Venkata Sai Sriram
Mosali
a,
Alan M.
Bond
ab and
Jie
Zhang
*ab
aSchool of Chemistry, Monash University, Clayton 3800, Victoria, Australia. E-mail: jie.zhang@monash.edu
bARC Centre of Excellence for Electromaterials Science, Monash University, Clayton 3800, Victoria, Australia
First published on 7th October 2022
Excessive reliance on fossil fuels has led to the release and accumulation of large quantities of CO2 into the atmosphere which has raised serious concerns related to environmental pollution and global warming. One way to mitigate this problem is to electrochemically recycle CO2 to value-added chemicals or fuels using electricity from renewable energy sources. Cu is the only metallic electrocatalyst that has been shown to produce a wide range of industrially important chemicals at appreciable rates. However, low product selectivity is a fundamental issue limiting commercial applications of electrochemical CO2 reduction over Cu catalysts. Combining copper with other metals that actively contribute to the electrochemical CO2 reduction reaction process can selectively facilitate generation of desirable products. Alloying Cu can alter surface binding strength through electronic and geometric effects, enhancing the availability of surface confined carbon species, and stabilising key reduction intermediates. As a result, significant research has been undertaken to design and fabricate copper-based alloy catalysts with structures that can enhance the selectivity of targeted products. In this article, progress with use of alloying strategies for development of Cu-alloy catalysts are reviewed. Challenges in achieving high selectivity and possible future directions for development of new copper-based alloy catalysts are considered.
Fig. 1 Monthly mean CO2 emissions from 1980 to May 2022 as reported by the NOAA global monitoring laboratory.2 |
A range of methods such as photochemical,4–7 biochemical,8–10 and photo-/thermochemical,11–15 have been explored for conversion of CO2 to value added chemicals or fuels. However, their success is highly sensitive to the reaction environment.16 Another possibility is use of the electrochemical CO2 reduction reaction (eCO2RR) which can be performed under ambient conditions using non-toxic materials and electrical energy from renewable sources, making it a highly desirable approach.17
The eCO2RR can be performed in aqueous and non-aqueous media where the former are attractive in terms of non-toxicity, low cost, good conductivity, high proton availability and eco-friendliness. Aqueous electrolytes used are generally alkali salts comprising Na+, K+, Li+, Rb+, or Cs+ cations and OH−, HCO3−, Cl−, H2PO4− or SO42− anions. Both cations and anions can have a strong impact on the efficiency of eCO2RR as well as product selectivity.18–21 Additionally, other factors, such as temperature and pH (more critically local pH, since it can differ substantially from the bulk value, especially under high current density/low buffer capacity conditions) also influence the product selectivity during the eCO2RR.22–24Table 1 provides several examples of eCO2RR products generated in aqueous media and their reversible potentials (E°′) vs. the reversible hydrogen electrode (RHE). Significant drawbacks of eCO2RR in aqueous solutions are the low solubility of CO2 resulting in low current densities, the presence of a competitive hydrogen evolution reaction (HER) which suppresses the overall eCO2RR efficiency, in addition to sluggish kinetics and hence a large overpotential required to produce commercially desirable products at a sufficient rate.17,21,25
Half-cell reaction | E°′ (V vs. RHE) |
---|---|
CO2 + 2H+ + 2e− → HCOOH(aq) | −0.12 |
CO2 + 2H+ + 2e− → CO(g) + H2O | −0.10 |
CO2 + 6H+ + 6e− → CH3OH(aq) + H2O | 0.03 |
CO2 + 8H+ + 8e− → CH4(g) + 2H2O | 0.17 |
2CO2 + 8H+ + 8e− → CH3COOH(aq) + 2H2O | 0.11 |
2CO2 + 10H+ + 10e− → CH3CHO (aq) + 3H2O | 0.06 |
2CO2 + 12H+ + 12e− → C2H4(g) + 4H2O | 0.08 |
2CO2 + 12H+ + 12e− → C2H5OH(aq) + 3H2O | 0.09 |
2CO2 + 14H+ + 14e− → C2H6(g) + 4H2O | 0.14 |
3CO2 + 16H+ + 16e− → C2H5CHO(aq) + 5H2O | 0.09 |
3CO2 + 18H+ + 18e− → C3H7OH(aq) + 5H2O | 0.10 |
To date, three types of electrolysers have been mainly used for the eCO2RR. The H-cell is most commonly used since it is easy to construct and facilitates rapid screening of catalysts and electrolyte composition. Fig. 2(a) shows a conventional H-cell for eCO2RR. In this, the aqueous electrolytes used in most cases in both anodic and cathodic compartments are pre-saturated with CO2. The cell is then sealed prior to performing electrolysis allowing gaseous products to be collected and quantified. The solubility of molecular CO2 in these electrolytes is around 30 mM at 1 bar and thus the rate of CO2 reduction is limited with a current density of typically less than 100 mA cm−2.26 To overcome the drawback of the low mass transport rate associated with the H-cell, electrolysers equipped with a gas diffusion electrode (GDE) have been developed. In these electrolysers, the CO2 gas is diffused through a thin electrolyte layer on the GDE, with the diffusion length being ∼50 nm.27 The much higher mass transport rate achieved under these conditions allows an industrially appropriate current density of above 0.5 A cm−2 to be achieved.27–29 Two types of such electrolysers are commonly used; the flow-cell and membrane electrode assembly (also known as zero-gap) electrolysers. To fabricate a GDE, a gas diffusion layer (GDL) is coated with a layer of catalyst. GDLs composed of carbon fibers possess high porosity.30,31 The structure of the GDE has an impact on the transport of both reactants and products, which affects chemical species adsorption and depletion at the catalyst surface, and hence the overall eCO2RR performance.31,32
Fig. 2 Illustration of (a) H-cell, (b) flow cell and (c) zero gap MEA electrolysers used for eCO2RR. |
Properties such as porosity, electrical conductivity and hydrophobicity of the GDLs are now being tailored to improve the cell performance.30 A typical flow-cell electrolyser is shown in Fig. 2(b) and consists of three channels, one for the CO2 gas flow and the other two for electrolyte (catholyte and anolyte). The GDE is placed in between the CO2 gas and catholyte channels. The catholyte and anolyte are separated by a polymer exchange membrane (PEM). The PEM used in the cell also greatly impacts the efficiency of the electrolyser. Cation exchange membranes (CEM) and anion exchange membranes (AEM) are the two types of PEM used depending on the reaction environment. Nafion, a CEM, was explored for eCO2RR by Delacourt et al.33 and found to suffer from excessive H2 production and lack of stability in long-term electrolysis. AEMs are better suited for eCO2RR circumstances because neutral or alkaline media can be used to minimize the competing H2 evolution reaction. Since anions are far less mobile than protons, both in solution and in the membrane, an additional driving force (i.e. larger overpotential) is needed to overcome the resistance effect. Further, the transport of bicarbonate ions away from the cathode also diminishes the eCO2RR performance.30,34,35 The ratio of CO2 and the proton source in a flow cell electrolyser is significantly higher than that found in an H-cell electrolyser. Furthermore, during the eCO2RR, CO2 is reduced and protons are consumed. Hence, the local reaction environment under the flow-cell conditions may be drastically different from that found under H-cell conditions due to its inherently higher CO2 mass transport rate and hence higher current density. Accordingly, the results obtained with an H-cell may not be applicable under flow-cell conditions. Consequently, eCO2RR performance ideally should be assessed under commercially relevant conditions using a flow cell-electrolyser. H-cell and flow-cell electrolysers are mainly used for fundamental research. The presence of a reference electrode in these configurations allows the anode and cathode to be investigated separately. In commercial applications, a reference electrode is not needed. A membrane electrode assembly (MEA) electrolyser is commonly used, inspired by developments of fuel cells, a closely related research area. A MEA electrolyser is composed of two channels, one for CO2 and the other for anolyte (Fig. 2c). The PEM is sandwiched between the anode and the cathode to minimize the cell resistance. Use of a MEA electrolyser allows the full cell performance, including energy efficiency, to be evaluated.
The catalyst employed in an eCO2RR is critical in determining the eCO2RR product selectivity.36 During electrolysis, the CO2 molecule commonly is adsorbed onto the catalyst present in the cathode. Accordingly, the CO2 reduction pathway usually depends on the binding strength of CO2 and its reduced intermediates over the catalyst surface, as depicted in Fig. 3 and 4. CO2 adsorbed on the catalyst is initially reduced to either carbonyl (*COOH) (an asterisk is used to indicate an adsorbed species) or formyl (*OCHO) intermediates in aqueous electrolyte which can be further reduced to CO and formate, respectively. CO is electroactive. However, if CO binds to the catalyst too strongly, then further reduction to desirable products and their subsequent desorption is inhibited. Poisoning of the catalyst also is likely to occur. Consequently, hydrogen evolution rather than CO2 reduction may become the favoured reduction pathway. This scenario occurs at Pt, Ni, and Fe electrodes.37 Alternatively, if the *CO binding ability to the catalyst is weak, as applies with metals such as Zn, Au, and Ag, then high selectivity for CO as the product is achieved.37 If the catalyst has moderate binding strength for *CO, further reduction is facilitated, leading to more reduced carbon products. Cu is the only catalyst that has moderate *CO binding energy, leading to unique prospects for formation of a wide range of products. Metals which bind eCO2RR intermediates too weakly are also likely to bind *H weakly. If these catalysts also possess strong *O affinity, an O-bound *OCHO intermediate will be formed during the initial reaction step which leads to formation of HCOOH.
As noted above, extensive eCO2RR studies with many metals in the periodic table have revealed that Cu has the unique capability of producing an extensive range of reduced carbon products containing one carbon (C1), two carbons (C2) or multi-carbons (i.e. more than two carbons; C2+). The 16 products detected using Cu catalysts during eCO2RR and their structures are shown in Fig. 5.38–41 This on one hand demonstrates the versatility of Cu catalysts, but on the other hand implies that poor product selectivity is a major challenge to address for commercial applications.
Fig. 5 Graphical representation of products detected during eCO2RR at a Cu electrode. Adopted and redrawn.38 |
In attempts to address the selectivity issue, nanostructuring, alloying, defect engineering, atomizing, and the synthesis of oxide, sulphide, and nitride-derived Cu as well as other methods have been employed.17,21,42–49
Since alloying is one of the most promising and fundamental strategies for manipulating the reaction pathways to obtain desired products, this review focuses on what has been achieved to date and what might be accomplished in the future by alloying Cu with other eCO2RR active metals. Several review articles17,25,50–54 have been published on the use of alloyed-Cu catalysts. However, they focus on a much broader aspect of eCO2RR with copper. In this review, the correlation between alloy type and selectivity is examined and future directions in the development of Cu based alloys as effective eCO2RR catalysts are proposed. Finally, in order to provide an overview of the status of findings with alloyed-Cu catalysts, comparisons of what has been achieved with state-of-the-art Cu alloys and other Cu based catalysts for each eCO2RR product also are provided in a table at the end of each section.
(a) Faradaic efficiency: faradaic efficiency (FE) is the indicator of the selectivity of the products obtained with a catalyst during eCO2RR. FE (reported in %) is calculated by using the formula given in eqn (1)
(1) |
(b) Overpotential: overpotential (η) is the difference between the applied potential (Eapplied) to obtain a product and the equilibrium potential (E°′). It is calculated by using the formula given in eqn (2)
η = |Eapplied − E°′|. | (2) |
(c) Current density: current density (j) denotes the rate/s of electrochemical reaction/s. It is the catalytic current (i) generated per unit area. Current densities presented in many studies are calculated with respect to the geometric area of the electrode. However, ideally an electrochemical active surface area should be used. Partial current density (jproduct) indicates the rate of formation of a given product and calculated by the formulae given in eqn (3) and (4)
(3) |
jproduct = jtotal × FEproduct | (4) |
(d) Stability: the stability is the lifetime of a catalyst for performing eCO2RR. In general, it is the time period for which a catalyst provides stable activity and selectivity.
The structure of the alloy or in other words, its atomic arrangement play a crucial role in determining the selectivity of the products formed during eCO2RR. Alloy systems can be ordered, intermetallic, disordered, core/shell structured, phase-separated and have high entropy structures (Fig. 6(a)). All these variations have been explored in the design of catalysts for eCO2RR. Intermetallic alloy systems having an ordered atomic arrangement are the most stable form of alloys. With thermodynamically stable alloy forms, the electronic structure and atomic co-ordination environment can be controlled precisely as they possess a specific crystal structure.57 By altering the alloy composition, bond lengths can be tuned, which effects the binding energies of molecules adsorbed on the surface of the catalysts and hence their catalytic behaviour. Disordered alloys are solid-solution mixtures with irregular arrangements of binary atoms. They are generally thermodynamically unstable and their local electronic and the geometry of their structures often changes during the electrolysis process.57,58
Fig. 6 (a) Alloy systems reported in the literature. (b) Illustration of binding of eCO2RR intermediates to Cu, Cu–M and Cu–M′ alloy catalysts where M is the guest metal with O-affinity and M′ is the guest metal with H-affinity, (c) H and O binding energies for designated metals, (d) M–H bond strength shown as a function of bond dissociation enthalpy of metal oxides, (e) classification of metals into different groups with respect to Cu, based on their O and H affinities. Reprinted with permission.25 Copyright 2018, Elsevier Inc. |
The catalytic properties in core/shell structured alloys are regulated by the strain and ligand effects that arise between the core and shell materials. Strain effects such as tensile stress on the surface of atoms arise due to the lattice mismatch between core/shell interface.59,60 Ligand effects are short-range and only exist within two or three atomic layers from the surface.61 Atomic vacancies and metal doping in the shell material also can occur. They are a function of the thickness of the shell and also impact on the product selectivity during electrolysis.62 In eCO2RR, the major benefit of such alloys is the creation of phase interfaces where the products formed at one phase can migrate through the interface and be further reduced to C2+ products. These catalytic Cu alloy forms are being extensively studied as a means of increasing the CO environment at Cu and facilitating the production of highly reduced C2+ products such as ethylene and ethanol.63–65 Details of the efficiency of phase separated alloy catalysts and the CO spill-over effect are discussed below.
High entropy alloys (HEAs) are a novel class of alloys recently being investigated for eCO2RR. These alloys are formed by about five metals, with each element constituting between 5–35% and with variation in the elemental ratios. Accordingly, a large variety of elemental combinations is possible in the alloy and consequently a wide range of catalytic capabilities can be achieved.66,67
Compared to bimetallic alloys, HEAs provide superior stability and durability particularly under harsh reaction conditions such as high temperature and high electrochemical potential due to their high mixing entropy and hence low Gibbs free energy as quantified by eqn (5).67,68
ΔGmix = ΔHmix − TΔSmix | (5) |
Grouping of metals according to their binding affinities for *H and *O relative to copper was proposed to facilitate the selection of a guest metal to form a binary alloy with Cu.25 In the alloy, the bimetallic catalyst material will have two binding sites available to achieve stabilisation of a desired intermediate. For example, metals which produce hydrogen have a more robust *H and *O affinity than Cu while those with weaker *O and *H binding strengths, produce CO via a *COOH intermediate. Metals having a lower H affinity but higher O-affinity than Cu favour formation of the *OCHO intermediate, which leads to formation of formate. Hence, metals with good O-affinity can be alloyed with Cu to stabilize O-bound intermediates, while those having H-affinity can stabilize H binding in the intermediates, as illustrated in Fig. 6(b). Fig. 6(c–e) show a comparison of O and H affinities of some metals with respect to copper and the grouping of metals. In the following sections, the factors that govern the selectivity of Cu alloy towards each product are summarized and discussed to guide the development of advanced alloy catalysts. Examples of the state-of-the-art Cu alloys and other Cu based catalysts for each product are also given.
Copper modified with Pd and Pd–H produces significant amounts of formate. For example, early studies by Fujishima and co-workers showed substantial formate production at the Pd modified Cu electrodes.100,101 Zhang et al.102 have synthesised a stannate derived Sn–Cu bimetallic catalyst via cation exchange with Na2Sn(OH)6 followed by annealing and electroreduction. This Cu–Sn catalyst produced 86% formate at −1.1 V vs. RHE in 0.5 M NaHCO3 with a partial current density of ∼11 mA cm−2. These authors also reported that a slight increase in formate faradaic efficiency could be achieved by incorporating the Cu–Sn catalyst into reduced graphene oxide along with a major increase in partial current density. The Hod group103 reported the favorable formation of formate with 3D structured Cu2S catalysts prepared by an electrochemically driven cation exchange mechanism. To prepare such electrodes, pre-synthesised CoSx nanosheets based 3D structures were taken as templates and a cation exchange reaction with Cu2+ (where Co2+ ions in CoS2 were replaced with Cu1+/2+ ions) was performed electrochemically in a 0.1 M LiClO4-dimethylformamide electrolyte solution containing 3 mM Cu(NO3)2 by applying a potential of −0.47 V vs. normal hydrogen electrode for varied time periods. The amount of Co2+ replaced with Cu2+ ions was varied by controlling the charge passed during the electrochemical reaction. The formate production was found to be somewhat dependent on the percentage of Co2+ ions exchanged by Cu2+. With a Cu2S sample obtained by passing 2.0 C of charge, 87.3% of formate was produced with a partial current density of 19 mA cm−2 at −0.9 V vs. RHE in 0.1 M NaHCO3 solution saturated with CO2. The authors reported that variation of the grain boundary obtained with the modulation of binary cations was the origin of the change in the product selectivity.
Ag is an excellent CO formation catalyst.104 However, AgCu alloy can be designed to favour formate formation. Recent work on CuAg bimetallic nanoarchitectures demonstrated the importance of electronic and geometric modulation of the catalyst with respect to product selectivity.73 The sponge-like Ag91Cu9 and coralline Ag65Cu35 nanoalloys prepared by anodising Ag52Cu39Sn9 alloy foil induced a dramatic change in the selectivity of CO versus HCOOH. Changing the composition of the AgCu alloy, modifies the electronic arrangement of Ag and Cu and hence the binding strengths of intermediates as demonstrated by in situ Raman spectroscopic measurements shown in Fig. 7. Thus, on increasing the Cu content of the AgCu alloy, the selectivity shifted from favouring CO to forming formate with a faradaic efficiency of 96%.
Fig. 7 (a and b) In situ Raman spectra obtained during electrolysis and (c) proposed reaction pathways on spongeous Ag91Cu9 and coralline Ag65Cu35 alloy nanoarchitectures. Reprinted with permission.73 Copyright 2020, Elsevier Ltd. |
Table 2 provides details of some state-of-the-art Cu alloys and other Cu based catalysts for selective formate production.
Catalyst | Electrolysis cell | Electrolyte | Potential (V vs RHE)a | FE (%) | j HCOO– (mA cm−2) | Ref. |
---|---|---|---|---|---|---|
a Potential at which maximum FE was obtained. | ||||||
Cu–Sn/rGO | H-cell | 0. 5M NaHCO3 | −1.0 | 87.4 | 20.7 | 102 |
Cu2S-2.0C | H-cell | 0.1 M NaHCO3 | −0.9 | 87.3 | 19.1 | 103 |
Coralline Ag65Cu35 | H-cell | 0.1 M KHCO3 | −1.0 | 91.8 | 16.8 | 73 |
CuCd@Cu-20,45 | H-cell | 0.5 M NaHCO3 | −1.1 | 70.1 | 26.8 | 99 |
Porous Cu6.26Sn5 | H-cell | 0.1 M KHCO3 | −1.1 | 97.8 | ∼30 | 136 |
CuBi-100 | H-cell | 0.5 M KHCO3 | −1.0 | 94.7 | 13 | 137 |
Cu/Au | H-cell | 0.5 M KHCO3 | −0.6 | 81 | 10.4 | 138 |
Sn–Cu | H-cell | 0.1 M KHCO3 | −0.95 | 92 | ∼10 | 139 |
Cu@Sn nanocones | H-cell | 0.1 M KHCO3 + 0.3 M KCl | −1.1 | 90.4 | 52 | 140 |
CuSn alloy@Cu doped SnO | H-cell | 0.5 M KHCO3 | −1.2 | 95.4 | 30.3 | 141 |
C11.5In88.5-OH | H-cell | 0.1 M KHCO3 | −1.1 | 85 | ∼10 | 142 |
MOF derived CuBi | H-cell | 0.5 M KHCO3 | −0.77 | 100 | 67.9 | 143 |
Cu2O/CuO/CuS | H-cell | 0.1 M KHCO3 | −0.7 | 84 | 20 | 144 |
CuBi NPs | H-cell | 0.1 M KHCO3 | −0.99 | 96 | 12.5 | 145 |
CuS | H-cell | 0.1 M KHCO3 | −0.8 | 80 | 18 | 146 |
HCS/Cu-0.12 | H-cell | 0.5 M KHCO3 | −0.81 | 82.4 | 26 | 147 |
S-doped OD-Cu | H-cell | 0.1 M KHCO3 | −0.8 | 73.6 | 13.9 | 148 |
Cu6Sn5/Sn | Flow cell | 1 M KOH | −1.0 | 86.7 | 103 | 149 |
CuSn/Sn | Flow cell | 1 M KHCO3 | −0.7 | 84.2 | ∼26 | 150 |
CuBi | MEA | 0.5 M KHCO3 | −1.07 | 98.3 | 55.6 | 151 |
In early studies in the 1990s, Watanabe et al.55,134 prepared CuZn alloys electrodes by electroplating onto gold electrodes and studied the eCO2RR efficiency in 0.05 M KHCO3 electrolyte. They reported that higher concentrations of Zn in the CuZn alloy led to an increase in CO generation compared to the use of pure Cu. The only other eCO2RR product found was HCOO−. The Berlinguette group135 studied eCO2RR using brass and bronze catalysts. They prepared 21 catalysts with varying Cu:Zn:Sn ratios by drop casting methanolic solutions of Cu, Zn, and/or Sn in the required proportions onto a titanium substrate followed by multiple exposures to near-infrared radiation and electroreduction to produce alloy films. Four-hour eCO2RR experiments with these film electrodes generated H2, CO and HCOOH as products. Cu–Zn–Sn alloys with Sn < 20% generated mainly syngas with the H2:CO ratio depending on the ratio between Cu and Zn. Hu et al.152 prepared Cu–Zn catalysts by annealing a brass substrate at 500 °C in an Ar atmosphere prior to electroreduction of ZnO formed in the Zn metal annealing step, as shown in Fig. 8(a). On annealing, uniform 200 nm particles formed over the brass substrate. Zn and Cu have different melting points of 419.5 °C and 1085 °C, respectively. Accordingly, Zn melts at 500 °C and migrates all over the surface and is then converted to ZnO on exposure to air while Cu remains unchanged. The resultant catalyst used for eCO2RR over a range of applied potentials produced syngas with a 2:1 ratio of CO:H2 (Fig. 8(b)) which is suitable for the Fischer–Tropsch process. The Broekmann group153 synthesised Zn94Cu6 alloy foams, as displayed in Fig. 8(c and d), by dynamic hydrogen bubble template assisted electrodeposition onto a Cu plate. eCO2RR with this catalyst produced 90% CO at −0.95 V vs. RHE in aqueous 0.5 M KHCO3 electrolyte (Fig. 8(e)). Lamaison et al.154 also designed Cu–Zn alloy foam catalysts by electrodeposition of Cu onto Zn plates in their eCO2RR application. These foams were composed of a dendritic CuZn alloy structure (Fig. 8(f and g). As shown in Fig. 8(h), bulk electrolysis experiments performed under eCO2RR conditions with 0.1 M CsHCO3 as the electrolyte revealed an increase in the H2:CO ratio with increasing Cu content.
Fig. 8 (a) Schematic illustration of the preparation of Zn–Cu catalysts from brass foil, (b) faradaic efficiencies, current density and CO/H2 ratio obtained during eCO2RR as a function of time. Reprinted with permission.152 Copyright 2018, Elsevier B.V. (c and d) SEM images of the Zn94Cu6 catalyst, (e) CO faradaic efficiencies obtained with designated CuZn and Zn catalysts. Reprinted with permission.153 Copyright 2018, the American Chemical Society. (f) SEM, (g) scanning transmission electron microscope – energy-dispersive X-ray spectroscopy (STEM-EDXS) mapping images of Zn90Cu10 catalyst (h) ratio of H2/CO faradaic efficiencies obtained with designated Zn100−xCux catalysts. Reprinted with permission.154 Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. |
CuAu alloys, like CuZn ones, have been extensively studied for eCO2RR generation of CO. For example, enhancement of CO at AuxCu100−x alloy catalysts has been reported by Christophe et al.155 In particular, a significant increase in the production of CO was observed with use of an Au50Cu50 catalyst. CO desorption at Au sites is promoted by the presence of adsorbed CO on Cu sites due to dipole repulsion between CO molecules at adjacent sites. Consequently, the presence of both Cu and Au in an optimal ratio enhances CO product formation. Bimetallic AuCu catalysts also induce an electronic effect which results in the shift in the d-band centre and hence changes the binding strength of the reaction intermediates. As a result of the shift in the d-band centre, the way the metals in the AuCu alloy interact with adsorbed intermediate also changes. The implication of this feature was examined by Kim et al.156 who found that the evolution of CO increases with increasing Au content in the AuCu alloy. In the case of Au3Cu, they reported a mass activity (i.e. current density normlised to the amount of catalyst loaded) exceeding 200 A g−1 at −0.73 V versus RHE, which represents a ten-fold increase in comparison with pure Cu. Surface valence band photoemission spectral analysis suggested that the improved selectivity towards CO with Au3Cu could also be due to the synergistic electronic and geometric effects associated with AuCu nanoparticle formation that occurred during the eCO2RR. The atomic arrangement of Au and Cu in AuCu bimetallic alloy nanoparticles (either ordered or disordered) also can have an impact. Among AuCu alloy nanoparticles, the nanoparticles with an ordered arrangement of Au and Cu (o-AuCu) exhibited a factor of two improvement in selectivity for CO (Fig. 9(a)).157 STEM (Fig. 9(b)) and X-ray absorption spectroscopy (XAS) (Fig. 9(c)) analyses revealed that a thin layer of segregated Au atoms was formed on the surface of o-AuCu nanoparticles during electrolysis. DFT calculations suggested that the lattice strain of 6% associated with the Au layer due to the underlying AuCu lattice affected the catalytic activity of o-AuCu nanoparticles.
Fig. 9 (a) Faradaic efficiencies for generation of H2 and CO using AuCu alloy nanopartcles with designated atomic ordering, (b) high-angle annular dark-field imaging – scanning transmission electron microscope (HAADF – STEM) analysis of ordered AuCu nanoparticles, and (c) XAS analysis of ordered and disordered AuCu nanoparticles. Reprinted with permission.157 Copyright 2017, the American Chemical Society. |
Intermetallic CuPd alloy catalysts provide another example of a Cu alloy that enhances CO generation during the eCO2RR process.159–161 Excellent selectivity for CO with a faradaic efficiency of 86% and a partial current density of 6.9 mA cm−2 was obtained on Pd85Cu15 alloy nanoparticles (Fig. 10(a and b)).158 On increasing the Cu content in the alloy, the efficiency decreased. X-ray absorption near edge structure (XANES) analysis shown in Fig. 10(c and d) for the PdCu alloy nanoparticles revealed that variation in Pd–Pd and Cu–Cu bond lengths alters the binding strengths of the CO2 reduction intermediate (*CO). Further, the formation of Pd and Cu oxides at higher Cu to Pd ratios leads to suppression of CO evolution during electrolysis. CuPd alloys with a 1.5:8.5 atomic ratio bind CO less strongly on the surface than with pure Pd, resulting in an enhancement of CO production.
Fig. 10 (a) CO faradaic efficiency PdxCuy nanoparticles, (b) CO partial current densities obtained using designated PdxCuy nanoparticles. (c) Pd K edge, and (d) Cu K edge XANES spectrum of a PdCu/C catalyst. Reprinted with permission.158 Copyright 2016, Elsevier Ltd. (e) A comparison of CO selectivity at PdCu NPs with specified compositions, (f) free energy diagram for CO2, *COOH and CO on a Pd terminated PdCu (111) surface. Reprinted with permission.159 Copyright 2016, the Royal Society of Chemistry. |
The study by Li et al.159 with mesoporous CuPd alloys also shows composition-dependent activity for CO2 conversion (Fig. 10(e)). An alloy with a Cu to Pd ratio of 3:7 gives high CO yields, which diminishes with an increase in the Cu content. DFT calculations predict that Pd terminated Pd7Cu3 alloy surfaces faceted with (111) planes will favour adsorption of a *COOH intermediate and weaken adsorption ability of *CO resulting in excellent CO selectivity (Fig. 10(f)). Due to the differences in electronegativity of Cu and Pd, the presence of Cu atoms adjacent to Pd in a Cu–Pd alloy increases the adsorption of *COOH and CO on Pd surfaces, making Pd sites the active sites for the eCO2RR.
Oxide derived Cu–M (M = In/Sn) catalysts have been surveyed for their ability to convert CO2 to CO. In and Sn in their metallic,162 oxide163–165 or chalcogenide166–168 forms produce HCOO− from the eCO2RR. However, when oxides of these post-transition metals are combined with copper, scaling relationships can be altered due to synergistic effects associated with the lattice mismatch between Cu/CuO and MO2 (M = In/Sn). The result is that product selectivity is shifted from formate to CO with remarkably high faradaic efficiencies. For example, the study by Li et al.,169 shows that the change in product distribution as a function of the thickness of the SnO2 coated onto Cu nanoparticles (Fig. 11(b and c)). With a thin SnO2 shell (0.8 nm), the product selectivity for CO had a FE of >90% at −0.6 V vs. RHE with little formate produced. In contrast, with a thick SnO2 shell (1.8 nm), the eCO2RR behaviour is similar to that at Sn and formate is the major product. According to the results of DFT calculations, the synergistic effect emerged because of the compression strain and self-doping of Cu into the thinner SnO2 shell layer during electrolysis which shifted selectivity from HCOO− to CO. A related study using Cu–In2O3 core–shell nanoparticles (Fig. 11(d))170 also showed that the thinnest shell provided a shift in selectivity to CO (Fig. 11(e and f)). With the thinner In2O3 shell, the lattice mismatch between Cu and In2O3 facilitates self-doping of Cu into the In2O3 shell resulting in alteration of binding energies of *COOH and *OCHO intermediates, which leads to the generation of CO and HCOOH, respectively. Cu oxide surfaces decorated with Sn also enhance conversion of CO2 to CO. Zhao et al.172 decorated electrochemically generated CuO nanowires with Sn. An optimal loading of Sn produced 90% of CO at −0.8 V vs. RHE in 0.1M KHCO3 electrolyte. Sarfraz et al.173 electrochemically deposited Sn on a Cu sheet. With a Sn loading of 3.9 mol cm−2, 90% of CO was obtained at −0.6 V vs. RHE in 0.1 M KHCO3. However, with higher Sn loadings, selectivity reverted to HCOO−. Zeng et al.174 used Cu–Sn foam based dendritic structures, where the Cu dendrites were decorated with small amounts of SnOx to form a Cu/CuOx–SnOx core/shell structure. This allowed 93–94% CO to be formed between −0.75 V and −0.9 V vs. RHE in 0.1 M KHCO3.
Fig. 11 (a) Electron energy loss spectroscopy (EELS) elemental mapping of Cu/SnO2 nanoparticles. (b) and (c) faradaic efficiencies of products obtained during eCO2RR with Cu/SnO2 NPs having a shell thickness of (b) 0.8 nm and (c) 1.8 nm. Reprinted with permission.169 Copyright 2017, the American Chemical Society. (d) EDS elemental mapping of C–Cu/In2O3 nanoparticles. (e) and (f) faradaic efficiencies of products obtained during eCO2RR with Cu/In2O3 nanoparticles having a shell thickness of (e) 0.4 nm and (f) 1.5 nm. Reprinted with permission.170 Copyright 2018, the American Chemical Society. (g) elemental mapping showing the formation of indium islands when undertaking eCO2RR with CuIn electrodes, (h) CO partial current density obtained during eCO2RR at −0.6 V with In2O3/Cu2O and In2O3/Cu electrodes as a function of island diameter and interfacial density. Reprinted with permission.171 Copyright 2018, Nature Publishing Group. |
Studies by the Takanabe group175,176 employing In coated on oxide derived (OD)-Cu175 and CuInO2 derived Cu–In alloy176 catalysts again revealed a drastic increase in CO selectivity. DFT calculations suggest that the replacement of a Cu atom with an In atom would suppress *H adsorption with *CO adsorption energy unchanged, resulting an increase in the FE of CO while decreasing HER. A study by Larrazábal et al.177 proposed that an in situ generated metastable In(OH)3 phase in the oxide plays an important role in the use of Cu–In catalysts. The core–shell structured Cu–In catalysts, with In(OH)3 as the shell evolved upon repeated voltammetric cycles of potential on the initial Cu–In in a 0.1 M KHCO3 solution, exhibited a significant increase in the CO selectivity. The same group also reported the dependence of CO selectivity on the nature of the Cu–In interfaces.171 In their study, the authors microfabricated In2O3 dots on Cu and Cu2O surfaces. In2O3 dots on Cu left the selectivity unaltered, whereas those on Cu2O led to a drastic improvement in selectivity for CO. When the interfacial density is low, both catalysts showed similar activity. However, when the density of metal–oxide interface was increased, the catalytic activity towards CO was enhanced with In2O3/Cu2O, as shown in Fig. 11(h). The irregular layer of Cu formed by reduction of Cu2O enhances the diffusion of In (Fig. 11(g)) giving rise to a synergistic effect at Cu–In interfaces. This study reveals that both the interfaces and Cu2O are essential for achieving synergistic effects that tune the selectivity pathway.
State of the art Cu alloys and other Cu based catalysts that favoured CO production are tabulated in Table 3.
Catalyst | Electrolysis cell | Electrolyte | Potential (V vs. RHE)a | FE (%) | j CO (mA cm−2) | Ref. |
---|---|---|---|---|---|---|
a Potential at which maximum FE was obtained. b Polymer electrolyte membrane electrolysis cell. c Cell voltage. | ||||||
Zn94Cu6 nanofoam | H-cell | 0.5 M KHCO3 | −0.9 | 90 | ∼4.5 | 153 |
o-AuCu NPs | H-cell | 0.1 M KHCO3 | −0.8 | ∼80 | 1.4 | 157 |
Mesoporous Pd7Cu3 | H-cell | 0.1 M KHCO3 | −0.8 | 80 | ∼1.5 | 159 |
Pd85Cu15/C | H-cell | 0.1 M KHCO3 | −0.9 | 86 | 6.9 | 158 |
C–Cu/SnO2-0.8 | H-cell | 0.5 M KHCO3 | −0.7 | 93 | ∼13 | 169 |
Cu–Sn10 nanowires | H-cell | 0.1 M KHCO3 | −0.8 | 90 | 4.5 | 172 |
Cu–Sn | H-cell | 0.1 M KHCO3 | −0.6 | 90 | 1.0 | 173 |
Cu–Sn foam | H-cell | 0.1 M KHCO3 | −0.8 | 94 | 3 | 174 |
PTFE-Cu | H-cell | 0.1 M KHCO3 | −0.4 | 71 | 1.5 | 178 |
Cu NPs-700 | H-cell | 0.1 M KHCO3 | −0.6 | 75.6 | 3.8 | 179 |
Cu–N2/GN | H-cell | 0.1 M KHCO3 | −0.5 | 81 | ∼2 | 180 |
Cu/CNT | PEM Cellb | 0.1 M KHCO3 | −3.5c | 75.7 | 12.2 | 181 |
CuPd nanosheets | Flow cell | 1 M KOH | −0.6 | 71 | 58 | 161 |
2.7 nm PdCu NPs | Flow cell | 1 M KOH | −1.0 | 82 | 80 | 182 |
Sb–Cu2O | Flow cell | 0.1 M KOH | −0.8 | 96 | 60 | 183 |
Electrochemical conversion of CO2 to methane has been explored on Cu–Zn catalysts. Cuenya and co-workers184 studied the eCO2RR product selectivity dependence on the composition and structure of CuZn nanoparticles derived from adoption of the inverse micelle encapsulation method. Their bulk electrolysis eCO2RR experiments with Cu100−xZnx revealed the presence of Zn from 10% to 50% increased CH4 formation while higher 50% Zn levels suppressed CH4 formation and increased that of CO (Fig. 12(a)). The XANES studies suggested that the Cu–ZnO interface is needed to generate CH4 or other hydrocarbons while CuZn (brass) alloy favours formation of CO/H2 mixtures. Due to the faster reduction rate with CuO than with ZnO, in samples with a lower Zn content (Cu100−xZnx, x < 50), the Cu–ZnO interface prevailed in the initial period of time during eCO2RR, when the hydrogenation of CO species was favoured at Cu sites. When both oxides were fully reduced, brass nanoparticles were generated so that Cu loses its ability to stabilise *CHy (y = 1–3) intermediates due to ligand effects and hence CO and H2 were released as the electrolysis products.
Fig. 12 (a) A comparison of faradaic efficiencies for the products obtained by eCO2RR with Cu100−xZnx catalysts, (b) Cu K-edge and (c)Zn K-edge XANES spectra of Cu50Zn50 nanoparticles: spectrum A – as-prepared sample, spectrum B – immediately after the onset of eCO2RR and spectrum C – after 7 h of electrolysis. Reprinted with permission.184 Copyright 2019, the American Chemical Society. Source: https://pubs.acs.org/doi/10.1021/jacs.9b10709. Further permissions related to the material excerpted should be directed to the ACS. |
Ceria (CeO2) is another material that facilitates CH4 production in the presence of Cu sites during eCO2RR.185 Doping CeO2 nanorods with Cu establishes a strong interaction between atomically dispersed Cu sites and CeO2. Accordingly, multiple oxygen vacancies are induced around Ce under eCO2RR conditions, which provides an effective site for electrocatalytic reduction of CO2 to CH4. CeO2 nanorods doped with 4% Cu produce ∼58% of CH4 at −1.8 V vs. RHE with a current density of 56 mA cm−2 in 0.1 M KHCO3, as shown in Fig. 13(b). The Buonsanti group186 has explained the importance of the interaction between CeO2 and Cu in facilitating CH4 formation by examining outcomes of the eCO2RR with Cu/CeO2 heterodimer (HD) nanoparticles as well as physical mixtures of Cu and CeO2 nanoparticles. In addition to the presence of CeO2 interactions with Cu nanoparticles described above, the size of the Cu nanoparticles was shown to plays a significant role in the improvement in CH4 selectivity. The 36 nm sized Cu/CeO2 HD nanoparticles enhanced eCO2RR performance over HER and the selectivity achieved for CH4 reached 54% at −1.2 V vs. RHE in 0.1 M KHCO3. In contrast, the Cu and CeO2 physical mixture resulted in only <10% of methane. XAS data (Fig. 13(d)) confirmed that during eCO2RR, Ce4+ in Cu/CeO2 HD is reduced to Ce3+. On the basis of DFT calculations, the eCO2RR intermediates (*COOH, *CHO, H2CO*, H3CO*) were postulated to be adsorbed onto both Cu and Ce sites with an O-vacancy in order to form CH4. Thus, oxygen vacancies played a crucial role in enhancing methane selectivity.
Fig. 13 (a) Comparison of product selectivity obtained with Cu doped CeO2 catalysts, (b) long term stability of a Cu–CeO2-4% (4% of Cu doped in CeO2) catalyst at −1.2 V vs. RHE. Reprinted with permission.185 Copyright 2018, the American Chemical Society. (c) Product selectivity comparison and CO2RR partial current densities obtained with Cu/CeO2−x HDs, Cu–CeO2−x mixture, Cu nanocrystals (NCs), and CeO2−x NCs, (d) Ce LIII-edge XANES spectra collected every eight minutes when using Cu/CeO2−x HDs under eCO2RR conditions at −1.2 V vs. RHE in 0.1 M KHCO3. Reprinted with permission.186 Copyright 2019, the American Chemical Society. |
Chang et al.187 have studied the impact of structural reconstruction of Cu100−xAgx catalysts on product selectivity using the techniques of in situ grazing-angle X-ray scattering/diffraction, X-ray absorption spectroscopy, and Raman spectroscopy. The authors showed that Cu68Ag32 nanowires underwent dynamic oxidation–reduction cycles on the nanowire surface resulting in an inter diffusion of Ag and Cu atoms prior to stabilization of the metallic states of Cu and Ag. The high methane FE of 60% is three times to that achieved with pristine Cu nanowire.
Albo et al.195 studied the use of Cu2O/ZnO catalysts for eCO2RR in a flow cell using electrodes prepared by airbrushing commercially available Cu2O and ZnO nanoparticles over carbon paper. They found that the ratio of Cu2O/ZnO has a significant effect on methanol generation during eCO2RR. Cu2O/ZnO with 2:1 ratio generated a high selectivity for methanol with a FE of 25.2% at −1.3 V vs. Ag/AgCl (sat. KCl) in 0.5 M KHCO3. Recently, Bagchi et al.192 reported excellent selectivity towards methanol could be achieved with an intermetallic CuGa2 electrocatalyst. An exceptional FE of 77.6% for methanol was achieved at −0.3 V vs. RHE. X-ray photoelectron spectroscopy (XPS) and in situ X-ray absorption fine structure (XAFS) analysis revealed the importance of surface and subsurface oxides of Ga at low potentials in enhancing methanol selectivity. However, when more negative potentials were applied, lattice expansion occurs, indicating reduction of Ga2O3, which resulted in the diminution of methanol selectivity.
Table 4 summarises state-of-the-art Cu alloys and other Cu based catalysts, which are selective for methane and methanol in H type and flow cells.
Catalyst | Product | Electrolysis cell | Electrolyte | Potential (V vs. RHE)a | FE (%) | j product (mA cm−2) | Ref. |
---|---|---|---|---|---|---|---|
a Potential at which maximum FE was obtained. | |||||||
Cu–ZnO NPs | CH4 | H-cell | 0.1 M KHCO3 | −1.35 | 70 | ∼40 | 184 |
4% Cu doped CeO2 | CH4 | H-cell | 0.1 M KHCO3 | −1.8 | 58 | 33.6 | 185 |
Cu68Ag32 nanowires | CH4 | H-cell | 0.5 M KHCO3 | −1.2 | 60 | ∼27 | 187 |
Cu foil | CH4 | H-cell | 0.5 M NaHCO3 + methyl carbamate | −2.1 | 85 | 31 | 188 |
Single atom Cu/GYD | CH4 | H-cell | 0.1 M KHCO3 | −1.3 | 66 | 20 | 189 |
Single atom Cu/TCNFs | CH3OH | H-cell | 0.1 M KHCO3 | −0.9 | 44 | 41 | 190 |
Cu (111) nanospheres | CH4 | Flow cell | 1 M KOH | −0.91 | 53 | 53 | 191 |
CuGa2 | CH3OH | Flow cell | 1 M KOH | −0.3 | 78 | ∼16 | 192 |
Fig. 14 (a) Schematic representation of the preparation of CuZn alloy nanoparticles, (b) faradaic efficiency of C2H4 obtained by eCO2RR using CuZn alloy, Cu–Zn mixtures and Cu catalysts, (c) mechanism proposed for C2H4 formation with CuZn catalysts. Reprinted with permission.201 Copyright 2018, the American Chemical Society; (d) schematic representation of use of a Cu–ZnO tandem catalyst to generate CO2 → C2+, (e) faradaic efficiency and current density obtained with Cu, Cu and ZnO mixtures and Cu–ZnO tandem catalysts. Reprinted with permission.203 Copyright 2020, Elsevier Inc. |
Garcia et al.202 have synthesized CuO/ZnO particles by a water/oil microemulsion method. Bulk electrolysis experiments using this catalyst in 0.1 M KHCO3 using a flow cell, produced ethylene with a faradaic efficiency of 91.1% at highly negative potentials (−2.5 V vs. Ag/AgCl). Recently, Zhang et al.203 prepared a Cu/ZnO tandem catalyst by airbrushing commercial Cu and ZnO nanoparticles over carbon paper and performed bulk electrolysis with a flow cell in 1 M KOH solution (Fig. 14(d)). With this Cu/ZnO tandem catalyst, the combination of ZnO acting as a CO generator and Cu as a C2 intermediate stabilizer enhances C2 and C2+ product formation. With an optimal loading of Cu and ZnO, Cu/ZnO tandem catalyst produces 48% ethylene at −0.7 V vs. RHE in 1 M KOH (Fig. 14(e)).
Use of uniformly distributed CuAg alloy nanowires prepared by electroplating in 3,5-diamino-1,2,4-triazole baths, significantly enhances C2 product generation (ethylene in particular) during eCO2RR in a flow reactor.204 The faradaic efficiency of ∼60% for ethylene at −0.7 V vs. RHE and a total current density of 300 mA cm−2 achieved with CuAg alloy nanowires containing 6% Ag, demonstrates superior CO2 reduction activity as shown in Fig. 15(a). Examination of in situ Raman spectra shown in Fig. 15(b and c) led to the conclusion that high stability of Cu2O layers under electrolysis conditions and the optimal availability of CO were the factors that resulted in high C2H4 selectivity. The importance of the geometric arrangement in tuning the product selectivity at copper-based bimetallic catalysts is emphasized by eCO2RR studies using alloys with different alloying configurations.64 CuPd alloys with ordered, disordered, and phase-separated geometric structures have been examined for CO2 reduction in 1 M KOH solution using a flow reactor. The phase-separated sample exhibited superior activity for C2 product selectivity. A total FE for C2 products of 63% was achieved at −0.8 V vs. RHE, in which the contribution from ethylene was ∼50%, as shown in Fig. 15(d). On the other hand, the selectivity for ethylene at disordered alloy sample is∼4 times lower than the phase separated sample. In contrast, the ordered CuPd alloy produced only C1 products, mainly CO.
Fig. 15 (a) Faradaic efficiencies for ethylene formation obtained during electrolysis at AgCu and Cu catalysts under different experimental conditions, (b and c) in situ Raman spectra showing (b) the Cu–CO stretch, and (c) the Cu–O stretch regions obtained from the eCO2RR using electrolysis AgCu and Cu catalysts at −0.7 V in 0.1 M KOH solution. Reprinted with permission.204 Copyright 2018, American Chemical Society. (d) Faradaic efficiencies of products obtained with three CuPd catalysts. Reprinted with permission.64 Copyright 2017, the American Chemical Society. (e and f) eCO2RR product distribution obtained using Cu catalysts (e) with and (f) without PdCl2 in the electrolyte, (g and h) XPS of (g) Pd 3d and (h) Cl 2p regions of PdCl2 and a Cu catalyst with PdCl2 after reduction, (i) schematic illustration of ethane formation from surface adsorbed ethylene and hydrogen using a Cu catalyst in the presence of PdCl2. Reprinted with permission.205 Copyright 2015, the American Chemical Society. |
The significant role of PdCl2 electrolyte in modifying the product selectivity achieved with Cu2O derived Cu catalysts was identified by Yeo and co-workers.205 With a pure Cu2O derived Cu electrode, a 32% yield of ethylene was achieved −1.0 V vs. RHE in an aqueous bicarbonate solution. With addition of PdCl2 into the electrolyte, a complete shift in the major product was observed with selectivity changing from C2H4 to C2H6. Now the highest faradaic efficiency was 30% for C2H6 at the same potential (Fig. 15(e and f)). Analysis of XPS data after electrolysis (Fig. 15(g and h)) revealed that PdCl2 provides the source of a sacrificial dopant by forming Pd0 on Cu2O during the electrolysis reaction. According to the authors, the Pd0 sites adsorb hydrogen effectively and provide a hydrogen source that facilitates reduction of C2H4 to C2H6, as shown in Fig. 15(i).
Tables 5 and 6 summarises the state-of-the-art Cu alloys and Cu based catalysts, which are selective for ethylene in H-type and flow cells, respectively.
Catalyst | Electrolysis cell | Electrolyte | Potential (V vs. RHE)a | FE (%) | j C2H4 (mA cm−2) | Ref. |
---|---|---|---|---|---|---|
a Potential at which maximum FE was obtained. b Custom made two compartment cell. c V vs. Ag/AgCl. | ||||||
OD-Cu + PdCl2 | H-cell | 0.1 M KHCO3 | −1.0 | 32 | ∼10 | 205 |
ZrO2/Cu–Cu2O | H-cell | 0.1 M KCl | −1.3 | 62.5 | 15 | 206 |
4H crystal phased Au/Cu nanoribbon | H-cell | 0.1 M KHCO3 | −1.1 | 44.9 | 14.4 | 207 |
Ag/Nafion-Cu2O | Single compartment cell | 0.1 M NaHCO3 | −1.9 | 80 | 16 | 208 |
Ag–Cu nanodimers | Custom made two compartment cell | 0.1 M KHCO3 | −1.1 | ∼40 | 10 | 209 |
Anodized Cu | PEEKb | 0.1 M KHCO3 | −1.08 | 38.1 | 7.3 | 210 |
Plasma activated Cu | Custom made two compartment cell | 0.1 M KHCO3 | −0.9 | 60 | >8 | 211 |
Cu2O derived Cu | Teflon cellb | 0.1 M KHCO3 | −1.0 | 42.6 | 13.3 | 212 |
Nano-defective Cu nanosheets | H-Cell | 0.1 M K2SO4 | −1.18 | 83.2 | ∼50 | 213 |
Cu2O film | Teflon cellb | 0.1 M KHCO3 | −0.99 | 37.5 | 12.9 | 214 |
Cu mesocrystal | PTFE cellb | 0.1 M KHCO3 | −0.99 | 27.2 | 7 | 215 |
B-doped Cu | H-cell | 0.1 M KCl | −1.1 | 52 | 36.4 | 216 |
Plasma treated Cu | H-cell | 0.1 M KHCO3 | −1.0 | 45 | 15.3 | 217 |
t-Cu2O NPs/C | H-cell | 0.1 M KHCO3 | −1.1 | 59 | 24 | 218 |
Cu2O NPs/C | H-cell | 0.1 M KHCO3 | −1.1 | 57.3 | 12 | 219 |
Cu3N nanocubes | H-cell | 0.1 M KHCO3 | −1.6 | 60 | 18 | 220 |
Pulsed-Cu | H-cell | 0.1 M KHCO3 | −1.0 | 48.6 | 20 | 221 |
CuOx | H-cell | 0.1 M KHCO3 | −1.3 | 53 | 14 | 222 |
UiO-66-derived amorphous ZrOx/Cu | H-cell | 0.1 M KHCO3 | −1.05 | 43.3 | 15.7 | 223 |
CuBr derived Cu nanodendrites | H-cell | 0.1 M KHCO3 | −0.85 | ∼40 | ∼9 | 224 |
Cu on Cu3N | H-cell | 0.1 M KHCO3 | −1.05 | 43 | ∼16.5 | 225 |
Reconstructed Cu–I | H-cell | 0.1 M KHCO3 | −1.09 | 59.9 | ∼15.7 | 226 |
Cu1.8Se nanowires | H-cell | 0.1 M KHCO3 | −1.1 | 55 | 8.3 | 227 |
A-Cu NWs | H-cell | 0.1 M KHCO3 | −1.0 | 69.79 | ∼18 | 228 |
Cu/TiNT | H-cell | 0.5 M KCl | −1.6c | 55 | 132 | 229 |
Catalyst | Electrolysis cell configuration | Electrolyte | Potential (V vs. RHE)a | FE (%) | j C2H4 (mA cm−2) | Ref. |
---|---|---|---|---|---|---|
a Potential at which maximum FE was obtained. b Cell voltage. | ||||||
Cu/ZnO | Plug flow reactor | 1 M KOH | −0.9 | ∼50 | ∼300 | 203 |
CuAg nanowires | Flow cell | 1 M KOH | −0.7 | 60 | ∼180 | 204 |
Phase separated CuPd NPs | Flow cell | 1 M KOH | −0.8 | ∼50 | ∼180 | 64 |
Ce doped Cu NPs | Flow cell | 1 M KOH | −0.7 | 53 | ∼80 | 230 |
Atomic Ni decorated Cu | Flow cell | 1 M KOH | −0.88 | 31.8 | 85 | 231 |
Cu (100) nanocubes | Flow cell | 1 M KOH | −0.7 | 60 | 120 | 191 |
Cu-DAT nanowires | Flow cell | 1 M KOH | −0.6 | 38.2 | 90 | 232 |
Cu NPs | Flow cell | 1 M KOH | −0.58 | 35 | 150 | 233 |
Graphite/carbon NPs/Cu/PTFE | Flow cell | 7 M KOH | −0.55 | 70 | ∼70 | 234 |
DVL-Cu | Flow cell | 1 M KCl | −0.81 | 84.4 | 92.5 | 235 |
Fluorinated-Cu | Flow cell | 0.75 M KOH | −0.89 | 65 | 1040 | 236 |
CuS/Cu–V | Flow cell | 1 M KOH | −0.92 | 21.1 | 84 | 237 |
Dendritic Cu | Flow cell | 0.1 M KHCO3 | −1.2 | 36 | 162 | 238 |
Nanoporous Cu | Flow cell | 1 M KOH | −0.67 | 38.6 | 252 | 239 |
PTFE-Cu NPS | MEA | 0.1 M KHCO3 | 3.8bb | 56.7 | 85 | 240 |
Molecular tuned Cu | MEA | 1 M KHCO3 | 3.65b | 72 | 230 | 241 |
Graphite/carbonNP/Cu/PTFE | MEA | 1 M KHCO3 | 4.2b | 46 | 92 | 242 |
Fig. 16 (a) Mechanism proposed for ethanol formation using CuZn catalysts. Reprinted with permission.243 Copyright 2016, the American Chemical Society. (b) Operando Raman spectra obtained during electrolysis of saturated CO2 in 0.1 M KHCO3 solution using Cu and CuZn catalysts. Reprinted with permission.244 Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. |
The Grätzel group244 reported that a CuO/ZnO core/shell structure derived CuZn bimetallic alloy catalyst for eCO2RR achieved 41% and 48% of C2+ product selectivity in H-cell and flow cell configurations respectively, with ethanol being the primary C2 product. To achieve this outcome, CuO nanowires were coated with ZnO (∼30 nm thick) by atomic layer deposition. Subsequently, electrochemical reduction was undertaken to form a CuZn bimetallic layer (∼90 nm) over the Cu nanorods. XRD analysis after electroreduction showed the presence of Cu and Zn/ZnO but not CuZn alloy. This CuZn catalyst was used to catalyze eCO2RR in 0.1 M KHCO3 using an H-cell. Ethanol was produced at −1.15 V vs. RHE with a FE of 32% and a partial current density of 10.5 mA cm−2. The mechanism proposed in this study on the basis of analysis of Raman spectra (Fig. 16(b)) is similar to that suggested by Ren et al.243 for the formation of ethanol using a phase segregated CuxZn catalyst.
An eCO2RR study by Lee et al.63 also reveals the importance of Cu–Ag biphasic boundaries in achieving high ethanol selectivity. The structure of this catalyst allows the CO concentration near the Cu–Ag phase blended catalyst surface to be increased, facilitating the insertion of CO into the Cu bound intermediates (*CH2) to form *COCH2. Upon further transfer of protons and electrons, *COCH2 is hydrogenated, resulting in the formation of ethanol via the acetaldehyde route. A recent eCO2RR study using CuOx coated Ag nanowires revealed an interesting relationship between the thickness of the CuOx layer and ethanol selectivity.245 While the Ag/CuOx-y (core/shell-y, where y represents the average thickness of the shell in nm) samples with lower CuOx thickness (Ag/CuOx-10) showed substantial ethanol formation (Fig. 17(c)), ones with thicker CuOx favoured ethylene. In situ XAS studies (Fig. 17(d)) revealed that Ag/CuOx-10 initially contained Cu2O which was stable in the potential range of −0.5 to −0.7 V vs. RHE (Cu–O(Cu2O) and Cu–Cu(Cu2O) bands preserved) but reduced to Cu0 at potentials more negative than −0.7 V resulting in formation of a Cu–Cu(Cu) bond (Cu–Cu band of metallic Cu). However, CuAg alloy was not formed during this reduction process. With Ag/CuOx-32, a less negative potential is sufficient to trigger this structural change. At a potential of −0.6 V vs. RHE, reduction of Cu2O to metallic Cu0 was observed. Further, in both cases this led to a stable atomic arrangement in the potential range of −0.7 to −1.3 V vs. RHE resulting in enhanced C2 product selectivity. The authors found that the existence of a majority of Cu in the +1 state facilitated the CO insertion mechanism and favoured ethanol production at Ag/CuOx-10 while the presence of more Cu0 in Ag/CuOx-32 favoured C–C coupling mechanism which led to the formation of ethylene. This result signifies that the transformation of Cu(I) to Cu(0) is controlled by the CuOx thickness, which in turn influences the eCO2RR product selectivity.
Fig. 17 (a and b) SEM images of Ag/CuOx-y (core/shell-y, where y represents the average thickness of the shell in nm) catalysts, (c) ethanol faradaic efficiencies and partial current densities obtained with Ag/CuOx-y catalysts, (d) in situ XANES and EXAFS data obtained with Ag/CuOx-10 and Ag/CuOx-32 catalysts. Reprinted with permission.245 Copyright 2019, the American Chemical Society. Source: https://pubs.acs.org/doi/10.1021/acscentsci.9b01142. Further permissions related to the material excerpted should be directed to the ACS. |
Tables 7 and 8 summarie the state-of-the-art Cu alloys and other Cu based catalysts, which are selective for ethanol in H type and flow/membrane electrode assembly (MEA) cells, respectively.
Catalyst | Electrolysis cell | Electrolyte | Potential (V vs. RHE)a | FE (%) | j C2H5OH (mA cm−2) | Ref. |
---|---|---|---|---|---|---|
N.R – not reported.a Potential at which maximum FE was obtained.b Custom made two compartment cell.c Potential (V) vs. SCE. | ||||||
Cu4Zn | H-cell | 0.1 M KHCO3 | −1.05 | 29 | 8.2 | 243 |
CuO/ZnO | PEEKb | 0.1 M KHCO3 | −1.15 | 32 | 10 | 244 |
Ag/CuOx-10 | H-cell | 0.1 M KHCO3 | −1.2 | 29 | 7.6 | 245 |
Cu3Sn | H-cell | 0.1 M KHCO3 | −1.0 | 64 | 5.7 | 246 |
Ag20Cu1 | H-cell | 0.1 M KHCO3 | −1.1 | 16.5 | 4.1 | 247 |
Au1Cu3 | H-cell | 0.5 M KHCO3 | −1.0 | 29 | 5.6 | 248 |
Cu5Zn8 | H-cell | 0.1 M KHCO3 | −0.8 | 46.6 | 2.3 | 249 |
Dendritic Cu–Cu2O | H-cell | 0.1 M KCl | −0.4 | 26 | 2.99 | 250 |
CuO NPs | H-cell | 0.2 M KI | −1.7c | 36.1 | N.R | 251 |
Cu/carbon nanospike | H-cell | 0.1 M KOH | −1.2 | 63 | ∼3 | 252 |
Cu–I | H-cell | 0.1 M KHCO3 | −0.9 | 25 | 11 | 253 |
Cu-GNC-VL | H-cell | 0.5 M KHCO3 | −0.9 | 70.5 | 9 | 254 |
Cu2/N0.14C | H-cell | 0.1 M KHCO3 | −1.1 | 51 | 14.4 | 255 |
UiO-66-derived amorphous ZrOx/Cu | H-cell | 0.1 M KHCO3 | −1.1 | 22.4 | 12.7 | 223 |
Cu on Cu3N | H-cell | 0.1 M KHCO3 | −0.95 | 18.4 | ∼8 | 225 |
Cu1.8Se nanowires | H-cell | 0.1 M KHCO3 | −1.1 | 24 | 3.5 | 227 |
Catalyst | Electrolysis cell | Electrolyte | Potential (V vs. RHE)a | FE (%) | j C2H5OH (mA cm−2) | Ref. |
---|---|---|---|---|---|---|
N.R – not reported.a Potential at which maximum FE was obtained.b Cell voltage. | ||||||
CuO/ZnO | Flow cell | 1 M KOH | N.R | 41 | 82 | 244 |
Phase separated CuPd NPs | Flow cell | 1 M KOH | −0.8 | ∼50 | ∼180 | 64 |
Cu9Zn1 | Flow cell | 1 M KOH | −1.0 | 26.2 | ∼60 | 256 |
CuAg nanowires | Flow cell | 1 M KOH | −0.7 | 25 | 75 | 204 |
Binding-site diverse Ag/Cu | Flow cell | 1 M KOH | −0.67 | 41 | 102 | 231 |
Ag decorated Cu/Cu2O | Flow cell | 1 M KOH | N.R | 19.2 | 304.5 | 257 |
Atomic Ni decorated Cu | Flow cell | 1 M KOH | −0.88 | 22.9 | 61.4 | 258 |
N–C/Cu | Flow cell | 1 M KOH | −0.68 | 52 | 156 | 259 |
Cu2O spheres | Flow cell | 2 M KOH | −0.61 | 26.9 | 71.8 | 250 |
Cu-DAT nanowires | Flow cell | 1 M KOH | −0.69 | 27.3 | ∼35 | 232 |
Fluorinated-Cu | Flow cell | 2.5 M KOH | −0.54 | 16 | 128 | 236 |
CuS/Cu–V | Flow cell | 1 M KOH | −0.92 | 24.7 | 99 | 237 |
Nanoporous Cu | Flow cell | 1 M KOH | −0.67 | 16.6 | ∼108 | 239 |
Cu-DS | Flow cell | 1 M KOH | −0.95 | 52 | 52 | 260 |
Cu2O NPs | Flow cell | 2 M KOH | −0.6 | ∼27 | 93 | 198 |
FeTPP[Cl]/Cu | Flow cell | 1 M KHCO3 | −0.82 | 41 | 124 | 261 |
FeTPP[Cl]/Cu | MEA | 0.1 M KHCO3 | 3.7b | 45 | 100 | 261 |
Cu3Sn | MEA | 1 M KOH | 3b | 40 | 361 | 246 |
PTFE-Cu NPS | MEA | 0.1 M KHCO3 | 4.2b | 17.1 | ∼40 | 240 |
Graphite/carbonNP/Cu/PTFE | MEA | 0.1 M KHCO3 | 4.2b | ∼15 | ∼30 | 242 |
Cu-DS | MEA | 0.1 M KHCO3 | 3.5b | 50 | 95 | 260 |
Acetate formation with reasonable selectivity (faradaic efficiency of 21% at −1.3 V vs. RHE), at 0 °C was achieved by the Meyer group by employing ultra-small (Cu)m,(Ag)n (m, n denote the atomic ratios of Cu and Ag) bimetallic nanoparticles immobilised on a polymer. In 0.5 M KHCO3 electrolyte containing 8 ppm benzotriazole at 0 °C, significant enhancement of acetate with a FE of 21.2% at −1.3 V vs. RHE was achieved on (Cu)m, (Ag)n compared to pure Cu where less than 1% of acetate was produced. The CO spillover mechanism from Ag to Cu was again proposed. As for ethanol production described above, CO inserted onto the Cu surface and then coupled with *CH2 to give the *COCH2 intermediate, but which in this case was further reduced to acetate via the acetaldehyde route. Since the availability of CO at copper is the limiting factor for C–C coupling, Lum et al.262 have introduced sequential Cu–Au and Cu–Ag catalysts. These were microfabricated electrodes where Cu and Au or Ag were organized in a sequential order with predefined thickness of each metal and the distance between the two metals as shown in Fig. 18(a–c). In these catalysts, Cu is in close proximity with Au or Ag and favours formation of CO during eCO2RR. The generated CO spills over to the Cu surface and facilitates production of more highly reduced carbon fuels. By this strategy, these authors have improved the formation of oxygenate products to 41.3% which surpasses the selectivity of 26.1% achieved for hydrocarbons. The microfabricated Cu–Au or Cu–Ag electrodes used in these studies consisted of integrated but independent Cu and Au/Ag electrodes arranged adjacent to each other (Fig. 18(a–c)). As a result, the CO generated during eCO2RR was accumulated near adjacent Cu electrodes. By varying the distance between the Cu and Ag probes, the oxygenate to ethylene ratio can be varied from 0.5 to 2.4, with the highest faradaic efficiency of oxygenates reaching 41.4%, with total C2/C2+ products being 65% (Fig. 18(d)). This study paves the way for new catalyst design strategies with controlled bimetallic arrangements for selective generation of C2/C2+ products other than ethylene. Morales-Guio et al.263 developed an Au/Cu tandem catalyst by depositing Au nanoparticles on Cu foil. This catalyst design provided more than a 100-fold increase in the rate of CO2 reduction to the higher reduced carbon products. The improvement in C2/C2+ alcohol production at low overpotentials was attributed to the enhanced local CO concentration provided by Au present in close proximity to copper.
Fig. 18 (a) Schematic diagram of a custom-made electrochemical device with microfabricated Au and Cu electrodes, (b) SEM and (c) EDX mapping of AuCu electrodes with 11% of geometric area occupied by Cu, (d) faradaic efficiencies and current densities obtained with AgCu electrodes. Reprinted with permission.262 Copyright 2018, the Royal Society of Chemistry. (e) Faradaic efficiencies of products obtained with a Mo8/Cu heterostructure catalyst (cyan: hydrogen, red: ethanol, green: acetate, yellow: methane, blue: ethylene, purple: ethane). Reprinted with permission.264 Copyright 2021, Elsevier B.V. |
Recently, the Wu group203 have reported that enhancement in oxygenate production also can be achieved with a Cu/ZnO tandem catalyst. Tandem catalysts with variable ZnO loading were prepared by air brushing ZnO onto a Cu electrode pre-synthesised on a GDL electrode and dried under vacuum. By controlling the ZnO loading, effective spatial management of CO transport was achieved to optimise CO utilisation. High selectivity for C2+ products with a faradaic efficiency reaching 80% is achieved as shown in Fig. 14(e) with a partial current density of ∼470 mA cm−2 obtained with Cu1.0Zn0.2, values which respectively are 1.2 and 3.4 times higher than obtained with pure Cu.
An excellent conversion yield for CO2 to acetate was reported recently at polyoxometalate modified Cu cubes. The Mo8/Cu heterostructure catalyst was rich in Cu–O–Mo interfaces, which led to the generation of acetate with a FE of 48.8% at −1.13 V (Fig. 18(e)) and a partial current density of 68.9 mA cm−2 in saturated NaHCO3.264 The selectivity for C2+ products was found to be dependent on the thickness of the shell in Cu/Pb core/shell nanocubes. Cu/Pb nanocubes with an optimal shell thickness of 0.7 nm gave 73.5% C2+ products with a partial current density of 294.4 mA cm−2 at −1.3 V in 1 M KOH solution under flow cell conditions (Fig. 19(a and b)).265 DFT calculations revealed that the synergistic effects due to the core/shell structure decreased the formation energies of *COOH and *OCCOH intermediates, thereby facilitating production of highly reduced carbon products as shown in Fig. 19(c and d). Azenha et al.266 reported the production of propane with a remarkable FE of 84.6% by employing Bi coated CuO nanowires on a filter press cell in 1 M KOH. Bidentate carbonate formation on the catalyst surface was an important requirement for the formation of propane. Detailed investigations revealed that the presence of Cu(I) sites and oxygen defects influenced the binding affinity of CO2 to the Bi/Cu NW catalyst in a manner which led to high selectivity for propane.
Fig. 19 (a and b) Faradaic efficiencies of C1 and C2 products obtained using a CuPb-0.7/C (where 0.7 represents the thickness of Pb shell) catalyst, (c) simulated data and (d) schematic illustration of the production of C2+ products via the eCO2RR route. Reprinted with permission.265 Copyright 2021, the American Chemical Society. |
Table 9 summarises the state-of-the-art Cu alloys and other Cu based catalysts, which are selective for other C2/C2+ products.
Catalyst | Product | Electrolysis cell | Electrolyte | Potential (V vs. RHE)a | FE (%) | j product (mA cm−2) | Ref. |
---|---|---|---|---|---|---|---|
N.R – not reported.a Potential at which maximum FE was obtained.b Custom made two compartment cell.c V vs. Ag/AgCl. | |||||||
(Cu)m, (Ag)n NPs | CH3COO− | H-cell | 0.5 M KHCO3 + 8 ppm benzotriazole | −1.33 | 21 | N.R | 267 |
Dendritic Cu–Cu2O | CH3COO− | H-cell | 0.1 M KCl | −0.4 | 40 | 4.6 | 250 |
4% Cu dots/Ag | Oxygenates | PEEKb | 0.1 M CsHCO3 | −1.0 | 41 | 4.4 | 262 |
Mo8/Cu | CH3COO− | Electrochemical batch cell | Saturated NaHCO3 | −1.13 | 49 | 57 | 264 |
CuBi | C3H8 | Filter press cell | 0.45 M KHCO3 + 0.5 M KCl. | N.R | 85 | 38 | 266 |
Single atom Cu/NPC | CH3COCH3 | H-cell | 0.1 M KHCO3 | −0.36 | 36.7 | N.R | 268 |
CuI derived Cu nanofibres | C2H6 | H-cell | 0.1 M KHCO3 | −0.73 | ∼30 | ∼9 | 224 |
CuSx-DSV | n-C3H7OH | H-cell | 0.1 M KHCO3 | −1.05 | 15.4 | ∼10 | 269 |
Graphene/ZnO/Cu2O | n-C3H7OH | H-cell | 0.5 M NaHCO3 | −0.9c | 30 | N.R | 270 |
MOF derived Cu | n-C3H7OH | Five port electrochemical cell | 0.1 M KHCO3 | −2.4c | ∼16 | ∼4 | 271 |
Commonly, the selectivity of the product is related to the binding energies of key reaction intermediates which can be tuned by varying the structure of the alloy. For example, in alloys with ordered atomic structures, the elemental composition influences the bond length between the elements. This, in turn, influences the binding energies of eCO2RR intermediates adsorbed onto the catalyst surface, providing a systematic route to tuning product selectivity. On the other hand, a disordered alloy system can lead to changes in electronic and geometric structures during electrolysis which also influences the products that are formed. In the case of a core/shell type alloy structure, it is the strain and ligand effects, atomic vacancies and doping that affect the binding strength of reaction intermediates, again providing systematic routes to selectively tuning the product formation pathway. Phase separated alloys enhance the product selectivity in a unique way. In this class of alloy there is a clear separation in the phases of the different elements in the alloy, allowing each phase to individually generate the product that is characteristic of the element in the particular phase. Migration of products or the coupling of intermediates at the boundary of different phases is possible. This helps in producing highly reduced carbon products such as ethylene or ethanol with enhanced selectivity generally facilitated by the CO spillover mechanism.
Strategic alloying can introduce specific synergistic effects that are useful to tune the selectivity towards each product. However, other inevitably coupled effects could diminish the product selectivity. Hence, despite the ability to manipulate the reaction pathways by altering the binding energies, complete selectivity for one product remains exceedingly difficult to achieve when an alloying strategy is used alone. Therefore, integration with other catalyst design strategies is highly recommended.
Alloy structures also tend to be unstable under electrolysis conditions. The lifetime for most of the catalysts used to date is typically within a day, which is far less than that required for commercial applications. To address the problem of stability, high entropy alloys are now receiving considerable attention. These are a new class of alloy formed with multiple elements. HEAs are highly stable under extreme conditions, such as high temperatures and prolonged electrolysis conditions due to an exceptionally high kinetic solid diffusion barrier. They are active for eCO2RR, and in principle the electronic properties and hence selectivity can be precisely tuned as they possess exceptional flexibility for adjusting the binding energy compared to primitive alloys. HEAs consist of a minimum of five different elements, resulting in a wide range of compositions and atomic arrangements. Hence, powerful machine learning tools272,273 are needed to identify the optimal composition and structure required to achieve selective generation of a desired eCO2RR product. Even though the suitable composition and structure of the HEAs are predicted theoretically, the currently used synthesis strategies for HEAs require high temperatures, pressures and an inert environment. Precisely controlling the microstructures or the local atomic arrangement also is highly complicated.274,275
The synergistic, so called “cocktail effect” displayed by HEAs with their multiple elemental combinations exhibit exceptional binding energies for reaction intermediates. The availability of different binding sites in these catalysts also can provide the possibility of wide range of binding energies for eCO2RR intermediates and thus provide the possibility to tune the selectivity of the products.274 This property can be advantageous for cascade reactions, where the product formed at one binding site is desorbed from the surface and re-bound at the other for further reaction. Hence the choice of elements and atomic arrangement in the catalyst should be given priority when designing these catalysts. Copper based high entropy alloys are a focus as they provide the possibility of producing highly reduced hydrocarbon fuels with improved selectivity by means of cascade reactions. Support from theoretical studies is critical to extricate the full potential of such alloys.66,72,276 Despite these highly attractive properties, application of HEAs in the eCO2RR is still limited.274,275 Nellaiappan et al.277 achieved the first experimental realization of the application of nanocrystalline HEAs for eCO2RR. Nanocrystalline AuAgPtPdCu produces 100% of gaseous products including CO, CH4 and C2H4 at a low potential of −0.3 V vs. RHE with high hydrocarbon (CH4 + C2H4) selectivity (∼70%) in aqueous 0.5 M K2SO4. First principle based DFT calculations attribute this outcome to the destabilisation of *OCH3 and highly enhanced stabilisation of *O intermediates. Pedersen et al.66 investigated the possibility of the formation of multicarbon products at HEAs by combining DFT with machine learning with (111) facets of CoCuGaNiZn and AgAuCuPdPt as model systems. The authors established a correlation between composition and the variation in H and CO binding energies.
In addition to achieving a more rational design of catalyst material, understanding the mechanism of eCO2RR at the atomic level should make it possible to establish more precise structure–property relationships. From this perspective, advanced in situ characterisation techniques are receiving increased attention.278–280 Several studies have shown that mixed oxidation states of Cu stabilise reaction intermediates that produce highly reduced C2 fuels.43,65,198,210,281 However, recent in situ studies through a combination of synchrotron based grazing incidence X-ray absorption and X-ray diffraction spectroscopy have suggested that Cu is present in the metallic 0 oxidation state during electrolysis.200 Such contradictory results may be due to structural differences in the catalytic materials. Clearly, morphological and structural changes in the catalyst during electrolysis need to be monitored. Development of more sensitive in situ spectroscopic and microscopic structural characterisation tools is needed.
Finally, the electrolysis system design must be optimised to enhance mass transport rates for both reactants and products. Recent research has demonstrated that use of gas diffusion electrodes and membrane electrode assemblies allow current densities to be increased drastically compared to those achievable with conventional H-shaped electrolysis cells.234,282,283 Improved cell designs and reaction conditions that allow much higher current densities, lower resistance and higher stability are essential to reach commercially viable efficiencies. So far, high purity CO2 gas has been commonly used in eCO2RR studies. To fully realize the potential of the eCO2RR for commercial application, the effect of impurities, such as SOx, NOx, which are commonly present in the industrial flue gases, should be investigated. However, research towards this direction is still limited.284
In summary, eCO2RR is a viable method for utilising CO2. However, even though eCO2RR can produce a variety of commercially valuable fuels, manufacturing of targeted products with high selectivity remains difficult. Using catalysts derived from alloying with Cu addresses some of the limitations since it allows considerable control over product selectivity by changing the guest element and the alloy structure. Despite many significant achievements emerging from the development of efficient Cu based alloy catalysts, fabrication of catalysts that have high stability and can generate significantly reduced carbon products in commercially viable quantities at competitive prices remains highly challenging.
This journal is © The Royal Society of Chemistry 2022 |