ELECTROREDUCTION OF CARBON DIOXIDE ON TRANSITION METAL OXIDE CATALYSTS

Abstract

Provided is a method for the electrolytic reduction of CO2 that comprises providing an electrolytic cell comprising at least one reaction chamber comprising at least one anode and at least one cathode placing at least one electrolyte solution between at least one anode and at least one cathode, wherein the at least one cathode comprises at least one catalyst surface comprising at least one transition metal oxide, providing CO2 in the electrolyte solution; and applying electrical potential to the electrolytic cell, so that CO2 undergoes at least one reduction reaction at the cathode to provide at least one product selected from the group consisting of methanol, methane, mcthanediol and formic acid. Also provided is an electrochemical dev ice for electrochemical reduction of CO2 that has at least one cathode comprising a transition metal oxide.

Description

FIELD

The invention is within the field of process chemistry, and specifically relating to the reduction of carbon dioxide with electrolytic methods, and new transition metal oxide catalysts therefor.

INTRODUCTION

Converting fossil resources into fuels and chemicals is an essential process in modern life. Presently more than 80% of global energy needs are met from fossil fuels and their daily-increasing use has dire consequences on global energy requirements. Over the past few decades, balance in nature has been lost because of the dramatic increase in CO₂ emission, mainly from the burning of carbonaceous fuels, resulting in global warming. In order to lower the CO₂ content of the atmosphere much effort on many levels have been made. One of the most important measures in this regard is to lower CO₂ emission in the first place and to convert it back to fuel or other chemicals. To pursue this goal, a novel catalyst that operates at ambient conditions and electrochemically reduces CO₂ to high-value-added products is the key technology (Whipple et al., J Phys Chem (2010) 1, 3451-3458). This catalyst could be used in a decentralized device powered by renewable energy sources (such as solar, wind, or geothermal power) consuming CO₂ as a reactant and directly converting it to fuel (Barton Cole et al., J Am Chem Soc (2010) 132, 11539-51). There are, however, a number of challenges that need to be overcome in order to develop such catalyst that is active, energy efficient and selective towards a particular product such as formic acid, methane or methanol (Ebbesen et al., J Power Sources (2009) 193, 349-58; Fu et al., Energy Environ Sci (2010) 3, 1382).

From many experimental works on the electrochemical CO₂ reduction reaction (CO₂RR) conducted on different metal catalysts, Cu has been confirmed to be the best pure metal catalyst for producing hydrocarbons with reasonable current efficiency, whereas other metal electrodes mainly form formate, CO or H₂ gas (Hori, Mod Asp Electrochem (2008) 42, 89-189; Kuhl et al., J Am Chem Soc (2014) 136, 14107-13). Cu is, however, not efficient enough for commercial applications since it requires a large overpotential. It also produces 15 different carbon-containing products as well as H2 gas and the separation of this wide range of products is a costly process. At around −1 V, Cu electrode produces methane (40%), methanol (0.1%), ethylene (25%), ethanol (10%) and propanol (4%). This shows that the Cu electrode is more efficient towards producing hydrocarbons than alcohols, and only insignificant amount of methanol is produced, which would be the most attractive product as transportation fuel.

A number of theoretical studies have been carried out over the last few years to model the CO₂RR using density functional theory (DFT) calculations. This is a quantum mechanical methodology that accounts for the interactions between electrons and nuclei; the input is the atomic structure of the system of interest, and the output the ground state energy of the system. Insight into the mechanism or the reaction pathway for reducing CO₂ to CH₄ on the stepped Cu(211) surface was first proposed by Peterson et al (Energy Environ Sci (2010) 3, 1311). There, the most important aspects of the overall system were calculated explicitly, or the free energy of adsorbed intermediates on the catalyst surface, also referred to as the thermochemical model (TCM). The TCM approach has been successfully applied on a number of electrochemical systems; including the water oxidation reaction on transition metal oxides (TMOs), the N₂ electroreduction reaction on the surfaces of transition metals, transition metal nitrides, as well as transition metal oxides, hydrogen evolution reaction (HER) on transition metal nitrides and CO₂RR on transition metals and TMOs. The effect of the applied potential can be included implicitly using the computational hydrogen electrode (CHE; Nøorskov et al. J Phys Chem B (2004) 108, 17886-92). However, other parts of the electrochemical environment (solvent effects, pH dependency, etc.) are usually not taken explicitly into account. Energy barriers of proton-electron transfer reactions are not included in the TCM approach but other studies have modeled that with various approaches. Those studies have improved the reaction mechanism and it has been concluded that energy barriers are needed to capture the trends in product distribution seen in the experiments on the pure metals. However, it has been concluded that the TCM-CHE approach is sufficient in obtaining a good estimate of the overpotential needed, both towards various products on Cu and towards methane on various metal electrodes.

Recently, it has been observed that the product distribution is different when TMOs are used as catalysts for CO₂RR than when the pure metal electrodes are used. Experimental works have shown that some particular TMOs, especially RuO₂, and RuO₂ in combination with other transition metal oxides such as IrO₂, are active for CO₂RR and more selective towards methanol formation than any of the pure metals tested so far. Current efficiency towards methanol formation has been measured to be between 2-76% depending on the catalyst composition/structure and reaction conditions such as electrolyte type, pH and applied potential (Bandi, J Electrochem Soc (1990), 137, 2157; Bandi & Kühne, J Electrochem Soc (1992), 139, 16045; Popia et al, (1997) 421, 105-10; Spataru et al, J Appl Electrochem (2003) 1205-10; Qu et al, Electrochim Acta (2005) 50, 3576-80). However, a mixture of other carbon-containing compounds is also formed, such as formic acid and methane, as well as hydrogen gas.

Karamad et al. (ACS Catal (2015) 5, 4075-81) used DFT calculations to get insight into the mechanism and reaction pathway for CO₂RR on RuO₂ using the TCM-CHE approach. Following that study, Bhowmik et al. (Chem Sus Chem (2016) 9, 3230-43) used the same methodology to study the effect of different TMO overlayers on the RuO₂ (110) surface. In both of these studies on the TMO surfaces, it is concluded that reaction intermediates of the carbon-containing species bind through the oxygen atom(s). This is in contrast to the intermediates on the metal surfaces where it is concluded that most intermediates bind through the carbon atom. This might be a justifying reason why metal oxide catalysts are more selective towards methanol, while pure metals are more selective towards methane and ethylene since there the oxygen atoms are reduced to water molecules. Therefore, metal oxides might open up new avenues for efficient liquid fuel production from CO₂.

SUMMARY

It has proven challenging to find good electrocatalysts for the reduction of CO₂. The present invention is based on the discovery that certain transition metal oxides are useful catalysts for CO₂ reduction.

The present inventors have found that certain transition metal oxide catalysts may be employed in the electrochemical reduction of carbon dioxide. This has lead to the present invention, that makes possible to produce high-value products from carbon dioxide under various conditions, including at ambient room temperature and atmospheric pressure.

In a first aspect, the invention provides a method for the electrolytic reduction of CO₂, the method comprising steps of: (i) providing an electrolytic cell comprising at least one reaction chamber that comprises at least one anode and at least one cathode, wherein the at least one cathode comprises at least one catalyst surface comprising at least one transition metal oxide; (2) placing at least one electrolyte solution between the at least one anode and the at least one cathode, so that the at least one anode and the at least cathode come into contact with the electrolyte solution; (iii) providing CO₂ in the electrolyte solution; and (iv) applying electrical potential to the electrolytic cell; whereby CO₂ undergoes at least one reduction reaction at the cathode to provide at least one product selected from the group consisting of methanol, methane, methanediol and formic acid.

In an embodiment, the transition metal oxide is selected from the group consisting of TiO₂, HfO₂, OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, PtO₂, ReO₂, ZrO₂, VO₂ and MoO₂.

The application of an electric potential that can preferably be a low applied potential such as a potential of less than −1.0 V, preferably less than −0.9 V, preferably less than −0.8 V, preferably less than −0.7 V, more preferably less than −0.6 V, more preferably at a potential of less than −0.5 V. As will be appreciated, the applied potential can also be varied depending on the transition metal oxide being used in the catalyst, and the desired product(s), so as obtain desired reaction products for any given catalyst surface.

Further, the cell can be operated at ambient conditions of temperature and pressure to produce the desired products, using protons provided by the electrolyte solution in the cell. Another aspect of the invention thus relates to process for the catalytic reduction of carbon dioxide, comprising steps of (i) introducing CO₂ to a solution comprising at least one electrolyte in an electrolytic cell so that the CO₂ comes into contact with at least one cathode electrode surface; and (ii) applying a potential to said electrolytic cell, whereby CO₂ reacts with protons to form at least one product selected from methanol, methane, methanediol and formic acid; wherein the cathode electrode surface comprises at least one catalyst surface comprising at least one transition metal oxide. The invention also provides a device for the reduction of CO₂. An aspect of the invention thus relates to an electrochemical device for the reduction of carbon dioxide to at least one reaction product, the device comprising at least one electrochemical cell that comprises an anode and a cathode, wherein the cathode comprises at least one cathode electrode having at least one catalyst surface comprising at least one transition metal oxide.

In the method, process and device according to the invention, the at least one transition metal oxide can be selected from the group consisting of HfO₂, IrO₂, TiO₂, OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, PtO₂, ReO₂, ZrO₂, VO₂ and MoO₂. In certain embodiments, the at least one transition metal oxide is selected from the group consisting of TiO₂, RhO₂, CrO₂, MoO₂, MnO₂, PdO₂ and PtO₂. In yet other embodiments, the at least one transition metal oxide is selected from the group consisting of RhO₂, CrO₂, MnO₂, MoO₂, PdO₂ and PtO₂. The at least one transition metal oxide can also be selected from the group consisting of RhO₂, CrO₂, MnO₂, MoO₂ and PdO₂ or RhO₂, CrO₂, MnO₂, and PdO₂. In yet other embodiments, the at least one transition metal oxide is selected from the group consisting of RhO₂, CrO₂, MoO₂ and MnO₂ or RhO₂, CrO₂, and MnO₂ In one embodiment, the at least one transition metal oxide is RhO₂. In one embodiment, the at least one transition metal oxide is CrO₂. In one embodiment, the at least one transition metal oxide is MnO₂. In one embodiment, the at least one transition metal oxide is MoO₂.

BRIEF DESCRIPTION OF THE FIGURES

The skilled person will understand that the figures, described below, are for illustration purposes only. The figures are not intended to limit the scope of the present teachings in any way.

FIG. 1 shows a model system of a (110) surface of a TMOs in its rutile structure. Large balls indicate metal atoms, small balls indicate carbon atoms and intermediate balls indicate oxygen atoms. CO is spectator and in all cases located on bridge site (except for IrO₂ where we results are shown when CO is on a bridge site or on a cus site) while CO₂RR takes place on other bridge and cus sites.

FIG. 2 shows in a) side view and b) top view of HCOOH on RuO₂ surface. c) Side view and d) top view of HCOO⁻+H⁺ complex on TiO₂ surface. In both cases CO is spectator. According to our simulation only TiO₂ and CrO₂ dissociate HCOOH to the HCOO⁻+H⁺ complex but other surfaces such as IrO₂, NbO₂, MoO₂, OsO₂, HfO₂, PtO₂, RhO₂, MnO₂, PdO₂ can form the undissociated HCOOH species on their surfaces similar to RuO₂.

FIG. 3 shows free energy diagram for formic acid production at 0 vs RHE. The more endergonic step for each TMO indicates the required reducing potential.

FIG. 4 shows theoretical volcano for formation of formic acid from scaling relations (lines). For each of the TMOs the explicit limiting potential values are included (diamond). Each line is for one electron-proton transfer step, indicated by the numbers in parenthesis after each chemical equation to show the sequence of the reaction towards formic acid.

FIG. 5 shows free energy diagram for methanediol production at 0 vs RHE. The most endergonic step for each TMO indicates the required reducing potential.

FIG. 6 shows theoretical volcano for formation of methanediol from scaling relations (lines). For each of the TMOs the explicit limiting potential values are included (diamond). Each line is for one electron-proton transfer step, indicated by the numbers in parenthesis after each chemical equation to show the sequence of the reaction towards methanediol.

FIG. 7 shows theoretical volcano for formation of methanol from scaling relations (lines). For each of the TMOs the explicit limiting potential values are included (diamond). Each line is for one electron-proton transfer step, indicated by the numbers in parenthesis after each chemical equation to show the sequence of the reaction towards methanol.

FIG. 8 shows theoretical volcano for formation of methane from scaling relations (lines). For each of the TMOs the explicit limiting potential values are included (diamond). Each line is for one electron-proton transfer step, indicated by the numbers in parenthesis after each chemical equation to show the sequence of the reaction towards methane.

FIG. 9 shows theoretical activity and selectivity volcano for methane, methanol, methanediol and formic acid production. Solid line on left and upper solid right on right indicate left and right sides of the volcano for formation of formic acid. Dashed line on left and lower solid line on right denote left and right sides of the volcano for formation of methane and methanol, and finally, solid line on left and lower solid line on right are for formation of methanediol. Two dashed lines (one in −0.34 eV and the other in −0.21) also separate the volcano into two main areas: the right area (ΔG_OH>−0.21 eV) that is selective toward formic acid and the left area (ΔG_OH<−0.34 eV) that is selective toward methane, methanol and methanediol. Furthermore, there is an area between −0.34 eV and −0.21 eV which is very active but without any specific selectivity.

FIG. 10 shows free energy diagram for OCHO, H and COOH on the TMO surfaces at 0 V. Negative values for H and COOH for each TMO are showing higher possibilities for getting poisoned by those species. For each metal oxide, the columns indicate free energy for OCHO (left column), H (right column) and COOH (middle column), respectively.

FIG. 11 shows scaling figure for COOH adsorption free energy vs. hydrogen binding free energy. If COOH is formed as a species on the surface, it can lead to CO formation or CO poisoning. A strong H binding energy will lead to H poisoning and H₂ formation whereas a weak H adsorption would prevent proton adsorption until at more negative potentials. Here, we can see an area which indicate weaker COOH and hydrogen binding energies. Based on our analysis, we concluded that if a catalyst is found there, it would be the most promising candidate for CO₂RR, since there is smaller chance of surface poisoning by CO and H.

FIG. 12 shows scaling figure for OH binding free energy vs. hydrogen binding free energy, where the OH is chosen as it is the descriptor on our volcano-graphs, and the catalysts with the lowest onset potentials for CO₂RR are located within the dashed lines. The hydrogen binding free energy can be used as an estimate of hydrogen evolution activity. Comparing hydrogen evolution activity with CO₂RR activity show that CrO₂ is more selective towards CO₂RR than HER and the only candidate located in this active and selective area.

FIGS. 13-20 show scaling relations for CO₂RR network studied in this work as illustrated in the above figures. This network includes OCHO (FIG. 13), HCOOH (FIG. 14), H₂COOH (FIG. 15), O (FIG. 17), CH₂O (FIG. 16), CH₃O+OH (FIG. 19), CH₃O (FIG. 18), OH, O+OH (FIG. 20) as intermediates. In these figures, it has been shown that the adsorption free energy of different intermediates correlate linearly with OH adsorption free energy.

FIGS. 21-33 show free energy diagrams for all the metal oxide catalysts investigated. Different reaction pathways for formic acid, methanediol, methanol, and methane products are also presented in these figures. Potential limiting step is only presented for methanol with highlighted line.

FIG. 34 shows four different surface types that are observed when water is included in the model. Each of the surfaces (a, b, c, d) correspond to the respective four types of surfaces as described in Table 7 herein, from top to bottom (i.e., a) shows the first category that includes RuO₂, NbO₂, MoO₂ and ZrO₂, b) shows the second category that includes IrO₂ and OsO₂, c) shows the third category that includes TiO₂, and d) shows the fourth category that includes PtO₂, RhO₂, CrO₂, VO₂, MnO₂ and PdO₂.)

FIG. 35 shows a theoretical volcano plot for formation of formic acid from scaling relations (lines) in the presence of water. For each of the TMOs the explicit limiting potential values are included (filled squares).

FIG. 36 shows a theoretical volcano for formation of methane and methanol from scaling relations (lines) in the presence of water. For each of the TMOs the explicit limiting potential values are included (filled squares).

FIG. 37 shows a theoretical volcano for formation of formic acid, methane and methanol from scaling relations (lines) in the presence of water. For each of the TMOs the explicit limiting potential values are included (filled squares).

FIG. 38 shows a scaling figure for COOH adsorption free energy vs. hydrogen binding free energy in the presence of water molecules. If COOH is formed as a species on the surface, it can lead to CO formation or CO poisoning. A strong H binding energy will lead to H poisoning and H₂ formation whereas a weak H adsorption would prevent proton adsorption until at more negative potentials. The area that indicates weak COOH and hydrongen binding energies is likely to indlude promising catalysts for CO₂RR.

FIG. 39 shows scaling figure for OH binding free energy vs. hydrogen binding free energy in the presence of water, where OH is chosen as it is the descriptor on the volcano graphs. Comparing hydrogen evolution activity with CO₂RR activity shows that MnO₂, CrO₂ and RhO₂ are more selective towards CO₂RR than HER.

FIG. 40 shows scaling figure for OH binding free energy vs. hydrogen binding free energy for both 25% and 50% coverage of CO as a spectator species, where OH is chosen as it is the descriptor on the volcano graphs. By varying the CO coverage, the activity and selectivity can be tuned and MoO₂ becomes active and selective towards methanol formation when the CO coverage is 50%.

DESCRIPTION

In the following, exemplary embodiments of the invention will be described, referring to the figures. These examples are provided to provide further understanding of the invention, without limiting its scope.

In the following description, a series of steps are described. The skilled person will appreciate that unless required by the context, the order of steps is not critical for the resulting configuration and its effect. Further, it will be apparent to the skilled person that irrespective of the order of steps, the presence or absence of time delay between steps, can be present between some or all of the described steps.

As used herein, including in the claims, singular forms of terms are to be construed as also including the plural form and vice versa, unless the context indicates otherwise. Thus, it should be noted that as used herein, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise.

Throughout the description and claims, the terms “comprise”, “including”, “having”, and “contain” and their variations should be understood as meaning “including but not limited to”, and are not intended to exclude other components.

The present invention also covers the exact terms, features, values and ranges etc. in case these terms, features, values and ranges etc. are used in conjunction with terms such as about, around, generally, substantially, essentially, at least etc. (i.e., “about 3” shall also cover exactly 3 or “substantially constant” shall also cover exactly constant).

The term “at least one” should be understood as meaning “one or more”, and therefore includes both embodiments that include one or multiple components. Furthermore, dependent claims that refer to independent claims that describe features with “at least one” have the same meaning, both when the feature is referred to as “the” and “the at least one”.

The present invention is based on the surprising discovery that on the surface of certain transition metal oxide catalysts, it is possible to reduce carbon dioxide under conditions of low temperature and pressure, including at ambient temperature and pressure, using a low applied potential. Given the surge in atmospheric CO₂ levels and the resulting impact on climate, the invention provides important advances in carbon neutral energy technology. Carbonaceous fuel, such as methanol or methane, can be used in current transportation systems without major investments in infrastructure or new technologies, and can also avoid or minimize use of battery-based energy storage.

The invention therefore provides an important advance in the development of technologies that can reduce CO₂ levels to produce organic feedstock such as combustible fuels.

Thus, the invention provides processes and systems for the electrochemical reduction of CO₂ at ambient temperature and pressure. Reaction products include methane, methanol, methanediol and/or formic acid. However, as will be apparent to the skilled person, the reduction of CO₂ can be geared to produce the desired product or mixture of products, by selecting the appropriate catalyst surface and/or adjusting the applied voltage to electrochemical cell.

In the process and system of the present invention, an electrolytic cell is used that can be any cell from a range of conventional commercially suitable and feasible electrolytic cell designs that can accommodate a special purpose cathode in accordance with the invention. Thus, the cell and system can in certain embodiment have one or more cathode cells and one or more anode cells.

An electrolytic cell in the present context is an electrochemical cell that undergoes a redox reaction when electrical energy is applied to the cell.

The skilled person will appreciate that chemical compounds as described herein are provided by their chemical formula irrespective of their phase or state. In particular, compounds that are present in their gaseous state when present in a pure and isolated form at room temperature (such as CO₂) are herein described by their chemical formula. For example, carbon dioxide is herein described as CO₂, whether present as a gas, as individual molecules, in clusters, bound to surfaces or present as a solute, and the same applies to other molecular species described herein.

Carbon dioxide (CO₂) can be provided by any one of bicarbonate (HCO₃⁻), carbonate (CO₃²⁻) and/or carbonic acid (H₂CO₃). For example, bicarbonate and carbonate can be provided as bicarbonate or carbonate salts, either in pure form or in a mixture into a solution, that can preferably be an aqueous solution. A mixture of any of bicarbonate, carbonate and carbonic acid will reach equilibrium in solution. Thus, as apparent to the skilled person, the relative concentration of these species will depend on pH of an aqueous solution. An alternate source of CO₂ is the gaseous form of the compound, CO_2(g). Gaseous CO₂ can be provided as a sole source of CO₂, or it can be provided as a supplement to other sources of CO₂ in the cell, including the aforementioned bicarbonate, carbonate and carbonic acid.

The proton donor in the electrochemical reactions taking place in the reduction of CO₂ can be any suitable substance that is capable of donating protons in the electrolytic cell. The proton donor can for example be an acid, such as any suitable organic or inorganic acid. The proton donor can be provide in an acidic, neutral or alkaline aqueous solutions. The proton donor can also, or alternatively, be provided by H₂ oxidation at the anode. I.e. hydrogen can be considered as a source of protons: H₂⇔2(H⁺+e⁻).

The electrolytic cell in general comprises at least three general parts or components, a cathode electrode, an anode electrode and an electrolyte.

The electrochemical reduction of carbon dioxide is the conversion of carbon dioxide to more reduced chemical species using electrical energy.

The different parts or components can be provided in separate containers, or they can be provided in a single container. The electrolyte can be an aqueous solution in which ions are dissolved. When provided as an aqueous solution, the aqueous solution can be a neutral, an alkaline or an acidic solution. In some embodiments, the aqueous solution is an acidic solution.

In general terms, the catalyst on the electrode surface should ideally have the following characteristics: It should (a) be chemically stable, it should (b) not become reduced or otherwise consumed during the electrolytic process, it should facilitate the formation of carbon-containing products, and (d) use of the catalyst should lead to the production of minimal amount of hydrogen gas. As will be further described, the catalyst oxides according to the invention fulfill these characteristics.

The catalyst can comprise one or more stabiliser that serves the role of preventing degradation of the catalyst. Suitable stabilisers should be more stable to degradation than the metal oxide(s) being employed, but otherwise are inert with respect to the catalytic reactions taking place on the electrode surface. Exemplary stabilisers include, but is not limited to, metal oxycarbides, metal oxynitrides, bimetallic oxides and the like.

The transition metal oxides can also be provided as a thin layer (e.g., as few layers or as a monolayer) on a stable and conductive surface.

An advantage of the present invention is that the process can be suitably operated using suitable electrolyte solutions. The electrolyte solution can be non-aqueous or aqueous. For example, the electrolyte solution can comprise, or consist of an aqueous solutions containing dissolved electrolytes (salts). Thus, in certain embodiments of the process and system, the electrolytic cell comprises one or more aqueous electrolytic solutions, in one or more cell compartments. Individual cell compartments can be separated by suitable barriers, such as membranes that allow electrolytes to pass through. Aqueous electrolyte solutions may comprise any of various typical inorganic or organic salts such as but limited to soluble salts of e.g. chloride, nitrate, chlorate bromide, etc. e.g. sodium chloride, potassium chloride, calcium chloride, ammonium chloride, and other suitable salts. The aqueous electrolyte solutions may also comprise any one, or a combination of, alkali or alkaline earth metal oxides, such as sodium hydroxide, potassium hydroxide, lithium hydroxide, calcium hydroxide, rubidium hydroxide and cesium hydroxide. The aqueous electrolyte solution can preferably comprise carbonate, bicarbonate or carbonic acid. The aqueous electrolyte solution can also further comprise one or more organic or inorganic acids. Inorganic acids can include mineral acids that include but are not limited to, hydrochloric acid, nitric acid, phosphoric acid, sulphuric acid, boric acid, hydrofluoric acid, hydrobromic acid, and perchloric acid. The electrolyte can alternatively be provided as a protic or aprotic, non-aqueous solution. For example, the electrolyte can be provided as an ionic liquid, i.e. as a molten salt, for example a sodium chloride salt.

As appears from herein, the essential feature of the present invention concerns the composition and structure of the cathode electrode. Transition metal oxides have a wide variety of surface structures which affect the surface energy of these compounds and influence their chemical properties. The relative acidity and basicity of the atoms present on the surface of metal oxides are also affected by the coordination of the metal cation and oxygen anion, which alter the catalytic properties of these compounds.

In certain embodiments of the process, method or device in accordance with the invention, the transition metal oxide catalyst on the cathode electrode surface comprises one or more of the following: ruthenium oxide (RuO₂), hafnium oxide (HfO₂), iridium oxide (IrO₂), titanium oxide (TiO₂), osmium oxide (OsO₂), rhodium oxide (RhO₂), chromium oxide (CrO₂), niobium oxide (NbO₂), manganese oxide (MnO₂), palladium oxide (PdO₂), platinum oxide (PtO₂), rhenium oxide (ReO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂) and molybdenum oxide (MoO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of of hafnium oxide (HfO₂), iridium oxide (IrO₂), osmium oxide (OsO₂), rhodium oxide (RhO₂), chromium oxide (CrO₂), niobium oxide (NbO₂), manganese oxide (MnO₂), palladium oxide (PdO₂), rhenium oxide (ReO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂) and molybdenum oxide (MoO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of of hafnium oxide (HfO₂), titanium oxide (TiO₂), osmium oxide (OsO₂), rhodium oxide (RhO₂), chromium oxide (CrO₂), niobium oxide (NbO₂), manganese oxide (MnO₂), palladium oxide (PdO₂), platinum oxide (PtO₂), rhenium oxide (ReO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂) and molybdenum oxide (MoO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of of hafnium oxide (HfO₂), osmium oxide (OsO₂), rhodium oxide (RhO₂), chromium oxide (CrO₂), niobium oxide (NbO₂), manganese oxide (MnO₂), palladium oxide (PdO₂), rhenium oxide (ReO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂) and molybdenum oxide (MoO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of titanium oxide (TiO₂), hafnium oxide (HfO₂), osmium oxide (OsO₂), rhodium oxide (RhO₂), chromium oxide (CrO₂), manganese oxide (MnO₂), molybdenum oxide (MoO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more hafnium oxide (HfO₂), osmium oxide (OsO₂), rhodium oxide (RhO₂), chromium oxide (CrO₂), manganese oxide (MnO₂), molybdenum oxide (MoO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of rhodium oxide (RhO₂), chromium oxide (CrO₂), manganese oxide (MnO₂), molybdenum oxide (MoO₂), zirconium oxide (ZrO₂), vanadium oxide (VO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of of rhodium oxide (RhO₂), chromium oxide (CrO₂), manganese oxide (MnO₂) and molybdenum oxide (MoO₂).

In one embodiment, the transition metal oxide catalyst comprises one or more of of rhodium oxide (RhO₂), chromium oxide (CrO₂), manganese oxide (MnO₂).

Depending on the substance composition of the catalyst, a suitable surface crystal structure may be preferred. Various different crystal structures exist for transition metal oxides and different structures can be obtained at different growth conditions. It is within scope of the skilled person to select appropriate surface crystal structures.

It may be preferable that the catalyst comprise at least one surface having a rutile structure. Other crystal structures known in the art (e.g., rocksalt structure, zincblende structure, anatase structure, perovskite structure) are also possible (see., e.g., International Tables for Crystallography; http://it.iucr.org).

Several different surface facets may exist for a given crystal structure (polycrystalline surfaces). The (110) facet of rutile exhibits the lowest surface free energy and is therefore in general thermodynamically most stable. Accordingly, in some embodiments, the transition metal oxides can be of rutile structure with a (110) facet providing the catalytic surface. Alternatively, the (100) and/or the (111) facets of the rocksalt structure can be chosen.

Thus, in some embodiments, the catalyst surface is a transition metal rutile surface. The surface can have any suitable facet, including but not limited to the (110) facet. In some embodiments, the surface facet comprises, or consists of, the (110) facet of a transition metal oxide selected from the group consisting of RuO₂, HfO₂, IrO₂, TiO₂,OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, PtO₂, ReO₂, ZrO₂, VO₂ and MoO₂. For example, the surface facet can comprise, or consist of, the (110) facet of a transition metal oxide selected from the group consisting of HfO₂, IrO₂, OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, ReO₂, ZrO₂, VO₂ and MoO₂.

In some preferred embodiments, the catalyst comprises the (110) facet of the rutile structure of one or more oxide selected from RhO₂, CrO₂, MnO₂, PdO₂ and PtO₂.

In some preferred embodiments, the catalyst comprises the (110) facet of the rutile structure of one or more oxide selected from RhO₂, CrO₂, MnO₂, MoO₂, and PdO₂.

In some preferred embodiments, the catalyst comprises a the (110) facet of the rutile structure of one or more oxide selected from RhO₂, CrO₂, MoO₂ and MnO₂.

In a preferred embodiment, the catalyst comprises a the (110) facet of the rutile structure of MnO₂. In another preferred embodiment, the catalyst comprises a the (110) facet of the rutile structure CrO₂. In another preferred embodiment, the catalyst comprises a the (110) facet of the rutile structure of one or more oxide selected from RhO₂. In another preferred embodiment, the catalyst comprises a the (110) facet of the rutile structure of one or more oxide selected from MoO₂.

A rutile metal oxide surface having a (110) facet contains metal atoms of two different coordination environments, where rows of sixfold coordinated metal atoms alternate with rows of fivefold coordinated metal atoms along the [001] direction. Whereas the sixfold coordinated metal atoms have approximately the same geometry as bulk, the fivefold coordinated metal atoms have an unsaturated bond perpendicular to the surface. Thus two distinct surface sites are found, coordinatively unsaturated sites (cus-sites) on top of the 5-fold coordinated metal atoms and bridging sites (br-sites) between two sixfold coordinated metal atoms.

In certain embodiments of the invention, the catalyst surface is provided as a pure transition metal oxide, i.e. the catalyst comprises a single transition metal oxide, i.e. the catalyst for example does not contain a mixture of transition metal oxides or one transition metal oxide that is coated by one or several layers of a second (or more) transition metal oxide.

Thus, as will be apparent to the skilled person, the catalyst according to the invention can comprise a single transition metal oxide. The catalyst can also comprise, or consist of, a mixture of two or more such oxides. Such mixed oxides can comprise a single structure, for example a rutile structure. The mixed metal oxides can also comprise a mixture of oxides that are of different crystal structures and/or oxides with different catalytic facets. Accordingly, such mixed oxides can further comprise a single, or a mixture of, facets. Mixed oxide catalysts can be grown or manufactured separately and then assembled into mixed catalysts comprising the different metal oxides, wherein the oxides in the mixture have the same or different crystal structures.

As described in more detail herein, running a current through the electrolytic cell leads to a chemical reaction in which carbon dioxide (CO₂) is reduced in a series of steps with protons to ultimately form one or more products, including methane, methanol, formic acid and methanediol. The running of current is achieved by applying a voltage to the cell. The invention makes possible electrolytic production of these products at a low electrode potential, which is beneficial in terms of energy efficiency and equipment demands.

The electrolytic cell can be operated at an ambient pressure of about 1 atmospheres. The electrolytic cell can also be operated at higher pressure, i.e. pressure that is greater than ambient pressure. For example, the cell can be operated at a pressure of up to 30 atmospheres, up to 20 atmospheres or up to 10 atmospheres. In some embodiments, the electrolytic cell is operated at a pressure that is in the range of 1 to 30 atmospheres, in the range of 1 to 20 atmospheres, in the range of 1 to 10 atmospheres, in the range of 1 to 5 atmospheres or in the range of 1 to 3 atmospheres. The electrolytic cell can also be operated at a pressure that is in the range of 2 to 20 atmospheres, in the range of 3 to 20 atmospheres, in the range of 4 to 20 atmospheres, or in the range of 5 to 20 atmospheres, such as at about 5 atmospheres, about 6 atmospheres, about 7 atmospheres, about 8 atmospheres, about 9 atmospheres, about 10 atmospheres, about 11 atmospheres, about 12 atmospheres, about 13 atmospheres, about 14 atmospheres, about 15 atmospheres, about 16 atmospheres, about 17 atmospheres, about 18 atmospheres, about 19 atmospheres or about 20 atmospheres.

The electric potential can be applied as a constant or variable electric potential. Pulsed electric fields generated by such pulsed potentials can be varied by adjusting a number of parameters, such as: electric field intensity, rise time of voltage pulses, number of pulses, frequency of pulses, pulse wave shape, treatment time (i.e. the time the pulse is applied, resulting in the product of the number of pulses and the duration of each pulse).

In certain useful embodiments of the invention, product can be formed at an electrode potential at less than about −0.7 V, less than about −0.6 V, less than about 0.5 V, less than about −0.4 V, or less than about −0.3 V. In some embodiments, product formation occurs at electrode potential in the range of about −0.7 V to about 0.0 V, such as in the range of about −0.5 V to about 0.0 V, or in the range of about −0.35 V to about 0.0 V. The upper limit (i.e., more negative potential limit) of the range can be about −0.3 V, about −0.4 V, about −0.5 V, about −0.6 V, or about −0.7 V. The lower limit (i.e. less negative potential limit) of the range can be about 0.0 V, about −0.1 V, about −0.2 V, or about −0.3 V.

The composition of products obtained in the reduction of CO₂ can be altered by selective adjustment of applied potential for any given catalyst surface. Thus, volcano plots show that, depending on the relative binding energies of adsorbed intermediates, the selectivity changes depending on the applied voltage.

For example, in some embodiments, formic acid can specifically be formed at an electrode potential that is in the range of about −0.3 V to about −0.1 V. In some embodiments, methanol, methane and methanediol can selectively be formed using an electrode potential of about −0.4 V to about −0.2 V. In some embodiments, any one or a mixture of formic acid, methanol, methane and methanediol can be formed at an electrode potential of about −0.3 V to about −0.2 V.

The preferred catalyst can be selected from MnO₂, MoO₂, RhO₂ and CrO₂. In certain embodiments, the catalyst comprises CrO₂. In certain other embodiments, the catalyst comprises RhO₂. In certain other embodiments, the catalyst comprises MnO₂. In certain other embodiments, the catalyst comprises MoO₂.

In a preferred embodiment, the catalyst surface is a MnO₂, MoO₂ or CrO₂ surface having a rutile structure, and the applied voltage for selective formation of methanol, methane and/or methanediol is in the range of about −0.4 V to about −0.2 V.

In another preferred embodiment, the catalyst surface is a MnO₂, MoO₂or CrO₂ surface having a rutile structure, and the applied voltage for selective formation of formic acid selectively is in the range of about −0.3 V to about 0.0 V, preferably in the range of about −0.3V to about −0.1V.

In a preferred embodiment, the catalyst surface is a MnO₂, MoO₂ or CrO₂ surface having a rutile structure, and the applied voltage for the generation of any one of formic acid, methanol, methane and/or methanediol can be formed at an electrode potential of about −0.3 V to about −0.2 V.

An advantage of the present invention is the efficiency of product (i.e., methanol, methane, methanediol and/or formic acid) formation over the side-product H₂ formation, which has been a challenge in prior art investigations and trials, due to the competing binding energies of hydrogen over oxygen on the catalyst surface. In certain embodiments, less than about 50% moles H₂ are formed compared to moles product formed, and preferably less than about 40% moles H₂, less than about 30% moles H₂, less than about 20% moles H₂, less than about 10% moles H₂, less than about 5% moles H₂, less than about 2% moles H₂, or less than about 1% moles H₂.

The pathway of CO₂ reduction depends on the relative energies of reaction intermediates. Thus, the pathway can depend on the system within which the reaction takes place, including for example the catalytic surface being used in the reaction.

The active part of an industrial heterogeneous catalyst is most commonly a solid surface, e.g. a metal or metal oxide. The surface offers a low-barrier energy path from reactants to products, by binding reactants and reaction intermediates. The binding energy of reactants to the surface must be strong enough to produce reaction intermediates, but weak enough to allow products to leave the surface, allowing more reactions to take place on the surface. This duality is the basis of the so-called Sabatier principle, which states that for a reaction, there is an optimum binding energy for an intermediate, such that both stronger and weaker binding leads to lower activity. The result is a volcano-shaped relationship, commonly referred to as a Volcano plot.

Catalyst activity can in general be modified by altering the local electronic structure by strain, ligand, substitution and/or alloying. These alterations lead to changes in binding energies of reaction intermediates, and thereby alter the thermodynamics of the overall reaction profile.

The concept of scaling relations is based on the linear relationship between binding energies of adsorbates that bind through the same type of atoms (e.g., C*, CH* and CH₂* binding through a C atom). Based on such relationships, the energetics of elementary steps of a chemical reaction pathway can be expressed by using a few binding energies as descriptors, allowing modelling of such pathways using fewer parameters.

The relationship between adsorbate binding energies and activity is not straight-forward for a complex reaction such as CO₂ reduction. It has however been found that there is a linear correlation between the adsorption free energy of different adsorbate intermediates involved in CO₂ reduction (i.e., OCHO, HCOOH, H₂COOH, O, CH₂O, CH₃O+OH, CH₃O and O+OH) and OH free energy. This simplifies the analysis, as illustrated by the following Examples, since this allows OH to be used as a descriptor for the different potential pathways.

The present invention can be described by the following non-limiting embodiments:

• • 1. A method for the electrolytic reduction of CO₂, the method comprising:
  • providing an electrolytic cell comprising at least one reaction chamber comprising at least one anode and at least one cathode, wherein the at least one cathode comprises at least one catalyst surface comprising at least one transition metal oxide;
  • placing at least one electrolyte solution between the at least one anode and the at least one cathode, so that the at least one anode and the at least one cathode come into contact with the electrolyte solution;
  • providing CO₂ in the electrolyte solution; and
  • applying electrical potential to the electrolytic cell;

whereby CO₂ undergoes at least one reduction reaction at the cathode to provide at least one product selected from the group consisting of methanol, methane, methanediol and formic acid.

• • 2. The method of embodiment 1, wherein the at least one transition metal oxide is selected from the group consisting of TiO₂, HfO₂,OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, PtO₂, ReO₂, ZrO₂, VO₂ and MoO₂, preferably HfO₂, OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, ReO₂, ZrO₂, VO₂ and MoO₂.
  • 3. The method of embodiment 1 or embodiment 2, wherein the at least one transition metal oxide is selected from the group consisting of RhO₂, CrO₂, MnO₂, PdO₂ and PtO₂, preferably RhO₂, CrO₂, MnO₂, MoO₂and PdO₂.
  • 4. The method of any one of the preceding embodiments, wherein the at least one transition metal oxide is selected from the group consisting of RhO₂, CrO₂, MoO₂ and MnO₂.
  • 5. The method of any one of the previous embodiments, wherein the catalyst surface comprises a single transition metal oxide.
  • 6. The method of any one of the preceding embodiments, wherein the catalyst surface is provided as a pure transition metal oxide.
  • 7. The method of any one of the preceding embodiments, wherein the catalyst surface comprises a mixture of transition metal oxides.
  • 8. The method of any one of the preceding embodiments, wherein the catalyst surface is provided on a cathode that comprises at least one metal, metal alloy or steel, including stainless steel.
  • 9. The method of any one of the preceding embodiments, wherein the CO₂ is comprised in a solution in the electrolytic cell, wherein the solution comprises at least one electrolyte that is provided between the anode and the cathode.
  • 10. The method of any one of the preceding embodiments, wherein the electrolyte comprises at least one source of carbon dioxide selected from bicarbonate and carbonic acid.
  • 11. The method of embodiment 9 or embodiment 10, wherein CO₂ is provided by a stream of CO₂ gas that is fed into the electrolytic solution.
  • 12. The method of any one of the preceding embodiments, wherein the electrolytic cell comprises an anode within one cell compartment and a cathode within another cell compartment.
  • 13. The method of any one of the preceding embodiments, wherein the process is carried out at a temperature in the range of about 0° C. to about 50° C., preferably about 10° C. to about 40° C., more preferably about 20° C. to about 30° C., even more preferably about 20° C. to about 25° C.
  • 14. The method of any one of the preceding embodiments, characterized in that the method is carried out at ambient pressure.
  • 15. The method of any one of the embodiments 1-13, characterized in that the method is carried out at a pressure in the range of 1 to 30 atmospheres, preferably in the range of 1 to 20 atmospheres, preferably in the range of 1 to 10 atmospheres, more preferably in the range of 1 to 5 atmospheres.
  • 16. The method of any one of the preceding embodiments, wherein the catalyst surface comrises at least one surface having a rutile structure.
  • 17. The method of any one of the preceding embodiments, wherein the catalyst surface comprises at least one surface having a (110) facet.
  • 18. The method of any one of the preceding embodiments, wherein an electrode potential that is less than about 0.5 V using a reversible hydrogen electrode (RHE) as a reference, is applied to the electrolytic cell.
  • 19. The method of any one of the preceding embodiments, wherein less than 50% moles H₂ are formed compared to moles CO₂ that is reduced.
  • 20. The method of any one of the preceding embodiments, wherein the at least one reduction product is selected from the group consisting of methane, methanol and methandiol.
  • 21. The method of the preceding embodiment, wherein an electrode potential that is in the range of about −0.4 V to about −0.2 V is applied to the electrolytic cell.
  • 22. The method of any one of the embodiments 1 to 18, wherein the at least one reduction product is formic acid.
  • 23. The method of the previous embodiment, wherein an electrode potential that is in the range of about −0.4 V to about −0.1 V is applied to the electrolytic cell.
  • 24. An electrochemical device for the reduction of carbon dioxide to at least one reaction product, the device comprising at least one electrochemical cell that comprises an anode and a cathode, wherein the cathode comprises at least one cathode electrode having at least one catalyst surface comprising at least one transition metal oxide.
  • 25. The electrochemical device of the previous embodiment, wherein the at least one reaction product is selected from the group consisting of methanol, methanediol, methane and formic acid.
  • 26. The electrochemical device of the previous embodiment, wherein the at least one transition metal oxide is selected from the group consisting of HfO₂, OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, ReO₂, ZrO₂, VO₂ and MoO₂, more preferably selected from the group consisting of RhO₂, CrO₂, MnO₂, MoO₂, and PdO₂, most preferably selected from the group consisting of RhO₂, CrO₂, MnO₂, MoO₂, PdO₂.
  • 27. The electrochemical device of any one of the preceding embodiments 24 to 26, wherein the catalyst surface comprises at least one surface having a rutile structure.
  • 28. The electrochemical device of any one of the preceding embodiments 24 to 27 wherein the catalyst surface comprises at least one surface having a (110) facet.
  • 29. The electrochemical device of any one of the preceding embodiments 24 to 28, wherein the electrolytic cell is further characterized in that the cathode electrode is disposed within a compartment that comprises an electrolyte solution.
  • 30. A process for the catalytic reduction of carbon dioxide, comprising:
  • introducing CO₂ to a solution comprising at least one electrolyte in an electrolytic cell so that the CO₂ comes into contact with at least one cathode electrode surface; and
  • applying a potential to said electrolytic cell, whereby CO₂ reacts with protons to form at least one product selected from methanol, methane, methanediol and formic acid;
  • wherein the cathode electrode surface comprises at least one catalyst surface comprising at least one transition metal oxide.
  • 31. The process of the preceding embodiment, wherein the at least one transition metal oxide is selected from the group consisting of HfO₂, OsO₂, RhO₂, CrO₂, NbO₂, MnO₂, PdO₂, ReO₂, ZrO₂, VO₂ and MoO₂, more preferably selected from the group consisting of RhO₂, CrO₂, MnO₂, MoO₂ and PdO₂.
  • 32. The process of any one of the preceding two embodiments, wherein the at least one transition metal oxide is selected from the group consisting of RhO₂, CrO₂, MoO₂ and MnO₂.
  • 33. The process of any one of the previous embodiments 30-32, wherein the catalyst surface comprises a single transition metal oxide.
  • 34. The process of any one of the preceding embodiments 30-32, wherein the catalyst surface comprises a mixture of transition metal oxides.
  • 35. The process of any one of the preceding embodiments 30-34 wherein the electrolyte comprises at least one of bicarbonate and carbonic acid as source of CO₂.
  • 36. The process of the previous embodiment, wherein the CO₂ is further provided by a stream of CO₂ gas that is fed into the electrolytic solution.
  • 37. The process of any one of the preceding embodiments, 30-36 wherein the electrolytic cell comprises an anode within one cell compartment and a cathode within another cell compartment.
  • 38. The process of any one of the preceding embodiments 30-37 wherein the process is carried out at a temperature in the range of about 0° C. to about 50° C., preferably about 10° C. to about 40° C., more preferably about 20° C. to about 30° C., even more preferably about 20° C. to about 25° C.
  • 39. The process of any one of the preceding embodiments 30-38, characterized in that the process is carried out at ambient pressure.
  • 40. The process of any one of the preceding embodiments 30-39, wherein the catalyst surface comrises at least one surface having a rutile structure.
  • 41. The process of any one of the preceding embodiments 30-40, wherein the catalyst surface comprises at least one surface having a (110) facet.
  • 42. The process of any one of the preceding embodiments 30-41, wherein an electrode potential that is less than about 0.5 V using a reversible hydrogen electrode (RHE) as a reference, is applied to the electrolytic cell.
  • 43. The process of any one of the preceding embodiments 30-42, wherein less than 50% moles H2 are formed compared to moles CO₂ that is reduced.
  • 44. The process of any one of the preceding embodiments 30-43, wherein the at least one reduction product is selected from the group consisting of methane, methanol and methandiol.
  • 45. The process of the preceding embodiment, wherein an electrode potential that is in the range of about −0.4 V to about −0.2 V is applied to the electrolytic cell.

46. The process of any one of the embodiments 30-45, wherein the at least one reduction product is formic acid. 47. The process of the previous embodiment, wherein an electrode potential that is in the range of about −0.4 V to about −0.1 V is applied to the electrolytic cell.

The above features along with additional details of the invention, are described further in the examples below, which are intended to further illustrate the invention but are not intended to limit its scope in any way.

EXAMPLES

The present invention will now be illustrated by the following non-limiting examples.

Example 1

Computational Methods and Model Systems

The electronic structure calculations were performed using DFT within the BEEF-vdW functional in the VASP software (Wellendorff et al. (2012) Phys Rev B 85). All the lattice parameters were optimized based on BEEF-vdW for ruthenium oxide (RuO₂), iridium oxide (IrO₂), niobium oxide (NbO₂), platinum oxide (PtO₂), titanium oxide (TiO₂), chromium oxide (CrO₂), manganese oxide (MnO₂), rhodium oxide (RhO₂), osmium oxide (OsO₂), hafnium oxide (HfO₂), molybdenum oxide (MoO₂) and palladium oxide (PdO₂) in their rutile crystal structure. The BEEF-vdW lattice parameters for these metal oxides are calculated and presented in the following lattice table:

Lattice table.
Shown are optimized lattice parameters for different metal
oxides studied.
	Oxide	a in (Å)	b (in Å)	c (in Å)
	RuO2	4.532	4.532	3.129
	IrO2	4.54	4.54	3.18
	TiO2	4.58	4.58	2.99
	NbO2	4.95	4.95	2.97
	MoO2	4.62	4.62	3.21
	OsO2	4.52	4.52	3.21
	HfO2	4.80	4.80	3.23
	CrO2	4.42	4.42	2.99
	RhO2	4.54	4.54	3.129
	MnO2	4.42	4.42	2.95
	PdO2	4.65	4.65	3.08
	PtO2	4.55	4.55	3.197

A plane wave basis set with a cutoff energy of 350 eV was used to expand the valence electron orbitals and the PAW method was used to represent core electrons (Blöchl, Phys Rev B (1994) 50, 17953-79). Monkhorst Pack grid was used in order to reduce the number of k-points, which were 4×4×1 in all cases. The atomic structure of the various reactants and products was found by minimizing the energy until atomic forces had dropped below 0.03 eV/Å.

The electrode was represented by a slab of four atomic layers with four metal atoms and eight oxygen atoms in each layer, and the slabs were separated with at least 16 Å of vacuum. Atoms in the bottom two layers were fixed while the atoms in the top two layers along with the adsorbed intermediates were allowed to fully relax. The dipole correction was used in all cases to decouple the electrostatic interaction between the periodically repeated slabs.

The rutile (110) surface contains metal atoms of two different coordination environments. Rows of six-fold coordinated metal atoms alternate with rows of five-fold coordinated metal atoms along the [001] direction. Whereas the six-fold coordinated metal atoms have approximately the same geometry as bulk, the five-fold coordinated metal atoms have an unsaturated bond perpendicular to the surface. Thus, two distinct surface sites are found, coordinatively unsaturated sites (cus-sites) on top of the five-fold coordinated metal atoms and bridging sites (br-sites) between two six-fold coordinated metal atoms (FIG. 1). We found that the br-sites generally bind adsorbates stronger than the cus-sites do. On the stoichiometric (110) surface, the br-sites are occupied by oxygen, while the cus-sites are vacant. The bridging oxygen atoms are under-coordinated and can be reduced from the surface. In order to create a realistic model system of our rutile (110) surfaces for the CO₂RR, we use the same assumptions as previously made by Karamad et al. (ACS Catal (2015)5, 4074-81)and Bhowmik et al. (Chem Sus Chem (2016), 9, 3230-43). According to Bhowmik et al., the most active sites for CO₂RR on their overlayers on RuO₂ are the bridge sites in the presence of CO spectator which is located on every other bridge sites. For the TMO surfaces considered in this work, we also find the bridge sites to be most active for CO₂RR, except for PtO₂ where some intermediates prefer to bind the cus sites. For some adsorbed species they adsorb bidentate on all the TMOs and there they adsorb on both a bridge site and on a cus site. The model system is thus the rutile (110) surfaces in their reduced form, but with ¼ ML of CO adsorbed on the bridge sites for all the TMOs considered here (see FIG. 1), with resulting free energ diagrams as shown in FIGS. 21-33. The only exception is IrO₂ (two different free energy diagrams; see FIGS. 22 & 23) since the differences between free energy of adsorbed CO when it is on the br-site and cus-site is 0.11 eV, and CO₂RR will happen while CO is on cus-site as spectator. In order to maintain generality, we also include IrO₂ when CO is on the bridge site. Throughout the figures, the bullet for IrO₂ is for when CO spectator is on a br-site, while the black bullets are for when the CO spectator is on a cus-site.

The computational hydrogen electrode (CHE) model is a tool to approximate the reaction free energy of an electrochemical reaction (Norskov et al. J Phys Chem (2004) 108, 17886-92) at a certain applied potential. The potential effects have been included by adding an implicit term, eU. The reaction free energy at an arbitrary potential U vs. the standard hydrogen electrode (SHE) is given by

ΔG_i(U)=ΔG_i(U=0)+eU   (1)

where e is the elementary charge. The ΔG_i (U=0) is calculated for each elementary step:

ΔG(U=0)=ΔE_DFT+ΔE_ZPE+ΔE_Sol+ΔH_0K→T−TΔS   (2)

where ΔE_DFT is calculated with DFT and ΔE_ZPE and ΔS are zero-point energy corrections and entropy differences which are calculated within the harmonic approximation for the adsorbed species, while the values for the gas phase species are taken from thermodynamical tables (Atkins & Paula, Atkins' Physical Chemistry (2009)). ΔE_Sol is the adsorbate stabilization term due to the solvent, which is not included in this study. ΔH_0K→T is changes in internal energy because of temperature.

The energy of molecular CO₂, H₂ and HCOOH are corrected by +0.3 eV, +0.1 eV and +0.15 eV, respectively, to correct for systematic DFT errors (Christensen et al. Catal Sci Technol (2015), 5, 4946-49). Reported reaction free energies and adsorbate binding free energies are referenced to the electronic energy of clean slab and CO_2(g), H₂O_(I) and H_2(g) free energies. Since CO₂RR is typically done in a 0.1 to 0.5 M solution of KHCO₃ or NaHCO₃, such solutions allow dilute formic acid to exist in a hydrated anion form (HCOO⁻). A free energy correction of −0.19 eV for deprotonation and solvation has therefore been included (Hansen, et al, Phys Chem Chem Phys (2016) 18, 9194-9201). The free energy of H₂C(OH)₂(aq) has been obtained from Hansen et al. (Catal Letters (2013), 143, 631-35).

Simulation Results

1. Methane, Methanol, Methandiol and Formic Acid Pathways and Corresponding Onset Potentials

In the first electron-proton transfer step of CO₂RR, two possible intermediates can be formed; OCHO (formate) or COOH (carboxyl). OCHO is found to be a more stable intermediate compared to COOH on all the TMO surfaces studied in this work. Therefore, all the reactions follow the pathways containing OCHO as an intermediate. The OCHO intermediate binds to the surface through both oxygen atoms (bidentate adsorption on one cus site and one bridge site) whereas COOH binds to the surface via its carbon atom to a bridge site.

A major competing reaction with CO₂RR is the HER. In Table 1, we estimate the onset potential for HER on the TMOs by calculating the binding free energy of a H adatom on the surfaces. In the discussion section, we will compare the onset potentials for CO₂RR towards different products with that of HER in order to estimate the selectivity between these two reactions.

In addition to HER, CO formation should also be taken into account, as it can result in CO poisoning. CO formation only happens via COOH intermediate where CO₂ is protonated to form carboxyl through the following reaction:

CO_2(g)+*+(H⁺+e⁻)→COOH   (6)

where * is an active site on the catalyst's surface. In this description we do not include the * for the adsorbed intermediates but the state of the reactant and desorbed products are indicated with (g), (aq) or (I). Further protonation leads to CO formation on the surface:

COOH+(H⁺+e⁻)→CO+H₂O(I)   (7)

which may poison the surface of the catalyst because we find that the CO reduction to CHO or COH has high thermochemical barriers on those TMOs. The CO admolecules that may be formed on the surfaces are therefore spectator species during CO₂RR. COOH binding free energy at U=0 V was found to be negative for RuO₂, IrO₂, OsO₂, MoO₂ and HfO₂, but positive for TiO₂, NbO₂, CrO₂, RhO₂, PtO₂, PdO₂. However, at the onset potentials for CO₂RR, the COOH adsorption is usually exergonic (but never as exergonic as the OCHO adsorption), and therefore a low CO coverage is expected to build up on the surfaces, while OCHO is the key intermediate of CO₂RR.

The first two protonation steps in CO₂RR may lead to the formation of aqueous formic acid or adsorbed HCOOH species on the surface. Käckell et al. (Appl Surf Sci (2000), 166, 370-75; Surf Sci (2000) 461, 191-8) have shown that HCOOH(aq) on TiO₂ surface should be in a dissociated form (HCOO⁻+H⁺) but according to the work done by Karamad et al. on RuO₂ (ACS Catal (2015) 5, 4075-81), HCOOH(aq) can adsorb on the surface in undissociated form (see RuO₂ in FIG. 2). We have included this possibility in this work and found that surfaces of TiO₂ and CrO₂ will rather dissociate the formic acid into the HCOO⁻ and H⁺ complex where HCOO⁻ binds to the bridge sites on the surface through its oxygen atoms but the proton binds to a surface oxygen atom (see TiO₂ in FIG. 2). However, the undissociated HCOOH species is the intermediate that forms on the RuO₂, IrO₂, HfO₂, OsO₂, and MoO₂. On the contrary, binding energy of HCOOH on PdO₂, PtO₂, MnO₂, and RhO₂ surfaces is so weak that that intermediate cannot form on those surfaces, and therefore it desorbs into the solution as HCOOH(aq). So, for those catalysts, in order to produce higher electron proton transfer products, the reaction may continue via activation of formic acid in aqueous form.

Forth protonation step results in methanediol formation and further protonation (six or eight) leads to methanol or methane products, respectively. Several possible pathways for methanol and methane formation are presented in Tables 2 and 3 (see also FIGS. 21-33). For RhO₂, PtO₂, MnO₂, and PdO₂, HCOOH binds weakly to the surface and formic acid is the main product before formation of methanediol, methane and methanol may take place. In order for the reaction to proceed to more reduced products, HCOOH in aqueous form must be activated and for this to happen the reaction needs to be run in acidic condition leading to lower methanediol, methane and methanol efficiencies.

Similar to the work done by Bhowmik et al. (ChemSusChem (2016) 9, 3230-43), OH adsorption free energy is also taken here as a descriptor for other adsorbate binding free energies to map the multi-demensional problem into one variable. With this we obtain activity volcanos (and to some extent selectivity volcanos) for different products on metal oxide surfaces. Our pathway for methane and methanol formation on RuO₂ is quite similar to the pathway used by Bhomwik et al (ChemSusChem (2016) 9, 3230-43). A small difference exists between the pathways presented herein and that reported by Karamad et al. (ACS Catal (2015) 5, 4075-81), which is presumably because Karamad's work is based on the RPBE functional whereas here we use the BEEF-vdW functional.

TABLE 1
Onset potentials towards various products, OH removal potentials and methane and methanol pathways along wi
potential limiting steps (number in the parenthesis, extracted from Tables 2 and 3)
								Methanol and methane
								production pathways (from
								Tables 2 and 3) along with
	Onset potential for different reactions in V vs RHE		potential limiting steps
		Formic					OH	(numbers)
Oxides	HER	acid	Methanol	Methane	Methandiol	OH removal	strength	Methanol	Methane
RuO2	−0.25	−0.98	−0.80	−0.80	−0.95	−0.56	Moderate	C (2)	F (2)
IrO2	−0.70	−1.27	−1.12	−1.02	−1.37	−1.02	Moderate	C (6)	G (8)
IrO2	−0.57	−0.97	−0.85	−0.76	−1.23	−0.76	Moderate	C (6)	F (8)
NbO2	−0.84	−1.13	−1.3	−1.30	−1.23	−1.3	Strong	C (6)	G (8)
PtO2	−0.16	−0.14	−0.27	−0.27	−0.27	−0.17	Moderate	A (3)	E (3)
TiO2	−0.19	−0.99	−0.95	−0.95	−1.09	−0.94	Moderate	C (2)	G (2)
OsO2	−0.97	−1.88	−1.49	−1.45	−1.98	−1.45	Strong	C (6)	G (8)
HfO2	−0.76	−2.41	−1.94	−1.94	−2.22	−1.94	Strong	D (6)	G (8)
MnO2	−0.63	−0.55	−1.2	−1.2	−1.2	0.33	Weak	A (3)	E (3)
RhO2	−0.06	−0.19	−0.23	−0.23	−0.23	−0.03	Moderate	A (3)	E (3)
CrO2	−0.29	−0.27	−0.31	−0.31	−0.31	−0.22	Moderate	B (3)	F (3)
PdO2	−0.29	−0.77	−1.1	−1.1	−1.1	0.4	Weak	A (3)	E (3)
MoO2	−0.41	−1.9	−1.51	−1.52	−1.86	−1.52	Strong	D (5)	G (8)
indicates data missing or illegible when filed

TABLE 2
Possible Pathways for Methanol formation
	(e− + H+ ) steps
Pathways	1	2	3	4	5	6
A	OCHO	HCOOH or HCOOH (aq)	H2COOH	H2CO	+ H2O(1)	CH3O	CH3OH(aq)
B	OCHO	HCOOH or HCOOH (aq)	H2COOH	O	+ CH3OH(aq)	OH	H2O(1)
C	OCHO	HCOOH or HCOOH (aq)	H2COOH	CH3O	+ OH	CH3O +	H2O(1)	CH3OH(aq)
D	OCHO	HCOOH or HCOOH (aq)	H2COOH	CH3O	+ OH	CH3OH(aq) +	OH	H2O(1)

TABLE 3

Possible Pathways for Methane formation

(e− + H+ ) steps

Pathways

1

2

3

4

5

6

7

8

E

OCHO

HCOOH

or

H2COOH

H2CO

+ H2O(1)

CH3O

CH4(g)

+ O

OH

H2O(1)

HCOOH

(aq)

F

OCHO

HCOOH

or

H2COOH

CH3O

+ OH

CH3O

+ H2O(1)

CH4(g)

+ O

OH

H2O(1)

HCOOH

(aq)

G

OCHO

HCOOH

or

H2COOH

CH3O

+ OH

CH4(g)

+ O + OH

O

+ H2O(1)

OH

H2O(1)

HCOOH

(aq)

Formic Acid and Methandiol Formation

Formic acid and methanediol are a two and four-step electron transfer products of CO₂RR, respectively. As mentioned earlier, RhO₂, PtO₂, MnO₂, and PdO₂ cannot adsorb the HCOOH intermediate on their surfaces and therefore these catalysts have the potential of producing aqueous formic acid. FIGS. 3 and 4 and Table 3 show that among these candidates, PtO₂ and RhO₂ have the smallest onset potentials towards formic acid, or −0.14 V and −0.19 V, respectively. In Table 1 the onset potentials for the HER are estimated from the adsorption free energy of a H adatom. For RhO₂, the hydrogen binding free energy is −0.06 eV whereas the binding free energy of OCHO is −0.19 eV, and therefore a higher efficiency towards formic acid is expected than towards H₂. On the other hand, for PtO₂ the hydrogen binding free energy is −0.16 eV but the binding free energy of OCHO is 0.14 eV, and therefore HER is expected to dominate over CO₂RR. In the other words, RhO₂ has low activity towards HER (in CO₂ saturated solution) but high activity towards formic acid, and thus these factor make RhO₂ the best candidate for production of formic acid. It should be noted here that without CO₂ in the solution, RhO₂ is predicted here to have very small onset potential for HER.

CrO₂ is another candidate located very close to the top of the formic acid volcano (FIG. 4), with predicted onset potential of −0.27 V (Table 1). The free energy of the adsorbed HCOOH species (actually as a HCOO⁻+H⁺ complex on CrO₂ as explained above) is only 0.01 eV lower than the free energy of HCOOH (aq) and therefore CrO₂ is expected to both evolve formic acid as well as more reduced products (discussed below). The hydrogen binding free energy on CrO₂ is 0.29 eV whereas the OCHO binding free energy is −0.33 eV, meaning that this catalyst should be much more selective towards CO₂RR than HER.

CO₂RR towards methanediol requires four proton-electron transfer steps. CrO₂ is also on the top of the volcano for methanediol formation (FIGS. 5 and 6) with predicted onset potential of −0.31 V (Table 1). RhO₂ and PtO₂ are also located at the top of this volcano but those TMOs are predicted to be more selective towards formic acid formation as discussed above.

3-Methane and Methanol Formation

CO₂RR towards methanol and methane requires 6 and 8 proton-electron transfer steps, respectively. Table 1 shows methanol and methane onset potentials and the thermodynamically potential limiting steps (PLS). The corresponding volcanos are presented in FIGS. 7 and 8, respectively. Since the mechanism to either of those products is not the same for all TMOs, we simplify these volcano plots by only including reactions steps that exist in the minimum free energy pathway towards the products. This divides our volcano plots into different regions, depending on our descriptor; the binding energy of OH on the surfaces. As an example, in both of these volcano plots the intersection of OCHO→HCOOH and OCHO→HCOOH(aq) lines is around −0.34 eV with respect to our descriptor. This can be seen from the change in slope of the red line at this value. Therefore, reduction of OCHO will form adsorbed HCOOH intermediate for the more reactive TMOs (binding OH stronger than −0.34 eV) while it will form desorbed HCOOH(aq) product on the less reactive TMOs (binding OH weaker than −0.34 eV). This may affect the subsequent steps, and therefore we only include the intermediates that are a part of the reaction network, for a given TMOs (or a region on the volcano plot). In several other cases, the lines are broken in a similar way.

For the left leg of the methanol volcano (FIG. 7) there are three PLS; OH removal to water, CH₃O→CH₃OH (aq), and OCHO→HCOOH. The OH removal and the CH₃O→CH₃OH(aq) step overlap completely, meaning that those two reaction steps are the same in magnitude (or at least very similar) for a given TMOs. The OCHO→HCOOH step, however, only becomes the PLS close to the top of the volcano, but there it overlaps with the aforementioned steps. For the left leg of the methane volcano (FIG. 8), we have the same PLS, except for the CH₃O→CH₃OH (aq) step, which is of course not a part of the reaction network. Thus, the OH removal potential may be considered as a representative of the left leg for both the methanol and the methane volcanos. For the right leg of both of the volcanos (where PtO₂, RhO₂, MnO₂, and PdO₂ are located), the HCOOH (aq)→H₂COOH step is PLS. However, as discussed above, those TMOs will start evolving formic acid at smaller potentials.

The top of the volcano for methane and methanol is in both cases around ΔG_OH−2−0.21 eV. CrO₂ is located at the top of both volcanos with predicted onset potential of −0.31 V (Table 1). The main conclusion here is that catalysts with ΔG_OH>−0.21 eV are selective towards formic acid, but those with ΔG_OH<−0.21 eV are selective towards methane, methanol, and methanediol. CrO₂ is, however, predicted to produce all these products (formic acid, methane, methanol, and methanediol) at small onset potentials of around −0.3 V in all cases (Table 1).

Experimental work for methanol formation on RuO₂, TiO₂, and IrO₂ have shown that these oxides are selective towards methanol at potentials between −0.5 V to −1 V vs. RHE. Our calculations confirm this and the predicted onset potentials we found for RuO₂ (−0.80 V), IrO₂ (−0.85 V), and TiO₂ (−0.99 V) are very close to the experimental values.

Discussion

Similar to the work done by Bhomwik et al, we observed that OH adsorption energy is a very good descriptor for CO₂RR on pure TMOs, as it was for the TMO overlayers on RuO₂. All the intermediates in the reaction network (Tables 1 and 2) bind to the surface through oxygen atom(s), except the CH₂O species, which bind through both the oxygen atom and the carbon atom. Subsequent protonation of CH₂O to the next possible intermediates (CH₂OH or CH₃O), always leads to the formation of the CH₃O intermediate for all the TMOs considered in this work. Therefore, CH₂OH is not a part of the reaction pathway to methane or methanol and not included in Table 1 or 2. The reaction network presented in both the work of Karamad et al. (ACS Catal (2015) 5, 4075-81) and Bhomwik et al. (ChemSusChem (2016) 9, 3230-43) is very similar, but not completely the same. In our work we include all the intermediates presented in their work and we find that the network presented in Karamad's work (for RuO₂) captures all the different lowest free energy reaction pathways on all the TMOs considered here. We find the scaling relations of all the intermediates included in the overall reaction network as a function of the binding energy of the OH species (see FIGS. 13-20). Those scaling relations are then used to construct volcano diagrams.

For the TMO surfaces to catalyze the CO₂RR continuously, the surfaces need to be cleaned of OH species, but in all of the reaction pathways presented in Tables 1 and 2 the OH removal is the last or the second last step. As OH removal potentials ranges from −2 V to 0.5 V for all the TMO catalysts, they can be categorized based on the strength of OH binding energy (Table 1). Due to high OH removal potential for some of these catalysts, the HER may become a dominant reaction compared to the CO₂RR, and this may affect the activity and selectivity of the catalysts. Even if the earlier intermediate steps in CO₂RR for producing methane or methanol are feasible at low overpotential, in order to remove OH, the reaction must be carried out at higher overpotentials where HER would be relatively fast.

FIG. 9 presents a “selectivity volcano” for formic acid, methanediol, methanol, and methane products. Here we have added all the volcanos (from FIGS. 4, 6, 7 and 8) into one volcano-plot for comparison. We conclude that for ΔG_OH>−0.21 eV the CO₂RR is more selective towards formic acid, but for ΔG_OH<−0.21 eV, the selectivity is more towards methane, methanol, and methanediol. On the left leg, the PLS towards formic acid and methanediol coincide each other, which means that the formation of those products need almost the same thermodynamic onset potentials. However, the overpotential towards methane and methanol is slightly lower and we predict here that those TMOs (from HfO₂ to RuO₂) will start forming methanol and methane at smaller onset potentials than needed to form methanediol or formic acid. CrO₂ and PtO₂ are on the top of the volcano for methane and methanol (blue line and dashed-blue line). In fact, PtO₂ is slightly on the right leg of this volcano which means it should be more selective towards formic acid. We have included the explicit values for the onset potentials towards different products on the volcano-plot and the onset potential towards formic acid on PtO₂ is −0.14 V whereas the onset potentials towards methanol, methane, and methanediol is −0.27 V. CrO₂ is located exactly at the top of the methanol-methane-methanediol volcanos and the formic acid volcano intercects those other volcanos where CrO₂ is located. Therefore, the onset potentials towards all these products for CrO₂ are very similar and all close to −0.3 V (see also Table 1).

The main conclusion about these overall volcanos (FIG. 9) is that for −0.34 eV<ΔG_OH<−0.21 eV all these three lines (or the PLSs) in this interval closely coincide on one line. Therefore, we cannot predict the selectivity here since the onset potentials are always similar towards all products (note that CrO₂ is the only candidate within this interval). ΔG_OH<−0.34 eV and ΔG_OH>−0.21 eV are the only selective parts of this overall volcano. The former is somewhat selective towards methanol and methane, and the latter is more selective towards formic acid. As discussed above, CrO₂ is located on the top of the volcano for methanediol, methanol, and methane and the onset potential towards formic acid is very similar, and therefore it is predicted to form all these four products of CO₂RR. The important thing about CrO₂ is that it can adsorb the formic acid as an intermediate on its surface and continue its reduction towards methane, methanol, and methanediol with much smaller applied potential than compared to RuO₂, TiO₂ and IrO₂. From these calculations and analysis it seems that in order to find an active and selective catalyst among TMOs towards methane, methanol, and methanediol formation, the binding energy of the OH species on the surface needs to be between −0.5 eV and −0.34 eV, where the OH binding energy is moderate. In addition of being in this interval, the possible candidate needs to have a low activity towards HER as well. Another criteria that needs to be considered here is the CO poisoning. The only pathway which may lead to CO poisoning is through the COOH intermediate. Minimizing the rate towards hydrogen evolution and CO poisoning along with binding OH in a moderate way are the three criteria that should be considered at the same time in order to find a promising candidate. The binding free energy of OCHO, COOH and H are compared in FIG. 10 and used in further analysis below. The candidates with weak H and COOH binding free energies will have low tendency to evolve hydrogen and poison the surface with CO. The candidates also need to bind OCHO moderately, between 0 and −0.5 eV which is similar as the interval for the OH binding free energy since OCHO and OH scale linearly, almost one-to-one.

In FIG. 11, we show that the COOH binding free energy and hydrogen binding free energy are correlated linearly. The region (and candidates very close to this region) with weak (positive) hydrogen and weak COOH binding free energy is where one should be looking for the promising candidate(s) because there the rates towards HER and the CO poisoning is lower. Moderate OH binding free energy is another restriction, which needs to be added into the picture. All the design parameters (moderate OH binding, weak hydrogen binding and weak COOH binding) are included in FIG. 12, where the OH binding free energy is plotted as a function of the H binding free energy (since COOH and H are correlated linearly we only include the H binding here). The most promising candidates for high CO₂RR selectivity and low HER activity should be in the regions denoted by solid horizontal lines. For a candidate to be selective towards methane, methanol, and methanediol, it needs to be located in the lowermost region (−0.5 eV<ΔG_OH<−0.34 eV; black solid lines). For a candidate to be selective towards formic acid, it should be located approximately between −0.21 eV<ΔG_OH<0 eV (light grey region). Candidates located between −0.21 eV<ΔG_OH<−0.34 eV (grey region) are predicted to evolve all four products. CrO₂ is the only candidate located in any of those regions and therefore it is promising for selective formation of CO₂RR products over HER.

In this work, all the analyses are based on the calculated thermodynamics of adsorbed intermediates on the TMO surfaces and the CHE is then used to vary the applied potential implicitly. In order to conclude whether these candidates are truly selective towards a given intermediate, one needs to consider kinetic and transport effects as well, which are beyond the scope of this study. Activation energies can be calculated for each of the proton-electron transfer steps (towards different intermediates and products) by setting up a charged double-layer model including a solid TMO electrode and an aqueous electrolyte solution as has been done for the metal-liquid interface for CO₂RR (Hussain et al, ACS Catal (submitted)). There, the charge difference in the double layer sets up the applied electric potential explicitly and therefore the proton-electron transfer barriers can be calculated as a function of explicitly varied potential. This approach is necessary in order to reproduce the experimentally observed product distribution as a function of both the metal type and the applied potential. However, it has been shown that the TCM-CHE approach is sufficiently accurate to predict the onset potentials for various reactions on different catalyst's materials, including CO₂RR on metal catalysts, both in terms of which product is formed on Cu and when the metal catalyst is varied.

CONCLUSIONS

The main goal of this work was to establish trends in CO₂RR activity and selectivity for production of formic acid, methanediol, methanol, and methane on the surface of pure TMOs. We construct volcano plots through scaling relations of adsorbed intermediates that show selectivity towards methane and methanol when ΔG_OH<−0.34 eV, but selectivity towards formic acid for ΔG_OH<−0.21 eV. We also observe that there is an interval for OH adsorption free energy between −0.34 eV<ΔG_OH<−0.21 eV, which is not selective towards any specific CO₂RR product, since the onset potentials towards all the products is predicted to be the same and very low. We found that CrO₂, PtO₂, and RhO₂ have smaller onset potentials compared to all other TMOs included in this work, or below −0.3 V. When the hydrogen evolution reaction is taken into account, CrO₂ and RhO₂ are predicted to be more selective towards formic acid, methane, methanol, and methanediol rather than evolving hydrogen, whereas PtO₂ would be more selective towards forming hydrogen gas than the CO₂RR products. Experimental works for CO₂RR on RuO₂, IrO₂, and TiO₂ have shown that these TMOs are rather selective towards methanol at potentials between −0.5 to −1 V. Our simulation confirms this and the calculated onset potentials for RuO₂, IrO₂, and TiO₂ are −0.80 V, −0.85 V and −0.99 V, respectively, or within these experimental values.

Example 2

Role of Coadsorbed Water

When transition metal oxides are exposed to an aqueous environment, water molecules can be adsorbed onto the metal oxide surface. These water molecules in turn can affect functional properties of metal oxide surfaces, including their catalytic potential.

To investigate the role of coadsorbed water in the catalytic reduction of CO₂, the calculations described above under Example 1 were repeated in the presence of water on the catalyst surface.

To address our system, we have carried out an ab initio molecular dynamic (Al MD) simulation of RuO₂ (110)-water. Born Oppenheimer molecular dynamic (MD) simulations and static geometry optimization calculation were done using a plane-wave based pseudopotential formalism with a generalized gradient approximation (GGA) to describe the exchange-correlation effects within BEEF-vdW functional implemented in the periodic DFT package VASP. A plane wave basis set with a cutoff energy of 350 eV was used to expand the valence electron orbitals and the PAW method was used to represent core electrons. The energy minimum structure of water bilayers were determined using 4×4×1 k-points until the energies were converged to within 10⁻⁴ eV. In our simulation the slab consists of four atomic layers with four metal atoms and eight oxygen atoms in each layer, and the exposed liquid phase was presented by 22 H₂O molecules. The system was subject to periodic boundary conditions in all direction. In our simulation atoms in the bottom two layers were fixed while the atoms in the top two layers along with the water molecules were allowed to reconstruct during MD run. The time step for MD runs is 0.5 fs. Canonical ensemble (NVT) conditions were imposed by a Nose-Hoover thermostat with a target temperature of 300 K. 1 ps of equilibration period was followed by 3˜4 ps of production period. The convergence of the vertical energy gap can be monitored by the time accumulative averages.

We performed the simulation of bulk water in contact with the RuO₂ surface. We observe a specific well-defined geometry in the simulation where a low coverage of water molecules are chemisorbed on the surface but the remaining water molecules have insignifant effect on the surface chemistry. Therefore, in the proceeding calculations, we only include those chemisorbed water molecules to obtain the effect of those on the CO₂RR catalysis.

Simulation Results:

• • 1. Formic Acid, Methandiol, Methanol and Methane Pathways and Corresponding Onset Potentials

TABLE 4
Methanol Possible Pathways
A	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH3O +	H2O(1)	CH3OH(aq)
B	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH3OH(aq) +	OH	H2O(1)
C	OCHO	HCOOH	H2COOH	O	+ CH3OH(aq)	OH	H2O(1)
D	OCHO	HCOOH	H2COOH	OH(cus) + OH(br)	+ CH3OH(aq)	OH(br) +	H2O(cus)	H2O(1)
E	OCHO	HCOOH	H2COOH	CH2O	+ H2O(1)	CH3O	CH3OH(aq)

TABLE 5
Methane Possible Pathways
F	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH3O	+ H2O(1)	CH4 (g)	+ O	OH	H2O(1)
G	OCHO	HCOOH	H2COOH	CH2O	+ H2O(1)	CH3O	CH4 (g)	+ O	OH	H2O(1)
H	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH4(g)	+ O + OH	O	+ H2O(1)	OH	H2O(1)
I	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH3O	+ H2O(1)	CH4(g)	+ OH(cus) + OH(br)	OH(br)	H2O(1)
J	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH3O	+ H2O(1)	CH4(g)	+ OH(cus) + OH(br)	OH(br)	H2O(1)
K	OCHO	HCOOH	H2COOH	CH3O	+ OH	CH4(g)	+ O + OH	OH(cus)	+ OH(br)	OH(br)	H2O(1)

TABLE 6
Onset potentials towards various products and OH removal potentials.
		Onset				Onset
		potential	Onset	Onset	Onset	potential
	OH,	for	potential	potential	potential	for
	removal	formic	for	for	for	hydrogen
	potential	acid V vs	methanediol	methanol	methane	evolution	Methanol	Methane
Oxides	V vs RHE	RHE	V vs RHE	V vs RHE	V vs RHE	V vs RHE	Pathway	pathway
RuO2	−1.08	−1.12	−1.02	−1.08	−1.08	−0.39	B	H
IrO2	−1.13	−0.94	−1.32	−1.13	−1.13	−0.77	B	H
TiO2	−1.2	−1.2	−1.18	−1.2	−1.18	−0.06	B	H
OsO2	−1.7	−1.54	−2.03	−1.7	−1.7	−0.84	B	H
RhO2	−0.39	−0.28	−0.39	−0.39	−0.39	−0.14	D	K
CrO2	−0.34	−0.24	−0.45	−0.45	−0.45	−0.41	B	H
NbO2	−1.43	−1.45	−1.62	−1.45	−1.45	−0.75	B	H
MnO2	0.09	−0.39	−1.33	−1.33	−1.33	−1.38	B	H
PdO2	0.15	−0.41	−0.73	−0.73	−0.73	−0.24	E	G
PtO2	−0.56	−0.21	−0.21	−0.56	−0.56	−0.21	D	K
ReO2	−1.26	−1.01	−1.03	−1.26	−1.26	−0.95	B	H
ZrO2	−1.48	−1.79	−1.64	−1.62	−1.62	−0.13	B	H
VO2	0	0.06	−1.1	−1.1	−1.1	−1.09	C	H
MoO2	−1.33	−1.65	−1.91	−1.33	−1.33	−0.21	B	H

2-Formic Acid Formation and Theoretical Activity Volcano

The first two protonation steps in CO₂RR may lead to the formation of aqueous formic acid or adsorbed HCOOH species on the surface. In the previous example it has been shown that HCOOH(aq) on TiO₂ surface should be in a dissociated form (HCOO⁻+H⁺). Here, by including water four different types of surfaces has been observed and they are presented in table 7 and corresponding FIG. 34-a, b, c and d. So, PtO₂, RhO₂, CrO₂, MnO₂, PdO₂ are the only catalysts that cannot adsorb the HCOOH intermediate on their surfaces and therefore these catalysts have the potential of producing aqueous formic acid. In our previous work we showed that CrO₂ has the potential to keep HCOOH on its surface but by including water, this catalyst shows different behavior and clearly desorbs HCOOH from the surface. For these candidates, corresponding onset potentials for HER and CO₂RR is shown in Table 6. According to this table hydrogen binding energy for RhO₂, CrO₂, and MnO₂ is more positive than OCHO binding free energy and therefore a higher efficiency towards formic acid is expected than towards H₂.

TABLE 7
Different behavior of formic acid on different surfaces
	Surfaces can	Surfaces can	Surfaces can
	keep	dissociate formic	dissociate formic
	undissociated	acid (dissociation is	acid (dissociation	Surfaces can
	form of	between spectator	is between	desorb
Oxides	formic acid	and formate)	surface and formate)	formic acid
RuO2	√	x	x	x
NbO2	√	x	x	x
MoO2	√	x	x	x
ZrO2	√	x	x	x
IrO2	x	√	x	x
OsO2	x	√	x	x
TiO2	x	x	√	x
PtO2	x	x	x	√
RhO2	x	x	x	√
CrO2	x	x	x	√
VO2	x	x	x	√
MnO2	x	x	x	√
PdO2	x	x	x	√

In FIG. 35, there is shown a theoretical volcano plot for the formation of formic acid using scaling relations (indicated by lines) in the presence of water. For each transition metal oxide, the limiting potential value is indicated by filled squares.

3-Methanol and Methane Formation and Theoretical Activity Volcano

FIG. 36 shows a theoretical volcano for formation of methane and methanol from scaling relations (lines) in the presence of water. For each of the transition metal oxides, the explicit limiting potential values are shwon by filled squares.

A comparable volcano plot is shown in FIG. 37, which shows formation of formic acid, methane and methanol from scaling relations (shown by lines) in the presence of water. For each of the transition metal oxide, the explicit limiting potential values are indicated by filled squares.

In FIG. 38 there is shown a scaling figure for COOH adsorption free energy vs. hydrogen binding free energy in the presence of water molecules. If COOH is formed as a species on the surface, it can lead to CO formation or CO poisoning. A strong H binding energy will lead to H poisoning and H₂ formation whereas a weak H adsorption would prevent proton adsorption until at more negative potentials. The area that indicates weak COOH and hydrongen binding energies is likely to include promising catalysts for CO₂RR.

A scaling figure for OH binding free energy vs. hydrogen binding free energy in the presence of water is shown in FIG. 39, where OH is chosen as it is the descriptor on the volcano graphs.

Comparing hydrogen evolution activity with CO₂RR activity shows that MnO₂, CrO₂ and RhO₂ are more selective towards CO₂RR than HER.

The overall trends of CO₂RR represented in the volcano diagrams in FIGS. 35-37 remain similar, with and without water included in the simulations. However, when HER is taken into account, the selectivity towards CO₂RR products and H₂ formation may change slightly. When water is included, MnO₂ and RhO₂ become even more promising than without water, and together with CrO₂, those are the most promising catalysts towards efficient CO₂RR, see FIGS. 38 and 39.

Claims

1. A method for the electrolytic reduction of CO2, the method comprising: providing an electrolytic cell comprising at least one reaction chamber comprising at least one anode and at least one cathode, wherein the at least one cathode comprises at least one catalyst surface comprising at least one transition metal oxide selected from the group consisting of HfO2, OsO2, RhO2, CrO2, NbO2, MnO2, ZrO2, and VO2;placing at least one electrolyte solution between the at least one anode and the at least one cathode, so that the at least one anode and the at least one cathode come into contact with the electrolyte solution;providing CO2 in the electrolyte solution; andapplying electrical potential to the electrolytic cell;whereby CO2 undergoes at least one reduction reaction at the cathode to provide methanol, or formic acid.

2. The method of claim 1, wherein the at least one transition metal oxide is selected from the group consisting of HfO2, NbO2, and ZrO2.

3. The method of claim 1. wherein the catalyst surface is provided as a pure transition metal oxide.

4. The method of claim 1. wherein the catalyst surface comprises two or more transition metal oxides.

5. The method of claim 1. wherein the method is carried out at a temperature in the range of about 0° C. to about 50° C.

6. The method of claim 1. characterized in that the method is carried out at ambient pressure.

7. The method of claim 1, characterized in that the method is carried out at a pressure in the range of about 1 atmospere to about 30 atmospheres.

8. The method of claim 1. wherein the catalyst surface comprises at least one surface having a rutile structure.

9. The method of claim 1. wherein the catalyst surface comprises at least one surface having a (110) facet.

10. The method of claim 1. wherein an electrode potential that is less than about −0.5 V using a reversible hydrogen electrode (RHE) as a reference, is applied to the electrolytic cell.

11. (canceled)

12. The method of claim 10, wherein an electrode potential that is in the range of about −0.4 V to about −0.1 V is applied to the electrolytic cell.

13. An electrochemical device for the reduction of carbon dioxide to methanol, or formic acid, comprising at least one electrochemical cell that comprises an anode and a cathode, wherein the cathode comprises at least one cathode electrode having at least one catalyst surface comprising at least one transition metal oxide selected from the group consisting of HfO2, OsO2, RhO2, CrO2, NbO2, MnO2, ZrO2, and VO2.

14. The electrochemical device of claim 13, wherein the at least one transition metal oxide is selected from the group consisting of HfO2, NbO2, and ZrO2.

15. The electrochemical device of claim 13, wherein the catalyst surface comprises at least one surface having a rutile structure.

16. A process for the catalytic reduction of carbon dioxide, comprising: introducing CO2 to a solution comprising at least one electrolyte in an electrolytic cell so that the CO2 comes into contact with at least one cathode electrode surface; andapplying a potential to said electrolytic cell, whereby CO2 reacts with protons to form methanol or formic acid;wherein the cathode electrode surface comprises at least one catalyst surface comprising at least one transition metal oxide selected from the group consisting of HfO2, OSO2, RhO2, CrO2, NbO2, MnO2, ZrO2, and VO2.

17. The process of claim 16, wherein the at least one transition metal oxide is selected from the group consisting of HfO2, NbO2, and ZrO2.

18. The process of claim 16, wherein the process is carried out at a pressure in the range of about 1 to 20 atmospheres and a temperature in the range of about 0° C. to about 50° C.

19. The process of claim 16, wherein the catalyst surface comprises at least one surface having a rutile structure.

20. The process of claim 16, wherein the catalyst surface comprises at least one surface having a (110) facet.

21. The process of claim 16, wherein the process is carried out at a temperature in the range of about 0° C. to about 50° C.

22. The process of claim 16, wherein the process is carried out at a temperature in the range of about 10° C. to about 40° C.

23. The process of claim 16, wherein the process is carried out at ambient pressure.