, with 20 times larger sol-
ubility than in water at ambient conditions. Nonaqueous solvents (DMF, DMSO, and
AN) were demonstrated to facilitate CO2,– dimerization with adsorbed CO2 mole-
cules to produce C2O42 on Sn, In, Pb, and Hg electrodes,38 whereas the solvent
of CH3OH increases CH4 selectivity on a Cu electrode.39 Manipulating the amount
of water in an organic electrolyte to tune proton availability can effectively control
faradic efficiency and selectivity for CO2 reduction.

Ionic Liquids
The use of ionic liquids (ILs), especially those featuring imidazolium cations, is bene-
ficial to: (1) lowering CO2 reduction overpotential, most likely by complexation to
reduce the energy of the CO2,– intermediate; (2) inhibiting the hydrogen evolution
reaction (HER); and (3) increasing the selectivity for CO formation.40,41

The ILs that are commonly used in ECR include 1-ethyl-3-methylimidazolium


tetrafluoroborate (EMIM-BF4), 1-butyl-3-methylimidazolium hexafluorophosphate
(BMIM-PF6), 1-ethyl-3-methylimidazolium trifluoromethanesulfonate (EMIM-TFO),

Chem 3, 560–587, October 12, 2017 565


1-butyl-3-methylimidazolium triflate (BMIM-OTF), and 1-ethyl-3-methyl-imidazo-
lium triflate (EMIM-OTF), among others.42 BMIM-BF4 was shown to exhibit the high-
est current density with large faradic efficiency for CH4 formation on an N-doped
graphene-like electrode among five different ILs, BMIM-BF4, BMIM-PF6, BMIM-
TFO, 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (BMIM-TF2N),
and 1-butyl-3-methylimidazolium dicyanamide (BMIM-DCA).43 Tuning the cation
structure of imidazolium-BF4 ILs showed that the C4 and C5 protons on the imidazo-
lium ring are vital for efficient CO2 reduction to yield CO on a Ag electrode.44

In order to overcome the relatively high cost and viscosity of ILs, they are usually dis-
solved in organic solvents or water for use as electrolytes. The concentrations of ILs
reported in the literature vary from as low as millimolar to neat ILs. Finding an opti-
mum ratio of ILs remains an important unanswered question. The ratio between an IL
and H2O affects the pH and viscosity of the electrolyte. Using an electrolyte
comprising 75 mol % water and 25 mol % EMIM-BF4 resulted in an enhanced CO2
reduction current density that was approximately five times higher than that of
pure EMIM-BF4.45 The decrease in pH caused by the formation of hydroxyl ions
(such as [BF3OH], [BF2(OH)2], or [BF(OH)3]) from EMIMBF4 hydrolysis led to a
greater proton availability, thus enhancing the rate of CO2 reduction. Likewise,
[Bmim]BF4–H2O binary electrolytes were observed to afford both large current den-
sity and high CH4 selectivity when the water content in the mixture was less than 5
wt %.43 High faradic efficiencies for CO formation were reported for electrolytes con-
taining R20 mol % EMIM-TFO at potentials more negative than 1.46 V versus Ag/
AgCl over a Ag electrode. The CO faradic efficiency reached 95.6% G 6.8% in a 50
mol % IL electrolyte.46 Despite IL-based electrolytes providing benefits of high CO2
solubility and low overpotentials, their high cost and poor stability in the presence of
H2O need to be addressed for practical CO2 electrolyzers.

Solid-Oxide Electrolytes
Molten carbonate or solid-oxide electrolytes are used in electrolyzers operating at
high temperatures (>673 K) for the co-electrolysis of H2O and CO2. Such electrolyte
is typically zirconia stabilized by yttrium oxide.47 A BaCeO0.5Zr0.3Y0.16Zn0.04O3-
d electrolyte also showed utility in converting CO2 into CO and CH4 in the presence
of H2 and H2O.48

Electrochemical Cells
Electrolyzer design has a profound effect on mass transport. There is no standard exper-
imental setup or methodology for ECR. A variety of electrolyzers and flow cells have
been developed. Two representative lab-scale prototypes are presented below.

H-Type Cells
A typical and commonly used lab-scale electrochemical cell (H-type) in ECR encom-
passes two compartments and three electrodes (i.e., working electrode, reference
electrode, and counter electrode).49 The cathodic and anodic chambers are sepa-
rated by a proton-conducting (e.g., Nafion 117) or an anion-exchange polymer
membrane. This setup allows ionic conductivity, preventing the transport of
cathodic products to the anode where they can be oxidized while keeping the elec-
trodes in close proximity. The cathode compartment can be directly connected to a
mass spectrometer or gas chromatography spectrometer for product analysis.

Continuous-Flow Cells
Microfluidic flow cells have been recently used in ECR. This type of setup comprises
two electrodes separated by a flowing liquid electrolyte50 and can test both cathode

566 Chem 3, 560–587, October 12, 2017


and anode performance by using an external reference electrode. Gas diffusion
electrodes (GDEs) are often used for better control of the three-phase boundary
where reactions occur. This continuous-flow reactor presents advantages for online
product detection. Other versions of flow reactors include filter-press and parallel-
plate type electrochemical cells.51 A scale-up ‘‘trickle bed’’ cell with a two-phase
flow of catholyte was designed by Li and co-workers, which can achieve a current ef-
ficiency up to 91% of the efficiency obtained in a lab prototype.52

Possible Reaction Pathways


Efforts have been devoted to exploring the reaction mechanism of CO2 activation
and reduction both theoretically and experimentally.25,53 However, most mecha-
nisms are hypothesized on the basis of the results of either the Tafel slope or den-
sity functional theory (DFT) calculations. Further experimental work combining DFT
calculations and in situ spectroscopy is needed to fully elucidate mechanistic
pathways.

Formation of Formate and Formic Acid


HCOO or HCOOH evolves possibly via an intermediate that binds to a transition-
metal electrode either through one oxygen atom (monodentate) or two oxygen
atoms (bidentate) (Figure 1A and top route in Figure 1B).25 This intermediate can
form through reaction with *H via CO2 insertion into the metal–hydrogen bond or
through direct protonation with H+ from solution. An alternative pathway is pre-
sumed to proceed via a CO2,– radical that reacts with a neighboring water or proton
to generate HCOO or HCOOH on p-block metals, such as In, Sn, Pb, and Hg (mid-
dle route in Figure 1B).25,54 In addition, the presence of HCO3 ions was found to
enhance HCOO production.54

Bocarsly and co-workers recently proposed that the formation of a surface-bound


tin carbonate is a key chemical step involved in the reduction of CO2 to HCOO on
Sn electrodes (Figure 1C).55 Specifically, a two-electron reduction of the electrode
from a native SnO2 to a SnII oxyhydroxide, which is the catalytic resting state, oc-
curs before the reduction of CO2. This species then reacts with CO2 to form a sur-
face-bound carbonate. The carbonate undergoes transfers of two electrons and
one proton to form HCOO—, which desorbs to return the surface to the SnII
oxyhydroxide.

Formation of Carbon Monoxide


The electrochemical reduction of CO2 to CO is a simple two-electron transfer pro-
cess (Figure 2). By transferring a concerted proton-electron (H+/e) from solution
to adsorbed species, a CO2 molecule is reduced to a carboxyl intermediate
*COOH. A second H+/e can subsequently attack the oxygen atom (OH) in the
*COOH to form H2O (l) and CO, which then desorbs from the electrode.53 The con-
version of COOH* to *CO was found to take place readily. But the initial step of CO2
conversion to *COOH is inhibited by weak COOH binding. Likewise, the last step of
CO desorption from the surface is hindered by strong CO binding. These two steps
were calculated to be rate limiting.56 Another possible route to generate *COOH is
through a decoupled electron and proton transfer, involving the formation of a
CO2,2– radical that is adsorbed at electrode surfaces (bottom route in Figure 2).
The two pathways were suggested to show different pH dependences. Increases
in pH or pressure likely promote the yield of *COOH over *CO.57 Metals such as
Au, Ag, Zn, and Pd bind *COOH tightly enough for further reduction to produce a
*CO intermediate. The *CO is weakly bound to the electrode surface, and CO de-
sorbs from the electrode as a major product.

Chem 3, 560–587, October 12, 2017 567


Figure 1. Possible Reaction Pathways for the Electrocatalytic Reduction of CO2 to HCOO or HCOOH
(A) Monodentate or bidentate intermediate route. 25
(B) CO 2 ,– radical intermediate route. 25,54
(C) Surface-bound carbonate intermediate route. 55

Formation of Formaldehyde, Methanol, and Methane (C1 Products)


*CO is likely a common intermediate for the production of HCHO, CH3OH, and CH4.25
DFT calculations of ECR on Cu (211) surfaces by Peterson et al. showed that the thermo-
dynamically most favorable pathway involves the initial formation of *CO and subse-
quent hydrogenation to *HCO, *H2CO (desorb as HCHO), and *H3CO (methoxy) (Fig-
ure 3A).58 CH3OH can be formed by *H3CO reduction. The methoxy intermediate can
be alternatively reduced to CH4 and *O. The *O is finally reduced to H2O. However,
C–H bond formation coupled with C–O dissociation to yield CH4 was calculated to
be kinetically prohibitive with a barrier of 1.21 eV, much higher than the barrier of
0.15 eV for CH3OH production.53 Another possible reaction path, as proposed by
Nie et al.,53 is through a *COH intermediate, which is reduced to an adsorbed C (Fig-
ure 3B). Such graphitic carbon species has been observed by in situ X-ray photoelectron
spectroscopy and Auger electron spectroscopy during ECR on Cu electrodes.59 The
surface carbon can be further reduced to *CH, *CH2, *CH3, and finally to CH4. The

568 Chem 3, 560–587, October 12, 2017


Figure 2. Possible Reaction Pathways for the Electrocatalytic Reduction of CO2 to CO
See Kortlever et al. 25

oxophilicity of the surface, as determined by the binding energy of Oads, could play an
important role in the reaction for determining selectivity between CH4 and CH3OH.
Only CH3OH was detected on the poorly oxophilic Au, whereas for the strongly
oxophilic Fe, only CH4 was identified.60 From this scenario, engineering a surface by
modifying the binding energy for Cads and Oads to favor or disfavor C–O bond breakage
allows one to improve selectivity toward one product.

Formation of Acetaldehyde, Ethylene, and Ethanol (C2 Products)


C2H4, acetaldehyde (CH3CHO), and C2H5OH have been produced at less negative
potentials than CH4, especially at high (alkaline) pH on Cu-based catalysts.26,61 The
formation of these C2 products is a complex process with many possible pathways
involving both electrochemical and chemical steps, which is still the focus of debate.
In a ‘‘carbene’’ mechanism, C2H4 is formed via either coupling of two *CH2 species53
or CO insertion in a Fischer-Tropsch-like step (Figure 4A), which has also been sug-
gested to be the pathway for the formation of C2H5OH.62 An alternative pathway in-
volves *CO dimerization with an electron transfer to form a *C2O2 key intermediate,
which is subsequently protonated to *CO–COH (Figure 4B).25,63 The intermediate of
ethenol can be further reduced to C2H4, or CH3CHO and then C2H5OH, with C2H4
being the kinetically favored product on Cu surfaces.64

Formation of Acetic Acid and Acetate


Another C2 product, CH3COOH or CH3COO, has been reported from ECR on Cu-
based electrodes65 and also on an N-doped nanodiamond/Si rod array electrode.66
The formation of acetic acid was attributed to a nucleophilic attack from adsorbed
CO2,– on the reduced –CH3 species (Figure 5).65,67 A faradic efficiency of 91.2%–
91.8% was achieved for CH3COO formation at 0.8 to 1.0 V (versus RHE) on
N-doped nanodiamond.66 Electrokinetic data in combination with in situ infrared
spectroscopy suggested the coupling of two adsorbed CO2,– to generate OOC–
COO, which is possibly an important intermediate to form acetate.

Formation of n-Propanol (C3 Product)


Electrochemical reduction of CO2 to yield long-chain products, such as C3H7OH, re-
mains a challenge. To date, only graphene/ZnO/Cu2O hybrid surfaces are known to

Chem 3, 560–587, October 12, 2017 569


Figure 3. Possible Reaction Paths for Electrocatalytic CO2 Reduction to Produce HCHO, CH3OH, and CH4 on Cu Electrodes
(A) A thermodynamic analysis. 58
(B) A combined thermodynamic and kinetic analysis.53

catalyze this process with maximum faradic efficiency of about 30%.28 It was hypoth-
esized that an adsorbed C2 intermediate could undergo intermolecular C–C
coupling with an adjacent C1 intermediate, followed by proton/electron transfers
to form propionaldehyde (CH3CH2CHO) (Figure 6).62 CH3CH2CHO is further
reduced to C3H7OH.

2D NANOSHEET CATALYSIS OF CO2 ELECTROREDUCTION


Searching for catalysts with low cost but high activity and selectivity remains
a challenge to promote the kinetically sluggish CO2 reduction process for
large-scale production. Recently, 2D nanosheets featuring a large fraction of
low-coordinated surface atoms have gained much attention for CO2 reduction
(Table 1).

570 Chem 3, 560–587, October 12, 2017


Figure 4. Possible Reaction Paths for Electrocatalytic CO2 Reduction to Produce C2H4, CH3CHO, and C2H5OH
(A) Coupling of two *CH 2 species or CO insertion in a Fischer-Tropsch-like step. 62
(B) *CO dimerization. 63

Formic Acid and Formate Selectivity


Metal Nanosheets
Xie and colleagues recently prepared freestanding 4-atom-thick Co sheets with and
without surface Co oxide by using a ligand-confined growth strategy (Figure 7).18
The authors showed an onset potential (Eonset) of 0.68 V (versus saturated calomel

Chem 3, 560–587, October 12, 2017 571


Figure 5. Possible Reaction Paths for Electrocatalytic CO2 Reduction to Produce CH3COO or CH3COOH
See Genovese et al. 65

electrode [SCE]) for formate production in a CO2-saturated 0.1 M Na2SO4 solution on


partially oxidized Co nanosheets (z0.84 nm), with an overpotential of only 0.07 V. A
high current density (j) of 10.59 mA cm2 and a formate faradic efficiency (Efaradic) of
90.1% were obtained at a low overpotential (0.24 V) over 40 hr, competitive with a
high-cost Pd-based catalyst with a j of 3.92 mA cm2 and Efaradic of 88% at an over-
potential of 0.15 V in 0.5 M aqueous NaHCO3.82 Compared with bulk Co, these
atomically thin layer catalysts achieved a 260-fold increase in catalytic activity . This
enhancement could result from a large number of active sites provided by increased
electrochemical surface area and the presence of a surface Co oxide with higher
intrinsic activity. The Tafel slope of these partially oxidized Co layers was estimated
to be about 59 mV dec1, suggesting a fast pre-equilibrium involving 1e transfer
to form CO2,– and subsequent slower chemical reactions as the RDS. It was thus in-
ferred that Co atoms confined in atomically thin layers accelerated CO2 activation
by stabilizing the CO2,– intermediate more effectively than the bulk counterpart.

An even cheaper Bi nanosheet catalyst was demonstrated to catalyze electrochem-


ical CO2 reduction to HCOO in 0.5 M KHCO3 solution with a much larger faradic
efficiency (92%) at 1.5 V (versus SCE) than that of commercial Bi powder
(55%).19 Such Bi nanosheets also exhibited 100 mV lower onset potential (1.3 V
versus SCE) but about five times larger current density (3.7 mA cm2 at 1.8 V versus
SCE) than commercial Bi.

Ultrathin metal layers can be highly active for CO2 reduction but can also be prone to
oxidation, causing a rapid decay in cyclability. To solve this issue, confining metal
sheets in graphene seems to be a good choice (Figure 8). Indeed, metallic Sn quan-
tum sheets, confined in few-layer graphene, retained stability in air up to 570 C,
whereas Sn nanoparticles (15 nm in diameter) without the protection of a graphene
interlayer started to be oxidized at  200 C.68 Such Sn quantum sheets displayed a
more positive onset potential (0.85 V versus SCE), larger current density (21.1 mA
cm2 at 1.8 V versus SCE), and better formate faradic efficiency (89%) than 15 nm

572 Chem 3, 560–587, October 12, 2017


Figure 6. Possible Reaction Paths for Electrocatalytic CO2 Reduction to Produce C3H7OH
See Hori et al. 62

Sn nanoparticles physically mixed with graphene (j  10.6 mA cm2; Efaradic


 61.4%), 15 nm Sn nanoparticles (j  8.4 mA cm2; Efaradic  59.3%), and bulk Sn
(j  1.6 mA cm2; Efaradic  44.5%). This enhancement could be due to a combination
of high electrochemical surface area favoring CO2 adsorption, rapid rate-limiting
electron transfer from CO2 to CO2,– benefiting from conductive graphene, and
low Sn-Sn coordination numbers enabling stabilization of CO2,–.

Metal-Oxide Nanosheets
Enhanced electrocatalytic activity is expected upon reduction in thickness as a result
of corresponding increases in active sites and electrical conductivity. This has been
confirmed by a recent work in which 1.72-nm-thick Co3O4 layers exhibited a CO2
reduction current density of 0.68 mA cm2 at 0.88 V (versus SCE), 1.5 and 20 times
higher than those of 3.51-nm-thick Co3O4 layers and bulk Co3O4, respectively.69

DFT calculations, along with CO2 adsorption isotherms and X-ray absorption fine
structure spectroscopy, demonstrated that oxygen (II) vacancies facilitated CO2
adsorption and also HCOO– desorption. Such oxygen (II) vacancies played a role
in lowering the rate-limiting activation barrier via stabilizing the HCOO–* intermedi-
ate, as reflected by the decrease of onset potential from 0.81 to 0.78 V (versus
SCE) and Tafel slope from 48 to 37 mV dec1.70 Co3O4 nanosheets (z0.84 nm)
rich in oxygen (II) vacancies exhibited a current density of 2.7 mA cm2 with
85% formate selectivity in 40 hr, showing promise in CO2 reduction catalysis.

Mesoporous SnO2 nanosheets were also reported as an efficient, selective, and du-
rable electrocatalyst for CO2 reduction.22 A high partial current density for HCOO
(45 mA cm2) at a moderate overpotential (0.88 V) with large faradic efficiency
(87% G 2%) was achieved (Figure 9), outperforming many GDEs in aqueous media.
This superior performance was correlated to the porous hierarchical structure, which
provided a large surface area and facilitated charge and mass transfer.

Heteroatom-Doped Graphene and Graphene-Based Hybrids


Despite that pristine graphene has low activity requiring high CO2 activation free en-
ergy (over 1.3 eV),83 doping graphene with heteroatoms (such as B, N, etc.) can
reduce the CO2 adsorption barrier markedly to catalyze the electrochemical reduc-
tion of CO2 to C1 fuels. Boron-doped graphene has been used as a metal-free cata-
lyst for ECR in 0.1 M KHCO3 aqueous solution.71 HCOO with a faradic efficiency of
66% was attained at 1.4 V (versus SCE). DFT calculations suggested that the pres-
ence of boron introduces asymmetric charge and spin density distribution
throughout the ground state geometry, resulting in a high spin density. Those B
and C atoms with positive spin densities were speculated to be catalytically active
sites, favoring the chemisorption of protonated CO2 (COOH*). An even higher

Chem 3, 560–587, October 12, 2017 573


Table 1. Summary of 2D-Based Materials Reported for Electrocatalytic CO2 Reduction
Catalysts Electrolytes Ja (mA$cm2) Onset Potential (V) or hb Main Products (FE)c Stability Reference
(V)
Partially oxidized Co 0.1 M Na2SO4 10.59 mA cm2 @ h: 0.24 V @ HCOO: 90.1% @ 40 hr Gao et al.18
4-atomic-layers 0.85 V (versus SCE) 10.59 mA cm2 0.85 V (versus SCE)
BiOCl-derived 0.5 M KHCO3 3.7 mA cm2 @ 1.8 V onset potential: 1.3 V HCOO: 92% @ 7 hr Zhang
Binanosheets (versus SCE) (versus SCE) 1.5 V (versus SCE) et al.19
Sn sheets confined in 0.1 M NaHCO3 21.1 mA cm2 @-1.8 V onset potential: HCOO: 89% @ 50 hr Lei et al.68
graphene (versus SCE) 0.85 V (versus SCE) 1.8 V (versus SCE)
Ultrathin Co3O4 0.1 M KHCO3 0.68 mA cm2 @ onset potential: HCOO: 64.3% @ 20 hr Gao et al.69
nanosheets 0.88 V (versus SCE) 0.82 V (versus SCE) 0.88 V (versus SCE)
Oxygen-deficient 0.1 M KHCO3 2.7 mA cm2 @ onset potential: HCOO: 87.6% @ 40 hr Gao et al.70
Co3O4 atomic layers 0.87 V (versus SCE) 0.78 V (versus SCE) 0.87 V (versus SCE)
Mesoporous SnO2 0.5 M NaHCO3 50 mA cm2 @ 1.6 V h: 0.88 V @ HCOO: 89% @ 24 hr Li et al.22
nanosheets (versus Ag/AgCl) 45 mA cm2 1.6 V (versus
Ag/AgCl)
B-doped graphene 0.1 M KHCO3 2.0 mA cm2 @ onset potential: 1.1 V HCOO: 66% @ 4 hr Sreekanth
1.4 V (versus SCE) (versus SCE) 1.4 V (versus SCE) et al.71
N-doped graphene 0.5 M KHCO3 7.5 mA cm2 @ onset potential: 0.3 V HCOO: 73% @ 12 hr Wang et al.20
0.84 V (versus RHE) (versus RHE) 0.84 V (versus RHE)
SnS2/rGO 0.5 M NaHCO3 13.7 mA cm2 @ h: 0.68 V @ HCOO: 84.5% @ 14 hr Li et al.72
nanosheets 0.75 V (versus RHE) 13.9 mA cm2 0.76 V (versus RHE)
Ag nanosheets 0.5 M NaHCO3 10 mA cm2 @ 0.8 V h: 0.29 V @ 5 mA cm2 CO: 95% @ 0.7 V NA Lee et al.73
(versus RHE) (versus RHE)
WSe2 nanoflakes 50 vol % EMIMBF4 330 mA cm2 @ h: 54 mV @ CO: 90% @ 0.764 V 27 hr Asadi et al.74
aqueous solution 0.764 V (versus RHE) 18.95 mA/cm2 (versus RHE)
MoSeS alloy 4 mol % EMIMBF4 43 mA cm2 @ 1.15 V onset potential: CO: 45.2% @ 1.15 V 10 hr Xu et al.75
monolayers aqueous solution (versus RHE) 0.55 V (versus RHE) (versus RHE)
Nb-doped vertically 50 vol % EMIMBF4 237 mA cm2 @ 0.8 V onset potential: CO: 82% @ 0.8 V NA Abbasi
aligned MoS2 aqueous solution (versus RHE) 0.142 V (versus RHE) (versus RHE) et al.76
N-doped graphene 0.1 M KHCO3 1.8 mA cm2 @ onset potential: 0.3 V CO:  85% @ 5 hr Wu et al.77
foam 0.58 V (versus RHE) (versus RHE) 0.58 V
(versus RHE)
Ni-N modified 0.1 M KHCO3 0.2 mA cm2 @ h: 0.54 V @ CO: > 90% @ 0.7  5 hr Su et al.78
graphene 0.65 V (versus RHE) 0.2 mA cm2 0.9 V (versus RHE)
Re-SURMOF thin 0.1 M 2.5 mA cm2 @ onset potential: 1.3 V CO:  93% @ 1.6 V NA Ye et al.79
films tetrabutylammonium 1.6 V (versus normal (versus NHE) (versus NHE)
hydroxide in hydrogen electrode,
acetonitrile with NHE)
5 vol %
trifluoroethanol
g-C3N4/MWCNTs 0.1 M KHCO3 0.55 mA cm2 (JCO) @ h: 0.64 V @ CO: 60% @ 0.75 V 50 hr Lu et al.80
0.75 V (versus RHE) 0.92 mA cm2 (versus RHE)
Mo-Bi bimetallic [BMIM]BF4/MeCN 12.1 mA cm2 @ h: 0.16 V @ CH3OH: 71.2% @ 5 hr Sun et al.81
chalcogenide 0.7 V (versus SHE) 1.0 mA cm2 0.7 V (versus SHE)
nanosheets
N-doped graphene- [BMIM]BF4 with 3.26 mA cm2 @ h: 1.16 V @ CH4: 93.5% @ 1.4 V 5 hr Sun et al.43
like materials 3 wt % H2O 1.4 V (versus SHE) 3.26 mA cm2 (versus SHE)
a
J refers to geometric current density or partial current density with specific subscript.
b
h refers to overpotential to generate a specific product.
c
FE refers to maximum faradic efficiency of the main products.

HCOO selectivity (73%) was reported on N-doped graphene at 0.84 V (versus


RHE) for the electroreduction of CO2 in 0.5 M aqueous KHCO3.20

N doping was supposed to modify the electronic properties of graphene, lowering


the free energy barrier of*COOH formation, which is the potential limiting step,
enhancing *COOH adsorption but weakening CO or HCOOH adsorption. However,

574 Chem 3, 560–587, October 12, 2017


Figure 7. Synthetic Scheme and Characterizations of Co Four-Atom-Thick Layers with and
without Surface Oxide
(A) Schematic formation processes of partially oxidized and pure-Co with four atomic layers.
(B) Lateral HAADF-STEM image. Scale bar, 1 nm.
(C) Corresponding intensity profile along the marked rectangle in (B) shows the four-atom thickness
of the layer.
(D and E) Corresponding crystal structures.
Reprinted with permission from Gao et al.18 Copyright 2016 Macmillan Publishers Limited.

the active sites and mechanism for selectivity remain to be explored. DFT calcula-
tions performed by Liu et al. showed that pyrrolic N performs the best for the ECR
to yield HCOOH as a result of its higher *COOH adsorption energy (3.24 eV)
and lower overpotential (0.24 V) than those of graphitic and pyridinic N configura-
tions in graphene.84 In contrast, Chai and Guo proposed that graphitic N-doped
edges have a low CO2 activation barrier (0.58 eV) and are likely the most active sites
for CO2 reduction relative to pyridinic and pyridinium N forms in graphene-based
materials.83 The authors found that the overpotential for a given product can be
tuned by a curvature effect. Recent work showed that tri-pyridinic defects in gra-
phene have suitable thermodynamic energies to undergo 2H+/2e reductions for
efficient CO2 conversion.85

2D SnS2 nanosheets supported on reduced graphene oxide were reported to be an


active catalyst for electrocatalytic reduction of CO2 into HCOO in aqueous bicar-
bonate media.72 Such 2D/2D hybrids provided an overpotential of 0.23 V for
HCOO formation and a HCOO faradic efficiency of 84.5% at an overpotential of
0.68 V, comparable with those of most state-of-the-art Sn or N-doped carbon-based
catalysts.

Carbon Monoxide Selectivity


Metal Nanosheets
Au, Ag, and Zn all bind CO weakly and exhibit relatively high CO2 reduction effi-
ciencies to CO, with edge-rich Au nanowires being the most active (CO Efaradic,
94% at an overpotential of 0.23 V) in aqueous solutions.4 CO2 can be selectively

Chem 3, 560–587, October 12, 2017 575


Figure 8. Schematic Illustration Depicts the Several Advantages of Ultrathin Metal Layers
Confined in Graphene for CO2 Electroreduction into Hydrocarbon Fuels
Reprinted with permission from Lei et al. 68 Copyright 2016 Macmillan Publishers Limited.

converted to CO (Efaradic R 95%) on both Ag and Bi cathodes at an overpotential


below 0.2 V in IL solutions.40 Pd nanoparticles (NPs) of 3.7 nm also provide a CO
faradic efficiency of 91.2% at 0.89 V (versus RHE) in 0.1 M aqueous KHCO3.86

There are few research works about metal nanosheets used for selective CO2 elec-
troreduction to CO. Recently, 50-nm-thick interconnected Ag nanosheets were
prepared by an electrochemical oxidation-reduction approach.73 Such nanosheets
provided a CO faradic efficiency of 95% at an overpotential of 0.29 V and a current
density 37 times larger than that obtained for polycrystalline Ag at 0.6 V
(versus RHE), which are among the best performances for aqueous CO2 reduction
to CO. A faradic efficiency of 93% at 1.6 V (versus SCE) was achieved for CO on
Zn nanoplates 1 mm long and 40 nm thick decorated with NPs (30–50 nm) in
aqueous NaCl, whereas the CO faradic efficiency is only 18% on bulk Zn in
NaHCO3 solution.87

Transition-Metal Dichalcogenide Nanosheets


2D nanoflakes (NFs) of transition-metal dichalcogenides (TMDs) are a new class of
catalysts that display remarkable CO2 reduction performances for CO generation
in ILs (such as 1-ethyl-3-methylimidazolium tetrafluoroborate [EMIM-BF4]).17,74
Vertically aligned WSe2, MoSe2, WS2, and MoS2 NFs all showed very high current
densities above 130 mA cm2 (up to 330 mA cm2 for WSe2 nanosheets) with 90%
CO selectivity (Figure 10),17,74,75 in contrast to 3.3 mA cm2 with 0% CO selectivity
for bulk Ag, 10 mA cm2 with 65% CO selectivity for 40 nm Ag NPs. WSe2 NFs gave
the best performance and exhibited a current density of 18.95 mA cm2, CO faradic
efficiency of 24%, and CO TOF of 0.28 s1 at an overpotential of 54 mV. The current
density decreased in the order of WSe2 > MoSe2 > WS2 > MoS2, which might corre-
late with their respective electron transfer properties. WSe2 has the lowest work
function and thus performs the best for CO2 activation in these TMDs. It was further
inferred that the metal edges of these TMDs had strong binding of COOH* and CO*
intermediates during the catalytic reaction to maintain a high turnover rate. These
TMD nanosheet/IL systems exhibit a synergy between metallic edge CO binding
and bulk IL CO2 solubility, thus providing promising TMD-based electrocatalysts
for CO production.

Increasing the number of exposed edges by doping with metal atoms to modify the
binding energies of intermediates can effectively enhance the CO2 conversion effi-
ciency of TMDs.29 In this regard, 5% niobium (Nb)-doped MoS2 nanosheets were
demonstrated to exhibit an order of magnitude higher CO formation TOF than

576 Chem 3, 560–587, October 12, 2017


Figure 9. Electroreduction of CO2 over Mesoporous SnO2 Nanosheets on a Carbon-Cloth
Electrode in CO2-Saturated 0.5 M NaHCO3 Solution
(A) Current density across the entire potential range.
(B) Corresponding faradic efficiency for HCOO , CO, and H 2 . Error bars represent standard error.
Reprinted with permission from Li et al.22 Copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA,
Weinheim.

pristine MoS2 in ILs at an overpotential range of 50–150 mV.76 This catalyst yielded a
very low onset overpotential of 31 mV for CO2 reduction.

Heteroatom-Doped Graphene, Metal-Organic Framework Thin Films, and 2D-


Nanosheet-Based Hybrids
N-doped graphene (NG) foam has been used as a catalyst for aqueous CO2 reduc-
tion.77 Such metal-free catalyst exhibited an overpotential as low as 0.19 V for CO
formation, a CO faradic efficiency of 85% at an overpotential of 0.47 V, and stability
over 5 hr, comparable with polycrystalline Ag and Au. The maximum faradic effi-
ciency for CO increased with the pyridinic N content, and the corresponding poten-
tial shifted anodically, which was, however, not observed for pyrrolic N and
graphitic N. DFT calculations based on a computational hydrogen electrode model
revealed that pristine graphene has the highest free energy barrier for the first step
of COOH* adsorption (Figure 11). Whereas the free energy barrier for COOH*
adsorption decreased substantially upon introducing N defects because the
dangling N bonds formed bind the COOH* intermediate much more strongly.
Triple-pyridinic N lowered the barrier of COOH* adsorption the most followed by
single pyridinic and graphitic N. COOH* adsorption at a pyrrolic N-defect site is
exergonic. However, an energy barrier of 0.6 eV is required to desorb chemisorbed
CO from pyrrolic N. Pyridinic N was suggested to be likely the most active site for
CO2 reduction to CO.

The introduction of transition-metal atoms coordinated with N into the sp2 networks
of graphene can create a synergy and further enhance CO2 reduction catalysis. For
example, Ni-N modified graphene (Ni-N-Gr) efficiently catalyzed the electrochemi-
cal reduction of CO2 to CO with a CO faradic efficiency exceeding 90% at 0.7 to
– 0.9 V (versus RHE).78 The TOFs for CO formation on per electroactive Ni atoms
of Ni-N-Gr reached 2,700 and 4,600 hr1 at 0.7 and 0.8 V (versus RHE) in neutral
solutions, respectively, comparable with those of the best Co-porphyrin-based
covalent organic framework catalysts for CO2 reduction.88

On the basis of DFT and microkinetics modeling, an adjacent graphene single-va-


cancy-supported Cu dimer was demonstrated to be catalytically active for CO2 elec-
troreduction to CO, with a reaction rate and CO partial current density comparable

Chem 3, 560–587, October 12, 2017 577


Figure 10. DFT Analysis
(A) Calculated free-energy diagrams for ECR to CO on Ag (111), Ag 55 NPs, MoS 2 , WS 2 , MoSe 2 , and
WSe 2 NFs at 0 V (versus RHE).
(B) Calculated partial density of states of the d band (spin-up) of the surface Ag atom of Ag 55 .
(C) Calculated partial density of states of the surface bare metal edge atom (W) of the WSe 2 NFs.
Reprinted with permission from Asadi et al.74 Copyright 2016 American Association for the
Advancement of Science.

with or even higher than those of Au electrodes.89 The selectivity for CO formation
was attributed to the fact that the positively charged metal sites and the electronic
interaction between Cu and graphene weakened CO adsorption.

Thin films of metal-organic frameworks (MOFs), such as Co-porphyrin MOF,90 have


been demonstrated as catalysts for electroreduction of CO2 to CO with CO selec-
tivity in excess of 76% in aqueous electrolytes. In particular, highly oriented MOF
thin films grafted with ReL(CO)3Cl (L = 2,20 -bipyridine-5,50-dicarboxylic acid) (Fig-
ure 12), grown on a fluorine-doped tin oxide (FTO) electrode displayed a faradic ef-
ficiency of up to 93% G 5% toward CO production at 1.6 V (versus NHE),
exceeding those reported for most electrocatalytically active MOF thin films.79 How-
ever, the stability of such Re-based MOF catalysts needs to be further improved for
future applications.

g-C3N4, an important metal-free 2D semiconductor material, has been widely used in


photocatalysis. Recently, Amal and co-workers first reported the potential of utilizing
g-C3N4/multi-walled carbon nanotube (MWNT) composites as a stable and selective
electrocatalyst for CO2 reduction in 0.1 M KHCO3. The as-prepared composites ex-
hibited a CO faradic efficiency of 60%, and no decay in catalytic activity was observ-
able after 50 hr. The covalent C–N bonds formed between g-C3N4 and MWNTs were
proposed as active sites for ECR. In addition, high specific surface area and improved
material conductivity of the composite led to enhancement in CO2 reduction.80

Methanol Selectivity
Metal oxides, such as Cu2O91 and RuO2,92 have demonstrated direct ECR to
CH3OH. However, using 2D nanosheets for ECR selectively to CH3OH has been
rarely reported. Han and colleagues81 showed that Mo-Bi bimetallic chalcogenide
(BMC) nanosheets were able to reduce CO2 to CH3OH with a faradic efficiency of
71.2% and a current density of 12.1 mA cm2, which outperform earlier reported cat-
alysts. In contrast, CH3OH was not generated when bulk MoS2/carbon paper (CP) or
Bi2S3/CP electrode was used. These Mo-Bi BMC nanosheets performed better than

578 Chem 3, 560–587, October 12, 2017


Figure 11. DFT Modeling of Electrocatalysis of CO2 on NG
(A) Free energy diagram of electrochemical reduction of CO 2 to CO on NG.
(B) Schematic of N configuration and CO 2 reduction pathway.
Reprinted with permission from Wu et al.77 Copyright 2016 American Chemical Society.

both Mo-Ag and Mo-Cu BMC nanosheets (Figures 13A–13C). This catalytic process
relies on an ionic liquid acetonitrile solution, wherein the IL, 1-butyl-3-methylimida-
zolium tetrafluoroborate, provides increasing faradic efficiency as the concentration
of IL is decreased from 1 M to 0.2 M and a maximum current density at about 0.5 M
(Figure 13D). This selectivity was attributed to a synergistic effect between Mo and
Bi, wherein the IL coordinated with CO2 for reduction to radical anion, which was
transformed by a second one-electron process to adsorbed CO. This nanosheet
result surpasses about 40% faradic efficiencies afforded by Ru/Cu (0.5 M aqueous
NaHCO3) and Cu (LiCl in ethanol/water), and 55% faradic efficiency from Mo
(0.2 M aqueous Na2SO4).

Incorporation of Ni-Cu dopant pairs into adjacent single vacancies in graphene (Ni-
Cu@SV) was calculated to lower the free energy barrier of the CO / CHO step from
0.99 eV on Cu (111) to 0.71 eV on Ni-Cu@SV, leading to an 0.3–0.4 V reduction in
overpotential.89 The O adsorption energy on the Ni-Cu dopant is also lower than
that of the Ni2 and Cu2 dopant pairs, making CH3OH production more thermody-
namically favorable than CH4 production, which promotes CH3OH selectivity.
More recently, Mn2 dimers supported on 2D expanded phthalocyanine nanosheets
were predicted to be the best catalyst for ECR to CH3OH among single Mn atom, or
Fe, Co, Ni, Cu dimers.93 The two Mn atoms in the dimer contribute to the bonding
between COOH* adsorbate and catalyst. The bridge adsorption of Mn-C-O-Mn
enhanced the metal-to-adsorbate p-back bonding, resulting in an easy transition
between C-end adsorbate and O-end adsorbate with lowered energy barrier in
CH3OH desorption.

Methane Selectivity
Nitrogen-doped graphene-like materials (NG) coated on CP electrodes were shown
to catalyze the electroreduction of CO2 to CH4 in an IL, 1-butyl-3-methylimidazolium
tetrafluoroborate.43 The faradic efficiency of CH4 formation increased significantly
with increasing N content, approaching 94% at 4.8 atom % of N. It was proposed
that pyridinic N and pyridonic or pyrrolic N species promoted CO2 adsorption;
the ionic liquid helped drive the transformation of adsorbed CO2 molecules to
CO2, radical anions. Meanwhile, a strong interaction between COads and the elec-
trode inhibited release of CO from the electrode, which is favorable for its further hy-
drogenation to form CH4. This process was found to retain selectivity when 3% of

Chem 3, 560–587, October 12, 2017 579


Figure 12. Schematic of the Fabrication Process of Re-surface-Grafted MOF on Functionalized
FTO Substrate in a Layer-by-Layer Fashion
(A) Zn acetate.
(B) Re-linker.
(C) Idealized structure of Re-surface-grafted MOF on FTO.
Reprinted with permission from Ye et al. 79 Copyright 2016 Royal Society of Chemistry.

water was added to the IL while improving the current density (from 1.4 to 3.3 mA
cm2). The current densities obtained with this NG system were about 6-fold higher
than that of copper foil used under similar conditions. Nitrogen doping is critical, as
only CO and H2 are generated in its absence.

2D IrTe2 (T), RhTe2 (T), PFeLi, and TiS2 (T) were calculated to have theoretical reduc-

tion potentials (E ECR) of 0.58 V, 0.61 V, 0.65 V, 0.69 V (versus RHE), respec-
tively. They all demonstrated lower overpotentials for CO2 conversion into CO

than the best metal catalyst Au (calculated E ECR = 0.71 V versus RHE), whereas

2D LiFeAs and ScS2 (T) were calculated to have strong CO affinity but low E ECR of
0.55 and 0.62 V (versus RHE) for CH4 formation, respectively, with lower overpo-

tential over 0.35 V than with Cu (calculated E ECR = 0.97 V versus RHE).94 Mn-Cu
dimers supported on graphene with adjacent single vacancies were demonstrated
to have strong *OH or *COOH adsorption, favoring further reduction of CO. In addi-
tion, such a catalyst can reduce the free energy barrier of the CO / CHO step to
facilitate CH4 generation.93 Despite such promising theoretical predictions, future
work needs to synthesize these 2D materials and further investigate their catalytic
properties for ECR.

Acetic Acid Selectivity


It is difficult to produce C2 and C2+ chemicals in CO2 electrochemical reduction
because of the higher energy required and slower kinetics for the formation of
C–C bonds than for the formation of C–H and H–H bonds. Generation of C2 products
has been rarely reported with 2D materials. Sun and co-workers demonstrated ECR
to CH3COOH over N-based Cu(I)/C-doped BN sheets.67 A high CH3COOH faradic
efficiency was obtained up to 80.3%, surpassing previously reported values. Such an
outstanding performance was attributed to a synergistic effect of C-doped BN
(BNC), the Cu metal center, the N-based ligand (N,N,N0 ,N0 -tetra(2-pyridyl)-2,6-pyr-
idinediamine), and the electrolyte ([Emim]BF4/LiI/H2O solution). Specifically, BNC
facilitated adsorption conversion of CO2 to CO2,. The Cu complex helped the for-
mation and protonation of *CO, *CHO, and *CH3O, and the C–C coupling in the
presence of LiI.

580 Chem 3, 560–587, October 12, 2017


Figure 13. Current Density and Product Faradic Efficiency at Different Applied Potentials in CO2-
Saturated 0.5 M [BMIM]BF4 MeCN Solution
(A) Mo-Bi BMC/CP electrode.
(B) Mo-Ag BMC/CP electrode.
(C) Mo-Cu BMC/CP electrode. Curve (a): current density; curves (b)–(e): faradic efficiency of (b)
CH3 OH, (c) CH 4 , (d) CO, and (e) H2 .
(D) CH3 OH faradic efficiency and current density at 0.7 V (versus SHE) as a function of [BMIM]BF 4
molar concentration in MeCN over Mo-Bi BMC/CP electrode.
Reprinted with permission from Sun et al. 81 Copyright 2016 Wiley-VCH Verlag GmbH & Co. KGaA,
Weinheim.

Note that only Cu27, NG quantum dots95 and graphene/ZnO/Cu2O hybrid28 have
been found to convert CO2 into C2+ products, the yields of which are however %
30% in both cases. 2D catalysts and strategies to afford C2+ products with higher ef-
ficiency need to be explored.

STRATEGIES FOR IMPROVING CO2 ELECTROCATALYTIC ACTIVITY OF


2D NANOSHEETS
Strategies that can be used to improve nanosheet electrocatalytic activity should
take into account the following: (1) enhancing current density at low overpotential
toward a single energy-rich product especially toward C2–C4 products, (2) slowing
down the competitor reactions such as HER, and (3) improving long-term stability.
This section provides basic information and understanding of some possible effec-
tive strategies.

Surface Engineering
Surface Modification
Surfaces of 2D nanosheets with a high level of exposed low-coordination atoms are
easily modified with metallic and non-metallic species. Alloying of different metals
can optimize the binding energies of reaction intermediates such as *CO, and might
also be good for breaking scaling relations between similarly adsorbed intermedi-
ates to reduce reaction overpotential.96 For instance, *CO tends to bind to metals
exclusively via the C atom, but embedding an oxophilic atom into the surface can
selectively stabilize *COOH and *CHO over *CO. Alternatively, doping with

Chem 3, 560–587, October 12, 2017 581


adatoms (Co, Mn, Bi, As, Li, H, C, S, N, B, P, O, F, and I) can substantially alter the
electronic and geometric properties of nanosheets. Such doping can tune the bind-
ing strengths of different intermediates and reaction energetics, thereby modulating
the activity and selectivity of CO2 conversion. Recently, we found that doping of Pd
with Te enables highly selective electrocatalytic reduction of aqueous CO2 to CO.
DFT calculations show that Te atoms preferentially bind at terrace sites of Pd,
thereby suppressing hydrogen evolution, whereas CO2 adsorption and activation
on high index sites of Pd give rise to CO (unpublished data). Chemical doping of
2D materials with controllable configuration and doping level remains a challenge.

Surface functional groups (such as amines) play a role in reaction activity and selec-
tivity. For example, amine functionalization of a catalyst can concentrate CO2 on the
electrode surface from the bulk solution, increasing its effective local concentration.
In addition, amine can stabilize CO2,– via a H-bond interaction, thus lowering the
onset potential for reducing CO2 to CO2,– by creating a stabilizing environment.
Likewise, amine can help the bonding of *CO, favoring subsequent reduction of
CO to methanol and C–C coupling for C2 formation.67

Pure graphene-like carbon materials exhibit small catalytic activities for CO2 activa-
tion, whereas doping with heteroatoms, such as B, N, and/or Ni, can drastically lower
CO2 adsorption barriers to facilitate electrocatalytic CO2 reduction. A remarkable
enhancement in carrier concentration can be achieved by quaternary N doping up
to 2.6 3 1013 cm2 (four times higher than that for pristine graphene) even at a
very small doping level of 0.6 atom %.97 We note that single- and co-doped gra-
phene materials with other p-block and d-block elements in addition to B, N, and
Ni, remain to be explored for ECR.

Surface-Structure Tuning
Structural tuning of 2D nanosheets is realized mainly by modification of four impor-
tant aspects: (1) surface defects, (2) surface porosity, (3) exposed crystal facets, and
(4) surface phase.

Surface tuning by an oxygen vacancy can increase the adsorption of CO2 on metal
oxides or hydroxides. CO2 molecules are prone to adsorb at oxygen vacancies
with one oxygen atom of CO2 situated by bridging oxygen vacancy defects,
thus decreasing the energy barrier for CO2 activation.98 Likewise, chalcogen va-
cancies on chalcogenide nanosheets can act as adsorption sites for CO2, facili-
tating CO2 reduction selectively to CO.17 The surface vacancy density can be
adjusted via high-temperature reduction or by doping with a lower-valent foreign
element.

Creating hierarchical micro-, meso-, and macroporous structure on 2D nanosheets


offers a way to enhance adsorption and capture of CO2 and to improve CO2 conver-
sion efficiency. Macropores (>50 nm) facilitate facile transfer and diffusion of reac-
tants and products over electrodes, whereas mesopores (2–50 nm) and micropores
(<2 nm) provide a large number of surface active sites with high dispersion. In addi-
tion, tuning the diameters of pores or channels according to the molecular sizes of
reactants or products can tailor selectivity correspondingly.

Crystal facets play a significant role in determining reaction activity and selectivity.
Different surface facets exhibit different Lewis acidity and polarizing power, thus
influencing CO2 adsorption and activation. Facets with high adsorption energies
and low activation barriers are preferred for electrocatalytic reactions.

582 Chem 3, 560–587, October 12, 2017


The catalytic activity and selectivity of TMDs and many metals are sensitive to surface
composition and order. The octahedral (1T) phase of TMD possesses inherently low
charge-transfer resistance. Active-site density increases as a result of strained lattice
distortion in comparison with the trigonal prismatic (2H) phase. Controlling ad-
sorbed ions near surface vacancies permits tuning of the electronic properties and
activity of catalysts. In particular, introducing the partially oxidized surface of metal
electrocatalysts, typically Sn, Bi, and Co, affords improved activity over that of pure
metals.18

Construction of Nanosheet Composites


Creation of nanosheet composites can enhance charge transport, imparting electronic
and physical coupling effects to facilitate CO2 conversion.80 Important parameters for
an interface between components include interfacial compositions and facets, areas,
interfacial defects, and electronic coupling. Interfacial charge transfer can be acceler-
ated by addition of conductive sheets such as graphene.72,89 When CO2 adsorption en-
ergy can be also increased, a synergy is created, promoting CO2 reduction.

SUMMARY AND OUTLOOK


Fossil fuels are likely to continue to be a major source of energy for the next few de-
cades. Alleviating the effects caused by waste CO2 emission remains a critical issue to
modern society. Electrocatalytic reduction of CO2 offers an intriguing way for CO2
mitigation, by which CO2 as a feedstock can be converted to fuels or value-added
chemicals. Notable results have been achieved in developing 2D materials with elec-
trocatalytic performances superior to their bulk counterparts. Partially oxidized 4-
atom-thick Co layers have provided large selectivity for HCOO, competitive with
the use of Pd nanoparticles supported on carbon (Pd/C). N doping of graphene
has been shown to offer high promise for metal-free catalysis of the ECR to produce
CO and HCOO in aqueous electrolytes, and CH4 in ILs, but the impact of quaternary
basal-plane nitrogen versus pyridinic nitrogen on edges has not yet been settled and
requires further analytical work. Interconnected Ag nanosheets have yielded one of
the best performances for aqueous CO2 reduction to CO so far. 2D TMDs have been
found to drastically promote CO2 reduction for CO generation in ILs, and the sele-
nides appear to be the most effective. Mo-Bi BMC nanosheets are capable of cata-
lyzing CO2 reduction to CH3OH, outperforming earlier reported catalysts.

For real-world CO2 electrolysis application, further improvement in electrocatalytic


activity is required, which should fulfill the following demand: (1) small overpotential
and high energy efficiency, (2) satisfactory selectivity and faradic efficiency, and (3)
good electrocatalytic stability (>100 hr). To promote CO2 adsorption and activation,
some useful strategies can be adopted by surface modification, surface-structure
tuning (surface defect, porosity, crystallographic orientation, composition, strain,
and curvature), and formation of multiphase hybrid nanostructures. Scaling relations
that exist between adsorption energies of different reaction intermediates need to
be circumvented, which can be realized possibly by selectively stabilizing the inter-
mediates through some external stimuli. Other technical aspects, including acidic or
alkaline working conditions, operating parameters (such as pressure, electrolyte
concentration, etc.), and electrolytic cells, must be optimized. For the sake of desir-
able overall cell efficiency, high-performance electrocatalysts to promote the anode
OER are required as well.

Large-scale production of 2D nanosheets is the key to their practical use in CO2


catalysis. Chemical vapor deposition and liquid-phase exfoliation methods hold

Chem 3, 560–587, October 12, 2017 583


promise for the preparation of layered materials. But scalable routes to prepare
atomically thin non-lamellar nanosheets, especially those being intrinsically catalytic
active, remain to be developed. The electronic properties of particular 2D materials
can be tuned by varying dimensionality (thickness and lateral dimension) to improve
efficiency of electron transfer and of CO2 reduction. Control in the number of layers,
flake dimensions, and defect levels, in addition to yield, throughput, and cost are
thus needed. Equally importantly, developing in situ (spectroscopic) characteriza-
tion techniques will help develop more precise theoretical models to provide
deeper insights into catalytic reaction pathways involving scaling relations among
reactive intermediates and structure-property relationships peculiar to these mate-
rials. The combination of theory and experiment will further aid catalyst design to in-
crease the number of active sites and intrinsic activity of each active site as well as
development of novel robust 2D catalysts for this clean energy reaction. Exploration
of low-cost, non-noble metal 2D nanosheets that can convert CO2 to target products
at sufficiently high reaction rate and efficiency continues to be a focus of research.
Porous carbon nanosheets modified with p-block and d-block elements, or with
NG quantum dots, 2D mixed-metal (phosphorus) chalcogenides, appear to be
promising candidates. Design of heterostructures comprising different 2D materials
(van der Waals heterostructures and lateral heterostructures) and/or nanoparticles
enables great tuning over the structural and electronic properties of such crystals
and hybrids, creating enormous diversity in properties and functionality. CO2 cata-
lytic conversion making use of strong electron interactions, confined space, and cre-
ation of sandwich structures offers much interest.

Initial positive results have been obtained from selectively producing CO and
HCOO by using 2D materials. Whereas the established Fischer-Tropsch process
is capable of taking generated CO and H2 to liquids, these kinds of two- to four-elec-
tron steps can be followed by the addition of extra catalytic phases through doping
or hybrid design for sequential reaction steps that lead to C2 and even higher-order
hydrocarbons and oxygenates. Research in this direction is an exciting arena.

AUTHOR CONTRIBUTIONS
Z.S. proposed the topic of the review and wrote the manuscript. T.M., H.T., and Q.F.
searched the literature and edited figures and the table. B.H. revised the manuscript.

ACKNOWLEDGMENTS
This work was supported by the State Key Laboratory of Organic-Inorganic Compos-
ites (oic-201503005), Fundamental Research Funds for the Central Universities
(buctrc201525), Beijing National Laboratory for Molecular Sciences
(BNLMS20160133), Key Laboratory of Photochemical Conversion and Optoelec-
tronic Materials (Technical Institute of Physics and Chemistry, Chinese Academy of
Sciences [CAS]), and Key Laboratory of Materials for High-Power Laser (Shanghai
Institute of Optics and Fine Mechanics, CAS).

REFERENCES AND NOTES


1. McCollum, D., Bauer, N., Calvin, K., Kitous, A., 3. Shi, J., Jiang, Y., Jiang, Z., Wang, X., Wang, X., 5. Habisreutinger, S.N., Schmidt-Mende, L., and
and Riahi, K. (2014). Fossil resource and energy Zhang, S., Han, P., and Yang, C. (2015). Stolarczyk, J.K. (2013). Photocatalytic
security dynamics in conventional and carbon- Enzymatic conversion of carbon dioxide. reduction of CO2 on TiO2 and other
constrained worlds. Climatic Change 123, Chem. Soc. Rev. 44, 5981–6000. semiconductors. Angew. Chem. Int. Ed 52,
413–426. 7372–7408.
4. Zhu, D.D., Liu, J.L., and Qiao, S.Z. (2016).
2. Davis, S.J., Caldeira, K., and Matthews, H.D. Recent advances in inorganic 6. Grodkowski, J., and Neta, P. (2001). Copper-
(2010). Future CO2 emissions and climate heterogeneous electrocatalysts for catalyzed radiolytic reduction of CO2 to CO in
change from existing energy infrastructure. reduction of carbon dioxide. Adv. Mater. 28, aqueous solutions. J. Phys. Chem. B 105, 4967–
Science 329, 1330–1333. 3423–3452. 4972.

584 Chem 3, 560–587, October 12, 2017


7. Wang, W., Wang, S.P., Ma, X.B., and Gong, J.L. as efficient electrocatalysts for the selective electrode in aqueous solution. Bull. Chem. Soc.
(2011). Recent advances in catalytic reduction of carbon dioxide to formate in Jpn. 55, 660–665.
hydrogenation of carbon dioxide. Chem. Soc. aqueous solution. Green. Chem. 18, 3250–
Rev. 40, 3703–3727. 3256. 35. Hori, Y., Murata, A., and Takahashi, R. (1990).
Formation of hydrocarbons in the
8. Klankermayer, J., Wesselbaum, S., Beydoun, 21. Liu, J., Guo, C., Vasileff, A., and Qiao, S. (2017). electrochemical reduction of carbon dioxide at
K., and Leitner, W. (2016). Selective catalytic Nanostructured 2D materials: prospective a copper electrode in aqueous solution.
synthesis using the combination of carbon catalysts for electrochemical CO2 reduction. J. Chem. Soc. Faraday Trans. 21, 2309–2326.
dioxide and hydrogen: catalytic chess at the Small Methods 1, 1600006.
interface of energy and chemistry. Angew. 36. Ogura, K., Ferrell, J.R., Cugini, A.V., Smotkin,
Chem. Int. Ed 55, 7296–7343. 22. Li, F.W., Chen, L., Knowles, G.P., MacFarlane, E.S., and Salazar-Villalpando, M.D. (2010). CO2
D.R., and Zhang, J. (2017). Hierarchical attraction by specifically adsorbed anions and
9. Grills, D.C., Matsubara, Y., Kuwahara, Y., mesoporous SnO2 nanosheets on carbon subsequent accelerated electrochemical
Golisz, S.R., Kurtz, D.A., and Mello, B.A. (2014). cloth: a robust and flexible electrocatalyst for reduction. Electrochim. Acta 56, 381–386.
Electrocatalytic CO2 reduction with a CO2 reduction with high efficiency and
homogeneous catalyst in ionic liquid: high selectivity. Angew. Chem. Int. Ed 56, 505–509. 37. Varela, A.S., Ju, W., Reier, T., and Strasser, P.
catalytic activity at low overpotential. J. Phys. (2016). Tuning the catalytic activity and
Chem. Lett. 5, 2033–2038. 23. Kauffman, D.R., Thakkar, J., Siva, R., Matranga, selectivity of Cu for CO2 electroreduction in the
C., Ohodnicki, P.R., Zeng, C., and Jin, R. (2015). presence of halides. ACS Catal. 6, 2136–2144.
10. Gong, J., Zhang, L., and Zhao, Z.J. (2017). Efficient electrochemical CO2 conversion
Nanostructured materials for heterogeneous powered by renewable energy. ACS Appl. 38. Vassiliev, Y.B., Bagotzky, V.S., Khazova, O.A.,
electrocatalytic CO2 reduction and related Mater. Interfaces 7, 15626–15632. and Mayorova, N.A. (1985). Electroreduction of
reaction mechanisms. Angew. Chem. Int. Ed carbon dioxide : Part II. The mechanism of
56, 11326–11353. 24. Lu, Q., Rosen, J., Zhou, Y., Hutchings, G.S., reduction in aprotic solvents. J. Electroanal.
Kimmel, Y.C., Chen, J.G., and Jiao, F. (2014). A Chem. 189, 295–309.
11. He, J., Dettelbach, K.E., Salvatore, D.A., Li, T., selective and efficient electrocatalyst for
and Berlinguette, C.P. (2017). High-throughput carbon dioxide reduction. Nat. Commun. 5, 39. Mizuno, T., Naitoh, A., and Ohta, K. (1995).
synthesis of mixed-metal electrocatalysts for 3242. Electrochemical reduction of CO2 in methanol
CO2 reduction. Angew. Chem. Int. Ed 56, 6068– at -30 C. J. Electroanal. Chem. 391, 199–201.
6072. 25. Kortlever, R., Shen, J., Schouten, K.J.P., Calle-
Vallejo, F., and Koper, M.T.M. (2015). Catalysts 40. Rosen, B.A., Salehi-Khojin, A., Thorson, M.R.,
12. Chernikov, A., van der Zande, A.M., Hill, H.M., and reaction pathways for the electrochemical Zhu, W., Whipple, D.T., Kenis, P.J.A., and
Rigosi, A.F., Velauthapillai, A., Hone, J., and reduction of carbon dioxide. J. Phys. Chem. Masel, R.I. (2011). Ionic liquid-mediated
Heinz, T.F. (2015). Electrical tuning of exciton Lett. 6, 4073–4082. selective conversion of CO2 to CO at low
binding energies in monolayer WS2. Phys. Rev. overpotentials. Science 334, 643–644.
Lett. 115, 126802. 26. Mistry, H., Varela, A.S., Bonifacio, C.S.,
Zegkinoglou, I., Sinev, I., Choi, Y.W., Kisslinger, 41. Alvarez-Guerra, M., Albo, J., Alvarez-Guerra,
13. Wang, H., Yuan, H., Sae Hong, S., Li, Y., and K., Stach, E.A., Yang, J.C., Strasser, P., and E., and Irabien, A. (2015). Ionic liquids in the
Cui, Y. (2015). Physical and chemical tuning of Cuenya, B.R. (2016). Highly selective plasma- electrochemical valorisation of CO2. Energ.
two-dimensional transition metal activated copper catalysts for carbon dioxide Environ. Sci. 8, 2574–2599.
dichalcogenides. Chem. Soc. Rev. 44, 2664– reduction to ethylene. Nat. Commun. 7, 12123.
2680. 42. Pardal, T., Messias, S., Sousa, M., Machado,
27. Lee, S., Kim, D., and Lee, J. (2015). A.S.R., Rangel, C.M., Nunes, D., Pinto, J.V.,
14. Tao, H.C., Zhang, Y.Q., Gao, Y.N., Sun, Z.Y., Electrocatalytic production of C3-C4 Martins, R., and da Ponte, M.N. (2017). Syngas
Yan, C., and Texter, J. (2017). Scalable compounds by conversion of CO2 on a production by electrochemical CO2 reduction
exfoliation and dispersion of two-dimensional chloride-induced Bi-phasic Cu2O-Cu catalyst. in an ionic liquid based-electrolyte. J. CO2 Util.
materials - an update. Phys. Chem. Chem. Angew. Chem. Int. Ed 54, 14701–14705. 18, 62–72.
Phys. 19, 921–960.
28. Geioushy, R.A., Khaled, M.M., Alhooshani, K., 43. Sun, X.F., Kang, X.C., Zhu, Q.G., Ma, J., Yang,
15. Deng, D.H., Novoselov, K.S., Fu, Q., Zheng, Hakeem, A.S., and Rinaldi, A. (2017). G.Y., Liu, Z.M., and Han, B.X. (2016). Very highly
N.F., Tian, Z.Q., and Bao, X.H. (2016). Catalysis Graphene/ZnO/Cu2O electrocatalyst for efficient reduction of CO2 to CH4 using metal-
with two-dimensional materials and their selective conversion of CO2 into n-propanol. free N-doped carbon electrodes. Chem. Sci. 7,
heterostructures. Nat. Nanotechol. 11, Electrochim. Acta 245, 448–454. 2883–2887.
218–230.
29. Hong, X., Chan, K.R., Tsai, C., and Norskov, J.K. 44. Lau, G.P.S., Schreier, M., Vasilyev, D., Scopelliti,
16. Tao, H.C., Gao, Y.A., Talreja, N., Guo, F., (2016). How doped MoS2 breaks transition- R., Gratzel, M., and Dyson, P.J. (2016). New
Texter, J., Yan, C., and Sun, Z.Y. (2017). Two- metal scaling relations for CO2 electrochemical insights into the role of imidazolium-based
dimensional nanosheets for electrocatalysis in reduction. ACS Catal. 6, 4428–4437. promoters for the electroreduction of CO2 on a
energy generation and conversion. J. Mater.
silver electrode. J. Am. Chem. Soc. 138, 7820–
Chem. A 5, 7257–7284. 30. Li, Y.W., and Sun, Q. (2016). Recent advances in
breaking scaling relations for effective 7823.
17. Asadi, M., Kumar, B., Behranginia, A., Rosen, electrochemical conversion of CO2. Adv.
45. Kumar, B., Asadi, M., Pisasale, D., Sinha-Ray, S.,
B.A., Baskin, A., Repnin, N., Pisasale, D., Energy Mater. 6, 1600463.
Phillips, P., Zhu, W., Haasch, R., et al. (2014). Rosen, B.A., Haasch, R., Abiade, J., Yarin, A.L.,
Robust carbon dioxide reduction on 31. Thorson, M.R., Siil, K.I., and Kenis, P.J.A. (2013). and Salehi-Khojin, A. (2013). Renewable and
molybdenum disulphide edges. Nat. Effect of cations on the electrochemical metal-free carbon nanofibre catalysts for
Commun. 5, 4470. conversion of CO2 to CO. J. Electrochem. Soc. carbon dioxide reduction. Nat. Commun. 4,
160, F69–F74. 2819.
18. Gao, S., Lin, Y., Jiao, X.C., Sun, Y.F., Luo, Q.Q.,
Zhang, W.H., Li, D.Q., Yang, J.L., and Xie, Y. 32. Schizodimou, A., and Kyriacou, G. (2012). 46. Neubauer, S.S., Krause, R.K., Schmid, B., Guldi,
(2016). Partially oxidized atomic cobalt layers Acceleration of the reduction of carbon D.M., and Schmid, G. (2016). Overpotentials
for carbon dioxide electroreduction to liquid dioxide in the presence of multivalent cations. and faraday efficiencies in CO2
fuel. Nature 529, 68–71. Electrochim. Acta 78, 171–176. electrocatalysis–the impact of 1-ethyl-3-
methylimidazolium trifluoromethanesulfonate.
19. Zhang, H., Ma, Y., Quan, F.J., Huang, J.J., Jia, 33. Murata, A., and Hori, Y. (1991). Product Adv. Energy Mater. 6, 1502231.
F.L., and Zhang, L.Z. (2014). Selective electro- selectivity affected by cationic species in
reduction of CO2 to formate on nanostructured electrochemical reduction of CO2 and CO at a 47. Ye, L.T., Zhang, M.Y., Huang, P., Guo, G.C.,
Bi from reduction of BiOCl nanosheets. Cu electrode. Bull. Chem. Soc. Jpn. 64, Hong, M.C., Li, C.S., Irvine, J.T.S., and Xie, K.
Electrochem. Commun. 46, 63–66. 123–127. (2017). Enhancing CO2 electrolysis through
synergistic control of non-stoichiometry and
20. Wang, H.X., Chen, Y.B., Hou, X.L., Ma, C.Y., and 34. Hori, Y., and Suzuki, S. (1982). Electrolytic doping to tune cathode surface structures. Nat.
Tan, T.W. (2016). Nitrogen-doped graphenes reduction of carbon dioxide at mercury Commun. 8, 17485.

Chem 3, 560–587, October 12, 2017 585


48. Xie, K., Zhang, Y.Q., Meng, G.Y., and Irvine, 62. Hori, Y., Takahashi, R., Yoshinami, A.Y., and A., Cerrato, J.M., Haasch, R., et al. (2016).
J.T.S. (2011). Electrochemical reduction of CO2 Murata, A. (1997). Electrochemical reduction of Nanostructured transition metal
in a proton conducting solid oxide electrolyser. CO at a copper electrode. J. Phys. Chem. B dichalcogenide electrocatalysts for CO2
J. Mater. Chem. 21, 195–198. 101, 7075–7081. reduction in ionic liquid. Science 353, 467–470.

49. Kang, X., Zhu, Q., Sun, X., Hu, J., Zhang, J., Liu, 63. Yang, K.D., Lee, C.W., Jin, K., Im, S.W., and 75. Xu, J.Q., Li, X.D., Liu, W., Sun, Y.F., Ju, Z.Y., Yao,
Z., and Han, B. (2015). Highly efficient Nam, K.T. (2017). Current status and T., Wang, C.M., Ju, H.X., Zhu, J.F., Wei, S.Q.,
electrochemical reduction of CO2 to CH4 in bioinspired perspective of electrochemical and Xie, Y. (2017). Carbon dioxide
ionic liquid using metal-organic framework conversion of CO2 to a long-chain electroreduction into syngas boosted by a
cathode. Chem. Sci. 7, 266–273. hydrocarbon. J. Phys. Chem. Lett. 8, 538–545. partially delocalized charge in molybdenum
sulfide selenide alloy monolayers. Angew.
50. Bevilacqua, M., Filippi, J., Miller, H.A., and 64. Bertheussen, E., Verdaguer-Casadevall, A., Chem. Int. Ed 56, 9121–9125.
Vizza, F. (2015). Recent technological progress Ravasio, D., Montoya, J.H., Trimarco, D.B., Roy,
in CO2 electroreduction to fuels and energy C., Meier, S., Wendland, J., Norskov, J.K., 76. Abbasi, P., Asadi, M., Liu, C., Sharifi-Asl, S.,
carriers in aqueous environments. Energy. Stephens, I.E.L., and Chorkendorff, I. (2016). Sayahpour, B., Behranginia, A., Zapol, P.,
Technol. 3, 197–210. Acetaldehyde as an intermediate in the Shahbazian-Yassar, R., Curtiss, L.A., and Salehi-
electroreduction of carbon monoxide to Khojin, A. (2017). Tailoring the edge structure
51. Alvarez-Guerra, M., Quintanilla, S., and Irabien, ethanol on oxide-derived copper. Angew. of molybdenum disulfide toward
A. (2012). Conversion of carbon dioxide into Chem. Int. Ed 55, 1450–1454. electrocatalytic reduction of carbon dioxide.
formate using a continuous electrochemical
ACS Nano 11, 453–460.
reduction process in a lead cathode. Chem. 65. Genovese, G., Ampelli, C., Perathoner, S., and
Eng. J. 207, 278–284. Centi, G. (2017). Mechanism of C-C bond
77. Wu, J.J., Liu, M.J., Sharma, P.P., Yadav, R.M.,
formation in the electrocatalytic reduction of
52. Li, H., and Oloman, C. (2007). Development of a Ma, L.L., Yang, Y.C., Zou, X.L., Zhou, X.D.,
CO2 to acetic acid. A challenging reaction to
continuous reactor for the electro-reduction of Vajtai, R., Yakobson, B.I., et al. (2016).
use renewable energy with chemistry. Green.
Incorporation of nitrogen defects for efficient
carbon dioxide to formate-part2: scale-up. Chem. 19, 2406–2415.
J. Appl. Electrochem. 37, 1107–1117. reduction of CO2 via two-electron pathway on
66. Liu, Y.M., Chen, S., Quan, X., and Yu, H.T. three-dimensional graphene foam. Nano Lett.
53. Nie, X.W., Esopi, M.R., Janik, M.J., and (2015). Efficient electrochemical reduction of 16, 466–470.
Asthagiri, A. (2013). Selectivity of CO2 carbon dioxide to acetate on nitrogen-doped
reduction on copper electrodes: the role of the nanodiamond. J. Am. Chem. Soc. 137, 11631– 78. Su, P., Iwase, K., Nakanishi, S., Hashimoto, K.,
kinetics of elementary steps. Angew. Chem. 11636. and Kamiya, K. (2016). Nickel-nitrogen-
Int. Ed 52, 2459–2462. modified graphene: an efficient electrocatalyst
67. Sun, X.F., Zhu, Q.G., Kang, X.C., Liu, H.Z., Qian, for the reduction of carbon dioxide to carbon
54. Zhang, S., Kang, P., and Meyer, T.J. (2014). Q.L., Ma, J., Zhang, Z.F., Yang, G.Y., and Han, monoxide. Small 12, 6083–6089.
Nanostructured tin catalysts for selective B.X. (2017). Design of a Cu(I)/C-doped boron
electrochemical reduction of carbon dioxide to nitride electrocatalyst for efficient conversion 79. Ye, L., Liu, J.X., Gao, Y., Gong, C.H., Addicoat,
formate. J. Am. Chem. Soc. 136, 1734–1737. of CO2 into acetic acid. Green. Chem. 19, M., Heine, T., Woll, C., and Sun, L.C. (2016).
2086–2091. Highly oriented MOF thin film-based
55. Baruch, M.F., Pander, J.E., White, J.L., and electrocatalytic device for the reduction of CO2
Bocarsly, A.B. (2015). Mechanistic insights into 68. Lei, F.C., Liu, W., Sun, Y.F., Xu, J.Q., Liu, K.T., to CO exhibiting high faradaic efficiency.
the reduction of CO2 on tin electrodes using Liang, L., Yao, T., Pan, B.C., Wei, S.Q., and Xie, J. Mater. Chem. A 4, 15320–15326.
in situ ATR-IR spectroscopy. ACS Catal. 5, Y. (2016). Metallic tin quantum sheets confined
3148–3156. in graphene toward high-efficiency carbon 80. Lu, X.Y., Tan, T.H., Ng, Y.H., and Amal, R.
dioxide electroreduction. Nat. Commun. 7, (2016). Highly selective and stable reduction of
56. Hansen, H.A., Varley, J.B., Peterson, A.A., and 12697. CO2 to CO by a graphitic carbon nitride/
Norskov, J.K. (2013). Understanding trends in carbon nanotube composite electrocatalyst.
the electrocatalytic activity of metals and 69. Gao, S., Jiao, X.C., Sun, Z.T., Zhang, W.H., Sun, Chem. Eur. J. 22, 11991–11996.
enzymes for CO2 reduction to CO. J. Phys. Y.F., Wang, C.M., Hu, Q.T., Zu, X.L., Yang, F.,
Chem. Lett. 4, 388–392. Yang, S.Y., et al. (2016). Ultrathin Co3O4 layers 81. Sun, X.F., Zhu, Q.G., Kang, X.C., Liu, H.Z., Qian,
realizing optimized CO2 electroreduction to Q.L., Zhang, Z.F., and Han, B.X. (2016).
57. Chaplin, R.P.S., and Wragg, A.A. (2003). Effects formate. Angew. Chem. Int. Ed 55, 698–702. Molybdenum-bismuth bimetallic chalcogenide
of process conditions and electrode material nanosheets for highly efficient electrocatalytic
on reaction pathways for carbon dioxide 70. Gao, S., Sun, Z.T., Liu, W., Jiao, X.C., Zu, X.L., reduction of carbon dioxide to methanol.
electroreduction with particular reference to Hu, Q.T., Sun, Y.F., Yao, T., Zhang, W.H., Wei, Angew. Chem. Int. Ed 55, 6770–6774.
formate formation. J. Appl. Electrochem. 33, S.Q., and Xie, Y. (2017). Atomic layer confined
1107–1123. vacancies for atomic-level insights into carbon 82. Min, X.Q., and Kanan, M.W. (2015). Pd-
dioxide electroreduction. Nat. Commun. 8, catalyzed electrohydrogenation of carbon
58. Peterson, A.A., Abild-Pedersen, F., Studt, F., 14503.
Rossmeisl, J., and Norskov, J.K. (2010). How dioxide to formate: high mass activity at low
copper catalyzes the electroreduction of overpotential and identification of the
71. Sreekanth, N., Nazrulla, M.A., Vineesh, T.V.,
carbon dioxide into hydrocarbon fuels. Energ. deactivation pathway. J. Am. Chem. Soc. 137,
Sailaja, K., and Phani, K.L. (2015). Metal-free
Environ. Sci. 3, 1311–1315. 4701–4708.
boron-doped graphene for selective
electroreduction of carbon dioxide to formic
59. DeWulf, D.W., Jin, T., and Bard, A.J. (1986). 83. Chai, G.L., and Guo, Z.X. (2016). Highly
acid/formate. Chem. Commun. 51, 16061–
Select this article electrochemical and surface effective sites and selectivity of nitrogen-
16064.
studies of carbon dioxide reduction to doped graphene/CNT catalysts for CO2
methane and ethylene at copper electrodes in 72. Li, F.W., Chen, L., Xue, M.Q., Williams, T., electrochemical reduction. Chem. Sci. 7, 1268–
aqueous solutions. J. Electrochem. Soc. 136, Zhang, Y., MacFarlane, D.R., and Zhang, J. 1275.
1686–1691. (2017). Towards a better Sn: efficient
electrocatalytic reduction of CO2 to formate by 84. Liu, Y.J., Zhao, J.X., and Cai, Q.H. (2016).
60. Kuhl, K.P., Hatsukade, T., Cave, E.R., Abram, Sn/SnS2 derived from SnS2 nanosheets. Nano Pyrrolic-nitrogen doped graphene: a metal-
D.N., Kibsgaard, J., and Jaramillo, T.F. (2014). Energy 31, 270–277. free electrocatalyst with high efficiency and
Electrocatalytic conversion of carbon dioxide selectivity for the reduction of carbon dioxide
to methane and methanol on transition metal 73. Lee, C., Zhao, Y., Wang, C.Y., Mitchell, D.R.G., to formic acid: a computational study. Phys.
surfaces. J. Am. Chem. Soc. 136, 14107–14113. and Wallace, G. (2017). Rapid formation of self- Chem. Chem. Phys. 18, 5491–5498.
organised Ag nanosheets with high efficiency
61. Hoang, T.T.H., Ma, S., Gold, J.I., Kenis, P.J.A., and selectivity in CO2 electroreduction to CO. 85. Saravanan, K., Gottlieb, E., and Keith, J.A.
and Gewirth, A.A. (2017). Nanoporous copper Sustainable Energy Fuels 1, 1023–1027. (2017). Nitrogen-doped nanocarbon materials
films by additive-controlled electrodeposition: under electroreduction operating conditions
CO2 reduction catalysis. ACS Catal. 7, 3313– 74. Asadi, M., Kim, K., Liu, C., Addepalli, A.V., and implications for electrocatalysis of CO2.
3321. Abbasi, P., Yasaei, P., Phillips, P., Behranginia, Carbon 111, 859–866.

586 Chem 3, 560–587, October 12, 2017


86. Gao, D.F., Zhou, H., Wang, J., Miao, S., Yang, for electrocatalytic reduction of carbon Vajtai, R., et al. (2016). A metal-free
F., Wang, G.X., Wang, J.G., and Bao, X.H. dioxide. J. Am. Chem. Soc. 137, 14129–14135. electrocatalyst for carbon dioxide reduction to
(2015). Size-dependent electrocatalytic multi-carbon hydrocarbons and oxygenates.
reduction of CO2 over Pd nanoparticles. J. Am. 91. Le, M., Ren, M., Zhang, Z., Sprunger, P.T., Kurtz, Nat. Commun. 7, 13869.
Chem. Soc. 137, 4288–4291. R.L., and Flake, J.C. (2011).
Electrochemical reduction of CO2 to CH3OH at 96. Jovanov, Z.P., Hansen, H.A., Varela, A.S.,
87. Quan, F.J., Zhong, D., Song, H.C., Jia, F.L., and copper oxide surfaces. J. Electrochem. Soc. Malacrida, P., Peterson, A.A., Norskov, J.K.,
Zhang, L.Z. (2015). A highly efficient zinc 158, E45–E49. Stephens, I.E.L., and Chorkendorff, I. (2016).
catalyst for selective electroreduction of Opportunities and challenges in the
carbon dioxide in aqueous NaCl solution. 92. Karamad, M., Hansen, H.A., Rossmeisl, J., and electrocatalysis of CO2 and CO reduction using
J. Mater. Chem. A 3, 16409–16413. Norskov, J.K. (2015). Mechanistic pathway in bifunctional surfaces: a theoretical and
the electrochemical reduction of CO2 on RuO2. experimental study of Au-Cd alloys. J. Catal.
88. Lin, S., Diercks, C.S., Zhang, Y.B., Kornienko, ACS Catal. 5, 4075–4081. 343, 215–231.
N., Nichols, E.M., Zhao, Y.B., Paris, A.R., Kim,
D., Yang, P., Yaghi, O.M., and Chang, C.J. 93. Shen, H.M., Li, Y.W., and Sun, Q. (2017). CO2
electroreduction performance of 97. Velez-Fort, E., Mathieu, C., Pallecchi, E.,
(2015). Covalent organic frameworks
phthalocyanine sheet with Mn dimer: a Pigneur, M., Silly, M.G., Belkhou, R.,
comprising cobalt porphyrins for catalytic CO2
theoretical study. J. Phys. Chem. C 121, 3963– Marangolo, M., Shukla, A., Sirotti, F., and
reduction in water. Science 349, 1208–1213.
3969. Ouerghi, A. (2012). Epitaxial graphene on 4H-
89. Li, Y.W., Su, H.B., Chan, S.H., and Sun, Q. SiC(0001) grown under nitrogen flux: evidence
(2015). CO2 electroreduction performance of 94. Shin, H., Ha, Y., and Kim, H. (2016). 2D covalent of low nitrogen doping and high charge
transition metal dimers supported on metals: a new materials domain of transfer. ACS Nano 6, 10893–10900.
graphene: a theoretical study. ACS Catal. 5, electrochemical CO2 conversion with broken
6658–6664. scaling relationship. J. Phys. Chem. Lett. 7, 98. Chang, X.X., Wang, T., and Gong, J.L. (2016).
4124–4129. CO2 photo-reduction: insights into CO2
90. Kornienko, N., Zhao, Y.B., Kiley, C.S., Zhu, C.H., activation and reaction on surfaces of
Kim, D., Lin, S., Chang, C.J., Yaghi, O.M., and 95. Wu, J.J., Ma, S.C., Sun, J., Gold, J.I., Tiwary, C., photocatalysts. Energ. Environ. Sci. 9, 2177–
Yang, P.D. (2015). Metal-organic frameworks Kim, B., Zhu, L.Y., Chopra, N., Odeh, I.N., 2196.

Chem 3, 560–587, October 12, 2017 587

You might also like