Hydrocarbon(s)

The present invention claims priority to AU2015904952 the contents of which are hereby

incorporated by reference in their entirety.

Field

The present invention relates to the production of hydrocarbon(s) such as methane or substituted hydrocarbon(s) such as methanol. In one embodiment, the hydrocarbon(s) can be formed using water and carbon dioxide as precursor materials.

Background

For many decades, oil has been the main feed stock for the production of hydrocarbons. Recently, however, concerns over increases in costs of fossil fuels and the effect of global warming have prompted the exploration of alternative more renewable feed stocks.

Carbon dioxide has received much attention as an alternative feed stock for the production of methane, because there is a drive to reduce carbon dioxide emissions to help slow global warming, and because it is cheap and readily available. Carbon dioxide can be converted into hydrocarbons such as methane by reacting it with hydrogen, for example via the Sabatier reaction. The

hydrocarbons produced can then be converted into other forms such as methanol.

While the process of converting carbon dioxide into hydrocarbons is relatively well understood, it has been an energy intensive process. For example, the hydrogen used for carbon dioxide conversion is usually produced from fossil fuels by steam reforming, and conversion of carbon dioxide to hydrocarbons typically requires relatively high temperatures. Catalysts are often employed to increase efficiencies, but they can add significant costs to the process. Using fossil fuels to produce hydrogen, which is then converted back into hydrocarbons, is known to be a relatively inefficient process. Alternative hydrogen feed stocks, such as water, can be used, but their use is a relatively energy intensive process.

Accordingly, there is a need to find a more sustainable way of producing hydrocarbons using more efficient and environmentally friendly methods. Summary of disclosure

According to a first aspect of the present invention there is provided a method for the production of hydrocarbon(s), such as methane, or substituted hyd rocarbons, such as methanol, the method comprising the steps of: contacting a catalyst with water and carbon dioxide in the presence of light in order to photocata lyse :

(i) the splitting of at least some of the water into hydrogen and oxygen; a nd

(ii) the reaction between hydrogen and carbon dioxide to prod uce hydrocarbon(s), such as metha ne, and/or substituted hydrocarbons, such as methanol; wherein the catalyst comprises at least gold and ruthenium, in the form of at least one nanocluster supported by a support substrate such as a titanium dioxide substrate.

According to a second aspect of the invention, there is provided a method for the production of hyd roca rbon(s), such as metha ne, or substituted hydrocarbons, such as methanol, the method comprising the steps of: a. contacting a first catalyst with water in order to photocatalyse the splitting of at least some of the water into hydrogen and oxygen;

b. contacting a second catalyst with a gas stream comprising carbon dioxide and

hydrogen, at least some of the hydrogen ca n be produced from step (a), in order to photocatalyse the reaction between the hydrogen and carbon dioxide to produce hydroca rbon(s), such as methane, a nd/or substituted hydrocarbons, such as methanol.

The first and second catalyst can be the sa me catalyst. The first catalyst a nd the second catalyst ca n be different cata lysts. The first a nd second cata lysts can comprise one or more nanoclusters. The first and second catalysts can be immobilized on the support. The first a nd second catalysts can be activated on the support. The nanoclusters can comprise gold and/or ruthenium nanoclusters. The nanoclusters can have an average cluster size of less than about 2 nm .

It should be understood that the splitting of at least some of the water into hydrogen and oxygen can include splitting the water into hydrogen and or oxygen containing species such as hydrogen radicals, hydronium a nd or hydroxyl radicals. Without wishing to be limited by hypothesis or theory, embodiments of the invention will now be summarised and then described based on the understanding of how the catalyst performs under various conditions.

A. Embodiments in which there is a first catalyst and a second catalyst

According to a third aspect of the present invention there is provided a method for the production of hydrocarbon(s), such as methane, or substituted hydrocarbons, such as methanol, the method comprising the steps of: a, contacting a first photocatalyst with water in the presence of light in order to

photocatalyse the splitting of at least some of the water into hydrogen and oxygen; wherein the first photocatatyst comprises gold nanoclusters supported by a titanium dioxide substrate;

b. contacting a second catalyst with a gas stream comprising carbon dioxide and at least some of the hydrogen produced from step (a) in order to catalyse the reaction between the hydrogen and carbon dioxide to produce hydrocarbon(s), such as methane, and/or substituted hydrocarbons, such as methanol;

wherein the second catalyst comprises ruthenium nanoclusters supported by a titanium dioxide substrate.

In an embodiment, the catalyst can comprise a first catalyst and a second catalyst. The first catalyst can be a photocatalyst. The second catalyst can be a photocatalyst.

Step (a): a first catalyst for the photocatalytic splitting of water into hydrogen and oxygen

The first photocatalyst preferred for use in step (a) above can comprise a substrate and an active metal component. The substrate can be graphene, graphite, carbon black, nanotubes, fullerenes, and/or zeolites. The substrate can be a carbon nitrate CxNy. The substrate can be a metal oxide or nitride. The substrate can be a titania, silica and/or alumina. The substrate can be barium titanate or perovskite. The substrate can be a titanium oxide. The titanium oxide support substrate can include anatase and/or the commercialiy available P25. The substrate can be a monolithic. The substrate can have a planar surface such as a plate or disc. The substrate can be particulate. The substrate can comprise nanoparticles. In one embodiment, the substrate comprises titanium dioxide nanoparticles.

Photocatalysts are activated by light. The light used can be determined by the specific type of photocatalysts. For example, some photocatalysts can catalyse a reaction using light with a wavelength over a broad range, while others may only catalyse the reaction with a specific wavelength e.g. 365 =/ ± 5 nm or 400 ± 5 nm. Depending on the reaction, it may be advantageous that the first photocatalyst comprises two or more types of photocatalyst where one can perform at a specific wavelength and the other can perform over a broad wavelength range Usually, the more intense the light, the more efficient the catalytic process is. However, in some circumstance the reactants and/or products may be degraded if the light source is too intense. Therefore, it can be advantageous to have a balance between rate of catalysis and the rate of degradation of the reactants/products.

A common wavelength range for photocatalysts are those in the ultraviolet range i.e. 200-400 nm. The source of ultraviolet light may be from a dedicated lamp or may be from a natural light source, such as the sun. Usually commercial ultraviolet light sources have a greater intensity compared to natural sources. Natural light sources can have a UV intensity (i.e. <400 nm) of approximately 4.63 mW cm ^"2 , while commercial sources can be many times more intense, such as >1000 mW cm ^"2 . Using a natural light source can be advantageous from an energy input perspective, and can make the process more environmentally friendly. If a natural light source is used, it may be supplemented with a commercial light source. Such circumstances may include during times of inclement weather and/or during times of reduced light activity, such as at night. In areas with plentiful natural light, e.g.

Australia, it may be advantageous to rely on the sun as a source of ultraviolet light during the day and a commercial Sight source during the night to allow constant photocata lytic activity over a 24 hour period. Concentrated solar sources, can provide energies in the range of from about 500 to about 1000 suns i.e. 2315 -4630 mW/cm ^"2 .

An advantage of using photocatalysts (when compared to other types of catalysts) is that they often do not require the use of heat to catalyse reactions. Not requiring heat can decrease operational costs, make the production of hydrogen more environmentally friendly, and make the production of hydrogen safer. Temperatures that can be used for photocatalysis are around room temperature e.g. about 20-30 "C, but may be as high at about 100-300 °C, for example 250 °C. Nevertheless, the photocatalyst could be used with the addition of heat, which may allow for a reduction in light energy input.

The active metal of the photocatalyst can be selected from gold, silver, copper, platinum, palladium, nickel, rhenium, ruthenium and/or titanium, and/or other transition metals and their corresponding oxides. In a embodiment the active metal is gold. It may be advantageous to have more than type of active metal, one of which could be gold. Whilst gold is exemplified herein, it should be understood that the invention is not so limited and other active metal nanoclusters could be prepared using the details disclosed herein.

The form in which the active metal is associated with the substrate can be determined by the reaction

--""ditions of the formation of the photocatalyst. For example, the active metal(s) could be present in the form of complexes, nanoparticles and/or clusters/nanoclusters. It may be advantageous to have more than one active metal where each metal has a different form. In a preferred embodiment, the active metal is present as a nanocluster.

By way of background, metal complexes have an active metal that is surrounded by one or more ligand(s). The type of ligand(s) can greatly affect the performance of the catalyst. One of the ligands can be immobilised on the surface of the substrate, which can help to prevent the complex from disassociating from the substrate. This can be advantageous, for example, in helping to recover the photocatalyst once a reaction is complete. Nanoparticles, on the other hand, can have an average size in a range of from about 5 to about 100 nm. The shape and arrangement of the nanoparticles can greatly affect the function as a photocatalyst. For example, a nanoparticle with a cuboid shape usually has a lower surface area compared to nanoparticles that are rods or ribbons, and a lower surface area is usually associated with a decrease in catalyst efficiency. Clusters or nanoclusters (referred to herein interchangeably unless the context makes clear otherwise), in yet a further form, refer to a collection or group of two or more active metal atoms, but usually contain less than approximately 200 atoms. Clusters typically differ from nanoparticles both structurally and electronically - unique packing of atoms not seen in larger metal particles and non-piasmonic (Au/Ag)/metallic. It terms of size, nanoclusters are usually considered as being between complexes and nanoparticles. It is to be understood that the number of atoms used to describe a nanocluster is the average number and there is typically a distribution associated with the average number. For example, nanoclusters containing more than 20 metal atoms can have a distribution of ± 10 or more percent e.g. M30±3, 55±5, 100±10. The metals that comprise the nanoclusters can comprise ligands. Similar to complexes, any ligands associated with the nanocluster can be used to stabilise the nanocluster and in some circumstances may help to improve the performance of the nanocluster when as a catalyst. In some cases, it is preferred to remove any ligands before the compound is used as a photocatalyst.

The first photocatalyst can have a support that is photoactive. The clusters can be deposited onto a support capable of adsorbing light of appropriate wavelength. The cluster plus the photoactive support forms the photocatalyst. The support can be particulate itself or is can be a bulk solid substrate. The bulk solid substrate can be a wafer such as a silicon wafer or a porous silica disk. The first photocatalyst in the form of a paste can be applied to the support. The thickness of the applied photocatalyst can be varied.

In one embodiment, the photocatalyst comprises a titanium dioxide substrate in the form of nanoparticles; the nanoparticles are associated with gold nanoclusters. The gold nanoclusters can comprise Au ₃ to Auioi. The gold clusters can be selected from (Ph ₃ Pau) ₃ OBF4, [(AuPPh ₃ ) ₃ 0]PF ₆ ,

Au ₅ (PPh ₃ )4CI ₍ Au ₆ (PPh ₃ (BF<) ₂ , Au _s {PPh ₃ lv1NOs)2, Au ₆ {PPh ₃ ) ₆ (PF ₆ )2, Au ₈ (PPh ₃ ) ₈ (N0 ₃ )_, Αυ ₈ (ΡΡϊι ₃ ) ₇ (Ν0 ₃ ) ₂ , Au ₉ (PPh ₃ ) _s (N0 ₃ ) ₃ , Auio(PPh ₃ ) ₅ (C ₆ Fs)4, AuuClsifm-CFaCeH^)^}?, Aun(PPh ₃ ) ₇ (PF ₆ ) ₃ ,

[Au ₁₃ {Pme ₂ Ph).oCl2](PF6) ₃ , Aui ₃ (PPh3)4[S(CH2)ii(CH ₃ )] _4/ [Au ₁₃ (PPH ₂ CH2pPH ₂ ) ₆ ](N0 ₃ )4,

AuBs(PH. C6H4S03Na.2H ₂ 0)i2Cl6, Au ₅ 5(P h ₃ )i2Ci ₆ , Aui ₀ i(PPh ₃ ) ₂ iCI ₅ , where "Ph" is phenyl and "Me" is methyl.

Once the size of the nanoclusters begins to increase over approximately 2 nm, the activity of the photocatalyst may decrease. In an embodiment, the nanoclusters have an average size of less than about 2.5, 2, 1.5 or 1 nm. For example, the average size of e.g. Auioican be approximately 1,6 nm.

The number of nanoclusters per substrate nanoparticle may depend on the type of active metal used. In one embodiment, the number of nanoclusters per nanoparticle is at least about 1, 2, 5, 10, 15, 20, 15 or 30. The percentage approximate coverage of the nanoparticles with nanoclusters can be in the range of from about 0.1 to about 10 % or more, or at least about 0.1, 0.5, 1, 17, 2, 3, 4, 5, 6 or 10 % or more as a percentage of the total available surface area. In one embodiment, the approximate coverage of the nanoparticles with gold nanoclusters is in the range of from about 0.17 to about 1.7 wt%.

The first photocatalyst can be pre-treated prior to use. Treatment methods can include calcining and/or acid treatment. Acid treatment can be performed with or without calcining. Where calcining is used, acid treatment can be performed before or after calcining. It is thought that acid treatment has an effect on the interaction between the catalyst substrate and the active metal during preparation of the photocatalyst.

Calcining can be performed at a temperature of at least about 50, 100, 200, 300 or 400 °C to remove any residual carbon contamination from the photocatalyst surface. Calcining can be performed under oxygen and/or hydrogen atmospheres and/or under vacuum. There is thought to be an improvement in i gas production as the first photocatalyst is treated under successively harsher conditions. This may be due to the removal of any ligands from the photocatalyst surface (leaving only the active metal clusters behind). It is hypothesised that in some embodiments, the removal of ligands and an increase in cluster size improves the catalytic performance of anatase-supported Au clusters.

To help ensure all contaminates (including adventitious carbons) are removed from the first photocatalyst prior to use, it can be advantageous to expose the first photocatalyst to a vacuum for an extended prior of time. Prior to use the first photocatalyst can be held under vacuum for at least about 1, 2, 5, 10, 12 or 15 hours. It is preferred that the photocatalyst is not exposed to the atmosphere once it has been held under vacuum.

The step of contacting the first photocatalyst with water can involve exposing all or some of the surfaceisl of nhotocatalyst with water in order to effect a reaction. The water can be from any source. The water can be substantially pure, or it can be a part of an aqueous solution. The water used to produce hydrogen can be in liquid form and/or vapour form. In one embodiment, the step of contacting the photocatalyst with water comprises immersing the photocatalyst in a body of water. The water can flow over the first photocatalyst. The flow can be continuous. When liquid water is used, the first photocatalyst may be homogenously or heterogeneously distributed in the body of water. Homogenous distribution may be performed by vigorously mixing the body of water and a first photocatalyst in a fine particulate form. The first photocatalyst can be an aggregate that can easily be separated from the body of water. Heterogeneous distribution may be achieved by immobilising the first photocatalyst on at least one stationary support. In one embodiment, the first photocatalyst is supported on rods that can be inserted into the body of water.

In one embodiment, the step of contacting the first photocatalyst with water includes allowing a water vapour to come into contact with the first photocatalyst. Bringing the water vapour into contact with the first photocatalyst can be performed in a variety of ways, for example, continuously flowing water vapour over the first photocatalyst. The pressure of the water vapour can be varied to achieve the desired result (optimum hydrogen production). Condensation of water vapour can occur if the pressure of the vapour is too high. To prevent condensation, the heat of the vapour may be increased, but applying too much heat to prevent condensation may be undesirable. The water vapour may be provided at below atmospheric pressures. In an embodiment, step (a) is performed under 20 Torr of water vapour. Additional gases may be included with the water vapour. The additional gas may be an inert gas. The inert gas can be argon (Ar). !n one embodiment, step (a) s performed under 280 Torr of argon (Ar).

During the production of hydrogen, oxygen is also produced according to the following equation (1):

2 H ₂ 0 ^ 2 H ₂ + 0 ₂ (1).

It can be advantageous to remove either hydrogen and/or oxygen to drive the reaction through favourable thermodynamics. The hydrogen and oxygen gases can be collected and stored for use in a subsequent reaction. The subsequent reaction can be the reaction of at least some of the hydrogen with carbon dioxide in an e.g. 4:1 molar ratio of hydrogen to carbon dioxide to produce hydrocarbons such as methane. In one embodiment, all of the hydrogen is passed to a further reaction to assist in the production of methane.

The amount of hydrogen that can be produced in step (a) can be at least about 15, 50, 80, 100, 150, 200, 250, 350, 450, 550, 1000, 1500, 2000 or 5000 μπιοΙ hr ¹ g^ cm ^-2 .

Step (b): a second catalyst for the catalytic reaction of carbon dioxide and hydrogen The hydrogen produced in step (a) can be used as feed for the production of unsubstituted hydrocarbons. Hydrocarbons can include Ci to Cio containing compounds such as methane, ethane, propane, butane, pentane, hexane, heptane, octane, nonane, decane, their various isomeric forms such as n-, iso-, sec- and tert-alkanes, and their respective oxides such as methanol and ethanol. More complex hydrocarbons such as aromatics may also be produced. The hydrocarbons produced can be greater than Cio. The hydrogen can also be used as a feed for the formation of a substituted hydrocarbon such as methanol, ethanol, propanol, and so on.

Hydrogen can be converted into methane using the Sabatier reaction shown in equation (2):

C0 ₂ + 4 H ₂ -> CH« + 2 H ₂ 0 (2) Hydrogen can be converted into methanol using the following equation (3):

C0 ₂ + 3 H ₂ -> CH3OH + H ₂ 0 (3).

Industrially, it is understood that these processes require the use of high temperatures i.e. about 200 to 500 °C and typically they require the presence of copper-, zinc oxide- and/or alumina-based catalysts.

The step of contacting the second catalyst with carbon dioxide and hydrogen can involve allowing the gas streams to flow over the surface. The amount of gas introduced to the surface of the second photocatalyst can be controlled (in terms of molar ratio) so as to ensure the desired reaction product. Steps (a) and (b) can be undertaken sequentially as two separate method steps, or they can be undertaken concurrently.

The second catalyst can be a photocatalyst. The photocatalyst can be activated by UV wavelengths of light. The second catalyst preferred for use in step (b) can comprise a substrate and an active metal component. The substrate can be as described above e.g. graphene, graphite, carbon black, nanotubes, fuilerenes, and/or zeolites. The substrate can be an oxide or a nitride. The substrate can be titania, silica and/or alumina and their oxides. The substrate can be a titanium oxide. The titanium oxide support can include anatase and/or the commercially available P25. The substrate can be a planar surface or it could be particulate. The substrate can comprise nanoparticles. In one embodiment, the substrate comprises titanium dioxide nanoparticles.

The active metal of the photocatalyst of step (b) can be selected from gold, copper, silver, platinum, palladium, nickel, rhenium, ruthenium and/or titanium, and/or other transition metals and their corresponding oxides and/or other transition metals and their corresponding oxides. In an embodiment the active metal is ruthenium. It may be advantageous to have more than type of active metal, where at least ruthenium is present. The second catalyst can be applied to a support. The support can be a particulate to increase the surface area, or the support can be solid substrate. The solid substrate can be a wafer such as a silicon wafer or a porous silica disk. The second catalyst can be applied to the support as a layer. The thickness of the layer can be varied.

The form in which the active metal is supported on the substrate can be determined by the reaction and/or the reaction conditions. For example, the active metal(s) may be present in forms of complexes, nanoparticles and/or nanoclusters. These forms of active metal are described in relation to step (a) above and that description also applies here. It may be advantageous to have more than one active metal, with each metal having a different form i.e. nano clusters and complexes. In an embodiment, the active metal is present as a ruthenium nanocluster.

In one embodiment, the second catalyst comprises a titanium dioxide substrate in the form of nanoparticles associated with ruthenium nanoclusters. The percentage of ruthenium nanoclusters loaded onto the nanoparticles can be at least about 0.1, 0.2, 0.5, 1, 2, 5 or 10 wt%.

The second catalyst can be pre-treated prior use. Treatment methods can include calcining and/or acid treatment. To help ensure contaminates are removed from the catalyst prior to use, it can be advantageous to expose the catalyst to a vacuum for an extended prior of time. Calcining can be performed for a period of at least about 1, 2, 5, 10, 12 or 15 hours. The pre-treatment can be at a temperature of at least about 50, 100, 200, 300 or 400 °C to remove any residual carbon

contamination from the photocatalyst surface. Calcining can be performed under oxygen and/or hydrogen atmospheres and/or under vacuum. This may be due to the removal of any ligands from the photocatalyst surface (leaving only the active metal clusters behind). It is hypothesised that in some embodiments, the removal of ligands and an increase in particle size improves the catalytic performance of anatase-supported u clusters.

A ruthenium-based catalyst may significantly reduce the temperatures and/or pressures required to produce methane and/or methanol. For example, temperatures less than about 100, 200 or 250 °C with pressure below a few atmospheres can be used with ruthenium-based catalysts to produce hydrocarbons (substituted or unsubstituted) from hydrogen. The efficiency of a ruthenium-based photocatalyst may also be improved by exposure to ultraviolet light. The support may assist in the photocatalytic production of hydrocarbon or substituted hydrocarbons.

The amount of methane that can be produced in step (b) can be at least about 350, 450, 550, 1000, 2000 or 5000 μιηοΙ hr ¹ g^crrf ² .

Apparatus and system The reaction of steps (a) and (b) may be performed in an apparatus (a reactor). The apparatus for step (a) can have an inlet for the introduction of water. The first photocatalyst of step (a) may be housed in a part of the apparatus and arranged so that the water can come into contact with the surface of the first photocatalyst. In some embodiments, the apparatus is seaiable once the water has been introduced. The water can be introduced as a liquid or vapour. If the water is a vapour it can be introduced under pressure. A light source can be arranged inside or outside of the vessel to allow activation of the first photocatalyst. The reaction may be allowed to proceed for as long as is necessary to produce as much hydrogen as is required (or as is stoichiometricaliy possible). The temperature and/or pressure within the reactor may be slowly increased to effect the optimal reaction. The gases evolved in the reactor may be collected from the apparatus from an outlet. The gases may be collected and separated.

A second apparatus may be provided for step (b). In step (b) carbon dioxide and hydrogen are mixed at the desired molar ratio in the presence of a second photocatalyst. The second photocatalyst may be housed in a part of the apparatus and arranged so that the gas streams can come into contact with the surface of the photocatalyst. In some embodiments, the apparatus is seaiable once the gases have been introduced. A light source can be arranged inside or outside of the vessel to allow activation of the second photocatalyst. The reaction may be allowed to proceed for as long as is necessary to produce as much product as is required. The temperature and/or pressure may be slowly increased in the apparatus to effect reaction. The gases evolved may be collected from the apparatus from an outlet. The gases may be collected and separated. In step (a) and step (b) the apparatus can be an autoclave. in one embodiment step (a) and step (b) are performed in the same apparatus. Because the production of hydrogen is photocatalytic, it may be possible to employ both the first photocatalyst and the second photocata!ysts to produce both hydrogen and hydrocarbons at the same time, sequentially. The two photocatalysts, first and second, may be independent of each other, or they may be associated. If the two catalysts are associated with each other, it may be that, for example, gold clusters and ruthenium nanoclusters are supported on the same titanium dioxide support. In some embodiments, there are gold ruthenium nanoclusters as described further below. Having one support with two active nanoclusters or one support with active Au-Ru nanoclusters may reduce the operational costs of the production of hydrocarbons and may make the process more

environmentally friendly.

In step (a) the molar ratio of hydrogen to carbon dioxide is always greater for any carbon dioxide produced during the production of hydrogen. As the molar ratio of hydrogen to carbon dioxide in Eq. 2 and Eq. 3 is always greater than 1:1, any carbon dioxide produced during the production of hydrogen is preferably supplemented with an additional source of carbon dioxide. If the production of hydrocarbons is coupled with a production that burns hydrocarbons e.g. for electricity, then the products from one process may be a feed stock for another.

B. Embodiments in which there is a single catalyst

In this embodiment, it is thought that steps (a) and (b) occur at the same catalyst site. The method of the present invention can be undertaken in the presence of a catalyst which can

(i) split at least some of water (into hydrogen and oxygen); and

(ii) react hydrogen and carbon dioxide to produce hydrocarbon(s), such as methane, and/or substituted hydrocarbons, such as methanol;

The catalyst can comprise a substrate and an active metal component. The substrate can be as described above with respect to the other catalysts. The substrate can be e.g. graphene, graphite, carbon black, nanotubes, fullerenes, and/or zeolites. The substrate can be titania, silica and/or alumina. The substrate can be a titanium oxide. The titanium oxide support substrate can include anatase and/or the commercially available P25. The substrate can be monolithic. The substrate can have a planar surface such as a plate or disc. The substrate can be particulate. The substrate can comprise nanoparticles. In one embodiment, the substrate comprises titanium dioxide nanoparticles.

The catalyst can be a photocatalyst that is activated by light. A common wavelength range for photocatalysts are those in the ultraviolet range i.e. 200-400 nm e.g. 365 nm =/± 5 nm. The source of ultraviolet light may be from a dedicated lamp or may be from a natural light source, such as the sun. Photocatalysts are described above, and all description made there applies here unless the context makes clear otherwise. An advantage of using photocatalysts (when compared to other types of catalysts) is that they often do not require the use of heat to catalyse reactions. Not requiring heat can decrease operational costs, make the production of hydrogen more environmentally friendly, and make the production of hydrogen safer. Temperatures that can be used for photocatalysis are around room temperature e.g. about 20-30 °C, but may be as high at about 100-300 "C, for example 250 °C. Nevertheless, the photocatalyst could be used with the addition of heat, which may allow for a reduction in light energy input.

The active metal of the catalyst can be selected from one or more of gold, copper, silver, platinum, palladium, nickel, rhenium, ruthenium and/or titanium, and/or other transition metals and their corresponding oxides. In an embodiment the active metal comprises only ruthenium. In an embodiment, the active metals comprise gold and ruthenium. The active metal can comprise gold and ruthenium bound together. The gold and ruthenium can have a bond distance in the range of from about 2.5 to 3 A. such as 2.7 1 2.8 A, or at least about 2.5, 2.7, 2.8 or 3 Angstrom (A). The gold x to ruthenium y ratio can be a bout 1 : 1.5, 1 : 2, 1 : 3. The active metal can be AuRu ₃ .,Au ₂ Ru ₃ and or AuaRu4.The AuRu ₃ can be Ru ₃ AuPPh ₃ CI{CO)i ₀ . The Au ₂ Ru ₃ can comprise [Au ₂ Ru ₃ (μ-Η) (μ ₃ -0ΟΜε) (μ-1 ₂ ) (C0 ₉ )] {where L ₂ = Ph ₂ P(CH ₂ )PPh ₂ }. The Au ₂ Ru ₄ can com prise [Au ₂ Ru ₄ (μ ₃ -Η) (μ-Η) (μ-Ph2ECH ₂ E'Ph ₂ ) (CO)i ₂ ] {where E = E' = As or P; E = As, E' = P} and or [Au ₂ Ru ₄ (μ ₃ -Η) (μ-Η) {n-l,2-Ph2PC ₆ H ₄ PPh ₂ ) (CO) ₁₂ ] and or [Au ₂ Ru ₄ (μ ₃ -Η) (μ-Η) (μ-dppf) (CO)i ₂ ] {where dppf - l,l'-i)/s(diphenylphosphino)ferrocene}. Sourced from: " Metal Clusters in Chemistry: Vol 1 Molecular Meta l Clusters", Editors: P. Braunstein, LA. Oro, P.R. Raithby. Wiley-VCH 1999. ISBN: 3-527-29549-6, the contents of which is incorporated in so fa r as the AuRu metal clusters are described a nd unless the context makes clear otherwise.

The active metal can be present in the form of complexes, nanoparticles and/or clusters

/nanoclusters. In a preferred embodiment, the active metals are present as a nanocluster. Clusters or nanoclusters (referred to herein interchangeably unless the context makes clear otherwise), in yet a further form, refer to a collection or group of two or more active metal atoms, but usually contain less than approximately 200 atoms. It terms of size, nanoclusters are usually considered as being between complexes a nd nanoparticles. The na nocluster can comprise more than 20 metal atoms with a distribution of ± 10 or more percent e.g. M ₃ o±3, M ₅ s±5, MmoilO. The metals that comprise the nanoclusters can comprise ligands. Similar to complexes, any ligands associated with the nanocluster can be used to stabilise the nanocluster and in some circumstances may help to improve the performance of the na nocluster when as a catalyst. In an embodiment, the nanocluster with ligands is of the form ula I n some cases, it is preferred to remove any ligands before the compound is used as a catalyst. In some embodiments, the ligands assist in the cata lytic activity.

Once the size of the nanoclusters begins to increase over approximately 2 nm, the activity of the photocata lyst may decrease. I n an embodiment, the nanoclusters have a n average size of less than about 2.5, 2, 1.5 or 1 nm. An active site for reaction can com prise more than one or more na noclusters.

The catalyst can be applied to a support. The support can be particulate itself or can be a solid substrate. The solid substrate can be a wafer such as a silicon wafer or a porous silica disk. The first cata lyst in the form of a paste can be applied to the support. The thickness of the applied cata lyst can be varied. The nanoclusters can be supported by e.g. titanium dioxide nanoparticles. The number of nanoclusters per substrate nanoparticle may depend on the type of active metal used . In one embodiment, the number of nanoclusters per nanoparticle is at least about 1, 2, 5, 10, 15, 20, 15 or 30. The percentage approximate coverage of the nanoparticles with nanoclusters can be at least in the ra nge of from about 0.1 to 10 % or more, or about 0.1, 0.5, 1, 1.7, 2, 3, 4, 5, 6 or 10 % or more as a percentage of the tota l availa ble surface area. In an embodiment, the method can comprise contacting a photocataiyst with water and C02 in order to photocatalyse the reaction of water with C02, wherein the photocataiyst comprises gold nanoclusters and ruthenium nanoclusters or mixed gold-ruthenium nanoclusters supported by a titanium dioxide substrate.

The catalyst can be pre-treated prior to use. Treatment methods can include calcining and/or acid treatment. Acid treatment can be performed with or without calcining. Where calcining is used, acid treatment can be performed before or after calcining. It is thought that acid treatment has an effect on the interaction between the catalyst substrate and the active metal during preparation of the catalyst.

Heterogeneous catalysts and photocatalysts are generally pre-treated in situ before testing, in order to remove advantageous hydrocarbons and other surface-adsorbed species, or to open up catalyst active sites by removal of ligands. Many different techniques for this can be undertaken for example including ozone treatment, calcination in C»2 or H2, and heating under a flow of inert gas. Preferably, any treatment does not have any damaging effect upon active metal clusters which might cause them or the substrates to which they are attached to aggregate into larger nanoparticles. The selection of an appropriate pre-treatment which removes adsorbed contaminants while still retaining intact clusters upon the surface for these materials is preferred. Calcining can be performed under oxygen and/or hydrogen atmospheres and/or under vacuum. Calcining can be performed at a temperature of not more than about 50, 100, 200, 300 or 400 °C. !n an embodiment, the calcining is undertaken at about 200 °C under vacuum. There is thought to be improvement in H ₂ gas production as the catalyst is treated under successively harsher conditions. This may be due to the removal of any ligands from the catalyst surface (leaving only the active metal clusters behind). It is hypothesised that in some embodiments, the removal of ligands and an increase in particle size improves the catalytic performance of anatase-supported clusters.

To help ensure all contaminates (including adventitious carbons) are removed from the catalyst prior to use, it can be advantageous to expose the catalyst to a vacuum for an extended prior of time. Prior to use the catalyst can be held under vacuum for at least about 1, 2, 5, 10, 12 or 15 hours. It is preferred that the catalyst is not exposed to the atmosphere once it has been held under vacuum.

The step of contacting the catalyst with water can involve exposing all or some of the surface(s) of catalyst with water in order to effect a reaction. The water can be from any source and the various ways in which the surface of the catalyst can contact water are described above, and also apply here unless the context makes clear otherwise. The catalyst is also exposed to carbon dioxide. Preliminary testing indicates a P _C O2:PH2O ratio of about 2, 3 or 4 is optimal for solar fuel production. In an embodiment, the Pcoi'-Pmo ratio is 3. In some embodiments, optimal production of CO and H ₂ was observed at a reagent ratio of 1:1, and C0 ₂ :H ₂ 0 ratios in the range of at feast about 0.5 to 4, preferably about 1 to about 3, give peak hydroca rbon production.

During the production of hydrogen, oxygen is also produced according to the following equation (1):

2 H ₂ 0 - 2 H ₂ + 0 ₂ (1).

The hyd rogen can be used for the production of unsubstituted hydrocarbons. Additional hydrogen can be injected into the system if desired. Hydrocarbons can include Ci to Cio containing compounds such as methane, ethane, propa ne, butane, pentane, hexane, heptane, octane, nonane, decane, their various isomeric forms such as n-, iso-, sec- and tert-alka nes, and their respective oxides such as methanol and etha nol. More complex hydrocarbons such as aromatics may also be prod uced. The hydrocarbons produced can be greater than Cio. The hydrogen ca n also be used for the formation of a substituted hydroca rbon such as methanol, ethanol, propanol, and so on.

Hydrogen can be converted into methane using the Sabatier reaction shown in equation (2): Hydrogen can be converted into methanol using the following eq uation (3):

C0 ₂ + 3 H ₂ - CHjOH + H ₂ 0 (3) .

It may be that the catalyst is able to stabilise intermediaries in reaction (1) such as hyd rogen rad icals, hydronium and or hydroxylradicals that go on to react with C0 ₂ .

The amount of hydrogen that can be produced by the Au-Ru catalyst can be at least about 70, 80, 90 or 100 μηηοΙ hr ¹ g^ cm ^"2 . The amount of methane, ethane, ethene, propane a nd/or propene that can be produced in step (b) can be at least about 350, 450, 550, 1000, 2000 or 5000 nmol hr ¹ g ^"1 cm ^"2 .

Apparatus and system

The reaction of steps (a) and (b) may be performed in an apparatus (a reactor). The apparatus can have an inlet for the introduction of water. The catalyst may be housed in a part of the apparatus and a rranged so that the water can come into contact with the surface of the catalyst. In some embodiments, the a ppa ratus is sealable once the water has been introduced. The water can be introduced as a liquid or vapour. If the water is a vapour it can be introduced under pressure. A light source can be arranged inside or outside of the vessel to allow activation of the cata lyst. The temperature and/or pressure within the reactor may be slowly increased to effect the optimal reaction. The apparatus can have an inlet for the introduction of ca rbon dioxide. The catalyst may be housed in a part of the appa ratus and arranged so that the carbon dioxide can come into contact with the surface of the catalyst. In some embodiments, the appa ratus is sealabie once the carbon dioxide has been introduced. Alternatively the carbon dioxide is continuously introduced into the appa ratus. The reaction temperature can be elevated to at least about 120, 150, 180 or 200 °C .The gases evolved in the reactor may be collected from the apparatus from an outlet. The gases may be collected and separated.

According to a second aspect of the invention there is provided an apparatus for the production of hydrocarbon(s) such as methane or substituted hydrocarbons such as methanol, the apparatus ada pted to undertake the method described herein.

According to a third aspect of the invention there is provided hyd rocarbons or substituted hydrocarbons when prod uced by a method as described herein, or when prod uced in an appa ratus herein described . According to a fourth aspect of the invention there is provided a catalyst when used in the method or apparatus of the invention.

Brief description of Figures

Embodiments of the invention will now be described, by way of example only, with reference to the accom panying non-limiting drawings, in which:

Figure 1: graph showing H ₂ gas yield for benchma rk Pt-TiC>2 photocata lysts and control experiments; the latter of which showed no i production.

Figure 2: Bar chart showing a compa rison of mean H2 peak production rates for samples that were exposed to vacuum in the reaction cell for 10 minutes, com pared with those that were evacuated for 12 hours. Acid-washed supports are denoted with the a/w abbreviation.

Figure 3: Graph showing the number of moles of O2 and C0 ₂ in the reaction cell throughout the course of an extended experiment, showing the consumption of 0 ₂ and peak C0 ₂ prod uction.

Figure 4: Ba r chart showing average H ₂ peak production rate for Au _s clusters supported on pure anatase nanopartides with various treatments.

Figure 5: Ba r chart showing average H ₂ peak production rate for Aug clusters supported on pure anatase and acid-washed P25 nanopartides with various treatments.

Figure 6: Ba r chart showing average H ₂ peak production rate for Auioi clusters supported on acid- washed P25, treatments, acid-washed anatase, and pure anatase nanopartides with various treatments Figure 7: Bar chart showing a comparison of H2 peak production rate for Aus, Aug, and Auioi clusters supported on anatase nanoparticies with various treatments.

Figure 8: Bar chart showing a comparison of H ₂ peak production rate for Aua, Aug, and Auioi clusters supported acid-washed P25 nanoparticies with various treatments.

Figure 9: Bar chart showing a comparison of h½ peak prod uction rate for 0.17% w/w Auioi, Au _g , and Aus clusters supported on Ti0 ₂ against 1.0% w/w Pt-P25 and 1.0% w/w Pt-anatase.

Figure 10: Graph showing hydrocarbon production following photocatalyst treatment at varying calcination temperature. The reaction temperature was set at 220 °C for each of the runs.

Figure 11: early experimental data on ruthenium nanoclusters (RU3) in step (b).

Figure 12: early experimental data on ruthenium nanoclusters {RU3) in step (b).

Figure 13: early experimental data on ruthenium nanoclusters (RU3) in step (b).

Figure 14: early experimental data on ruthenium nanoclusters in step (b).

Figure 15A-H: early experimental data on gold nanoparticies in step (a).

Figure 16A: Peak production rates of (left) CH _A and (right) H ₂ by various photocatalysts tested. Average production rates are shown from three independent tests, with error bars representing the standard error in the mean of these. Sta ndard reaction conditions were used for all tests: pre-treatment under vacuum at 200 °C, reaction at 200 °C, PCOZ:PH2O = 3.

Figure 16B: Peak production rates of longer-chain hydrocarbon prod ucts by various photocatalysts tested . Standa rd reaction conditions were used for all tests.

Figure 17A Peak production rates of (left) H ₂ and (right) CH* products by various

photocatalysts tested, normalised to total precious meta l content of co-catalysts.

Figure 17 B: Peak production rates of longer-chain hydrocarbon products by various photocatalysts tested, normalised to total precious metal content of co-catalysts.

Figure 18A: Peak production rates of (left) H2 a nd (right) CH4 by all cluster-deposited titania materials tested here. Standard reaction conditions were used for a ll tests.

Figure 18B: Peak prod uction rates of longer-chain hydrocarbon prod ucts by all cluster- deposited titania materials tested here. Figure 19: Peak production rates of Ha by all cluster-deposited titania materia ls, in atmospheres of H ₂ 0/Ar and HzO/CC /Ar. Standard reaction conditions were used for ail tests, with CO2 neglected in the case of H ₂ 0/Ar atmospheres.

Figure 20A: Peak production rates of hydrogen and methane by AURU3-T1O2 as a function of combined material pre-treatment a nd reaction tem perature. Pcoi'.Pmo = 3 for all tests here.

Figure 20B: Peak production rates of longer-chain hydrocarbon products by AuRuj-TiC as a function of combined material pre-treatment and reaction temperature. PCO2:PH2O = 3 for all tests here.

Figure 21A: Peak production rates of hydrogen, methane and CO by AURU3-T1O2 as a function of reaction tem perature. Pcoi'-Pmo - 3, pre-treatment temperature of 200 °C for a ll tests here.

Figure 21B: Peak production rates of longer-chain hydrocarbon products by Au RurTi0 ₂ as a function of reaction temperature. PCO2'.PHIO = 3, pre-treatment temperature of 200 °C for all tests here.

Figure 22 shows bond distances in Ru3(p-AuPPh ₃ )(p-CI)(CO)i ₀ . Examples of embodiments of the invention

Em bodiments of the invention, and other embodiments, will now be described with reference to the accompa nying non-limiting exam ples. Any % referred to herein may be wt% unless the context makes clear otherwise.

Example 1 Formation of the gold nanoclusters

An aqueous stock solution of 50 m gold chloride anions (AuCU ^" ) in a glass vial was made by dissolving HAuCU 3 H2O with the same molar amount of HCI, ensuring stability for more than several months. An aqueous stock solution of 50m borohydride anions (BH ^" ) in a glass beaker was made by dissolving Na BF gra nules with the same molar amount of NaOH, guaranteeing stability for several hours at room temperature.

For the smallest nanoparticies of 3.2 nm in diameter, we added 100 μΙ_ of the AuCU ^" /H ⁺ solution to a glass vial with water and later injected 300 xl of the BFU ^' /OH ^" solution all at once, while stirring on a mechanical shaker for uniform mixing. The total weight of the aq ueous solution was controlled to be 10 g so that the concentration of gold ions is 0.50 mM. The solution changed colour from light yellow to orange immediately, a nd then to red while the vial was stirred for 1 min to release hydrogen gas molecules. For nanoparticies of other sizes, the amount of the BH/f/OH ^" solution was increased from 300 to 650 uL followed by heating for 2-3 min at the boiling temperature of water on a hot plate. The average diameter of gold nanoparticles was precisely controlled from 3.2 to 5.2 nm. The amount of the BH4 ^' /O H ^" solution was changed from 200 to 1300 μί. during the search for the "sweet zone" before heating.

Nanoparticles can be prepared by this method as described in the paper entitled: Cha rged Gold Nanoparticles in Non-Polar Solvents: 10 Minute Synthesis and 2D Self-Assembly, LANG M U IR, 26(10) pp7410-7417 (2010), the entire contents of which are hereby incorporated by reference in their entirety. If there are any inconsistencies between this document and the incorporated document, this document shall take precedence unless the context ma kes clear otherwise.

Example 2 Photocatalytic Performance of PM/O2 for Water-Splitting (step (a))

In order to establish a benchma rk (control) for photocatalytic experiments of Au/Ti0 ₂ , photocatalytic water-splitting experiments were undertaken using platinised P25 nanopa rticles (1.0 wt%Pt/Ti 0 ₂ ) a nd platinised anatase nanoparticles (1.0 wt Pt/anatase).

In addition, va rious control experiments were also performed to ensure that the water vapour was the source of H ₂ production. Experiments were performed at 28 °C with 20 Torr of H ₂ 0 vapour and 280 Torr of Ar in the reaction cell at the start of the experiment, with 20.7 mW cm ^"2 of UV light irradiating the sample disc, eq uivalent to ~4.5 suns worth of UV intensity (assuming UV <400 nm) .

Selected results of these experiments are presented in Figure 1. The graph shows that both Pt-P25 and Pt-a natase began to produce Ha gas after the reaction cell was irradiated with UV light. The blank microfiber disc a nd Pt-P25 without water vapour show no production of H2 after irradiation with UV light, indicating that there is only H2 prod uction when Pt is present on T1O2 or when Pt-Ti0 ₂ has access to water vapour. Non-platinised P25 and anatase do not show any measurable levels of H2 production (not shown).

For all Ti0 ₂ samples, there is also production of C0 ₂ , but no measurable levels of 0 ₂ prod uction. This is a consequence of the well-known capacity for T1O2 to photo-degrade carbonaceous species in the presence of Oz.

Pt-P25 and Pt-anatase have average i production rates of 77.1 ± 9.9 and 45.6 ± 12.7 μηΊθΙ hr ¹ g ^"1 cm ^"2 , respectively. These results show the well-known effectiveness of Pt co-catalysts in enabling TiC to photocata lytica lly split water. It has been widely accepted that this is due to decreased electron- hole recombination by allowing for greater charge separation via migration of the photo-excited electron to Pt. The unplatinised samples do not produce any notable amounts of H ₂ as the rate of electron-hole recombination is too high to afford a ny detectable levels of H ₂ , as Ti0 ₂ cannot split water photocatalytica lly without co-catalysts. The increased performance observed for Pt-P25 compared with Pt-anatase could be due to the mixed polymorphs of anatase, rutile, and amorphous T1O2 present in these nanoparticles, which has been demonstrated to provide a greater degree of charge separation during photo-excitation, as well as possible synergistic effects between anatase and rutile.

Example 3 Effects of Sample Exposure to Vacuum

Over the course of running control and benchmark experiments, it was discovered that the photocatalytic performance of the catalysts was improved when they were prepared under vacuum for an extended period. Examples of the difference in peak H2 production rates for samples exposed to vacuum for 10 minutes, compared to those exposed to vacuum for 12 hours, are shown in Figure 2. This effect is most pronounced for the Au/TiC samples, such as Aug/acid-washed P25, which has a H ₂ production rate of 166.9 ± 42,3 μιηοΙ hr ^-1 g ^"1 cm ^'2 for those samples that were exposed to vacuum for 10 minutes, compared with 511.4 ± 51.1 pmol hr ¹ g ^"1 cm ^"2 for those that were exposed for 12 hours.

Given that exposing a sample in the reaction cell to vacuum for 12 hours prevented the use of the experiment for other samples, attempts were made to prepare samples under vacuum in a secondary stainless steel cell preparation cell, evacuated overnight, using the same vacuum line as the reaction cell. This secondary cell did not have the features of the primary reaction cell and was only used for the preparation of samples under vacuum. The samples would then be transferred from the secondary cell and into the main reaction cell as rapidly as possible, taking approximately 5 minutes for sample changeover. However, this still resulted in decreased catalytic performance due to the brief exposure to the atmosphere during sample transfer. There must therefore be some effect on the catalysts after exposure to an oxidising environment, even for a short period, compared with those samples that were evacuated within the reaction cell overnight. After this discovery was made apparent for a number of samples, all future samples were prepared for photocatalysis experiments by placing them in the reaction cell and evacuating the cell overnight, then performing experiments without exposing the sample to the atmosphere. Only those samples that have been prepared in this way have been included in the results presented herein. This is similar to most literature studies that undergo rigorous sample preparation procedures, such as extended flushing of reaction cells with Ar or baking samples under UHV for prolonged periods.

Exam pie 4 Effects of clus ter size

Once the size of the nanociusters begins to increase over approximately 2nm, the activity of the photocatalyst appears to decrease. It is thought that as the size of the Au nanocluster increases, the energy levels required for hydrogen production begin to match those of the substrate. This can be seen in Figures 15a-h. Various Au nanoparticles samples, with a size of approximately 3-5 nm in size respectively, and which have had various pre-treatments were tested. For the various Au nanoparticles, pre-treatment before hydrogen production includes no pre-treatment (Figure 15a, f), calcining at 200 °C followed by vacuum (Figure 15b-d), calcining at 200 °C (Figure 15e), and calcining 200 °C in the presence of oxygen (Figure 15g, h). The rate of hydrogen production was usually less than approximately 160 μηιοΐ hr ¹ g ^-1 cnrr ² .

The Auioi nanoclusters used in the following experiments have a size of approximately 1.4 nm and have a much increased hydrogen production yield.

Example 5 Production of COz and Consumption of Oi

During the water-splitting photocatalysis experiments, the increase in H ₂ present in the reaction ceii upon UV irradiation was accompanied by an increase in CO2, and a decrease in O2, as shown in Figure 3. It should also be noted that the increased catalytic performance observed between different samples manifest as both an increase in the H ₂ production rate and an increase in the CO2 production rate.

Studies that deal with the photo-oxidation of organic contaminants present on the Ti0 ₂ surface have shown the production of CO ₂ , and this has been suggested to be due to photo-activated oxygen. These studies have shown that even after rigorous steps were taken to clean the Ti0 ₂ surface, that even under UHV conditions, there are sti!l carbon contaminants present, which react with oxygen in the system when irradiated with UV light to produce C0 ₂ (vide infra).

One of the most likely sources of carbon in the reaction cell is that of adventitious carbon. This is usually a thin layer of carbonaceous molecules that are found on the surface of any material or vacuum system exposed to the atmosphere. It consists primarily of short chain hydrocarbons and small amounts of single and double bonded, functiona!ised groups

Example 6 Photocatalytic Performance of Au Clusters on TiOifor Water-Splitting (step (a))

Aua, Au ₉ , and Auioi clusters were supported on P25 and anatase nanoparticles with various treatments as summarised in Table 1. Table 1: A table summarising the different pre- and post-treatments applied to the supported Au clusters used in photocatalytic experiments.

Cluster Support Pre- Post-Treatments Clusters Per Approximate Xy e Treatment Nanoparticie Coverage (%)

Aus Anatase None Untreated 25.56 5.21

0 ₂ 200 °C

0 ₂ + H ₂ 200 °C

AUg Anatase None Untreated 22.71 4.62

0 ₂ 200 °C

P25 Acid washed Untreated 14.31 4.13

Heat treated 200 °C

0 ₂ 200 °C

Auioi Anatase none Untreated 2.02 0.81

0 ₂ 200 °C

0 ₂ + H ₂ 200 °C

Anatase Acid Untreated 2.02 0.81 washed 0 ₂ 200 °C

0 ₂ + H ₂ 200 °C

P25 Acid Untreated 1.28 0.72 washed

Heat treated 200 °C 0 ₂ +

H ₂ 200 °C Table 2: A table sowing a summary of the key trends in ligand loss and agglomeration observed for Au8, Au9, Aull and AulOl on acid-washed P25 and pure anatase supports under the various post-treatment conditions.

Treatment Acid washed P25 Nanoparticle Supported Pure Anatase Nanoparticle Supported

Au8 A119 Aun AulOl Aug Au9 ulOl

Untreated •Virtually Unchanged ^• Small increase in particle size ^• Partial ligand removal ^• Partial ligand removal

•Half of clusters remain intact, while the •Unknown if agglomeration other half undergo partial agglomeration but likely

Washed at •Removal of a fraction of clusters from the surface •Unknown changes to size N/A

100 °C •A portion ofclustcrs remain virtually unchanged •Loss ofa signi ficant amount of clustcrs

•Some removal of ligaiids from the surface

•No signi ficant agglomeration •No Au-0 formation

•Formation of Au-0 bonds, to a greater extent foi' Au ] j

Healed at •Agglomeration ofa poitton ofclustcrs while still Sigand- ; -Further agglomeration ofclustcrs, but N/A

ticulate in nature

200 "C protccted (these arc still smaller than AulO ! ) ■ still nanopar

•Other portion of clusters lose some ligands and form : -Removal of some ligands and formation

Au-0 and agglomerate to larger particles ^: ofAu-O bonds

•Of the portion thai loses ligands, some clusters may not

agglomerate

•Removal of ligands is less effective for Au9

Calcined 0 ₂ at Increased agglomeration and size distribution •Removal of ligands ^• Removal of ligands •Removal of ligands has progressed 200 °C and agglomeration and agglomeration has further

has progressed progressed further • Unknown if agglomeration occurs, further •No fraction ofclustcrs but likely

^• Half of clustcrs maintain their size

maintain their size

Calcined 0 ₂ /H ₂ N/A ^• Complete ligand removal •Complete ligand removal at 200 °C ^• Agglomeration has progressed further, no •Unknown if agglomeration evidence ofany clusters remaining intact. progresses further, but likely.

Table 2 summarises the key changes to the physical properties of these catalysts due to the various treatments. In general, there is a trend of ligand loss and agglomeration with successively harsher post-treatment conditions. This effect is far more pronounced for clusters supported on pure anatase nanoparticles than on the acid-washed P25 nanoparticles, showing the strong effect of acidic pre-treatment on the interaction between the Ti0 ₂ surface and Au clusters. For samples on either support, there is general evidence for two cluster states after post-treatment, with one portion remaining unchanged, while the other undergoes some level of agglomeration.

Example7: Photocatalytic Performance of the Au ₈ cluster

The peak H ₂ production rates for Aus/anatase with various treatments are shown in Figure 4. The Aus/anatase samples have peak H ₂ production rates of 17.92 ± 3.22, 51.74 ± 5.17, and 71.12 ± 7.11 μηιοΙ hr ¹ g^ crrf ² for the untreated, calcined at 200 °C under 0 ₂ , and calcined at 200 °C under 0 ₂ +H ₂ treatments, respectively. There is a clear improvement in H ₂ gas production as the clusters are treated under successively harsher conditions.

Calcination at 200 °C under 0 ₂ for Aus/anatase results in almost complete removal of ligands and agglomeration (See Table 2), while harsher calcination under 0 ₂ followed by H ₂ results in complete removal of ligands. Given this information, it is the exposed Aug clusters that are more effective catalysts for photocatalytic water-splitting, compared to untreated Aus/anatase, of which there is only partial removal of ligands. However, the loss of ligands may not be the primary cause of increased catalytic activity, given that these calcination treatments bring about agglomerated clusters, which do not maintain their Aus size. It could therefore be argued that the small size of the Au ₈ clusters are not beneficial for photocatalytic water-splitting, with larger Au nanoparticles on the anatase surface yielding the best catalytic environment. This is further supported by the deposition of Aus on anatase without any treatment, which also results in some loss of ligands, and a small fraction of Aus clusters agglomerating. Given the low, but still present catalytic activity of these untreated samples, this is further evidence that the agglomerated Aug clusters are the catalytically active sites.

Example 8: Photocatalytic Performance of the Aug cluster

The average H ₂ production rates for Aug supported on anatase and acid-washed P25 nanoparticles with various treatments are shown in Figure 5. The Au ₉ /anatase samples have H ₂ production rates of 33.5 ± 3.35 and 112.9 ± 12.3 μιηοΙ hr ^"1 g ^"1 cm ^"2 for untreated and calcined under 0 ₂ samples, respectively. The acid-washed P25 supported samples yield H ₂ production rates of 82.7 ± 8.27, 511.4 ± 51.1, and 75.3 ± 7.53 μιηοΙ hr ¹ g^ cm ^"2 for the untreated, heat treated under vacuum, and calcined under 0 ₂ samples respectively. There is a clear improvement in productivity for the Aug/anatase clusters after calcination under an O2 atmosphere, whereas for the Aug/acid-washed P25 clusters, there is a large increase in performance after heat-treatment at 200 °C under vacuum, followed by a decrease in performance after further calcination at 200 °C under an O2 atmosphere.

Calcination at 200 °C under O2 for Aug/anatase results in a large degree of ligand removal and agglomeration, with only a fraction of the clusters maintaining their size. Given that this support is not acid-washed, this effect is likely more severe than the same treatment on the acid-washed P25 nanoparticles. This treatment for the anatase-supported Aug clusters yields a higher production rate than the untreated or calcined under O2 at 200 °C treated, acid-washed P25 supported Aug clusters, possibly due to the increased size and removal of ligands.

It is therefore interesting that the untreated Aug/acid-washed P25 has a greater performance than untreated Aug/anatase, as the former results in virtually no change in the size or ligand coverage of the Aug clusters after they are supported. Heat treatment at 200 °C results in agglomeration of a portion of the Aug clusters while still ligand-protected, while the other portion lose some ligands, forming Au-0 bonds, and begin to agglomerate. There is also evidence that of the portion that loses ligands, some may not agglomerate. Further calcination at 200 °C under 0 ₂ for Aug/acid-washed P25 results in increased agglomeration according to HRTE , but there is no XPS data to provide details about ligand removal or bond formation between the cluster and the surface. Without wishing to be limited by theory, it is speculated that the decrease in performance observed for this treatment could imply that there is an ideal size for the Au clusters when supported on acid-washed P25, whereby a small amount of Aug agglomeration and ligand loss is necessary for ideal performance. Alternatively, it could be the small portion of Aug clusters that have lost ligands without

agglomerating that are the most effective, as these are most likely lost with further calcination at 200 °C under O2. It is difficult to determine if the large differences in performance between the anatase and P25 series is due to the pure vs mixed polymorph nature of the support, or if it is due to the acid-wash pre-treatment, without further characterisation studies.

Example 9 Photocatalytic Performance of the Auioi cluster

The average H ₂ production rates for Auioi supported on anatase and P25 nanoparticles with various treatments are summarised in Figure 6. For the Auioi/untreated anatase series, the production rates are 50.93 ± 13.3, 94.3 ± 15.7, and 190.4 ± 29.6 μηιοΙ hr ¹ g ^"1 cm ^"2 for the untreated, calcined under 0 ₂ , and calcined under O2 and H2 samples respectively. For the Auioi/acid-washed anatase series, the production rates are 534.8 ± 53.5, 112.2 ± 11.2, and 122.7 ± 11.2 μιηοΙ hr ^"1 g ^_1 cm ^"2 for the untreated, calcined under 0 ₂ , and calcined under 0 ₂ and H ₂ samples respectively. For the Auioi/acid- washed P25 series, the production rates are 238.3 ± 16.5, 437.5 ± 43.7, and 188.8 ± 34.0 μηηοΙ hr ¹ g ^"1 cm ^"2 for the untreated, heat treated, and calcined under 0 ₂ samples respectively.

The trend of increasing H ₂ production for Auioi/untreated anatase can be attributed to the increased amount of ligand removal and possible agglomeration under successively harsher calcination conditions. This effect is the same as that discussed previously for both Aus and Au ₉ , supporting the hypothesis that removal of ligands and certain cluster size improves the catalytic performance of anatase-supported Au clusters.

For the Auioi/acid-washed P25 series, there is a clear maximum in performance for the heat-treated sample, followed by a decrease with harsher calcination under 0 ₂ , similar to the trend observed for Aug previously. Nevertheless, there is no evidence for a bimodal distribution of clusters; instead, all clusters have ligands removed and agglomerate, while forming Au-0 bonds to the surface. Further calcination under 0 ₂ results in a decrease in H ₂ production, with an increase in particle size evidenced by H TEM. Also of note is that the untreated samples are more effective than those supported on pure anatase, but less effective than those supported on acid-washed anatase. There is evidence to suggest that support of untreated Auioi on acid-washed P25 results in virtually no change to the size or ligand coverage of the clusters.

For the Aum clusters supported on acid-washed anatase, the highest H ₂ production rate is observed for the untreated Auioi sample, which is the highest of ali samples in this series. The H ₂ production rate drops significantly after harsher calcination treatments. There is no characterisation data available for this series of clusters, therefore it is unknown if the untreated Auioi clusters are maintaining their size after deposition on acid-washed anatase, similar to what occurs for untreated Auioi on acid-washed P25, or if there is partial removal of ligands as seen for untreated anatase supports. Given that pre-treatment of the Ti0 ₂ surface with acid should help to reduce

agglomeration, the former seems more likely. If this is the case, then there is a clear performance advantage to keeping the Auioi cluster intact on the acid-washed anatase surface. Nonetheless, there is the possibility of a small amount of ligand removal (given the likelihood of this occurring on anatase) comparable to that observed for heat-treated Auioi on acid-washed P25, resulting in the similar production rate observed for clusters that maintain size, but have partial ligand removal. This will need to be confirmed with further characterisation experiments in the near future. Example 10 Overall Comparisons of the Photocatalytic Performance Between Clusters

Figure 7 shows a comparison between Au ₈ , Au ₉ , and Auioi clusters supported on anatase nanoparticles. Similar trends are observed for all three clusters as successively harsher post- treatments are applied. When samples are calcined under an 0 ₂ atmosphere, their H ₂ production rate increased compared to their untreated counterparts. When samples are calcined under 0 ₂ and H ₂ , a harsher and prolonged calcination, their H ₂ production rate is increased beyond that of samples calcined under O2 alone. There is no data available for Aug calcined under 0 ₂ and H2, although it can be assumed that it would follow the same trend as the other clusters, given that Aug calcined under 0 ₂ has a production rate within experimental error of the production rate for Auioi calcined under 0 ₂ .

As shown in Table 2, these successively harsher treatments result in increasing ligand removal and agglomeration for all clusters on the anatase surface. It is clear that it is the large, unligated Au particles that have lost their defined size, which are the most effective photocatalysts on the unwashed anatase nanoparticles. Since the overall performance also increases when comparing the set of Aue to Aug to Auioi samples, this is further evidence that it is the largest Au particles that are the most effective photocatalysts on this support.

Figure 8 shows the similar trends in H2 production rates for both clusters after treatment, whereby 200 °C heat treatment of the clusters results In a large increase in performance compared to the untreated samples, followed by a decrease in performance for the calcined under 0 ₂ samples.

For the untreated samples, it is known that the clusters remain virtually unchanged after being supported on acid-washed P25; therefore, the intact Auioi clusters are more effective photocatalysts for water photolysis than Aug. This could be due to their larger particle size, similar to the effect observed for Auioi supported on pure anatase nanoparticles.

Comparison between these two clusters also reveals that the untreated Auioi clusters do not form any Au-0 bonds with the surface; untreated Aug has a portion of clusters forming Au-0 bonds with the surface according to XPS, while untreated Aug has a portion of clusters forming Au-0 bonds.

The production rate of the heat treated samples are within the experimental error of each other, and the size measurement by the H TE are also within sampling error of each other (2.4 ± 1.7 vs 3.2 + 1.7 nm for Aug and Auioi, respectively). Therefore, the similar production rate measured for these two clusters on acid-washed P25 with the same treatment could be because the two samples are of similar size after agglomeration, while still being protected by a comparable number of ligands.

The large drop in production rate for Aug calcined under O2 is surprising given that the size of the nanoparticle is now the same as that of heat-treated Auioi according to HRTEM (3.1 ± 2.1 vs 3.2 ± 1.7 nm for Aug and Auioi respectively). There is no XPS data for this treatment, but it could be assumed that the extent of Au-0 bond formation has increased, following the increase observed for the heat- treated samples. Since untreated Auioi has no Au-0 bonds and performs better than untreated Au ₉ with Au-0 bonds, and that Auioi calcined under 0 ₂ may have an increased amount of Au-0 bond formation, it is feasible that high levels of Au-0 bond formation is detrimental to the photocataiytic performance. This is further evidenced by Aum calcined under O2 performing worse than the untreated Auioi samples.

Example 11 Photocataiytic Performance of Au¾ Au ₉ , and Auioi Compared to Pt-Ti0 ₂

Given that Ti02 is cheap and relatively abundant, while both Au and Pt are expensive elements in the current marketplace, it is fair to assume that the major cost of these catalysts would come from the procurement of these two rare elements. Comparison between the production rates for 1.0 wt% Pt- P25 and 1.0 % Pt-anatase to that of the 0.17 wt% Au supported on P25 or anatase nanoparticles can be made by normalising for the amount of precious metal present in the catalyst instead of by the total mass of the catalyst; these H2 production rates normalised by precious metal mass are shown in Figure 9.

This comparison shows the increased efficacy of AU/T1O2 compared to typical Pt-Ti02 photocatalysts. For the anatase-supported series, Auioi calcined under O2 and H2 is ~20 times more effective than Pt- anatase for the same amount of precious metal present in the catalyst, while for the P25 series, heat-treated Aug is ~66 times more effective than Pt- P25 for the same amount of precious metal present in the catalyst. Even the least effective Au catalyst in the series, untreated Au ₈ on pure anatase, is more effective than both Pt- P25 and Pt-anatase when compared using this normalisation scheme.

Conclusions from Examples 1 to 11

Benchmark photocataiytic water-splitting experiments were undertaken using Pt-P25 and Pt-anatase nanoparticles to ensure the newly designed experimental apparatus was performing adequately. During these studies, it was discovered that there Is a decrease in the photocataiytic performance of samples when repeating experiments using the same catalyst material. Further preliminary studies of AU/T1O2 found a similar effect of performance degradation. This degradation in the performance of samples was accompanied by a colour change in the samples. The degradation and colour change of samples was attributed to accumulation of carbon deposits during the oxidation of organic compounds, and may be related to the known photo-induced agglomeration effects of ambient light on the Au ₉ and Auioi clusters evidenced by XANES and HRTEM data. Improved performance was also observed for ail samples prepared in the reaction cell with 12 hours of exposure to vacuum, compared to relatively short vacuum exposure times of 10 minutes.

The production of H2 from photocatalytic water-splitting experiments was accompanied by the production of C0 ₂ and consumption of 0 ₂ . The C0 ₂ by-product arises from the well-known capacity for TiC>2 to photo-oxidise organic contaminants, and consumes the stoichiometrically evolved O2 from the water-splitting reaction throughout the experiment. The source of carbon in the reaction cell is most likely from unavoidable adventitious carbon that is present in all vacuum systems and samples exposed to atmosphere, in addition to the possible contribution from oil back-streaming from the rotary pump. Various carbon based sealant material used in the reaction cell and adsorbed CO that is difficult to evacuate during sample preparation may also contribute to the source of carbon.

O2 present in the reaction cell at the beginning of the experiment due to low vacuum is likely rapidly consumed by quenching defect states within the Ti0 ₂ nanoparticles and by photo-adsorption of 0 ₂ to the T1O2 surface over the initial hour of experiments. This initial O2 presence could also include O2 molecules adsorbed to the T1O2 surface at ambient temperature, or those adsorbed to the walis of the reaction cell. The formation of surface Oz ^~ and Of species during this period by molecular 0 ₂ likely behaves as electron traps or hole scavengers after photo-excitation, increasing electron-hole separation, which could explain the decrease in both H ₂ and CO2 production after the excess 0 ₂ in the reaction cell is consumed.

Example 12 RU3 nanoctusters on titania

Ru clusters have interesting properties when it comes to catalysis and these are mostly unexplored. The materials explored in this example are ligand stabilised clusters on a titania support. All the experiments were conducted at 2 bar with a 4:1 ratio of H2 to CO2. All Ru clusters were loaded at 0.17% on titania.

For Ru3 a series of experiments were performed testing different H ₂ calcination temperatures. A reactor was filled to 1.5 bar with H ₂ , and heated to the target temperature at a rate of 10 °C per min. A series of different reaction temperatures were tested, a series of different masses on a substrate were tested and finally a series of varying evacuation techniques. Figure 10 shows the production rates at varying calcination temperatures for CO, methane and ethane (note the axis is on the right for CO). For each of these samples XPS data was also gathered, and saw partial oxidation at higher calcination temperatures. The species of u are yet to be clearly identified but it is known from XPS that a RuO forms. . Initial tests were carried out over several hours with Ru nanoclusters on titania. The Sabatier mix used was 3:1 ratio of H2 to CO2 with the aim to produce methanol.

After identifying a consistent calcination temperature, a series of samples at different reaction temperatures were explored. Surprisingly relatively low temperatures are required, contrary to literature where values of 300 °C and above are reported. Figure 11 shows there is a clear saturation after 220 "C, ideal temperature being 250 °C.

Example 13 Ru ₄ nanoclusters on titania

At a calcination temperature of 200 °C and a reaction temperature of 250 °C the gases produced are 379 μηΊθΙη ^-1 g ^"1 of methane, 4649 μπιοίΓΓ ¹ of CO and 149 μι ^η οΙΙτ ¹ of ethane.

Ruthenium nanoparticles at a 3% loading produced in the range of 2000-3000 ηΊθΐΓΓ^ ^"1 of methane, but had 20 times more Ru than the cluster samples. Production rate normalised to Ru mass shows that Ru clusters out-perform the ruthenium nanoparticles by almost 4 times as much. .

To compare Ru ₄ with Ru ₃ nanoclusters, the same series of hydrogen calcinations was completed at the same temperature. These are plotted in Figure 14. CO production was similar for both, but RU3 had a general better production rate for methane and ethane.

Example 14 The effect of the thickness of the photocatalyst in step (b)

Experiments were performed in order to determine the effect of the thickness of the deposited photocatalyst on a silicon wafer substrate and whether it had an effect on the hydrocarbon gas production rates. For this experiment the weight of catalyst on the wafer was determined. Figure 12 shows that lower loadings on the Si wafer produced more CO, decreasing with weight. Whilst methane was pretty stable throughout, yet still increasing slightly as the mass increased.

Prior to this, all samples were left overnight in the reactor pumped overnight using a rotary vane pump; reaching pressures of approx 1.10 ^"2 mbar. A series of different conditions were tested, as shown in Figure 13. After each evacuation, the samples were calcined at 200 °C in H ₂ at 1 bar. The reaction temperature was 250 °C.

Example 15 Photocatalytic Studies ofAuRuz Deposited upon Anatase TIOi The cluster was deposited upon anatase T1O2, (hereinafter referred to as "AURU3-T1O2") and was evaluated for photocatalytic solar fuel production in the gas-phase, using a heterogeneous batch reactor apparatus. H2 and methane were detected as the major products of these reactions, with longer-chain hydrocarbons up to C4 species observed as minor products under certain conditions.

Example 15.1 Synthesis of Ru ₃ AuPPh ₃ fa-CI)(CO) ₁₀

The was synthesised as follows along the lines of the technique described in the paper entitled Synthesis and Structural Characterization of a New Ruthenium-Gold Cluster Complex: Inorganic Chemistry, Vol. 23, No. 5, (1984). In typical synthesis, 310 mg of u3(CO)i2 and 240 mg of AuPPh ₃ CI were dissolved in 50 mL of dry dichloromethane. The solution was stirred and refluxed (at 50°C) under N ₂ atmosphere overnight. After the reaction mixture was cooled to room temperature, silica gel 60 was added. The solvent was removed under vacuum for about an hour. The reaction mixture was chromatographed on silica gel 60. Elution with toluene-petroleum ether (1-1) afforded yellow band of Ru3(CO)i2 - Further elution with pure toluene afforded violet band of After solvent removal in vacuo, dichloromethane- hexane (1-5) was added. The solvent was removed under reduced pressure using rotary evaporator to obtain the crystals of the violet solution.

Drying anatase

A 12 g (12.0749 g) of anatase was dried in vacuo at 200°C for 5 hours with stirring. After cooling to room temperature, dry anatase was kept in desiccator overnight. A 1.9% weight loss was found according to moisture content.

Preparation of Ru3AuPPh ₃ (p-Ci)(CO)io stock solution

A 300 mg (302.13 mg) of Ru ₃ Au h3(p-Cl)(CO)io crystals was dissolved in small amount of dichloromethane. The solution was transferred into 25-mL volumetric flask following by making volume up to 25 mL by adding dichloromethane to obtain Ru3AuPPh ₃ CI(CO)io stock solution.

Deposition Ru3AuPPh ₃ (p-CI)(CO)w on anatase

X grams of dry anatase (see Table below) was suspended in 20 mL of dichloromethane in a Schlenk tube. After vigorously stirring (750 rpm) under N2 for 30 min, a Y μί of Ru3AuPPh ₃ Cl(CO)io stock solution was injected into the Schlenk tube. The solution was stirred (750 rpm) at room temperature under 2 for 90 min. The solvent was carefully removed under vacuum for around an hour to obtain deposited on anatase. The final catalyst was sonicated and then transferred into a sealed vial.

Example 15.2 Photocatalytic Benchmarks and Comparisons

AuRus/TiO _h Pt/TiOj & Bare TiOi

To assess the relative photocatalytic activity of Au U ₃ /TiO- towards CO2 reduction, two other materials were tested as benchmarks. Bare anatase T1O2 nanoparticles (Sigma-Aldrich) were tested as-purchased, as well as 1 wt% Pt nanoparticles deposited upon P25 Ti0 ₂ . Preliminary testing indicated that pre-treatment and reaction temperatures of 200 °C and a Pzoi-Pmo ratio of 3 were optimal for solar fuel production, and so these conditions were used for testing all samples. These will hereafter be referred to as "standard conditions".

H2 and methane were detected as major products, as well as trace amounts of C2-C3 alkane and alkene species. To ensure that the source of these products was indeed the reagent gases and not other carbonaceous contaminants, control reactions were conducted (i) without catalyst, in the presence of UV irradiation and reagent gases; (ii) in the absence of UV irradiation, with catalyst and reagent gases, and (iii) with catalyst under UV irradiation, but with argon buffer gas used in place of the reaction mixture. The former two tests yielded negligible amounts of the products of interest here, however the third blank test gave off trace levels of C1-C3 hydrocarbons. Further investigation showed that these residual hydrocarbon levels scaled linearly with the mass of catalyst used, and is likely due to the photo-induced breakdown of surface-adsorbed advantageous hydrocarbons or tigands. This background hydrocarbon production was normalized to total catalyst mass, and subtracted from all subsequent photocatalytic tests.

Photocatalytic production rates of methane and hydrogen by anatase Ti0 ₂ , AuRU3/Ti02 and Pt/Ti0 are shown in Figure 16A. Production rates of minor, longer-chain hydrocarbon products are then shown in Figure 16B. Deposition of the AuRu ₃ cluster improves the turnover of both major and minor hydrocarbon products relative to bare anatase, with methane production increasing by ~3 ^χ and ethane by a factor of two. In the case of the minor hydrocarbon products, the large uncertainties make interpretation of this data difficult. These errors are predominantly due to the extremely low levels of these products generated ("10-100 ppb), giving poor signal-to-noise ratios in the GC-FID. However, it can be said with confidence that generation of C ₃ products propane and propene is greater for the AuRu3-deposited sample than the bare titanta, as these production rates do not agree even under these large experimental errors. Additionally, negligible H ₂ production is observed over the bare a atase substrate, whereas the AuRu_/Ti0 ₂ system generates 68.5 μηησΙ hr^g ^"1 .

When comparing the activities of AuRu ₃ /Ti0 ₂ with Pt/TiOz, the latter shows higher production rates for both methane and hydrogen. This is unsurprising considering the greater content of co-catalyst on platinized sample than on the cluster-deposited sample (vide infra). However, greater amounts of saturated, longer-chain hydrocarbon products ethane and propane are generated by AURU3/T1O2 than Pt/Ti0 ₂ . The same cannot be said for unsaturated products; Pt-Ti0 ₂ generates more ethene than AuRu3/Ti0 ₂ , and the levels of propene generated by these two catalysts agree under experimental error. Evidently, Pt/Ti0 ₂ has a much higher selectivity for the formation of unsaturated hydrocarbon products than AuRu ₃ /Ti0 ₂ .

To account for the different loadings of co-catalyst upon AuRu3-Ti0 ₂ and Pt-Ti0 ₂ , Figures 17A and 17B show the same production rates discussed above, but instead normalized to the total precious metal content (Pt or Au/Ru) deposited upon the Ti0 ₂ nanoparticles. As can clearly be seen when compensating for total co-catalyst content, the cluster-based AuRu ₃ /Ti0 ₂ out-performs Pt/Ti0 ₂ in the generation of all products detected here. As platinum nanoparticles can be highly-active co-catalysts for C0 ₂ photo-reduction this is extremely promising for potentially further improving the efficiency of these reactions by use of sub-nanometer clusters instead of nanoparticles. However, as the cluster and nanoparticle co-catalysts compared here have entirely different elemental compositions, it must be acknowledged that the improved activity of AuRu ₃ /Ti0 ₂ may be due to selection of a more appropriate metal for this reaction as well as (or instead of) the benefits of sub-nanometer co- catalysts over nanoparticulate species. Example 16 A comparison between the catalysts: AuRus/TI02, RU3/T1O2 and RU4/HO2

Both the u3(CO)i2 precursor of AuRu ₃ and a H4 u4(CO)i2 cluster deposited upon Ti0 ₂ have previously been characterized as catalysts for water-splitting, ethene hydrogenation and Sabatier CO2 reduction. Therefore, these clusters were deposited upon anatase Ti0 ₂ and tested for C0 ₂ photo-reduction in the same manner as AURU3/T1O2. Figures 18A and 18B compare the

photocatalytic activities of these three cluster species deposited on Ti0 ₂ for solar fuel production under UV irradiation.

Methane and H ₂ are detected were major products across all three species, with C ₂ -C ₃ hydrocarbons as minor products. Both Ru ₃ -Ti0 ₂ and RU4-T1O2 also gave off low amounts of CO under UV irradiation; however, the quantities of this varied between scans (potentially due to de-ligation of carbonyl ligands convoluting this signal), and so have been excluded from analysis here. As is clearly evident in Figure 19, AURU3/T1O2 exhibits the highest rate of H2 production out of all three cluster-based systems, with Ru3/TiC>2 and RU4/T1O2 giving very similar production rates ~3x lower than this. In terms of methane production, Ru3/Ti0 ₂ again gives the lowest catalytic efficiency, with Ru ₄ /Ti0 ₂ and AuRu ₃ /Ti0 ₂ yielding higher production rates. Within the experimental errors of these measurements, both AURU3/T1O2 and RU4/T1O2 give comparable methane production rates. Similar trends are observed for most of the minor hydrocarbon products, with large experimental uncertainties in production rates of ethane, propane and propene again preventing further conclusions from being drawn. The exception to this is ethene, for which Ru ₄ /Ti0 ₂ generates appreciably greater amounts than either of the other two photocatalysts. However, hydrocarbon production rates for both AuRu3/TiC>2 and RU4/T1O2 are greater than for Ru ₃ /Ti0 ₂ . From the above results, it is evident that addition of a single metal atom to the U3 cluster - either gold or ruthenium - improves its overall rate of solar fuel production when supported upon anatase titania. However, substitution of gold for a ruthenium atom reduces the H ₂ generation rate to that of RU3-T1O2. Therefore, it is likely that the improved H ₂ production rate of AuRu ₃ /Ti0 ₂ is due to electronic interactions of the gold atom, possibly shifting the cluster's electronic structure to favor H ^' reduction or OH oxidation. As the improvement in hydrocarbon production occurs regardless of the added element, this is less likely to be a purely electronic effect, and could more plausibly be due to the increased cluster size allowing for more effective reagent binding.

To assess the potential competitive nature of the water-splitting and C0 ₂ photo-reduction reactions recorded here, these three cluster-based materials were also tested for water-splitting, i.e. in the absence of CO2. Figure 19 shows the relative H ₂ production rates from these materials observed under these conditions, alongside the same production rates with CO2 present. Within the experimental uncertainties here, all three clusters exhibit comparable hydrogen production rates both with and without CO2 present in the reaction mixture. This indicates that the presence of CO2 does not hinder the water-splitting reaction for these materials in any way; or conversely, that CO2 does not effectively compete with water for reduction at the photocatalyst surface.

Example 17 Optimising Pre-Treatment Temperature for Au-Ru catalysts

Heterogeneous catalysts and photocatalysts are generally pre-treated in situ before testing, in order to remove advantageous hydrocarbons and other surface-adsorbed species, or to open up catalyst active sites by removal of !igands. Many different techniques for this can be undertaken for example including ozone treatment, calcination in O2 or H2, and heating under a flow of inert gas. However, the inventors work demonstrates that many of these treatments have damaging effects upon clusters deposited upon T1O2, often causing aggregation to larger nanoparticles. This is undesirable in developing cluster-based catalytic materials, as it removes the size-specific nature of the cluster co-catalysts and complicates the correlation of catalytic activity to particle size. Hence, selection of an appropriate pre-treatment which removes adsorbed contaminants while still retaining intact clusters upon the surface for these materials is paramount.

Heating under vacuum was selected for catalyst pre-treatment, as it was shown to have the least aggregative properties of material treatments studied. All photocata lytic materials discussed above were heated to 200 °C while pumping under vacuum for 20 minutes. However, a range of temperatures from 50 - 250 °C (the limit of the apparatus) were also tested for AURU3/T1O2. Figures 20A and 20B show the dependence of photocata lytic activity upon this pre-treatment temperature for major and minor products, respectively. It should be noted that throughout these experiments, the reaction temperature was kept the same as the pre-treatment temperature. This was done to ensure that no samples were tested at higher temperatures than they were pre-treated, such that any observed change in chemical state could be ascribed to the treatment and not the reaction. Using constant, lower reaction temperatures of 50-100 °C could also have achieved this; however as can be seen below, reacting at these temperatures yielded little to no activity. Therefore, it must be acknowledged that the resultant photocata lytic production rates shown in Figures 20A & 20B are convoluted also by effects of reaction temperature. The dependence upon reaction temperature will be discussed in Example IS.

Despite this aforementioned convolution of pre-treatment and reaction temperature, an optimal pre-treatment temperature of 200 °C is evident for all products of interest. Production of methane and C2-3 hydrocarbons peak at this temperature, with near-baseline production at most other temperatures. Hydrogen production displays a different behaviour, increasing almost linearly from 50-200 °C, before then decreasing slightly beyond 200 °C. Therefore, a pre-treatment temperature of 200 °C a pears to be justified here, as it gives optimal activity towards all products of interest.

Example 18 Optimising Reaction Temperature for Au-Ru catalysts

To de-couple the effects of pre-treatment and reaction temperature, a series of photocatalytic tests on AuRu ₃ /Ti02 were run over a range of reaction temperatures, while keeping a constant material pre-treatment temperature. 200 °C was chosen as the standard temperature for pre-treatment due to the results from Example 17. Figure 21A shows the dependence of major product turnovers on reaction temperature, while Figure 21B shows the same for minor hydrocarbon products detected. Notably, in addition to methane and H ₂ , carbon monoxide was also detected as a major product by GC analysis at some reaction temperatures. As the CO parent ion at m/z - 28 overlaps with a fragmentation peak of C0 ₂ , RGA analysis could not be used to verify this quantification of CO.

Peak generation rates of most products of interest were observed at a reaction temperature of 150 °C. At near-ambient temperatures of 50 °C, no appreciable levels of hydrocarbon products are detected, with only H ₂ and CO observed in low amounts. The first hydrocarbons are detected at a temperature of 100 °C, corresponding to a decrease in CO production and a very slight increase in H ₂ production. Most products follow a very similar trend, increasing in yield from 50-150 °C, before decreasing on further raising the temperature to 200 °C. The exceptions to this are C3 products such as propane, which have such large experimental uncertainties that no reasonable conclusions can be made; and CO, which shows a decrease from 50-100 °C, before peaking at 150 °C like all other products. Several previous works in this field have postulated that CO may be an intermediate species in the reduction of CO2 to hydrocarbons. Hence, the more complex relationship with reaction temperature that CO exhibits here may be due to it being consumed to produce hydrocarbons such as methane or ethane.

A distinctly non-linear relationship with temperature is observed for all products of this

photocatalytic reaction. The reaction may be limited by adsorption-desorption effects upon the Ti0 ₂ surface, where the rate-limiting step is desorption of products at lower temperatures, and adsorption of reagent molecules at higher temperatures. Reacting at 150 °C may achieve an optimal equilibrium between reagent adsorption and product desorption. At lower temperatures, poor desorption of products or intermediates from water reduction could simultaneously limit the H ₂ production rate and proton transfer to C0 ₂ . As the reaction temperature then increases these reduced states of water would then be mobilized and more readily desorbed, allowing for formation of C-H bonds and giving greater H2 production rates. However, on rising above 150 °C the limiting factor could then become reagent adsorption, with the excess thermal energy in the system causing molecules to desorb from the T1O2 surface before completing photocatalytic transformations and hence decreasing overall production rates.

Conclusions from Examples 15 to 18

When comparing AURU3/T1O2 to u3(CO)i2 or H Ru(CO)i2 clusters deposited upon titania, the bimetallic-deposited species shows the greatest affinity for H2 production from water, while both fVU-based clusters show improved hydrocarbon turnover when compared to RurTiC>2. Optimal turnover is observed when the photocatalyst was pre-treated under vacuum at 200 °C and reacted at 150 °C and higher partial pressure ratios of H2O to CO2 improve hydrocarbon production rates. Optimal production of CO and H ₂ was observed at a reagent ratio of 1:1, and COarhhO ratios in the range of 0.5 - 4 gave peak hydrocarbon production.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in ail respects as iilustrative and not restrictive.

In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.