"USE OF SEMICONDUCTORS TO CONTROL THE SELECTIVITY OF ELETROCHEMICAL REDUCTION OF CARBON DIOXIDE"

Technical field

The present invention reports the use of at least one semiconductor to control the selectivity of electrochemical reduction of CO ₂ at high current densities in gas- and liquid-phase electrolysers for electrochemical reduction of carbon dioxide (ERCO ₂ ) to organic molecules (e.g. CH ₄ , CH ₃ OH, HCOOH, etc.) and CO.

Particularly, the present invention addresses the mitigation of the competing hydrogen evolution reaction at the cathode side by delivering electrons at a higher energy level. This is achieved by tailoring the energy level needed for the electroreduction of CO ₂ , at the catalyst surface, with protons transported from the anode to the cathode; the disclosed invention hinders/mitigates the proton recombination to molecular hydrogen and increases ERCO ₂ faradaic efficiency.

Background art

The electrochemical reduction of carbon dioxide (ERCO ₂ ) is considered as the most promising strategy for CO ₂ conversion into valuable chemical products (Power-to-X) using the surplus from renewable electricity sources. Compared to other CO ₂ conversion approaches, ERCO ₂ is a process controllable by electrochemical potentials, pH, pressure and temperature. The system is compact, modular, on-demand and easily scalable. The ERCO ₂ to organic molecules is a reaction that requires intricate electron/proton coupling steps leading to low current densities, poor selectivities and large overpotentials. Most of the advances in the last few years were on fundamental and mechanistic studies of ERCO ₂ . Others were on reducing the overpotential of the ERCO ₂ through the development of electrocatalysts. Rd, Pt, Pd-Cu, Ru/Ti bimetallic oxide, Mo complexes and copper-based catalysts, have shown to be promising for ERCO ₂ to produce organic molecules. However, most of them are not realistic for industrial applications due to high cost and poor abundance. Until now, copper-based is the most used electrocatalyst to reduce CO ₂ to chemical feedstock and organic substances due to its good activity to produce value-added chemicals while mitigating the opportunistic hydrogen evolution reaction (HER). Despite its ability to convert CO ₂ electrocatalytically at relatively low overpotentials, several attempts have been made to increase overall efficiency and copper selectivity to high-value products by: a) alloying copper with other metals to shift the binding energy with the reactants, b) modifying copper structure, oxidation state and surface area. However, copper and other commonly used metal-based catalysts still suffer from limitations such as meagre selectivity, short-term stability, deactivation and restructuring during long-term electrolysis .

The use of semiconductors for CO ₂ reduction is very important in photo-electrochemical systems to serve either as photo- electrocatalyst or promoter (alloys and co-catalyst). Thermodynamically, a semiconductor can catalyze ERCO ₂ with water if it possesses the conduction-band more negative than that the ERCO ₂ redox potential and, at the same time, if the valence-band edge is more positive than the redox potential for the oxygen evolution reaction (OER). Thus, the photogenerated electrons are used to reduce CO ₂ and the holes are used to oxidize H2O to O2. Figure 1 shows the band-edge positions of some semiconductors as well as the redox potentials of HER, OER and ERCO ₂ to produce carbon monoxide, methane, methanol, formic acid and formic aldehyde. ¹ Based on which electrode acts as the photoabsorber material, three different photo-electrochemical configurations can be envisioned: i) photocathode and dark anode, ii) photoanode and dark cathode, and iii) photocathode-photoanode. The fact that each photoelectrode can combine multiple absorber layers, for making more efficient use of the solar spectrum, complicates the process. The use of a diode or a solar cell, based on a buried junction concept, covered by one or more electrocatalyst (s) that is contact with the electrolyte acts as a semiconductor, being considered as a promising integrated solution.

Two major types of ERCO ₂ electrolysers have been reported, depending on reaction media: liquid-phase and gas-phase configurations. Regardless of the reaction media, an electrolyser is composed by an ion-exchange membrane between two electrodes in a zero-gap configuration, composing a membrane electrodes assembly (MEA). Porous liquid/gas diffusion layers and bipolar plates are the remaining components. The electrons flow from cathode to anode, through the external circuit while ions are conducted through the membrane. Both types of ERCO ₂ electrolysers can be arranged in a zero-gap cell configuration.

Zero-gap cell configuration has great potential to increase the energy efficiency of the process due to its lower electrical resistances at the interfaces and higher volumetric power density due to the compact design. Moreover, the use of materials already developed for fuel cells and water electrolysers is expected to accelerate ERCO ₂ market adoption. What determines the type of ERCO ₂ phase is how the CO ₂ is fed to the cathode, i.e., in the form of CO ₂ saturated aqueous/non-aqueous catholyte - liquid-phase electrolyser configuration or gaseous CO ₂ stream - gas-phase electrolyser. Gas-gas and liquid-gas are other more thorough configurations that depend on the anode feed; hydrogen/steam or liquid electrolytes are fed, respectively, and they will generate ions to compensate the charge for CO ₂ reduction. Liquid-phase type ERCO ₂ electrolysers are like redox flow batteries with recirculation of liquid electrolytes on both sides. CO ₂ is absorbed into aqueous electrolytes supplied to the cathode side (CC>2-saturated catholyte). Generally, buffer electrolytes such as potassium carbonate (KHCO3) and hydroxide potassium (KOH) are used to maintain the pH at the electrodes. In the anodic chamber, a similar electrolyte is typically pumped to avoid pH unbalance that would require permanent electrolyte replacement. In liquid-phase electrolysers, apart from requiring pumps and periodical electrolyte replacement, the operation at high current densities is hindered by excessive mass diffusion overpotentials; the solubility of CO ₂ in water solutions is limited and the separation of products is also complex. Finally, the alkalis react with the pumped protons at the cathode side requiring a continuous pH correction.

In gas-phase configuration, the electrolyser is fed with gaseous CO ₂ and uses a MEA separating the anodic and cathodic compartments in a similar way to the ones used in proton exchange membrane (REM) fuel cells. The membrane acts as a barrier for electrons and gases involved in the ERCO ₂ . Nevertheless, in gas-phase configuration the use of PEMs generates high proton concentrations at the cathode leading to a highly acidic environment. This promotes the parasitic hydrogen evolution reaction (HER) and low ERCO ₂ selectivity. In gas-phase electroliser configurations, the ERCO ₂ to organic molecules and/or CO requires H ₂ O electrolysis or hydrogen oxidation followed by proton conduction from the anode to the cathode side through a proton exchange membrane (REM). At the cathode, protons and electrons react with CO ₂ at the catalyst surface such as a copper-based catalyst that favors the reaction. ERCO ₂ kinetics is slower than water or hydrogen oxidation using state-of-the-art catalysts. This generates high proton concentrations at the cathode, promoting HER. For this reason and impossibility to maintain a moderate proton concentration, the highest faradaic efficiency reported so far in gas-phase ERCO ₂ using a PEM is 2 % and using an anion exchange membrane (AEM) is 12 % of CO. ²

Other chemistries have been used to achieve similar efficiencies for gas-phase as those of liquid-phase. By feeding aqueous KOH or NaOH solutions to the anode, PEMs may also permeate other cations that allow to decrease proton concentration at the cathode side; however, the ionic conductivity of these cations in PEMs is lower than that of protons, thus generating ohmic losses. Another strategy is replacing PEMs by anion exchange membranes (AEMs). Instead of protons, anions that may be a result of water reduction (OH-) are transported from the cathode to the anode side to be oxidized and generate oxygen, water and electrons. More recently, bipolar membranes (BPM) have also been used since their composition serves itself as a buffer; the alkaline- acidic ionomer junction fundamental behavior can be seen as an ionic analogue of an electronic semiconductor p-n junction: the cation conductive "acidic" ionomer provides positive mobile charge carriers, such as protons; and, the anion conductive "alkaline" ionomer provides negative mobile charge carriers, such as hydroxyl ions.

The faradaic efficiency of ERCO ₂ is affected by several parameters. Regarding the catalyst itself, it is important to consider the chemical composition, microstructure, morphology and particle size. In the case of composite catalysts, the composition, microstructure and morphology of the support, and the concentration and dispersion of the active phase are important; if the active phase of the composite catalysts is made of crystals, their crystallinity and crystal orientation are also critical.All these features should be optimized for maximizing the protons reduction overpotential and to decrease the CO ₂ reduction overpotential .

On the other hand, the faradaic efficiency is also highly influenced by operating conditions. If the cathode is submersed in a liquid solution, the relevant operating conditions are temperature, solvent composition, pH, CO ₂ dissolved concentration and electrode potential. In the case of the cathode being fed with gaseous CO ₂ , the relevant operating conditions are temperature, CO ₂ partial pressure, relative humidity and electrode potential. CO ₂ is a quite stable linear molecule with a strong C=O bond (750 kJ mol ^-1 ). Its electrochemical conversion is difficult and intricate requiring a high activation barrier, that leads to substantial overpotentials. Moreover, the ERCO ₂ involves multi-electron/proton transfer processes, which together with intermediate reactions and the wide portfolio of possible products makes the process highly complex.

The most common electrolyser configuration used for the ERCO ₂ is the liquid-phase configuration (Figure 2). In this type of electrolyser, liquid electrolytes are circulated over both electrodes, separated by a polymer electrolyte membrane; the electrodes may be covered by a gas-liquid diffusion layer. In gas-phase configuration (Figure 3), CO ₂ is supplied in gas phase and the electrodes are also separated by an ion exchange membrane. The electrodes and membrane form a MEA, which is then sandwiched between two gas diffusion layers (GDLs) that serve also as current collectors .

For liquid- and gas-phase electrolyser configurations (Figures 2 and 3), three types of membranes can be considered: proton exchange membrane (PEM), anion exchange membrane (AEM) and bipolar membrane (BPM). PEM can transport protons (H ⁺ ) and other cations from the anode to the cathode side, while prevent anions and other reactants to cross. If using an AEM, hydroxide ions (OH-), which are produced from the water reduction, or bicarbonate ions (HCO3-), from the reaction equilibrium between carbon dioxide and hydroxyl ions, are transported from the cathode to the anode. A BPM combines the chemistry and ion conduction of both PEM and AEM, delivering simultaneously H ⁺ to the cathode and OH- (or HCO3-) to the anode, respectively. ³

In the liquid-phase electrolyser configuration, ion conduction is made across the membrane and the supporting electrolyte. Ionic transport through the membrane occurs by vehicular mechanism where ions are solvated and bonded to groups to the membrane with opposite charge. At the supporting electrolyte, proton and hydroxyl ions transport is made by Grotthuss mechanism in which ions diffuse through hydrogen bonds with water due to concentration gradients; migration and molecular diffusing transport mechanisms apply to all charged species at the support electrolyte. Thus, both membrane and supporting electrolyte are responsible for ohmic losses related to the charge transport, which depends on several factors such as membrane composition and thickness and, electrolyte molarity and composition.

Liquid-phase electrolysers operate with both liquid anolyte and catholyte, where the catholyte contains dissolved CO ₂ . CO ₂ solubility in water and reactivity towards the electroreduction that is highly dependent on the pH. The optimal pH of the catholyte is normally slightly alkaline. The local pH, nearby the surface of the electrode, may differ from the bulk pH as the CO ₂ reduction reaction progresses. This effect can be mitigated by using buffering electrolytes. In aqueous electrolytes, CO ₂ forms the following buffer, especially in non-acidic solutions:

In alkaline electrolytes, it forms the following buffer:

Furthermore, for more alkaline medium, the bicarbonate buffer can produce at the electrode surface CO ₃ ² :-

The local pH nearby the electrode surface depends on different parameters namely, current density, side-products, electrolyte buffering and mass-transport of OH~, CO ₂ , HCO3- and CO ₃ ^2- . ⁴

The mitigation of HER, in liquid-phase electrolysers, is critical to increase the energy efficiency. The local pH, besides affecting the CO ₂ electroreduction reaction kinetics and selectivity, also affects the competing HER. As such, the choice of pH buffers in the catholyte (e.g., potassium/sodium bicarbonate salts) is of utmost importance. On the other hand, the pH also conditions the majority ionic charge carriers in the electrolyser. This charge carrier can be H ⁺ for acid to neutral medium and OH~ for neutral to alkaline medium. Accordingly, PEM or AEM membranes should be considered. The BPM should be used for acidic to neutral conditions .

In liquid-phase electrolysers, strategies such as buffering the catholyte besides using a suitable catalyst and membrane allow to achieve high ERCO ₂ selectivities. Nevertheless, these electrolysers operate at low current densities that hinder the volumetric power density and high production rates needed for industrial applications. Moreover, it requires complex separation units when methanol or formic acid are produced and need to be separated from aqueous electrolyte. These highly valuable products are liquid at normal conditions of pressure and temperature.

ERCO ₂ in a gas-phase electrolyser (Figure 3), includes two electrodes (anode and cathode), separated by an ion-exchange membrane preferably in a zero-gap configuration. The cathode is typically fed with humidified CO ₂ , essential for the electrode and the membrane humidification. The electrode humidification allows that a thin water film coats the electrocatalyst particles allowing the mobility of protons, dissolved carbon dioxide and other ionic species, some of them intermediates in CO ₂ electroreduction. The membrane humidification is needed for improving the ion conductivity. ⁵ At the anode side, different feed streams are used to supply electrons and H2 to reduce CO ₂ at the cathode side. Water and H2 may be fed to deliver protons and electrons to the cathode when using PEMs and acidic ionomer (equations (5) and (6), respectively) . Hydroxyl ions are the source of electrons in alkaline medium provided by the AEM. In this case, hydroxyl ions are generated by the reduction of CO ₂ with water at the cathode and conducted through the AEM to the anode to be oxidized (equation (9)).

The components of gas-phase electrolysers, such as catalyst, GDLs and bipolar plates must be carefully chosen. The pH character of these electrolysers is given by the ion-exchange membrane. Proton exchange membranes provide an acid character; the hydrogen oxidation at the anode side should be catalysed by a platinum-based catalyst, while iridium- based catalysts are used to oxidize water. In the case of water oxidation, catalyst support, GDL and bipolar plate must be corrosion resistant; typically, titanium is used due to its resistance to corrosion and sufficient electronic conductivity. In alkaline medium (using AEMs) corrosion is not that critical; transition metals (Ni, Fe, Co, etc.) are used to catalyze the reduction of the hydroxyl ions - equation (9). The remaining components, GDLs and bipolar plates, are normally carbon-based.

Gas-phase configuration enables to operate at high current densities due to easier removal of liquid products of the CO ₂ electroreduction (such as methanol and formic acid) and avoids the need of pumps to make the liquid electrolytes to circulate. Until now, the portfolio of products obtained in the gas-phase configuration is much more limited than that using a liquid-phase electrolyser. The highest ever reported faradaic efficiency to produce CO of a gas-phase ERCO ₂ was 2 % when the electrolyser is equipped with a PEM, and 12 % when equipped with an AEM. ² In general, ERCO ₂ can produce a wide portfolio of organic molecules (e.g., CH4, CH3OH, HCOOH) and CO, since they can be formed by reducing CO ₂ with H ⁺ (e.g., equation (4)) or with H2O (e.g., equation (7)). The species that are formed on the electrodes surface, in the gas-phase electrolyser configuration, depend on several operating and design parameters such as temperature, relative humidity, electrode materials and pH. The flux of protons (or other ions) impact on the pH at the electrode surface and on the reactions that are taking place.

In the case of acidic environment, when CO is produced from the ERCO ₂ the following reactions occur: ³ ' ⁶

In the case of alkaline environment, when CO is produced from the ERCO ₂ , the following reactions occur: ³ ' ⁶

The half-electrochemical reactions and electrode potentials for the evolution of CH4, CH3OH and HCOOH by ERCO ₂ are listed below : ⁷

Thermodynamic reduction potential of CO ₂ to relevant products is close to zero V versus standard hydrogen electrode (V vs SHE) and then close to the HER reduction potential, in alkaline or acidic media - cf. equations (10), (12) and (14). However, the overpotentials of the CO ₂ electroreduction are considerably high. For example, a commonly intermediate species in the electroreduction of CO ₂ is free radical CO ₂ ~ *, which is formed at -1.9 V vs SHE. The use of electrocatalysts allows to decrease this overpotential, but it stills too high (the absolute value) compared with the parasitic electroreduction of H ⁺ to molecular hydrogen.

Summary

The present invention relates to a method to control the selectivity of the electrochemical reduction of carbon dioxide to organic molecules and CO in gas- or liquid-phase in an electrolyser, comprising the use of at least one semiconductor material on the electrode of the cathode or anode side of the electrolyser that is suitable to increase the faradaic efficiency of the reaction.

In one embodiment at least one semiconductor material acts as an electrocatalyst.

In one embodiment at least one semiconductor material further comprises particles of metals selected from Cu, Sn, Ag, Ru, Bi, Pt and Pd. In one embodiment the semiconductor material is selected from binary transition metal-oxides cP such as TiO ₂ , ZnO and

WO ₃ ; materials based on transition metal cations with d ⁿ electronic configurations such as CU ₂ O, Fe ₂ O ₃ , Co ₃ O ₄ and NiO; ternary metal-oxides such as BiVO ₄ , CuWO ₄ , CuFeCO ₂ , CaFeO ₂ ,

ZnFe ₂ O ₄ and BaSnO ₃ ; transition metal hydroxides such as

Mg (OH) ₂ and Ni (OH) ₂ ) ; layered double hydroxides such as NiFe-

LDH, CuCr-LDH, ZnTi-LDH; the family of (oxy) nitritrides such as TaON and Ta ₃ N ₅ ; the family of metal chalcogenide such as

Cu (In, Ga) Se ₂ , CdS, MoS ₂ , WSe ₂ ; carbon-based semiconductors such as g-C ₃ N ₄ ; silicon-based semiconductors; and III-V class of semiconductors such as GaP, GaInP ₂ , InP.

In one embodiment at least one semiconductor material is in the form of a thin film or nanoparticulated structures.

In one embodiment the semiconductor material is in the form of a porous layer of nanoparticulated structures with a thickness between 1 μm and 50 pm.

In one embodiment the semiconductor material in nanoparticulated structures are spherical nanoparticles with a size distribution between 2 nm and 1.5 μm.

In one embodiment the semiconductor material is in the form of a thin film with a thickness between 2 nm and 200 nm.

In one embodiment at least one semiconductor material is in at least one diode mounted in series or parallel configuration in the external electrical circuit of the electrolyser . In one embodiment at least one diode is selected from a 0.4 V diode or a 0.7 V diode.

In one embodiment the membranes used in the electrolyser are selected from cation exchange membranes, anion exchange membranes, proton exchange membrane, or bipolar membranes.

In one embodiment the reaction temperature is between 0 °C and 100 °C.

In one embodiment the reaction pressure is between 1x10 ⁵ Pa and 1x10 ⁷ Pa.

The present invention also relates to the use of a semiconductor material suitable to increase the faradaic efficiency on a method for the electrochemical reduction of carbon dioxide as disclosed in any of the preceding claims.

General description

The principles of the present innovation are focused on the use of semiconductors to control the selectivity of electrochemical reduction of CO ₂ to organic molecules at high current densities in gas- and liquid-phase electrolysers.

Brief description of drawings

Understanding the main advantages of the present invention may be achieved by those skilled in the art by reference to the accompanying figures. The following figures provide preferred embodiments for illustrating the description and should not be seen as limiting the scope of present disclosure . Figure 1 shows the band-edge positions of some typical semiconductor photocatalysts relative to the energy levels of the redox couples involved in the reduction of CO ₂ ; ¹

Figure 2 shows the schematic representation of the ERCO ₂ electrolyser configuration in accordance with an embodiment of the present invention;

Reference numbers:

(1) - Stands for the ERCO ₂ electrolyser containing the semiconductor;

(2) - Stands for the Pumps;

(3) - Stands for the Reservoirs;

(4) - Stands for the Power supply.

(5) - Stands for the Diode.

Figure 3 shows the electrolyser configuration considering the use of semiconductor nanoparticles as catalyst support in accordance with an embodiment of the present invention; Reference numbers:

(1) - Stands for the cathode bipolar plate;

(2) - Stands for the cathode current collector/gas diffusion layer;

(12) - Stands for the cathode catalyst layer with semiconductor nanoparticles as catalyst support;

(4) - Stands for the ion exchange membrane;

(5) - Stands for the anode catalyst layer;

(6) - Stands for the anode current collector/gas diffusion layer;

(7) - Stands for the anode bipolar plate.

Figure 4 shows the electrolyser configuration considering the semiconductor film coated on the current collector/gas diffusion layer in accordance with an embodiment of the present invention;

Reference numbers:

(I) - Stands for the cathode bipolar plate;

(10) - Stands for the cathode current collector/gas diffusion layer coated with a semiconductor film;

(II) - Stands for the cathode catalyst layer;

(4) - Stands the ion exchange membrane;

(5) - Stands for the anode catalyst layer;

(6) - Stands for the anode current collector/gas diffusion layer;

(7) - Stands for the anode bipolar plate.

Figure 5 shows the electrolyser configuration considering the semiconductor film coated on the catalyst layer in accordance with an embodiment of the present invention; Reference numbers:

(1) - Stands for the cathode bipolar plate;

(2) - Stands for the cathode current collector/gas diffusion layer;

(3) - Stands the cathode catalyst layer coated with a semiconductor film;

(4) - Stands for the ion exchange membrane;

(5) - Stands for the anode catalyst layer;

(6) - Stands for the anode current collector/gas diffusion layer;

(7) - Stands for the anode bipolar plate.

Figures 6 and 7 show the production rates of methane and methanol as a function relative humidity and cell potential, respectively. The anode, loaded with a Pt/C electrocatalyst, was fed with hydrogen and loaded with a Pt/C electrode. The cathode, composed by Ru/C electrocatalyst, was fed with carbon dioxide. The cell was operated at 80 °C and 10 ⁵ Pa.

Figure 8 shows the faradaic efficiencies and current density at -0.8 V of cell potential. A copper foam with and without a 5 nm TiCt film was used as cathode, fed with gaseous CO ₂ . The anode was composed by Pt/C electrode and fed with hydrogen. Nafion 117 was used as membrane. The cell was operated at 80 °C and 10 ⁵ Pa.

Figure 9 shows the faradaic efficiencies and current density of CuZnGa-based gas diffusion electrodes tested in aqueous- phase. Dependences on sequential post-treatments namely untreated (CZG_U), calcinated (CZG_C) and calcinated reduced (CZG_CR) are displayed. The cell was operated at room conditions and -1.8 V vs Ag/AgCl.

Figure 10 shows the faradaic efficiencies and current density of CuZnGa-based gas diffusion electrodes tested in aqueous- phase. Dependences on in-situ activation with potential cycling, namely untreated activated (CZG_UA), calcinated activated (CZG_CA) and calcinated reduced activated (CZG_CRA) are displayed. The cell was operated at room conditions and -1.8 V vs Ag/AgCl.

Detailed description of the embodiments

Before any embodiments of the present invention are detailed, it should be comprehended that the embodiments may not be limited in application per the details nor of the structure or the function as set forth in the following descriptions or illustrated within the figures. Different embodiments may be capable of being practiced or conducted in different ways. Moreover, it is to be understood that the terminology used herein is for the purpose of description and should not be regarded as limiting. The use of terms such as "including", "comprising, " or "having" and variations thereof herein are meant to encompass the item listed thereafter and equivalents thereof as well as additional items. Further, unless otherwise noted, technical terms may be used according to conventional usage. It is further contemplated that the reference numbers describe similar components and the equivalents thereof.

The objective of the present invention is using at least one semiconductor material to increase the faradaic efficiency of the electrochemical reduction of carbon dioxide (ERCO ₂ ) in gas- and liquid-phase electrolysers as well as mitigate the evolution of hydrogen in the reaction. These semiconductor materials can be used in the form of thin films, nanoparticulated structures or in an independent device like a diode. Considering the bandgap and the position of conduction and valence bands, the energy of electrons that will reduce the CO ₂ molecule can be controlled to promote a reaction mechanism that could be more difficult and produce desired products. Additionally, impedance ascribed to the semiconductor limits protonic conduction through the membrane that can generate a highly acidic environment at the cathode; hydrogen evolution is hindered when there is a shortage of protons.

The most typical ERCO ₂ electrolyser, shown in Figure 2, is similar to a redox flow battery where two liquids are recirculated both at the negative and positive electrode of an electrochemical cell or stack. Typically, peristaltic pumps provide energy to the liquids which are water at the negative electrode and an aqueous electrolyte containing absorbed CO ₂ . External electric power needs to be ensured to drive the process using a typical power supply. According to Figures 3, 4 and 5 the electrolyser is composed by two bipolar plates (1, 7) that are compressed with bolts. They contain flowfields provide a better liquid distribution through the complete active area of the diffusion layers (2, 6). Both components typically are carbon based or made of metals to withstand corrosion when using acidic liquids or high potentials. In the center, the MEA is composed by an ion-exchange membrane (4) and two catalyst layers (12, 3, 11, 5). At the negative electrode, iridium is typically the choice to oxidize water in acidic media. In alkaline conditions, mostly Ni-based electrodes are used to oxidize hydroxyl anions. Whereas for the positive electrode, a wide portfolio of metals can be used to reduce CO ₂ to produce valuable products. Copper is widely used metal to produce methanol and other light alcohols.

To solve the existing technical problems, it is proposed a method to control the selectivity of the ERCO ₂ to organic molecules and CO in gas- and liquid-phase electrolysers, comprising the use of at least one semiconductor material on the electrode of the cathode or anode side, the semiconductor material being suitable to increase the faradaic efficiency of the reaction. The semiconductor can be present in the form or electrocatalyst support (Figure 3), in a thin film deposited on the gas diffusion layer (Figure 4) and in thin film deposited directly over the catalyst layer.

The selectivity towards a given product is highly affected by the electrode material and the cell conditions, rather than the product itself. Therefore, the introduction of a semiconductor layer combined with the electrocatalyst, whose alignment of the band-edges is suitable for the reduction of CO ₂ , will better tune the energy of the electrons that are conducted from the anode to the cathode active sites. This arrangement favors the correct energy to generate a certain product while mitigating the parasitic reactions.

To ensure proper functionality from the semiconductor it just needs to be connected in series in the electrical circuit of the system. But its function will be dependent on the composition of it as a component (multiple semiconductors can be used at the same time) and its morphology such as thin-film or nanoparticles, and crystalline structure such as nanospheres, nanotubes or nanorods.

Once the electron possesses the correct energy and reached the electrocatalyst, the function of the semiconductor is done despite of it generating some additional ohmic losses. The semiconductor morphology and crystalline structure effects are also important since they affect charge carriers transport and transfer processes. The semiconductor can be deposited as a thin and planar film or as a nanoparticulated structures. Thin films are advantageous for carrier collection, but nanostructured morphologies with controlled hickness, porous layers with 1 and 50 μm of thickness, play a crucial role in improving the surface area to facilitate electron transport for higher performances.

Nanoparticles of semiconductors can be also used as electrocatalyst support to stabilize small particles of active metals, the semiconductor further comprising a metal selected from Cu, Sn, Ag, Ru, Bi, Pt and Pd, while serving as the function of controlling the energy of electrons that reach the cathode.

The use of at least one diode is another way to include and benefit from using semiconductors with an independent component that does not interfere with the structure of the cell or stack.

In one embodiment, the semiconductor material is used in the form of a thin film, nanoparticulated structures or in at least one diode in the external electrical circuit of the electrolyser.

In one embodiment, the semiconductor thin film has a thickness between 2 and 200 nm.

In another embodiment, the semiconductor material is selected from the classic binary transition metal-oxides (cP oxides), such as TiO ₂ , ZnO and WO ₃ to the ones based on transition metal cations with cP electronic configurations, such as CU ₂ O, Fe ₂ O ₃ , Co ₃ O ₄ and NiO. Ternary metal-oxides are also promising due to the rationally engineer of the bandgap through cation combination (e.g. Sn ²⁺ , Bi ³⁺ , V ⁵⁺ , Ti ⁴⁺ , Nb ⁵⁺ , etc.); BiVCy, CUWO4, CuFeCO ₂ , CaFeO ₂ , ZnFe ₂ O ₄ and BaSnO ₃ are good examples. Transition metal hydroxides (such as Mg (OH)2 and Ni(OH)2) or layered double hydroxides (such as NiFe-LDH, CuCr-LDH, ZnTi-LDH, etc.) can be suitable. In addition to tuning the bandgap of oxides using metal-cation orbitals, substituting a nitrogen atom for oxygen is also a possible strategy; therefore, the family of (oxy)nitritrides, such as TaON and Ta ₃ N5 are potential semiconductor materials. The family of metal chalcogenide (e.g. Cu (In,Ga)Se ₂ , CdS, MoS ₂ , WSe2), carbon-based semiconductors such as g-C ₃ N ₄ , silicon- based semiconductors and III-V class of semiconductors (such as GaP, GaInP2, InP) also offer tunability and exceptional charge-transport characteristics for CO ₂ reduction.

In one embodiment, the semiconductor material is in the form of spherical nanoparticles with size distribution between 2 nm and 1.5 pm, more preferably between 30 and 200 nm.

With this method it is possible to perform the ERCO ₂ to organic molecules and CO in gas- and liquid-phase at temperatures between 0 °C and 100 °C, and pressures between 1x10 ⁵ Pa and 1x10 ⁷ Pa, if water is in the reaction remains in liquid state.

In one embodiment the membranes can be selected from the membranes used in the electrolyser are selected from cation exchange membranes, anion exchange membranes, proton exchange membrane, or bipolar membranes.

The expected impact from this method using at least one semiconductor is considerably higher in gas-phase than in liquid phase.

Supplying gaseous CO ₂ to the electrolyser brings advantages like higher power densities and more compact systems but it poses difficulties with kinetics. Since the pH nearby the electrocatalyst cannot be easily controlled like in aqueous systems, the use of semiconductors is a breakthrough to overcome the easier hydrogen evolution when compared to more demanding reaction mechanisms. The depletion of protonic concentration at the cathode deeply impacts on the magnitude of hydrogen evolution and thus the faradaic efficiency of the electrolyser.

In Figure 1, the band-edge positions of some semiconductors as well as the redox potentials of HER, OER and ERCO ₂ for the production of carbon monoxide, methane, methanol, formic acid and formic aldehyde. The ideal semiconductor to catalyze ERCO ₂ should display its conduction-band edge more negative than that the ERCO ₂ redox potential and, at the same time, the valence-band edge should be more positive than the OER redox potential. If so, photogenerated electrons will reduce CO ₂ and holes will oxidize H2O to O2. However, ERCO ₂ also uses an electrocatalyst that can display better performance if the coming electrons have their energy tuned by the semiconductor. In this regard, several other semiconductors and arrangements of semiconductors may result in beneficial.

Experimental results in Figure 7 indicate that applying a small potential difference to run the ERCO ₂ in gas-phase with an electrolyser equipped with a PEM and hydrogen at the anode side, produces nothing but molecular hydrogen. Increasing the potential difference up to 1.8 V does not increase the CO ₂ electroreduction selectivity, which is still zero. However, using thicker membranes or saturating the cation exchange membrane with K ⁺ , makes the electrolyser to display a measurable selectivity towards CH4 and CH3OH at ca. 1.8 V. The potassium exchanged membrane presents a much lower proton conductivity. The cathode electrocatalyst is then lean in protons, loaded with CO ₂ , and receiving high energy of electrons. It is then the balanced presence of CO ₂ , H ⁺ and high energy electrons that allows high selectivities. Such experiments were carried out feeding hydrogen to a gas diffusion electrode loaded with platinum at the anode and another gas diffusion electrode loaded with ruthenium at the cathode. The cell was kept at 80 °C and 1x10 ⁵ Pa under varying relative humidity. Nafion membranes with different thicknesses were used (50 pm, 135 pm and 170 pm).

Another experiment (Figure 8), using a copper foam coated with a thin film deposited by spray pyrolysis of TiCh (thickness of ca. 50 nm). Compared with the same cell without the semiconductor at the cathode, ERCO ₂ faradaic efficiency increased from 0 % to 70 % (light alcohols). The semiconductor caused a small drop in the current density at -0.8 V of cell potential, but hydrogen evolution was unprecedently mitigated in a gas-phase electrolyser. Nafion 117 was used as membrane without ionic exchange.

In liquid-phase electrolysers equipped with PEM membranes, the balance between these species is reached since the catholyte is normally slightly alkaline, which allows to control the catalyst concentration on H ⁺ despite the proton flowrate to the cathode; this imposed by the required high potential differences to have available electrons with sufficient energy to allow the electroreduction of CO ₂ . This approach, however, consumes alkaline reactants. A semiconductor was used in an electrocatalyst composed by copper, zinc and gallium. The catalyst was loaded in a gas diffusion electrode and assembled in a cathode of an electrolyser, as shown in Figure 2. The main difference from the catalysts used in Figures 9 and 10 is assigned to posttreatments after the synthesis. These treatments included calcination, reduction, and in-situ potential cycling. Without any of the treatments, the catalyst does not display sufficient stability and there is no interaction between the three metals to enhance ERCO ₂ kinetics. After calcination, the catalyst is fully oxidized and the most important metal, copper, has very little ERCO ₂ activity in its most oxidized form. Upon reduction with hydrogen the catalyst surface is reduced, and metallic sub-oxides (semiconductors) become dominant. The faradaic efficiency is greatly increased since hydrogen evolution occurs with lower extent. Additionally, in-situ potential cycling stabilizes the metallic sub-oxides and promotes reversible re-dox cycles that activate and generate more semiconductor sites deeper in the microstructure. As such, hydrogen evolution drops to less than 20 % of the faradaic efficiency while generating higher current density. In conclusion, the use of semiconductors not only is able to increase ERCO ₂ efficiency but also increase the power density of the electrolyser.

Copper-based catalysts, proved to be of interest for ERCO ₂ to CO, hydrocarbons and alcohols. On the other hand, it has been used as a photoelectrocatalyst in its oxidized form of CU ₂ O, a non-toxic p-type metal-oxide semiconductor with a high Hall mobility (90 cm ² V ^-1 ) and a direct bandgap of 1.90 - 2.17 eV. For photoelectrochemical applications, the CU ₂ O photocathode reached almost 90% of its theoretical photocurrent density limit (14.7 mA cm ^-1 ). Moreover, it delivered a photovoltage of 1 V and stability beyond 100 h using a nanostructured CU ₂ O absorber layer covered with a Ga ₂ O ₃ intermediate layer, a TiO ₂ protection layer and a RuOx HER catalyst. ⁸

This patent application discloses a method strategy for controlling the ion flowrate through the ion-exchange membrane and the electron energy in an independent and efficient way. The energy of the electrons reaching the cathode can be also controlled, independently of the flowrate of ions through the membrane. This is achieved by introducing a diode or combination of diodes mounted in series or a parallel configuration in the external the electrical circuit of the electrolyser .

The use of a semiconductor material in at least one diode with a given bandgap (potential) in the external circuit makes that only electrons above this bandgap threshold reach the cathode. The flowrate of protons - electrolyser equipped with PEM - or of hydroxyl ions - electrolyser equipped with AEM - can then be controlled increasing the potential difference above the diode bandgap. This fine tune allows to increase the electrolyser current density but also affects the selectivity to CO ₂ electroreduction products and products flowrate. The use of a diode, with a selected bandgap, allows to maximize productivity and selectivity in an independent way. In one embodiment, at least one diode is selected from a 0.4 V diode or a 0.7 V diode. Other diodes can also be considered for the present invention.

The use of at least one semiconductor material at the electrode of the cathode or anode side, preferably in the cathode side, has the same effect of the diode described above.

In one embodiment, the semiconductor material is deposited over the current collector. This semiconductor material supplies electrons to the electrocatalyst deposited over it with a minimum energy level. The use of CU ₂ O as the semiconductor material allows that only electrons with a potential of ca. -1.20 V versus reversible hydrogen electrode (V vs RHE) or above (in absolute terms) are supplied to the electrocatalyst. This semiconductor may, itself, be an electrocatalyst. The advantage of this approach is when photoelectrochemical cells are used and the photoelectrode is a p-type semiconductor, i.e. a photocathode.

The electron potential at the cathode side catalyst can also be regulated increasing the resistance to the ion transport through the ion exchange membrane. As the ion exchange membrane resistance increases, for a given ionic transport rate, the electron is delivered with more energy. The membrane resistivity can be tuned, tuning the membrane thickness and/or the concentration of the membrane ionic functional groups.

The above strategies can be used separately or combined. For optimizing the selectivity and the current density to a given product of the CO ₂ electroreduction, the following critical variables must be optimized: a) electrocatalyst, which should minimize the overpotential and maximize the exchange current density to the target product and increase the overpotentials to other products and to the HER; b) the energy of the delivered electron; c) the membrane type, by choosing for example REM, AEM, BPM; d) the pH at the catalyst surface (between 3 and 8) and; e) the temperature (0 - 100 °C). For gas-phase electrolysers, the CO ₂ partial pressure must also be optimized.

References

(1) Xie, S.; Zhang, Q.; Liu, G.; Wang, Y. Photocatalytic and Photoelectrocatalytic Reduction of CO 2 Using Heterogeneous Catalysts with Controlled Nanostructures. Chem. Commun . 2016, 52 35-59. https://doi.org/10.1039/C5CC07613G.

(2) Patru, A.; Binninger, T.; Pribyl, B.; Schmidt, T. J.

Design Principles of Bipolar Electrochemical Co-Electrolysis

Cells for Efficient Reduction of Carbon Dioxide from Gas

Phase at Low Temperature. J. Electrochem. Soc. 2019, 166

(2), F34-F43. https://doi.Org/10.1149/2.1221816jes.

(3) Delacourt, C.; Ridgway, P. L.; Kerr, J. B.; Newman, J.

Design of an Electrochemical Cell Making Syngas ( CO + H2 ) from C02 and H20 Reduction at Room Temperature. J. Electrochem. Soc. 2007, 155 (1), B42. https://doi .org/10.1149/1.2801871.

(4) Pupo, M. M. de S.; Kortlever, R. Electrolyte Effects on the Electrochemical Reduction of C02. ChemPhysChem 2019, 20 (22), 2926-2935. https://doi.org/10.1002/cphc.201900680.

(5) BINNINGER, T.; PATRU, A.; PRIBYL, B.; Schmidt, T. J.

Co-Electrolysis Cell Design for Efficient Co2 Reduction from Gas Phase at Low Temperature. EP3434810A1, January 30, 2019.

(6) Larrazabal, G. 0.; Martin, A. J.; Perez-Ramirez, J. Building Blocks for High Performance in Electrocatalytic C02 Reduction: Materials, Optimization Strategies, and Device Engineering. J. Phys. Chem. Lett. 2017, 8 (16), 3933-3944. https://doi.org/10.1021/acs.jpclett.7b01380.

(7) Bard, A. J.; Parsons, R.; Jordan, J. Standard Potentials in Aqueous Solution; CRC Press, 1985.

(8) Pan, L.; Kim, J. H.; Mayer, M. T.; Son, M.-K.; Ummadisingu, A.; Lee, J. S.; Hagfeldt, A.; Luo, J.; Gratzel, M. Boosting the Performance of Cu 2 0 Photocathodes for Unassisted Solar Water Splitting Devices. Nat. Catal. 2018, 1 (6), 412-420. https://doi.org/10.1038/s41929-018-0077-6.

This description is of course not in any way restricted to the forms of implementation presented herein and any person with an average knowledge of the area can provide many possibilities for modification thereof without departing from the general idea as defined by the claims. The preferred forms of implementation described above can obviously be combined with each other. The following claims further define the preferred forms of implementation.