Ethanol electrooxidation reaction in alkaline media for direct ethanol fuel cells

Fuel cell technology dates back to 1839 when Grove demonstrated that chemical energy from hydrogen and oxygen can be converted directly into electrical energy with high efficiency.¹  Therefore, by principle, fuel cells represent a promising technology for clean generation of power from chemicals compared to combustion Carnot engines. Fuel cells convert chemical energy (fuel) into electrical energy with higher energy efficiencies, i.e., 45% in electrical energy, 90% in total energy (electricity and heat production) compared to combustion engines with total efficiency of up to 40%.²  In addition, fuel cells have low-to-none emission of pollutants.^2,3  There are several good review articles^4–12  and book chapters^13–15  reporting the status of direct ethanol fuel cells (DEFCs) at different periods since 2000. By 2002 DEFCs were conceived based on the significant progress done on proton exchange membrane fuel cells (PEMFCs). The motivation was that direct alcohol fuel cells (DAFCs) were to eliminate the need for the bulky and expensive reformers hence allowing the deployment of the technology for mobile applications.¹²  However, this shift to DAFCs encountered other challenges such as: (i) the complex and incomplete oxidation of alcohols leading to low fuel efficiencies, (ii) the alcohol crossover, in particular methanol, through the proton exchange membrane, etc. These challenges became the focus of active research since then. Most investigators focused on direct methanol fuel cells (DMFCs) and DEFCs in acidic media. It was recognised that DEFCs were more promising than DMFCs if the C–C bond would be cleaved.¹²  Cleaving the C–C bond was found to be favourable in alkaline conditions than in acidic media because the OH species readily accepts a proton which facilitates ethanol dehydrogenation. This finding sparked interest for direct alcohol alkaline fuel cells (DAAFCs). In 2005, Varcoe and Slade reported alkaline anion-exchange membranes followed by reports on Platinum-free low temperature fuel cells by Tsivadze et al. in 2007.^16,17 

DEFCs have several advantages compared to the most studied hydrogen and methanol fuel cells:^5,18  (i) ethanol is a non-toxic liquid, which lowers the investment of handling facilities because the current infrastructure for gasoline can be largely used, (ii) ethanol can be conveniently produced from biomass, hence is carbon neutral which mitigates the increasing atmospheric CO₂, (iii) ethanol is the smallest alcohol with the C–C bond, hence can serve as a model for the electro-oxidation of bigger compounds containing the C–C bonds, (iv) ethanol has a higher energy density than methanol if completely oxidized to CO₂ since it can deliver 12 electrons per molecule following the anodic reaction in eqn (1):⁵ 

In DEFCs, eqn (1) is counterbalanced at the cathode by the oxygen reduction reaction, generating a theoretical cell voltage of 1.14 V. However, in practice ethanol is known to be partially oxidized to acetic acid (acetate in alkaline media) giving a maximum of 4 electrons as shown in eqn (2):

The standard potential of acetic acid formation indicates that the reaction path leads to inevitable potential losses of about 0.4 V (the difference between ideal potential for CO₂ and acetic acid “production”).¹⁹ 

The progress in the development of DEFCs is well documented since the concept was conceived in earlier 2000. At that time, McLean et al. reviewed the state of the art of alkaline fuel cell (AFC) technology based on publications covering the past twenty five years up to 2002.²⁰  Although popular in the 1970s and 1980s, the AFC had fallen out of favour with the technical community in the light of the rapid development of proton exchange membrane fuel cells (PEMFCs). AFCs had been shown to provide high power densities and achieve long lifetimes in certain applications, and appeared to compete favourably with ambient air PEMFCs. In the review, McLean et al. examined the overall technology of AFCs, i.e., the power density, the lifetime performance, and the potential solutions were discussed. They presented a rough cost comparison between ambient air AFCs and PEMFCs. Overall, they showed that AFCs had potential to succeed in certain market niche applications, but lacked research and development support to refine the technology into successful market offerings.

The mechanistic understanding of ethanol electrooxidation reaction was reviewed in 2008 by Koper et al.¹⁵  They highlighted that the synergy between single-crystal in-situ studies and DFT calculations were beginning to unravel the kinetic and thermodynamic factors, the reaction pathways, and the structure sensitivity issues in electrocalysis with a special focus in their own work. Shortly after, Lamy et al. reviewed the working principles for DEFCs with a particular focus on solid alkaline membrane fuel cell.¹⁴ 

By 2010 there was realization in the community that direct alkaline fuel cells had many advantages compared to acidic counterparts.¹¹  The success leading up to this major shift from acidic to alkaline was the development of alkaline anion-exchange membranes (AEMs).^16,21–23  The use of AEMs have several advantages over conventional AFCs: (i) the enhanced electro-kinetics of the reaction, (ii) the potential to use non-noble metal catalysts, (iii) the use of higher energy density fuels such as ethanol, ethylene glycol, and glycerol, (iv) no carbonate precipitation since there was no mobile cation (Na⁺ or K⁺) which mitigates the issue of progressive carbonation of the alkaline electrolyte, (v) no electrolyte weeping (flooding), (vi) the reduced alcohol crossover, (vii) the simplified water management due to the fact that the water is produced at the anode and consumed at the cathode, and (viii) the reduced corrosion when working in alkaline media compared to acidic media.^11,24  Varcoe et al. have pointed out the importance and breakthrough of designing membrane electrode assembly (MEA) without metal cations (e.g. K⁺, Na⁺) present in alkaline fuel cells in which CO₂ is supplied to or generated at the electrodes to avoid undesirable carbonate precipitation, a major problem with traditional aqueous potassium hydroxide (KOH) electrolyte AFCs.²²  Antolini and Gonzalez reviewed the progress of the catalysts and membranes tested for alkaline direct alcohol fuel cells fuelled by methanol, ethanol, and ethylene glycol as of 2010.¹¹  The same year (2010) Zhao et al. presented a comprehensive review on the development of AEM DEFCs including the aspects of catalysts design, AEMs, and single-cell design and performance.²⁵ 

In 2012, Yu et al. reviewed developments in AFCs, considering different types of fuels, novel catalysts and AEM.⁹  They showed AFC systems and configurations particularly the new designs for portable devices. They pointed out that further development of DAFCs will rely on: (i) the improved AEMs with good ionic conductivity and stability, (ii) the low cost non-Pt catalysts with high activity, and (iii) the catalyst stability towards various fuels and oxidants. Rabis et al. presented a perspective summarizing the most outstanding contributions covering ten years (2002 to 2012) in terms of activity and durability of the catalyst materials for ethanol oxidation and oxygen reduction reaction, respectively.¹⁰  They provided an outlook towards the development of new catalyst support materials with improved performance/stability, the use of advanced characterization techniques, and the fundamental studies of reaction mechanisms and degradation processes as areas deserving attention from researchers.¹⁰ 

In 2013, Almeida and Andrade reviewed the trends in DEFCs with special attention to: (i) the systematic study toward the preparation of effective catalyst formulations by use combinatorial method, (ii) the oxidation of ethanol in amorphous alloys containing low amounts of Pt, and (iii) the use of non-noble materials as catalysts.¹³  Singh et al. reviewed the status of the efforts in developing low cost and efficient electrocatalysts (the preparation and structural characterization catalysts) so as to decrease the over-potential for alcohol oxidation reaction and oxygen reduction reaction.²⁶  Brouzgou et al. reviewed the comparison in performance of PEM-DEFCs and AEM-DEFCs.⁸  They pointed out that Pt-containing or Pt-free PEM-DEFCs that use acid proton-exchange membranes (typically Nafion type) exhibited relatively low performance, while AEM-DEFCs exhibited better performance values. They noted that the best value ever reported (peak power density was 360 mW cm⁻² at 60 °C) had been obtained in a very promising alkaline-acid direct ethanol fuel cells (AA-DEFCs).

In 2014, Rao et al. reviewed the progress in ethanol electrooxidation reaction focusing on the thermodynamic process, the reaction mechanism, and the advantages and disadvantages of different electrocatalysts.²⁷  They discussed the factors affecting the reaction activity and selectivity such as supports, nanoparticle sizes, catalyst structure, and alloying of metals. Sharaf et al. although focusing on hydrogen fuel cells provided a very good and concise review of fuel cell ranging from the fundamentals, history developments, the competing technologies, to the system evaluation factors.⁶  They used the most current data from industry and academia to highlight the relations between fuel cell fundamentals and applications.

In 2015, Wang et al. presented an overview of the advances in the study of ethanol electrooxidation mechanism and the electrocatalytic materials with a focus on Pt- and Pd-based catalysts.²⁸  They discussed the mechanistic understanding of ethanol oxidation reaction (EOR) on Pt and Pd surfaces. They reported that consensuses from the mechanistic studies are that sufficient active surface sites to facilitate the cleavage of the C–C bond and the adsorption of water or water residue were critical for obtaining a higher activity. They showed how this understanding had been applied to achieve improved performance on various Pt- and Pd-based catalysts. This was achieved by optimization of electronic and bifunctional effects, as well as by tuning the surface composition and structure of the catalysts. Badwal et al. reviewed various types of DEFCs currently under development with emphasis on ethanol sources and production methods, the fuel cell construction materials and their operating regime, the performance and life time issues and market applications.²⁹  An et al. reviewed the comparison of acidic and alkaline DEFCs, i.e., their working principles, cell performance, system efficiency, reaction products, and the cost.⁵  Similarly, recently Akhairi and Kamarudin published an overview of the acidic and alkaline DEFCs.⁴  The review focused on the work done on platinum and palladium as strong competitors and highlighted the outstanding problems such as the incomplete oxidation of ethanol to carbon dioxide, the need to optimize the performance of DEFCs at standard conditions, the discovery of suitable catalysts for higher tolerance to surface poison, the stability of the catalysts, the promotion of better diffusivity between the membranes and the electrodes, and the need to control the selectivity of the reaction.

From the above quick survey, it is evident that alkaline DEFCs are superior to their acidic counterparts. Therefore, the detailed review and focus of this chapter is on alkaline DEFCs. The literature review starts with the various prototypes of DEFCs which have been proposed to highlight current state-of-the-art of DEFCs, followed by an overview of catalyst designs with a focus on monometallic, bimetallic catalysts and metal oxide supports. Then we review the current understanding of EOR mechanism with emphasis to the reaction in alkaline conditions. The review is ended with a summary of outstanding challenges for DEFCs and proposal(s) of the strategies to deal with them.

There have been five types of DEFC designs reported in the literature to date,^5,8,30  (i) the proton-exchange membrane fuel cells (PEMFCs), (ii) the Anion-exchange membrane fuel cells (AEM-FCs), (iii) alkaline-anode membrane acid-cathode fuel cells (AA-FCs), (iv) direct alkaline fuel cells (DAFCs), and (v) direct ethanol solid oxide fuel cells (DE-SOFCs). The first three are shown in Fig. 1 and the last two are shown in Fig. 2 and 3, respectively.

The reactions for PEM-based, AEM-based, and AA for DEFCs are shown in eqn (3–5), (6–8), and (9–11), respectively.

PEM-based DEFCs

AEM-based EOR

AA based EOR

where M=Na or K

Fig. 1a shows the working principle of the first conceived fuel cell directly fed with the alcohol at the anode. The design had the advantage of avoiding the use of the bulk fuel reformer used in PEMFCs.¹²  Typically an aqueous solution of ethanol is circulated through the anodic compartment, and oxygen (or air) is circulated in the cathodic compartment. Later, it was realized that there were several challenges for design such as:⁸  (i) the acidic electrolyte membranes (mostly Nafion-based) were expensive, (ii) the incomplete oxidation of ethanol to CO₂, instead acetaldehyde and acetic acid which liberates only 2 and 4 electrons, respectively greatly reduced the Faradaic efficiency of the fuel cell, (iii) the sluggish reaction kinetics for EOR in acid media, leading to a large activation loss, (iv) ethanol crossover from the anode to the cathode within the PEM which lead to a parasitic current generation, (v) the need to use a cathode catalyst tolerant to ethanol, (vi) the durability of the state-of-the-art catalysts employed such as PtRu/C and PtSn/C was in question. Therefore, there was need to develop other fuel cell designs to avoid the use of acidic electrolyte as presented below.

The prospect of using an anionic membrane or a solid polymer electrolyte (SPE) in an alkaline fuel cell was first looked at by Agel et al. in 2000.³¹  Their goal was to extend the concept of using cheaper SPE which were well developed in lithium-ion batteries into the fuel cell design. They characterized SPE for ionic conductivity, transport numbers, water content and assembled a prototype alkaline fuel cell to show the viability of the new design. They reported that the performance of the prototype fuel cell was greatly improved while using an interfacial solution between the electrodes and the membrane.³¹  Fig. 1b shows the working principle for AEM-DEFCs. This concept was rapidly explored in the community as evidenced by the articles and patents reviewed by Varcoe and Slade in 2005.¹⁶  The advantages of this configuration are presented in the introduction.

The challenges for AEM-based DEFCs are: (i) the incomplete oxidation of ethanol to CO₂ remain an issue, (ii) the activity and durability of the Pd-based catalyst (the top candidates in the literature) for the EOR in alkaline media needs to be further enhanced, (iii) enhancing the catalytic activity of non-Pt catalysts at the cathode to make them comparable to that of Pt is required. Currently, Ag-based cathode catalysts for the ORR in alkaline media are the leading candidates, (iv) a significant improvement is needed to upgrade the OH⁻ conductivity, chemical, mechanical, and thermal stability of the existing AEMs. The OH⁻ conductivity can be improved by increasing the amount of charged groups in the membrane; however, there is a trade-off with the mechanical properties. A loss of the mechanical properties by promoting excessive water uptake is the result of increasing the concentration of the charged groups. The thinness of the AEM is an important requirement related to mechanical stability. To keep good mechanical stability when immersed in water, an AEM as thin as ∼50 µm is necessary. AEM suffers also of a poor chemical stability in alkaline media, stemming from the hydroxide attack on the cationic group. The result of this degradation is an important loss in the number of anionic exchange groups, and a decrease of the ionic conductivity. (v) Improvements of the ionic conductivity and the thermal and chemical stability of the ionomers present within the catalyst layers are required.

Fig. 1c shows the working principle for alkaline-acid DEFCs. It consist of an alkaline anode, a membrane, and an acid cathode employing hydrogen peroxide as oxidant which boost the theoretical voltage from 1.14 V to 2.52 V.^30,32  Although, this design has been reported to deliver the highest power density (360 mW cm⁻²), it has two major issues; (i) the species crossover, and (ii) the hydrogen peroxide decomposition.³³ 

In direct alkaline fuel cell (DAFC) designs, an aqueous solution of ethanol and KOH/or NaOH is used in a flow type arrangement without the need for a membrane to separate the anode and the cathode as shown in Fig. 2.³⁴ 

The electrode reactions are similar to AEM fuel cells, eqn. (2.6)–(2.8). This design has been mostly studied by Verma and Basu.^34–37  The best performance for this system was obtained with electrolyte concentration of 3 M KOH and 2 M Ethanol. However, it was reported that ethanol oxidation in this configuration proceeded to only acetaldehyde which involves only two electron. Therefore, considerable effort is required to optimise this technology that is promising for stationary power supply.

Last but not least is the direct ethanol solid oxide fuel cells (DE-SOFC) configuration shown in Fig. 3.^38,39  This design marks the efforts towards using liquid fuels such as ethanol directly in the SOFCs.

The development and design of efficient catalysts for breaking the C–C bond during ethanol electrooxidation is a central question in electrocatalysis. Many factors are known to influence the catalyst activity and selectivity such as chemical composition, morphology, size and shape of the catalyst in addition to the reaction conditions.^40–46  Therefore, the precise control of these parameters is crucial for the rational design of efficient and stable electrocatalysts for DEFCs. Fig. 4, gives a summary of the common factors/parameters that control the catalytic performance of the catalyst.⁴³ 

The efforts in the rational catalyst design strategies include the search for optimal formulations, the catalyst supports, and in the methodologies for catalyst preparations to tune the sizes, the morphologies and the surface composition. In this section, we first give an overview of the various methodologies reported, followed by a review of mono-, and bi-metallic catalysts for EOR in alkaline media.

There has been excellent progress in the synthesis of electrocatalysts with different morphologies, mono- and multi-metallic nanoparticles with various compositions and well-controlled shapes.^41,47  It is now possible to rationally design catalysts at the atomic-level to enhance electrocatalytic performance, hence making it possible to correlate the nanoparticle structure with activity. Fig. 5 shows the facets for different nanoparticle structures accessible for Au and Pd bimetallics.⁴⁷ 

During the nanoparticle synthesis, various approaches are used to reduce the metal ions as summarised in Table 1 (see Appendix) such as:¹³  sodium borohydride reduction, polyol (alcohol–reduction), formic acid reduction, tannic acid, zinc, sodium silicate (Na₂SiO₃), gas phase reduction, thermal decomposition of polymeric precursors (Pechini method), and microwave-assisted heating method. A number of synthesis methods have been adopted such as: (i) Impregnation, in which the support is mixed with a suitable metal precursor solution before the reduction of the metal ions. The method is best suited for monometallic catalysts.^32,48,49  (ii) Sequential impregnation/Colloidal, in which a colloid precursor is first synthesised preferably in an organic solvent in the presence of a suitable surfactant before adding the supports.⁵⁰  The approach is better suited for the synthesis of polymetallic systems of uniform composition. (iii) Micro-emulsion, in which the first step is the formation of the nanoparticles via a water-in-oil micro-emulsion reaction followed by a reduction step.^51–53  A micro-emulsion is formed by vigorous stirring or sonication and is thermodynamically stable. Nano-sized particles can spontaneously form within the micron size water droplets as a thermodynamically stable microemulsion.²⁶  This method provides the ability to control the metallic composition and particle size with a narrow distribution.⁵³  (iv) Sol–Gel Derived, in which the catalysts are prepared by the hydrolysis of acetylacetonate of the metal salt precursors in the presence of tetra methyl ammonium hydroxide followed by solution evaporation to form xerogel then thermal treatment under controlled atmosphere.⁵⁴ 

Examination of the available literature for EOR reveals that there are four strategies commonly adapted to design catalysts: (i) the use of high surface area carbon supports and/or reducible metal oxide supports, (ii) the tuning of the catalyst structure and morphology which includes mesoporous, two- and three-dimensional structures with preferential facets, (iii) the addition of the second or third ad-atoms on the catalyst surface, and (iv) the tuning of the reaction conditions such as electrolyte pH, cations and anions.

Catalyst supports are reported to play a significant role towards morphology, dispersion, activity, and selectivity of the catalysts.^55–57  Carbon is the most widely used support in the fuel cell research because it has high electrical conductivity and excellent structural properties which are important for fuel cell application.⁵⁸  Carbon has been widely used in PEMFCs and alkaline fuel cells for fabrication of the bipolar plate, the gas-diffusion layer and as a support for the active metal in the catalyst layer. Antolini has reviewed the application of carbon supports for Pt-based catalysts in fuel cells.⁵⁹  The novel carbon materials presented showed improved electrocatalytic properties and stability during fuel cells operation. Carbon nanotubes and aerogels have been investigated for use as catalyst support leading to the fabrication of more stable and active catalysts by reducing the undesirable carbon corrosion and degradation.⁶⁰  Graphene or graphene oxide materials are attractive alternative supports for dispersion and stabilization of the catalyst nanoparticles. Graphene is one atom thick nano-carbon materials which has attracted considerable attention in various applications including electrocatalysis.⁶¹ 

There are several reports showing the application of metal oxide supports such as CeO₂,^62–73  SnO_2,^{62,64,74–82}  TiO₂,^{62,64,83–86}  Mn_xO_y,^70,71,87  WO_x,^64,88  MoO_x,^64,89  RuO₂,⁷⁸  ZrO₂,^64,90  CaSiO₂,⁹¹  MgO,^92,93  NiO/foam,^{70,71,94–96}  and CoO_x^70,71  as promising supports for EOR catalysts. These metal oxide supports have a significant effect on the catalytic activity of the catalysts because of the interaction phenomenon known as “strong metal-support interaction” which was recognized by Tauster et al.^97,98  and advanced by Sanches and Gazquez.⁹⁹ 

Platinum is one of the default catalyst metals for many reactions, as such it has been considered for EOR. Katayama et al. investigated the role of adsorbed OH⁻ species on Pt catalyst for EOR.¹⁰⁰  They did a comparative study between ionomer-coated Pt and highly oxophilic CeO₂ modified Pt electrode using in situ ATR-FTIR to monitor adsorption behaviour of adsorbed OH⁻. They observed a distinct change in adsorption behaviour of adsorbed OH⁻ in blank KOH solution, which was attributed to the activity enhancement for EOR. This activity increase was not observed under acidic conditions. Hence, the pH has a significant effect not only on the reaction kinetics but also on the equilibrium properties of both solution and surface species. During EOR in alkaline media, the OH⁻ species are consumed which alters the local pH at the electrode surface, decreasing the reaction kinetics. Figueiredo et al. have shown the evidence of the local changing pH for EOR on Pt electrodes in alkaline media.¹⁰¹  They used rotating ring-disc electrode experiments to monitor the local pH change during EOR. The current at the ring when polarized at the onset of hydrogen evolution (0.1 V vs RHE) served as a measure of the local pH in the vicinity of the electrode. Their results showed that the current at the ring at 0.1V became more negative during EOR, owing to a change in the equilibrium potential of the hydrogen evolution reaction caused by a change in the local pH. Lai et al. investigated EOR on Pt in electrolytes of varying pH and composition using electrochemical and surface-enhanced Raman spectroscopy (SERS) techniques.¹⁰²  The reaction activity increased significantly when the pH of the electrolyte was above 10 (as shown in Fig. 6).

According to their report, the reaction selectivity strongly depends on the nature of the electrolyte, but to a smaller extent on the electrolyte pH.¹⁰²  These findings opened up the door to exploration of various electrolyte compositions. Of great interest from Lai et al. investigations, was the observation that the cleavage of the C–C bond was only observed on Pt in the absence of strongly adsorbed anions, which was attributed to the competition for the active sites. A comparative study of EOR on Pt electrode in acidic and alkaline media using DEMS was conducted by Cremers et al.¹⁰³  They reported that in the acidic environment the initial oxidation of ethanol was via acetaldehyde formation, which proceeded rather easily and was found to be highly reversible. They pointed out that the challenge in implementing DEFCs was to consequently oxidize acetaldehyde, a step that was particularly difficult as it must not proceed via acetic acid which cannot be oxidized further to CO₂.¹⁰⁴  In alkaline media, they found EOR to proceed rather faster and lead to a complete oxidation to CO₂, a promising approach for DEFCs.

Electrochemically reduced Pt oxide films were reported to be 29 times more active for EOR than Pt films.¹⁰⁵  The superior activity was attributed to higher electrochemical active surface area and the existence of residual oxygen based on CV and XPS measurements. The concentration dependence for EOR on Pt in alkaline medium was studied using electrochemical and DEMS techniques by Bayer et al.^106,107  They showed that selectivity for EOR to CO₂ was favoured at lower concentrations and was only observed during CV and not during CA, indicating that formed intermediate(s) were playing a key role. They did a comparison to ethylene glycol, which showed significant CO₂ formation during CV and CA experiments with the tendency that low concentrations and low potentials yielded higher CO₂ current efficiencies. Although in acidic medium both alcohols exhibited a comparable electrochemical performance, in alkaline medium the current densities for ethylene glycol were substantially higher.¹⁰⁷ 

The single-crystal Pt structural effects on EOR have been studied by several authors.^108–110  Buso-Rogero et al. used single-crystal Pt electrodes to show the effect of different facets (111, 110, and 100) for EOR using electrochemical and IRRAS techniques.¹⁰⁸  Although, the Pt(111) electrode displayed the highest currents and also the highest onset potential in CV, the CA showed that the activity decreased in the order of 110>100>111. Surprisingly, their IRRAS data showed that the C–C bond cleavage was not favoured in alkaline media. Lai and Koper studied irreversible adsorption of ethanol on Pt single crystal in alkaline solution using SERS.¹⁰⁹  They reported that EOR was very sensitive to the electrode surface structure, i.e., a higher concentration of low-coordination sites increased the current, lowered the over-potential required and lowered the deactivation rate. They found that the terrace length affected the quantity and nature of the adsorbed species, i.e., on Pt(110) only adsorbed CO was observed whereas adsorbed CH_x was only found on Pt(111) terrace sites.¹⁰⁹  Tripkovic et al. studied EOR on Pt(111), Pt(755), and Pt(332) surfaces in NaOH solution with a special focus on the oxygen-containing species generated and adsorbed on the surface.¹¹⁰  They suggested the existence of reversible and irreversible adsorbed OH⁻ and PtO species in the potential region relevant for EOR. They suggested the role of these species in the reaction and proposed a dual path reaction mechanism as discussed in the mechanism section.¹¹⁰ 

There are studies which have looked at Pt nanoparticles for EOR in alkaline media.^111,112  Buso-Rogero et al. investigated EOR on Pt nanoparticles with different shapes and loadings using electrochemical and spectroscopic techniques.¹¹¹  The nanoparticles with a large amount of (100) ordered domains showed higher current densities compared to nanoparticles with higher (111) domains. They reported that acetate was the main product with negligible amounts of CO₂, regardless of the type of Pt nanoparticles used. Sun et al. investigated the nanoparticle size effect for Pt nanoparticles supported on sulfonated graphene (Pt/sG) for EOR in alkaline solution.¹¹²  They prepared five catalysts with various average particle sizes. They reported that 2.5 nm catalyst had the highest current density peak for EOR.¹¹² 

The effect of the support on Pt activity has been explored by many research groups. Xu and Shen did a comparative study for Pt/C and Pt–CeO₂/C.^72,73  They reported that the electrode with a weight ratio of Pt to CeO₂ of 1.3 to 2.1 and a Pt loading of 0.30 mg cm² had the highest activity. Xu et al. studied EOR on MgO promoted Pt/C catalysts in alkaline media.⁹³  The promoted electrocatalysts were superior to pure Pt and the electrode with a weight ratio of Pt to MgO of 4 : 1 showed the highest activity for EOR. Bai et al. compared EOR on Pt–ZrO₂/C with Pt/C(20 wt.% E-TEK) using CV, Tafel plot, and impedance spectroscopy in alkaline conditions.⁹⁰  They reported that molar ratio of Pt : ZrO₂ of 1 : 4 had the best catalytic activity for EOR. Recently, a comparative study of the effect of metal oxide support (support=TiO₂/C, ZrO₂/C, SnO₂/C, CeO₂/C, MoO₃, and WO₃) on Pt nanoparticles for EOR in alkaline media was reported.⁶⁴  Godoi et al. used in-situ XAS to show that Pt-support interaction produces changes in the Pt 5d band vacancy, which correlated to the EOR catalytic activity.⁶⁴  They observed the highest and lowest activities for Pt nanoparticles on TiO₂/C and CeO₂/C, respectively.⁶⁴  Using the IRRAS technique, they reported that acetate was the main product and traces of CO₂ with different amounts for each support. They showed good correlation between fuel cell performances with electrochemical data.

Palladium is the strongest competitor to platinum catalysts so far based on reports in literature. In particular, Pd-based catalysts show high activity for EOR in alkaline media, hence has been extensively studied in the last decade. The influence of halide ions on EOR on Pd was reported by Kumar and Buttry, who found that halide ions decreased the peak currents monotonically as a function of increasing halide concentration.¹¹³  The extent of poisoning, which also shifted the oxidation peak potential in more positive direction, was in the order of I⁻>Br⁻>Cl⁻. This study highlighted the importance of thoroughly cleaning the nanoparticles prepared from palladium halide salt precursors. The effect of concentration has been studied using 8 wt.% Pd, on Vulcan XC-72 in passive alkaline DEFC and it was reported that improved performance and stability was observed when the [hydroxyl]/[ethanol]=1.¹¹⁴  Carrera-Cerritos et al. investigated the performance and stability of Pd nanostructures (nanopolyhedral, nanobar and nano-rod particles) in DEFCs.¹¹⁵  They studied the effect of the operation parameters, i.e., temperature and ethanol concentration on the maximum power density (MPD) and open circuit voltage (OCV). They reported that OCV values increased with increasing temperature for all of the catalysts at low ethanol concentration. Although, the MPD increased with temperature for all of the catalyst independent of the ethanol concentration, the effect of the temperature on the MPD for each Pd structure results in different slopes due to the different crystal faces.¹¹⁵ 

A study on single crystal Pd has been reported by Wang et al. who demonstrated the effectiveness of an electrochemical treatment consisting in cycles of constant potential oxidation and reduction of polycrystalline Pd surface in the enhancement of EOR.¹¹⁶  The rise of the activity after the treatment was ascribed to the increase of both the surface area and density of low coordination surface atoms. With the aid of IRRAS, they showed that a change in the reaction products distribution also occurred, resulting in some cases, in an increased tendency to cleave the C–C bond.

Most of the studies have focused on the use of Pd nanoparticles. Assaud et al., reported the use of three dimensional Pd clusters grown on TiO₂ nanotubes by atomic layer deposition for EOR.¹¹⁷  They found that there existed not only a direct correlation between the catalytic activity and the particle size but also a steep increase of the response due to the enhancement of the metal-support interaction when the crystal structure of the TiO₂ nanotubes was modified by annealing at 450 °C in air. Rohwer et al. have reported a comparison of microwave and non-microwave treated Pd nanoparticles for EOR in alkaline medium.¹¹⁸  Microwaved Pd nanocatalyst showed higher electrochemical active surface area, aggregation/uniformity dispersion, higher amounts of palladium oxides, and had remarkable activity for EOR. The morphological effect of Pd catalyst for EOR was investigated by Cerritos et al., who studied three different structures; nanoparticles (NP/C), nano-bars (NB/C) and nano-rods (NR/C) with preferentially exposed crystal faces supported on carbon black.¹¹⁹  They reported considerable differences with the performance trend of peak oxidation potential of: NB/C<NP/C<NR/C<commercial Pd/C, indicating that NB/C catalyst enclosed by Pd(100) facets was the best catalysts. Cherevko et al. used high surface area Pd foams with roughness factors of more than 1000 and a specific surface area of 60 m² g⁻¹ obtained by electrodeposition and reported them to have high activity towards the EOR.¹²⁰ 

The effect of the catalyst support for Pd has been shown is several studies. For instance, Monyoncho et al. reported the promotional role of metal oxide supports (CeO₂, SnO₂, TiO₂) for EOR on Pd in alkaline media.⁶²  They monitored in-situ electrooxidation products using the PM-IRRAS which revealed that the supports influence the selectivity the reaction. They reported superior selectivity towards breaking the C–C bond to produce CO₂ on Pd/CeO₂. Acetate was the major product evident on all the catalysts, but at different ratios.⁶²  Safavi et al. have proposed the use of immobilized Pd nanoparticles in a well-structured composite of hydroxyapatite and carbon nanotubes EOR.¹²¹  They demonstrated that the use of hydroxyapatite-carbon nanotubes composites lead to remarkable enhancements in the electrocatalytic activity, the kinetic parameters, and the durability of the catalyst. They attributed catalytic improvements to the synergetic effects between the immobilized Pd nanoparticles and the functionalities on the carbon nanotube-hydroxyapatite.¹²¹  To avoid the use of metal oxides, Zhian et al. have conducted a study of EOR on bis(dibenzylidene acetone)palladium(0), Pd(DBA)₂, complex shown in Fig. 7.¹²²  They reported that Pd(DBA)₂ had higher tolerance against poisoning intermediate/products for EOR, which was successfully employed as an anode catalyst in a passive air breathing DEFCs achieving a maximum power densities of 30, 31, 25 and 18 mW cm² for ethylene glycol, ethanol, glycerol and methanol, respectively.¹²² 

Chen et al. conducted a comparative study of DEFCs build with a 2 μm thick layer of TiO₂ nanotube arrays doped with Pd nanoparticles (1.5 mg Pd cm⁻²) and reported a maximum power densities were 210, 170, and 160 mW cm⁻² at 80 °C for fuel cells fed with 10 wt.% aqueous solutions of ethanol, ethylene glycol, and glycerol, respectively in 2 M aqueous KOH.¹²³  The use of highly porous 3D-Graphene nanosheets synthesized using the sacrificial support method as a support for Pd nanoparticles was presented by Serov et al.¹²⁴  Their approach allowed the preparation of nanoparticles with smaller particle size distribution, higher surface area and showed good electrochemical activity and durability for EOR. Silva et al. described the use of Pd nanoparticles supported on physical mixtures of C+TiO₂ for EOR in alkaline media.⁸³  They prepared C/TiO₂ mass ratios of 100 : 0, 80 : 20, 60 : 40, 40 : 60, 20 : 80, and 0 : 100. They reported that Pd/C+TiO₂ (40 : 60) as the most promising mixture ratio. Chen et al. studied EOR on Pd/C promoted with CaSiO₃ in alkaline medium and demonstrate that the Pd/CaSiO₃ and C in wt.% 50 : 50 had higher current density (1408 mA mg⁻¹) than that of the Pd/C catalyst (743 mA mg⁻¹).⁹¹  Li et al. investigated the effect of adding MgO to Pd/C catalyst for EOR in alkaline medium and reported a significant improvement in activity and in the poisoning resistance.⁹²  They reported that a catalyst with a weight ratio of Pd to MgO of 2 : 1 had the best performance, the onset potential was negative shifted by 80 mV and the peak current density was 3.4 times higher than Pd/C.⁹²  Li et al. reported EOR on Pd nanoparticles supported on multi-wall carbon nanotubes synthesized on a carbon fiber paper (MWCNTs/CFP) in alkaline media which gave higher activity and stable performance than the commercial Pd/C and Pd/CFP.¹²⁵  The promotion role of oxide phases on Pd for the EOR was studied by Martinez et al. who presented the evidence for the difference between an intrinsic effect obtained from an alloyed system and a synergistic effect produced by the presence of an oxide phase (SnO₂).⁷⁴  They interestingly showed that at 1M KOH, SnO₂ acted as a co-catalyst to provide OH⁻ ions to the interface layer which increased the turnover rate. However, acetate was reported to be the main product instead of the desired complete electrooxidation of ethanol to CO₂.⁷⁴  Pd nanoparticle-deposited MoO_x/C catalyst (Pd–MO/C) were considered by Lim et al. who showed a 35% higher mass activity compared to Pd/C catalyst for EOR.⁸⁹  They attributed the performance of Pd–MO/C to the high active surface area and the higher resistance to adsorbed CO. Uhm et al. synthesized well-ordered arrays of free-standing Pd–CeO₂ nano-bundles in an anodic alumina template via occlusion electrodeposition which showed dramatically enhanced activity for EOR compared to pure Pd.⁶⁸  Chu et al. reported EOR activity enhancement on palladium-indium oxide supported on carbon nanotubes (Pd–In₂O₃/CNTs) composites prepared via chemical reduction and hydrothermal reaction process.¹²⁶  The composite electrode with the mass ratio of Pd : In₂O₃ equals to 10 : 3 (with Palladium loading of 0.20 mg cm⁻²) showed the highest electrocatalytic activity for EOR.¹²⁶ 

Comparative studies between Pt and Pd for EOR in alkaline media have been conducted to determine the best candidate.^{70,79,95,127–130}  Xu et al. showed that Pd/C has a higher catalytic activity and better steady-state performance for EOR than Pt/C in alkaline media and the addition of oxides (CeO₂, NiO) significantly promoted the activity.⁹⁵  They found better performance for Pd or Pt supported on CeO₂ and NiO with weight ratio of 2 : 1 and 6 : 1, respectively. In another study, they reported EOR Pt and Pd electrocatalysts supported on carbon microspheres (CMS).¹³⁰  The results showed that nanoparticles supported on carbon microspheres gave better performance than those supported on carbon black and pointed out that although Pd was not a good catalyst for methanol oxidation; it was excellent catalyst for EOR in alkaline media than Pt. Hu et al. prepared Pt/C and Pd/C electrocatalysts supported on NiO by intermittent microwave heating (IMH) method and tested them for EOR in electrolyte with and without the presence of CO.⁹⁶  They reported that EOR on Pd–NiO/C electrocatalyst was better than Pt–NiO/C electrocatalyst. Bayer et al. studied EOR on Pt and Pd in alkaline medium using DEMS.¹²⁹  They reported that the reaction products and their current efficiencies depended strongly on the metal used. Acetate was the major reaction product while the current efficiency for CO₂ was low for both. However, the amount of acetate was higher for Pd electrode. At higher ethanol concentrations, ethyl acetate was formed on the Pt electrode but was absent on the Pd electrode. Cantane and Lima have studied EOR on electrodeposited layers of Pd and Pt in alkaline electrolyte and monitored reaction products by online DEMS.¹³¹  The DEMS evidenced similar amounts of CO₂ for Pd and Pt but Pd presented a higher production of ethyl acetate (acetic acid) and EOR on the Pd surface occurred to a greater extent, in agreement with previous reports. They proposed a mechanism as presented in the study.

It is well known that bimetallic, trimetallic, and quaternary catalysts are better than monometallic catalysts.^42,132–142  We are going to focus our attention on the bimetallic work to understand the synergetic effect of the metals involved to determine the rationally for catalyst optimization for EOR in alkaline media. Table 1 (appendix) gives a summary of bimetallic catalysts and their corresponding monometallic catalysts tested for alkaline DEFCs.

There are many reports that Pt–M (M=Pd,^143–150  Sn,^151–158  Ru,^{60,155,158–164}  Mo,^159,165  W,¹⁵⁸  Bi,^166–168  Au,^{102,128,148,169–173}  Pb,^{145,167,174–177}  Rh,^178,179  Cu,¹⁸⁰  Co,¹⁸¹  and Ag,^182,183 ) etc. catalysts are significantly more active for the EOR than Pt alone. Here we examine a number of these studies to highlight what has been done.

Pt–Pd catalysts: The fabrication of Pt–Pd alloy nanoparticles on graphene nano-sheets (PtPdNPs/GNs) have been described by Chen et al.¹⁴³  They reported that varying the molar ratio of the starting precursors, nanoparticles with different shapes such as spherical (Pt₁Pd₁NPs), nanoflowers (Pd@PtNFs) and nanodentrites (Pt₃Pd₁NPs) could be produced on graphene nanosheets. Based on these observations, they proposed a plausible growth mechanism of PtPdNPs/GNs as shown in Fig. 8.

The electrocatalytic properties of PtPdNPs/GNs for EOR exhibited higher activity and better tolerance to poisoning intermediates compared to PtPdNPs supported on carbon black (PtPdNPs/C).¹⁴³  Zhu et al. reported 3D PdPt bimetallic alloy nano-sponges which exhibited enhanced activity and stability towards EOR in an alkaline medium.^144,148  They developed a method to synthesize well-defined PdPt alloy nanowires, which exhibited significant activity enhancement towards EOR. Yang et al. reported an electrochemical method to synthesize PtPd alloy nanoparticles on Nafion-graphene film and demonstrated that the catalyst had good tolerance against poisoning by the reaction intermediates generated during EOR.¹⁴⁶  Datta et al. used poly-vinyl carbazole (PNVC), a conducting polymer composite matrix, cross linked with vanadium pentoxide (V₂O₅) and embedded with PtPd nano-crystallites for EOR leading to higher currents compared to carbon supported counterpart.¹⁴⁷  Lin et al. reported a spontaneous reduction method to prepare PtPd with high activity for EOR.¹⁴⁹  They showed that Pd₇₇Pt₂₃ had the highest activity followed by Pd₈₇Pt₁₃, Pd, and Pt.

Pt–Ru catalysts: PtRu bimetallic is the oldest and most studied in the literature. Therefore, it is more appropriate here to refer the interested reader to a comprehensive review by Petri who discusses the three periods of Pt–Ru research:¹⁶²  (a) the initial period after discovery (1963–1970); (b) the observation and classification of basic tendencies (like the effects of compound segregation, structural features on the activity; up to 1990); and (c) the nano-structural studies and molecular level consideration of electrocatalytic phenomena in combination with advanced applied studies of materials, mechanistic, and applied aspects (after 1990 to 2008). The review focuses on the balance of various aspects of Pt–Ru electrochemical related to material science and electrocatalysis as well as to remember the early basic results being of importance for future understanding of Pt–Ru functional properties.¹⁶²  Gralec et al. have studied the role of the Kegging-type phosphomolybdate (PMo₁₂O₄₀³⁻) ions adsorbed on C-supported PtRu and PtRu/C for EOR using CV, DEMS, and XPS.¹⁵⁹  They showed that modification of PtRu/C nanoparticles with phosphomolybdate ions lead to the suppression of the formation of surface Ru oxides which resulted into more than 40% activity increase for EOR at potentials >700 mV.¹⁵⁹ 

Pt–Sn catalysts: This bimetallic system has been extensively studied and here we mention a few selected studies. The structure-to-property relationship for EOR on PtSn in alkaline and acidic environments have been studied by Artyushkova et al.¹⁵¹  They observed that transitioning from acidic to an alkaline environment, changes the material structure and electrochemical reaction mechanisms. Electrocatalysts containing larger particles with larger relative amounts of metallic Pt and Sn performed better in acid media which they attributed to the inner-sphere electron transfer reaction on active PtSn alloy phase.¹⁵¹  PtSn electrocatalysts containing larger amounts of oxidized Pt and Sn performed better in alkaline, which they suggested indicated that hydroxyl species that are natively present on oxidized Pt and Sn were promoting an outer-sphere electron transfer. Du et al. sought to explain why Pt–Sn nanoparticles are active electrocatalysts for EOR but inactive for splitting the C–C bond to CO₂ using microelectrode to monitor the amount of CO₂.¹⁵²  They reported that the composition and crystalline structure of the Sn element played an important role in the CO₂ generation. The non-alloyed Pt₄₆–(SnO₂)₅₄ core/shell particles demonstrated a strong capability for breaking the C–C bond than pure Pt and intermetallic Pt/Sn. The effect of ethanol concentration on the DEFCs with PtSn anode performance and products distribution was studied by Assumpcao et al. using in-situ single fuel cell/ATR-FTIR setup.¹⁵³  They performed experiments at 80° using commercial Pt₃Sn/C as anodic catalyst and the concentrations of ethanol solution were varied from 0.1 to 2.0 M. An increase in power density was observed with the increase of ethanol concentration up to 1.0 M, and the FTIR spectra band intensities revealed an increase of acetic acid/acetaldehyde ratio with increasing concentration of ethanol. Baranova et al. studied EOR on PtSn/C nanoparticles in alkaline media synthesized using a polyol reduction method.¹⁵⁴  They formed bi-phase PtSn electrocatalysts where one group was composed of disordered PtSn alloys and the other group composed of PtSn alloys intimately mixed with SnO_x. They reported that all catalysts were active during CV experiments but the bi-phase PtSn+SnO_x nanoparticles had significantly higher current densities at lower over-potentials compared to the pure alloy PtSn catalysts.¹⁵⁴  They correlated the catalyst bulk and surface structure with the observed EOR in alkaline media demonstrating that 1 M KOH was the best when electrocatalyst contained higher amounts of both Pt and oxides. They reported that alloying of Pt with Sn improves intrinsic Pt catalytic activity and plausibly prevents Pt oxidation.¹⁵⁴ 

Pt–Rh catalysts: For these catalysts, Calderon-Cardenas et al. have studied the effect of the composition and thermal treatment in H₂ of Pt–Rh/C materials with atomic ratios of Pt : Rh 3 : 1, 1 : 1 and 1 : 3 and metal loading of 40 wt.% for EOR in alkaline media.¹⁷⁸  They reported that thermally treated Pt–Rh catalysts in a hydrogen atmosphere showed greater stability and higher current densities and suggested the necessity of exploring the effects of thermal treatments of the catalysts for EOR. Shen et al. prepared PtRh/C catalysts and compare their catalytic activities with that of Pt/C in alkaline media and reported that the peak current density on Pt₂Rh/C was about 2.4 times of that on Pt/C.¹⁷⁹  They ascribed the enhanced activity to the improved C–C bond cleavage in the presence of Rh and to the accelerated oxidation kinetics of adsorbed CO to CO₂ in alkaline media.

Pt–Pb catalysts: Gunji et al. synthesized Pt₃Pb(core)–PtPb(shell) nanoparticles on carbon black by converting nano-crystalline Pt to an ordered intermetallic compound with the reduction of Pb ions and tested them for EOR in alkaline media.¹⁷⁴  The nanoparticles exhibited enhanced catalytic activity and relatively stable cycle performance towards EOR in an alkaline solution. They attributed the improved performance to both the enhancement of ethanol dehydrogenation and the higher concentration of surface adsorbed OH⁻ on the modified PtPb surface in the Pt₃Pb–PtPb core-shell NPs. The mechanism for EOR on a Pt electrode modified with an irreversibly-deposited layer of Pb in alkaline solution was proposed by Christensen et al. based on in-situ IRRAS insights.¹⁴⁵  He et al. described an approach for the selective cleavage of the C–C bond using a solution-born co-catalyst based on Pb(iv) acetate, which they suggested controlled the mode of ethanol adsorption so as to facilitate the direct activation of the C–C bond, as shown in Fig. 9.¹⁷⁶ 

Matsumoto studied the electrocatalytic activities of a wide range of intermetallic bulk compounds for EOR in alkaline media including PtPb which had the lowest onset potential for ethanol oxidation of 20–30 mV less than that of pure Pt and Pd.¹⁶⁷  The current densities for PtPb were ≥17 times larger than those of pure Pt and Pd. Yang et al. synthesized Pt–PbO_x nanocomposite catalyst with a mean size of 3.23 nm with a much higher catalytic activity and a longer durability than Pt nanoparticles and commercial Pt black catalysts for EOR in alkaline media.¹⁷⁵  In-situ IRRAS data revealed that breaking the C–C bonds on Pt–PbO_x was 5.17 times higher than that of the Pt nanoparticles.

Pt–Au catalysts: Mourdikoudis et al. synthesized PtAu hetero-nanostructures comprising the dimer (Pt–Au) and core-satellite (Pt@Au) configurations by means of a seeded growth procedure using Pt nano-dendrites as seeds.¹⁶⁹  They reported that the prepared PtAu bimetallic nanostructures were highly efficient catalysts for EOR in alkaline solution. Dutta et al. synthesized PtAu alloyed nanoparticles and reported improved half-cell activity for EOR and a considerable increase in the peak power density (>191%) in an in-house fabricated DEFCs.¹⁷⁰  da Silva et al. tested PtAu/C electrocatalysts in different atomic ratios and reported that the 50 : 50 as the most promising ratio for half-cell tests for EOR in alkaline media, while single fuel cell suggested a 70 : 30 ratio.¹⁷²  They attributed the discrepancy to the electrode architecture since 50 : 50 ratio yielded a much thicker electrode than the 70 : 30 catalyst because the Pt load was the same. Song et al. reported the preparation of hollow Au@Pt core–shell nanoparticles and used them for EOR in alkaline media which showed high current density in the forward scans.¹⁷³  Cherevko et al. prepared highly ordered Pt decorated Au nanowire arrays, Pt/Au NWA and studied the effect of shell materials for EOR in alkaline media and reported up to 4-fold increase in the ethanol oxidation peak current.¹²⁸ 

Pt–Bi catalysts: Figueiredo et al. reported the enhancement of EOR activity on Pt/C by simple adsorption of Bi on the surface.¹⁶⁶  They reported that Bi promoted the cleavage of the C–C bond. Matsumoto et al. reported PtBi and PtBi₂ intermetallic compounds as promising electrocatalysts among the various bulk electrodes they examined which had lower onset potentials for the EOR and exhibited extremely stable oxidation currents of 4.8 and 3.3 mA cm⁻² during the constant-potential electrolysis.¹⁶⁷  Tusi et al. prepared Pt/C, Bi/C and PtBi/C (Pt : Bi atomic ratios of 90 : 10, 70 : 30 and 50 : 50) electrocatalysts and showed that PtBi/C had significant increase of performance for EOR in alkaline compared to Pt/C.¹⁶⁸  They noted that the performance of PtBi/C electrocatalysts for EOR was superior in alkaline medium compared to acid medium.

Pt–others: Li et al. synthesized a series of Mo-doped Pt/C catalysts with a microwave assisted technology and investigated the effects of Mo content on using CV, CA, and EIS.¹⁶⁵  They reported that the Pt₂Mo/C showed the highest current density and the slowest deterioration from intermediates/products poisoning. Kepenier et al. fabricated graphene supported PtCo catalysts (Pt : Co=1 : 1, 1 : 7 and 1 : 44) by the rapid microwave heating method and reported that the molar ratio of 1 : 7 had highest activity.¹⁸¹  Jin et al. prepared Pt/C catalysts modified by the potentiostatic deposition of Ag and reported a significant improvement in the activity of PtAg/C for ethanol oxidation in alkaline solution.¹⁸²  El-Maksoud et al. investigated the electrocatalytic effect of Pb, Tl, and Cd ad-atoms on Pt electrode for EOR in alkaline medium.¹⁷⁷  They reported that all three metal ad-atoms enhanced activity and Pb and Tl ad-atoms increased the oxidation rate by a factor of about 15, whereas Cd ad-atoms shifted the polarization curves negatively by a factor of about 5 at lower over-potentials.

Pd is considered the most active metal for EOR in alkaline media; hence a lot of effort has been directed towards improving its catalytic activity as evidenced from recent reviews.^1–4  Several bimetallic catalysts have been tested so far Pd–M (M=Ni,^17,184–203  Ru,^{17,60,204–209}  Au,^{17,128,148,173,208,210–216}  Ir,^217,218  Bi,^219–221  Sn,^{75,185,188,216,222}  V,²²³  W,^224,225  Ag,^49–52  Cu,^226–232  Co,^17,233  Fe,¹⁷  Mn,⁵⁷  Ti,^234,235  Rh,^236,237  Sb,²³⁸  Te,²³⁹  La,²⁴⁰  and Pb.²⁴¹ ). Herein we highlight a few of them.

Pd–Ni catalysts: PdNi combination has been extensively studied because Ni is very cheap. Obradovic et al. synthesized Pd–Ni/C using NaBH₄ reduction method which they reported to be up to three times more active for the EOR compared to Pd/C.¹⁸⁴  They found that maximum activity was attained after fifty cycles with the positive potential limit of 1.2 V/RHE regardless of whether they were performed in the electrolyte with or without ethanol, hence proposed that potential cycling induces reorganization of the catalyst surface bringing Pd and Ni sites to a more suitable arrangement for ethanol electrooxidation.¹⁸⁴  Moraes et al. have reported performance enhancement for alkaline DEFCs using non-functionalized and functionalized Vulcan carbon supported Pd, PdNi, and PdNiSn anodic electrocatalysts produced by impregnation-reduction.¹⁸⁵  They reported that alkaline DEFCs with PdNiSn supported on functionalized Vulcan had the best performance which they attributed to improved textural properties.¹⁸⁵  Chen et al. prepared PdNi nano-catalysts supported on multi-walled carbon nanotube (MWCNT) using a modified polyol method.¹⁸⁷  They reported that the surface oxygen content in PdNi/MWCNT was higher than in Pd/MWCNT and Ni existed mainly in the form of hydroxides which were attributed to be responsible for the improved poison resistance. Wang and co-workers have demonstrated in a number of studies that de-alloying can be used to improve PdNi electrocatalysts.^186,189,192  They used Pd–Ni–P film prepared via electro-deposition on Au substrate and de-alloyed it by repetitive potential cycling in acidic media to leach out most of the Ni and P components and the resulting film showed significantly enhanced and durable activity for EOR. They used in-situ ATR-SEIRAS for reaction insights which revealed that the enhanced electrocatalysis correlated well with the enhanced formation of adsorbed CO and acetate.¹⁸⁶  This was an extension of their earlier work on Pd–Ni–P where they showed that Pd–Ni–P have double the number of electrocatalytically active sites (12.03%) compared with the Pd–Ni (6.04%) and Pd-black (5.12%) samples.¹⁹²  Dutta and Datta have investigated EOR on Pd_xNi_y/C in alkaline medium synthesized by simultaneous reduction of metal precursors using NaBH₄ method.¹⁹⁰  They attributed the improved catalytic activity on NiO present in the binary catalyst matrix. Ahmed and Jeon studied a series of graphene supported Ni_xPd_y binary alloyed catalysts for EOR and reported activities in the order Ni₇₅Pd₂₅/G>Ni₀Pd₁₀₀/G>Ni₂₅Pd₇₅/G>Ni₅₀Pd₅₀/G as shown in Fig. 10.¹⁹¹ 

Sheikh et al. synthesized Pd–Ni/C catalysts by impregnation-reduction method and reported that Pd₄₀Ni₆₀/C had the best catalytic performance for EOR in alkaline medium, which they attributed to Ni hydroxides (Ni(OH)_x).¹⁸⁸  Lee et al. prepared highly monodisperse 5 nm Pd–Ni alloy nanoparticles by the reduction of Pd(acac)₂/Ni(acac)₂ mixtures with tert-butylamine-borane complex in the presence of oleic acid and oleylamine which exhibited higher activity and stability for EOR.¹⁹⁶  Miao et al. used to electroless co-plating to coat Pd–Ni nanoparticles on Si nanowires for EOR.^195,199  They reported that Pd–Ni/SiNWs electrode had higher activity and better long-term stability in an alkaline solution.¹⁹⁵  The work was an extension of their previous work on using silicon microchannel plates modified with Ni–Pd nanoparticles.¹⁹⁹ 

Shen et al. performed a quantitative product analysis of EOR in an anion-exchange membrane DEFC that consisted of a PdNi/C anode and found that Pd₂Ni₃/C leads to a significant increase in the cell performance compared to Pd/C but did not improve the selectivity towards CO₂.^193,201  They observed that among the operating conditions tested (temperature, discharge current, and ethanol concentration) the operating temperature was the most significant parameter that affect the CO₂ selectivity: increasing the temperature from 60 to 100 °C increased the CO₂ current efficiency from 6.0% to 30.6% with the Pd/C. This work was an extension of their earlier studies.⁷  Roy et al. prepared spherical Pd nanoparticles and dip-coated them Ni-foil and found them to be superior electrocatalysts for EOR compared to the Ni-supported Pd electrode despite of them having less Pd⁰ loading.¹⁹⁴  Qi et al. used de-alloying method to prepare Pd₄₀Ni₆₀ alloy from a ternary Al₇₅Pd₁₀Ni₁₅ in a 20 wt.% NaOH solution under free corrosion conditions.¹⁹⁸  They reported that Pd₄₀Ni₆₀ had enhanced electrocatalytic performance for EOR in alkaline media than nanoporous Pd. Zhang et al. prepared Pd_xNi_y/C through a solution phase-based nanocapsule method and showed that onset potential for EOR on Pd₄Ni₅/C was negative shifted by 180 mV and the exchange current density was 33 times higher compared to Pd/C.¹⁹⁷  They proposed that surface Ni promoted a refreshing of the Pd active sites, thus enhancing the overall reaction kinetics. Maiyalagan and Scott prepared Pd–Ni nanoparticles supported on carbon nanofibers by NaBH₄ reduction method and reported negative onset potential shift of 200 mV and four times increased peak current density for EOR on Pd–Ni/CNF compared to Pd/C.²⁰² 

Pd–Ru catalyst: Monyoncho et al. prepared PdRu nanoparticles supported on carbon Pd_xRu_1−x/C (x=1, 0.99, 0.95, 0.90, 0.80, 0.50) using a polyol method and reported that the resulting bimetallic catalysts were primarily a mix of Pd metal, Ru oxides and Pd oxides.²⁰⁴  They found that addition of 1–10 at.% Ru to Pd not only lowered the onset oxidation potential for EOR but also produced higher current densities at lower potentials compared to Pd/C by itself. In particular, they singled out Pd₉₀Ru₁₀/C and Pd₉₉Ru₁/C which gave up to six times higher current densities than Pd/C at −0.96 V and −0.67 V vs MSE, respectively. Similarly, Ma et al. studied PdRu/C catalysts with various Pd : Ru atomic ratios synthesized by impregnation method and tested them in AEM-DEFCs.^205,206  They reported that the anode with Pd₃Ru/C showed a maximum power density as high as 176 mW cm² at 80 °C, which was about 1.8 times higher than that of the Pd/C catalyst. Anindita et al. reported that addition of Ru to a Pd-0.5wt%C composite electrode increased the electrocatalytic activity greatly, attaining a maximum at 20wt% i.e. (Pd-0.5wt%C-20wt% Ru) for EOR in alkaline media.²⁰⁷ 

Pd–Au catalysts: Cai et al. synthesized Pd nanotubes covered by high-density Au-islands that increased the mass activity by up to six times for EOR in alkaline media compared to Pd/C as shown in Fig. 11.²¹⁰  They proposed a model to explain the relationship between the structure and the catalytic activity.

Hong et al. have demonstrated a rapid synthetic process for alloyed dendritic PdAu nanocrystals that are active for EOR in alkaline media.²¹¹  The process involves mixing Na₂PdCl₄, HAuCl₄, polyvinylpyrrolidone and hydroquinone and heating at 50 °C for 15 min. Smiljanic et al. examined the catalytic properties of Pd/Au(111) nanostructures obtained by spontaneous deposition of Pd using PdSO₄ and PdCl₂ salts.²¹²  They reported that Pd/Au(111) nanostructures obtained using PdCl₂ salt had higher activity which they ascribed to the thinner and smoother Pd deposits on the surface, hence more convenient sites for the adsorption of ethanol and its subsequent oxidation steps. Song et al. prepared hollow Au@Pd core-shell nanoparticles using galvanic displacement with Ag which showed highest current density in forward scan for EOR in alkaline media.¹⁷³  Xu et al. prepare Pd–Au alloy electrocatalysts using dimethylformamide co-reduction method under an ultrasonic process and reported Pd₃Au/C exhibited an enhanced catalytic activity and stability for EOR compared to monometallic Pd/C catalyst.²¹³  Cheng et al. prepared highly ordered PdAu nanowire arrays using a combination of anodized Al oxide template-electrodeposition and Pd nanowire arrays reacting with HAuCl₄.²¹⁴  They found that the PdAu nanowires shifted onset oxidation potential by 123 mV more negative compared with that on the Pd nanowires.

Recently, Assaud et al. reported Pd nanoparticles of controlled particle size deposited by Atomic Layer Deposition (ALD) on electrochemically grown TiO₂ nanotubes (TNTs).^117,242  The particle size was controlled by the number of ALD cycles (Fig. 12). They showed by TEM that catalysts fully cover the inner and outer walls of the three-dimensional nanostructured TiO₂. The influence of TiO₂ nanotube support was demonstrated through the modification of the crystalline structure of the TNTs anatase TiO₂ phase obtained after annealing is more conductive than the amorphous TiO₂. Catalysts with a different number of ALD cycles were prepared and studied for EOR. Among the prepared electrocatalysts (N=400–900 ALD cycles), the 500 ALD Pd/TNTs system showed the best catalytic activity and satisfactory stability in alkaline media (Fig. 13).

Zhu et al. decorated carbon-supported gold nanoparticles with monolayer of Pd atoms with different Pd : Au atomic ratios using chemical epitaxial seeded growth method and showed that PdAu nanoparticles had higher specific activities than Pd/C for EOR in alkaline media.²¹⁵  He et al. prepared carbon-supported Pd₄Au and Pd_2.5Sn nanoparticles using a chemical reduction method and examined the kinetics for EOR using impedance spectroscopy and Tafel plots which showed that the reaction kinetics were somewhat more sluggish on the Pd-based alloy catalysts than on commercial Pt/C, but the alloy catalysts had higher tolerance to surface poisoning.²¹⁶  Pd₄Au/C displayed the best catalytic activity among the series of prepared catalysts for EOR in alkaline media.

Pd–Cu catalysts: Serov et al. used a sacrificial support method in combination with the thermal reduction of metal precursors to prepare unsupported uniformly-distributed PdCu catalysts with ratios of 1 : 3, 1 : 1, and 3 : 1.²²⁶  They found that PdCu and Pd₃Cu electrocatalysts showed improved EOR activity, which they attributed to the presence of surface Cu sites favouring adsorbed OH⁻ species as confirmed by their DFT calculations in the paper.²²⁶  Mao et al. prepared a series of surface Pd rich Cu_xPd_y/C catalysts and reported Cu₁Pd₂/C stood out from the four sets tested for EOR in alkaline.²²⁷  Cai et al. reported catalyst with Cu core-shell structure prepared by the galvanic replacement between Pd²⁺ ions and Cu particles (Cu@PdCu/C) which showed greatly improved durability, poisoning tolerance, and current density of 2.78 times higher than Pd/C for EOR.²³⁰  Zhao et al. have demonstrated a one-pot, room temperature aqueous synthesis of submicrometer-sized PdCu networks as superior catalysts for EOR in alkaline medium.²²⁹  Their composition-optimized Pd₇₃Cu₂₇ network showed superior performance for EOR and better tolerance of CO-like poisoning species compared to commercial Pd/C. Wang & Kang et al. have worked on the development of high performance PdCu/C catalysts to enhance EOR performance.^231,232 

Pd–Sn catalysts: Mao et al. used impregnation reduction method to prepare carbon-supported PdSn–SnO₂ with higher catalytic activity for EOR in alkaline solution compared to Pd–Sn/C and Pd/C catalysts.⁷⁵  They attributed the higher activity to easy adsorption-dissociation of OH⁻ over the SnO₂ surface which changed the electronic effect and accelerated the adsorption of ethanol on the surface of Pd. Du et al. prepared a series of carbon-supported Pd–Sn binary alloyed catalysts using a polyol method among which Pd₈₆Sn₁₄/C catalyst showed much enhanced current densities.⁸  They supplemented their study with DFT calculations, which confirmed that Pd–Sn alloy structures lead to lower reaction energies for ethanol dehydrogenation compared to pure Pd crystals.

Ni-based bimetallic catalysts without Pt or Pd have been tested for EOR in alkaline media.^243–252  Zhan et al. synthesized well-dispersed mesoporous NiCo₂O₄ fibres using an easy-controlled template-free method with specific surface area of 54.469 m² g⁻¹ and average pore size of 13.5 nm.²⁴³  The catalysts exhibited significantly high EOR activity with higher current densities and lower onset potential compared to those of Co₃O₄ and NiO. Ren et al. reported a three-dimensional free-standing Ni nanoparticle aerogel with a graphene sheet network formed through the self-assembly aggregation of graphene accompanied by nickel nanoparticle in-situ loading on the graphene sheet during the hydrothermal reduction of graphene oxide and Ni ions.²⁴⁴  The three-dimensional composite architecture revealed excellent EOR activity. Hassan and Hamid electrodeposited Ni–Cr₂O₃ nanocomposite supported on carbon electrodes for EOR and showed that the catalytic activity of the fabricated electrodes increased with increasing the volume fraction percent (V_f%) of Cr₂O₃ in the deposited film up to 7 V_f%.²⁴⁵  The Ni–Cr₂O₃/C (7 V_f%) electrode displayed significantly enhanced catalytic activity and stability towards EOR compared with Ni/C electrode. Yi et al. compared nanoporous Ni electrode synthesized by electrodeposition into alumina template with smooth Ni electrode and reported that nanoporous Ni electrode had dominant (111) facets and self-regulated NiOOH rich surface in KOH solution, which they ascribed to be active for EOR.²⁴⁶  Tarasevich, Tsivadze, and co-workers conducted studies on anodic (RuNi/C) and cathodic (PtCo/C and CoN₄/C) catalysts, polybenzimidazole membrane, and membrane-electrode assemblies for alkaline ethanol-oxygen fuel cell.^247,250  They reported optimized atomic percent of Ru : Ni=68 : 32 and the metal mass on carbonaceous support of 15–20% which was superior to commercial Pt/C and RuPt/C catalysts when calculated per unit mass of the precious metal. Using chromatographic analysis of the products, they reported the highest CO₂ yield at low electrolysis overvoltage and elevated temperature. The use of non-platinum group metals such as Au^{102,208,253–258}  and Rh^259,260  have also received attention, especially Au-based, but they are expensive for commercial applications, hence will not be discussed further.

The understanding EOR mechanism is critical for the rational design of catalysts and in the reaction optimization for DEFCs. The synergy between experimental techniques and theoretical simulations has been employed to achieve this objective. A combination of pure electrochemical methods with state of the art in-situ analytical methods to monitor adsorbed intermediates/products is necessary to visualize the reaction paths. See the methodology chapter for more details on the techniques used.

Ethanol electrooxidation mechanism has many pathways leading to controversial debates on the details in the literature. Nevertheless, there is a general consensus that EOR mechanism exhibits a “dual pathway”.^{110,257,261,262}  Several reaction mechanism schematics have been presented to explain the mechanistic details.^{102,110,145,155,170,176,261,263–272}  Currently, there are two schematics, which in our opinion are inclusive of all the other schematics reported in the literature. One of the schematics is more general and is based on electrochemical experiments without proper identification of intermediates/products.^110,261  The second schematic is more detailed and applicable to both acidic and alkaline conditions.¹⁵⁵  The first and more general schematic is shown in Fig. 14 and shows a dual path reaction where in one path the reactive intermediates are weakly bound to the catalyst surface, which leads to incomplete oxidation “C₂ pathway”.^110,261  The other pathway involves strong bond intermediates, which can be fully oxidized to CO₂ in the “C₁ pathway” if the right conditions are met, otherwise they can block the catalyst surface.

The arrows in Fig. 14 show the interplay between the two pathways. In the dominant C₂ pathway, the C–C bond does not break and ethanol is oxidized to products such as acetaldehyde, acetic acid, acetate, germinal diols etc. depending on the electrolyte used. In the C₁ pathway, the C–C bond is broken and the fragments are oxidized into CO and eventually CO₂.

The second and more comprehensive schematic presented to date is shown in Fig. 15.¹⁵⁵  It was based on cumulative experimental data from different groups and intuition. In the schematic, the lighter highlighted species were identified with NMR^155,273  while the darker highlighted species had been identified by DEMS, chromatography, and infrared spectroscopy. The reaction steps marked with stars (*) are those in which OH_ads are involved and catalytic sites are regenerated due to reaction with OH_ads. Note that although, the schematics show the reaction paths in a stepwise manner it is possible some steps would happen in a concerted manner as suggested in literature.²⁷⁴  The schematic is universal in a sense that it can be used to explain observations in both acidic and alkaline conditions. Therefore, the schematic in Fig. 15 stands out among the many schematics^{102,110,145,155,170,176,261,263–272}  in highlighting the experimental progress made in understanding EOR without theoretical insight. It is important to note that EOR intermediates/products are influenced by the nature of the catalyst, the electrolyte, and the applied potential. Therefore, not all the intermediates/products shown are observed in each reaction.

The schematic in Fig. 15 tells us that there are three possible routes for ethanol electrooxidation. The first two pathways are due to dehydrogenation which would lead either to adsorbed CH₃CHOH or CH₃CH₂O in the first step. In the second pathway, acetaldehyde is the central molecule. Following the first steps in these two pathways in EOR, there are many other possibilities as shown in the schematic of which the details are not necessarily accurate as presented in the Fig. 15. The third pathway is due to loss of the water molecule to form adsorbed CH₃CH₂ which can be reduced to ethane. Path three is specific to acidic media where ethane was observed as one of the products in low/cathodic potential. The key points we would like to highlight from Fig. 15 are: (i) Acetyl (CH₃CO) pathway leads to the cleaving of the C–C bond to form CO₂ but it faces a strong competition to formation of acetic acid and/or ethyl acetate instead based on quantitative analysis of the products, (ii) the NMR technique was critical in identifying products such as ethyl acetate (CH₃COOCH₂CH₃), germinal diol (CH₃CH(OH)₂), and CH₃CH(OH)OCH₂CH₃, (iii) the link of the reactions leading to some products is not explicitly presented such as CH₃CH(OH)₂→CH₃COOH, (iv) the schematic does not incorporate first-principle insights available in literature. The last two points represents the weaknesses of the schematic which will be addressed in a different forum. Herein we focus our attention to what has been done for EOR in alkaline media particularly in addressing the question as to why it is difficult to break the C–C bond.

The understanding of EOR mechanism in alkaline conditions is in its infancy, for only a few studies provide molecular information.^{104,269,275–280}  The reason for the scarcity of EOR mechanism details was the fact that the produced CO₂ forms soluble carbonates in the presence of aqueous alkaline electrolyte, which makes it difficult to study using FTIR or model DEMS systems. Rao et al. overcame this obstacle by using alkaline polymer electrolyte membranes which gave them opportunity to observe CO₂ produced during EOR using fuel cell effluents coupled to DEMS system.^104,279  Hence, DEMS was the first technique to reveal that in alkaline conditions the C–C bond cleavage in ethanol is more efficient than in acidic media. Rao et al. demonstrated that CO₂ current efficiency was around 55% at 0.8 V/RHE at 60 °C for alkaline MEA compared to only 2% for acidic MEAs.^104,279  This fact was confirmed by Cremers et al. who reported that the kinetics for ethanol oxidation in alkaline media were higher than in acidic media under the same conditions.¹⁰³  In a subsequent study, Cremers et al. made some interesting observations:²⁸¹  (i) in alkaline medium ethanol adsorbates can only be desorbed in form of carbon dioxide and methane, (ii) pre-adsorbed CO could not be reduced to methane at Pt in alkaline conditions, hence they ascribed methane formation from ethanol adsorbates to an adsorbed CH_x or CO_xH_y species, (iii) ethanol adsorbates in alkaline media can be oxidized in two potential regions, i.e., below and above 0.9 V/RHE, (iv) the calculated number of electrons per molecule of CO₂ evolved in the potential region below 0.9 V/RHE was found to be two, independent of the adsorption potential. They claimed that a form of adsorbed CO_ads species was present on the electrode. They suggested that the higher calculated number of electrons per molecule of CO₂ in the potential region above 0.9 V/RHE pointed to the co-existence of more than one adsorbate species. The deviation of the number of electrons per molecule of CO₂ from the value of two in the potential region below 0.9 V/RHE for stripping experiments started in cathodic direction indicated an alteration of the adsorbate in the form of CO_ads, initially being present, or the co-existence of two adsorbate species. Their observation that no CO₂ formation was observed in the potential region above 0.9 V/RHE for stripping experiments started in cathodic direction, lead to their conclusion that the adsorbates which are oxidized in the potential region above 0.9 V/RHE are the ones which can be reduced to methane.²⁸¹  They later determined that in alkaline media CO₂ was only formed in the potential region of the oxidation of adsorbed CO.¹⁰⁷  Therefore, they inferred that in alkaline conditions, CO₂ was produced from adsorbed ethanol and not from bulk ethanol. (v) The efficiency of breaking the C–C bond is lower with an increase in the concentration of ethanol,²⁸¹  and comparison between Pd and Pt showed that on Pd, ethanol oxidation in alkaline media is almost selectively towards acetate formation. Similarly, on Au electrodes, acetate and ethyl acetate seemed to be the exclusive products with no cleaving of the C–C bond. They reported Ni to be a poor catalyst for EOR.¹²⁹ 

It has been recognized that the inconsistencies of EOR details in literature both from experimental and computational studies were due to the sensitivities of the reaction to the surface structure of the electrode and the adsorbed ions.²⁷¹  Melke et al. summarized the available literature data up to 2010 into the reaction schematic shown in Fig. 16.^{103,272,282–284} 

Fig. 16, shows that there exist two main EOR paths governed by the reaction conditions and the structure/morphology of the catalyst used. Acetyl serves as a bridging intermediate between the two reaction paths, which means it will have a very small life-span once generated during the reaction due to the strong competition to either break the C–C bond or to the formation of acetic acid. The reaction path towards acetic acid (acetate in alkaline media) is preferred on close-packed surfaces such as Pt(111) and/or at high surface coverage (left side in Fig. 16). On the other hand, full oxidation to CO₂ prefers open or stepped surfaces like Pt(110), Pt(100), and Pt(211) (right side in Fig. 16) and low surface coverage. It has been suggested that the bond breaking takes place either within adsorbed CHCO or CH₂CO species,⁷⁷  and CH_x and CO are the strongly adsorbed intermediates.²⁸¹ 

Fang et al. reported the mechanism for EOR on a Pd electrode in alkaline solution using cyclic voltammetry and in-situ infrared spectroscopy.²⁸⁵  They observed the best performance at the pH=14 (1M NaOH) and acetate was the main product for concentrations higher than 0.5 M NaOH. They reported that the C–C bond cleavage to form CO₂ occurred at pH≤13, which was in agreement with online mass-spectrometry results from Cantane and Lima evidencing CO₂ production over Pt and Pd in 0.01 M NaOH where acetic acid formation was almost absent.¹³¹  Christensen et al. have shown that the interfacial pH drops at higher potentials due to the high consumption of OH⁻ which is not completely counterbalanced by the OH⁻ diffusion from the bulk-phase.^269,278,286  This phenomenon leads to a transition from alkaline to acidic conditions at the interphase. The transition potential varies with the diffusion rate of OH⁻ which is dependent on the temperature and mass flow-rate. They reported that during electrooxidation reaction, ethanol is converted to acetate in alkaline pH but above the transition potential, acetic acid and traces of CO₂ are formed.²⁸⁷ 

A comparative study of EOR on Pd, Pt, and Rh in alkaline electrolyte through online DEMS experiments have shown similar amounts of CO₂ for the three metals but Pd electrode produced higher amounts of ethyl acetate (which they attributed to acetic acid formation).¹³¹  The authors reported that on Pt and Rh the formation of CO₂ occurred mainly via oxidation of either the adsorbed CO or CH_x species formed after dissociative adsorption of ethanol or the oxidation of the ethoxy species that takes place only at low potentials. This argument was based on the observation of methane for Pt and Rh electrodes during potential excursions to lower potentials which was lacking for the case of Pd electrode. These insights implied that the dissociative adsorption of ethanol or ethoxy species is inhibited at higher potentials on Pt and Rh.¹³¹  For the case of Pd electrode, the reaction may be occurring via non-dissociative adsorption of ethanol or ethoxy species at lower potentials followed by oxidation to acetaldehyde and to acetic acid. Alternatively, they proposed a parallel reaction path where acetaldehyde molecules adsorbed on the Pd surface can be deprotonated, yielding a reaction intermediate in which the C–C bond can be easily broken to produce CO₂ after potential excursions to higher potentials.¹³¹ 

The question(s) on why it is difficult to break the C–C bond and at what intermediate does it occur during ethanol electrooxidation was recently tackled by the authors.²⁸⁸  We found that at lower potential (e.g. 0.56V/RHE), it is possible to break the C–C bond on Pd/C in alkaline media to form CO₂ as shown in Fig. 17b. However, the selectivity is poor due to competition towards the formation of acetate and other side products. At higher potentials (e.g. 0.72 V/RHE), the selectivity towards C–C cleavage gets worse, as shown in Fig. 17e. In that work, DFT calculations were used to complete the reaction pathway details using the computational hydrogen electrode approach. The calculations highlighted the pivotal role of the CH₃CO intermediate that can either undergo a C–C bond scission yielding CO and then CO₂ or that can be oxidized towards CH₃COO⁻. The latter is a dead end in the reaction path towards CO₂ production, since it cannot be easily oxidized or broken into C₁ fragments in agreement with CV data provided therein showing that acetic acid in KOH was not active. Unfortunately, CH₃CO is not the most favored intermediate formed during EOR on Pd, hence limiting the production of CO₂. Strategies to overcome these limitations are discussed in the conclusions and outlook below.

The top priority challenge for the successful development of DEFCs depends on the detailed understanding of the reaction mechanism which would pave way for the rational design of the catalysts capable of cleaving the C–C bond in ethanol. A breakthrough in this endeavour will increase the overall DEFCs efficiency from the current 14% to 43%, hence making them the strongest competitor to hydrogen fuels cells, which have an efficiency of 54%.²⁹  Other issues which are more in the engineering development part include: membrane and ionomer improvements, water and ethanol transport management, carbon dioxide regulation and electrolyte development.

Although, ethanol electrooxidation reaction kinetics are faster in alkaline conditions, an efficient catalyst for the complete oxidation of ethanol to CO₂ remains an outstanding challenge in the reviewed literature. Even after three decades of active research, there is still no selective catalyst for breaking the C–C bond. Therefore, efforts in the fundamental understanding of the reaction mechanism are required to pave the way for the rational design of efficient catalysts. On the cathode side, the challenge is how to enhance non-Pt catalysts to make them comparable to Pt for ORR.

Anion-exchange membrane can be grouped into two categories: polyelectrolytes and alkali-doped polymer membranes.³¹  Membrane improvements in parameters such as composition, ionic conductivity, ethanol permeability, thermal and chemical stability are discussed elsewhere.^16,25,289  The main challenges with alkaline anion-exchange membranes are the low stability in OH⁻ and low OH⁻ conductivities. Therefore, efforts are required not only in developing new membrane formulations, but also the development of tools for characterising membrane properties such as water uptake, ethanol permeability, water diffusivity, and electro-osmotic coefficient.

Ionomers are critical components of the fuel cell that help to bind discrete catalyst particles, which must form a porous conduction layer for the transfer of ions, electrons, and reactants/products.²⁹⁰  Therefore, similar to membranes, there is need to improve ionic conductivity, thermal and chemical stability, and make them soluble in nontoxic and cheap solvents.

Water management is a critical requirement for the long-term operation of DEFCs. Water is produced at the anode and consumed at the cathode, leading to a high water crossover from the anode to the cathode. Although, the water crossover phenomenon has the advantage of improving the ionic conductivity, too much water crossover would lead to the cathode flooding and hinder oxygen transport. Similarly, a low water crossover can facilitate oxygen transport but leads to mass transport loss for ORR, which would result into high cathode activation loss.

Maintaining proper circulation of the fuel (ethanol) is critical to obtaining the optimum current density and avoiding fuel waste due to incomplete reaction.²⁵  If ethanol concentration in the anode is too high will lead to the reduction of the coverage of OH⁻ and increase the anode activation loss. Secondly, high ethanol concentration will increase ethanol crossover that can reduce fuel utilization. On the other hand, too low ethanol concentration level in the anode will increase mass transport loss and reduce the optimum current. Therefore, efforts in the design of the anode flow field as well as determining the optimum operating conditions, such as ethanol concentration supplied to the flow field and the ethanol solution flow rate in the flow field and temperature are required.

It is cheaper to use air rather than oxygen in the fuel cells but this presents a challenge for alkaline fuel cells because air contains around 0.039% (volume fraction) of CO₂. Under standard conditions, this CO₂ will react with the OH⁻ generated by the ORR to form carbonate (CO₂³⁻), which may affect cell performance in two aspects:²⁵  (i) by decreasing the pH level in the cathode, thus affecting the kinetic of the ORR, and (ii) by reducing the ionic conductivity in both the cathode and membrane, increasing the cell resistance. Hence, the problem associated with CO₂ from air is an issue that needs to be addressed in the future.

The fact that KOH reacts with CO₂ produced at the anode or from the air at the cathode presents a serious challenge. As discussed above KOH reacts with CO₂ to form carbonate which precipitates and blocks the pores of the membranes and the electrodes. At the cathode, KOH can reduce the hydrophobicity of the gas diffusion layer, thereby breaking the balance of mass transport between water and oxygen. Therefore, efforts in consideration of other possible electrolyte formulations such as ionic liquids to mitigate the use of the KOH will be welcomed in the future studies.

From the survey of the literature presented herein, it is clear that in the past two decades tremendous efforts have been devoted to developing DEFCs from both the fundamental understanding of the reaction to the prototype fuel cell design development. From a fundamental perspective, many analytical techniques have been extended towards in-situ identification and quantification of the reaction products. This is a significant advance for they now provide the opportunity to visualize the progress of the reaction in real-time, hence providing much needed insight into understanding the reaction mechanism. A range of tools (electrochemical techniques, mass spectroscopy, surface enhanced Raman spectroscopy, sum frequency generation spectroscopy, X-ray absorption spectroscopy, chromatography, and nuclear magnetic resonance spectroscopy) and protocols for using them to probe electrocatalytic reactions in-situ are now available in the literature. Besides the electrochemical techniques, infrared spectroscopy, nuclear magnetic resonance spectroscopy, and X-ray absorption/photoemission spectroscopy are recommended as complementary techniques to capture complete details of the reaction mechanism. To complement these experimental tools is the use of first-principles calculations such as DFT for atomic insights to facilitate the screening of the best candidate catalysts for experimental testing. Combining the contributions of these various techniques for the past two decades has provided good understanding of the EOR mechanism, although it is very complex. Special credit goes to the application of NMR,^155,273  which has disclosed intermediates such as CH₃CH(OH)₂, CH₃CH(OH)OCH₃, CH₃CH(OH)OCH₂CH₃, and CH₃COOCH₂CH₃, which were impossible to distinguish with the other techniques used so far.

When it comes to catalyst design, excellent progress has been made which allows tailoring the nanoparticle structures and morphology from the atomic level. Now it is possible to make nanoparticles with preferred facets (111, 100 etc.), dimensions (1D, 2D, 3D such as nano-sheets, nano-wires, nano-tubes, nano-cages, nano-boxes, nano-spheres etc.), and compositions (mono- or multi-metallic). All these possible structures provide great opportunities for structure-activity relationship studies which are yet to be optimized for EOR. These possibilities also present a great challenge for there are many to be optimized. In addition to these possibilities, there is great consensus in literature that supports have a significant influence on the catalytic properties of the nanoparticles. Of particular interest are the metal-oxide supports such as SnO₂, CeO₂, NiO/foams, and TiO₂. In terms of selectivity towards cleaving the C–C bond, CeO₂ stands out as the best candidate support. On the other hand SnO₂, although it significantly improves the reaction kinetics, the selectivity for breaking the C–C bond is very poor. However, it presents the opportunity of considering running the EOR in DEFCs with the benefit of getting value added chemicals instead of CO₂. In designing the catalyst structures, it seems the efforts should be focused on 3D and/or mesoporous materials. This is because, from computational insights, ethanol is reported to preferentially adsorb via the oxygen lone pair of electrons, which allows the activation of the alpha-carbon (α-C). Therefore, it would be interesting to confine ethanol molecule in thin-cavity catalysts, which would allow the beta-carbon (β-C) to be activated too.

Five prototype DEFCs have been proposed and tested in literature to date. They are proton-exchange membrane fuel cells, anion-exchange membrane fuel cells, alkaline-anode acid-cathodes fuel cells, direct alkaline fuel cells without membranes, and solid oxide fuel cells. Each design has its own advantages and disadvantages which mean that there should be simultaneous development and optimization hence presents a great challenge. Nevertheless, these provide the opportunity to start identifying for what application each design is best suited for and the potential returns. Regardless of which prototype design is the best, the various components of the fuel cells still needs combined efforts of both scientists and engineers. These components include the development of membranes, ionomers, catalysts (anode and cathode), water transport and ethanol transport management, carbon dioxide management, and electrolyte formulations that favour the reaction selectivity towards complete oxidation. For the electrolyte formulations, there is need to use buffer solutions to mitigate the effect of the changing pH at electrolyte/electrode interfaces. Changing the pH when working, especially in alkaline conditions, alters the reaction kinetics, as reported in the literature. The other option would be to consider the use of ionic liquids with ions, which would not block the catalyst active sites for ethanol oxidation. Overall, the development of DEFCs looks promising in the near future.