Field of the Invention

The present invention relates to an iridium-based oxygen evolution catalyst and its use as an oxygen evolution reaction (OER) catalyst, for example in a water electrolyser.

Background

The electrolysis of water to produce high purity hydrogen and oxygen can be carried out in both alkaline and acidic electrolyte systems and practical devices using both types of electrolyte systems exist as commercial products. Those electrolysers that are acid electrolyte-based typically employ a solid proton-conducting polymer electrolyte membrane (PEM) and are known as polymer electrolyte membrane water electrolysers (PEMWEs). A catalyst-coated membrane (CCM) is employed within the cell of a PEMWE, which comprises the PEM with two catalyst layers (for the anode and cathode reactions) applied on either face of the PEM. To complete the electrolysis cell, current collectors, which are typically metal meshes, are positioned either side of the CCM. Suitable CCMs for water electrolysers are described in WO2018/115821 (Johnson Matthey PLC). During water electrolyser operation the oxygen evolution reaction (OER) takes place at the anode and the hydrogen evolution reaction (HER) takes place at the cathode, approximated by the following eguations:

H ₂ O O ₂ + 4H ⁺ + 4e- (OER) 4H ⁺ + 4e- 2H ₂ (HER)

PEM fuel cells have a similar arrangement to a PEM water electrolyser but work in a different manner with the above reactions reversed. During operation of a fuel cell, oxygen is reduced at the cathode and hydrogen is oxidized at the anode, generating water and an electrical current.

The cathode catalyst used in both PEM water electrolysers and PEM fuel cells is generally a platinum catalyst, usually platinum on carbon. Iridium and iridium oxide are well known for their properties as excellent OER catalysts and are preferred materials for the oxygen evolution reaction on the anode side of a water electrolyser. Iridium-based OER catalysts may also be incorporated into fuel cell anodes, to improve cell reversal tolerance. PEM water electrolysers generally require a higher loading of OER catalyst on the anode side compared to a PEM fuel cell.

Due to the rising global demand for hydrogen and the scarcity of iridium, there is a need to develop OER catalysts with improved performance and/or equivalent performance using lower loadings of metal. One way to reduce the iridium content is to provide the catalyst on an inert support such as an inorganic oxide, for example as described in W02005/049199 (Umicore AG & Co KG).

It is known to produce iridium oxide by carrying out a precipitation reaction between an iridium salt and base to produce a mixture of iridium hydroxide and iridium oxide and converting the mixture to iridium oxide by calcination. The iridium oxide produced can vary from amorphous to crystalline (rutile structure) depending on the calcination conditions.

The crystallinity of iridium oxide has an important impact on performance of the catalyst in the OER. A recent summary of the factors is given in the paper “Effect of Base on the Facile Hydrothermal Preparation of Highly Active IrOx Oxygen Evolution Catalysts” ACS Appl. Energy. Mater. 2020, 3, 800-809. In general, increasing the annealing temperature leads to a more crystalline material. While crystallinity has some benefits, increasing crystallinity is generally accompanied by a reduction in surface area and a reduction in OER activity. The authors describe the preparation of amorphous iridium oxide by a modified hydrothermal method involving reaction between I rC hydrate and a base (U2CO3, LiOH, Na2COs, NaOH, K2CO3 or KOH) at a 1 : 8 molar ratio Ir : base. The choice of base had a strong influence on the morphology of the final material, with LiOH and U2CO3 bases producing a porous spongelike morphology with high surface area.

The PhD thesis “Catalysis to produce solar fuels: From the production of hydrogen via water splitting, to hydrogen conversion to methanol by its reaction with CO2” (J. Esquius, Cardiff University 2019), hereafter “Esquius”, describes some of the results reported in the paper ACS Appl. Energy. Mater. 2020, 3, 800-809 and additionally includes examples of iridium oxide prepared by a similar hydrothermal route but including a calcination step after precipitation and drying. A standard procedure for preparing iridium oxide catalysts is provided in section 2.2.1 of Esquius and involves the following steps:

(i) combining 1 mmol of IrC hydrate and 8 mmol of base (Li, Na or K hydroxide or carbonate) in 10 mL of deionized water;

(ii) stirring for 16 h at room temperature;

(iii) diluting with a further 10 mL of deionised water and heating at reflux for 3 h;

(iv) recovering the precipitate by filtration;

(v) washing the precipitate with 2 L of hot deionised water; and

(vi) drying the precipitate at room temperature overnight.

The resulting catalysts are formulated as inks by dissolving the catalyst in a solution comprising water, ethanol and Nation® and then sonicating. The inks are then applied onto a working electrode by drop-casting to produce a catalyst loading on the electrode of 100 pg cm ^-2 (see section 2.3.1). In Chapter 3 the author concludes that iridium oxides prepared using lithium carbonate or lithium hydroxide had enhanced activity compared to the other bases tested.

Esquius also investigates the role of lithium-doping on the OER activity of iridium oxide. To produce a lithium rich structure, a series of materials were produced using a modified procedure by omitting the washing step (v) in order not to wash out excess lithium carbonate (see section 5.2). The dried, unwashed material was annealed at 500 °C for 3 hours and then washed with 2 L of deionised water after calcination. This material offers a good balance between stability and activity in the oxygen evolution reaction.

There is a need for alternative materials having improved stability and which are simpler to manufacture at scale.

Summary of the invention

The present inventors have found that the procedure reported in Esquius has a number of downsides when carried out at large scale. In particular, the step of recovering the precipitate is very slow and large amounts of water are needed for the washing step. The process is therefore time intensive and produces a large amount of aqueous effluent. The large amount of effluent is particularly problematic because iridium is a scarce metal and needs to be recovered from the effluent; the more dilute the solution the greater the cost to recover the metal.

The present inventors have found that both of the above problems can be solved by reducing the amount of lithium salt used in the precipitation step. In Esquius the procedure uses 8 mmol of U2CO3 for each mmol of IrCh, i.e. the molar ratio of Ir : Li is 1 : 16. The present inventors have found that this ratio can advantageously be dropped to 1 : 1 to 1 : 10. Reducing the equivalents of lithium carbonate simplifies the filtration step, reduces the amount of water needed for the washing step, which in turn makes the recovery of iridium from the filtrate more economical.

Surprisingly, these changes also produce a catalyst which has different chemical and physical properties from the catalysts described by Esquius. The relative proportion of I rC>2 phase is increased and the catalyst has a higher BET surface area compared to a catalyst prepared by the Esquius method. The catalyst prepared by the method of the present invention offers better activity in the oxygen evolution reaction.

In a first aspect the invention relates to an oxygen evolution catalyst comprising a mixture of I rC>2 and LisIrCU phases, wherein: the XRD pattern of the catalyst includes reflections at 20 = 18° and 43° for U3I rCU; the XRD pattern of the catalyst includes reflections at 20 = 28° and 54° for I rC>2; and the N2-BET surface area of the catalyst is > 50 m ² /g.

The catalyst may be formulated as an ink for application to produce the anode catalyst layer of a catalyst coated membrane of a water electrolyser.

The catalyst may also be incorporated into a fuel cell, particularly in a fuel cell anode, to improve cell reversal tolerance.

In a second aspect the invention relates to a method of manufacturing an oxygen evolution catalyst, comprising the steps of:

(i) preparing an aqueous mixture comprising an iridium salt and a lithium salt, wherein the molar ratio of Ir : Li in the mixture is from 1 : 1 to 1 : 10;

(ii) heating the mixture at a temperature above 80 °C to effect a precipitation and produce a precipitate; (iii) isolating the precipitate;

(iv) calcining the precipitate;

(v) washing the precipitate with deionised water.

The oxygen evolution catalyst is preferably a catalyst according to the first aspect of the invention.

In a third aspect the invention relates to an oxygen evolution catalyst produced or producible by the method according to the second aspect.

In a fourth aspect the invention relates to a catalyst coated membrane comprising an oxygen evolution catalyst according to the first or third aspects.

In a fifth aspect the invention relates to the use of a catalyst according to the first or third aspects as the anode catalyst in a water electrolyser.

Description of the Figures

Figure 1a is an overlay of the XRD patterns of catalysts CE1 , CE2 and E1.

Figure 1b is an overlay of the XRD patterns of catalysts CE1 , CE2 and E1 compared with Ir.

Figure 1c is an overlay of the XRD patterns of catalysts CE1 , CE2 and E1 compared with lrC>2 (rutile).

Figure 1d is an overlay of the XRD patterns of catalysts CE1 , CE2 and E1 compared with LisIrCU.

Figure 2 shows a plot of jmass / A g ^-1 i _r vs E / V for catalysts CE1 , CE2 and E1.

Figure 3 shows the cyclic voltammograms for catalysts CE1 , CE2 and E1.

Detailed Description

Any sub-headings are included for convenience only and are not to be construed as limiting the disclosure in any way. As used herein, the term “room temperature” means ambient temperature of 15-25 °C.

Manufacturing method

In one aspect the invention provides a method of manufacturing an oxygen evolution catalyst, comprising the steps of:

(i) preparing an aqueous mixture comprising an iridium salt and a lithium salt, wherein the molar ratio of Ir : Li in the mixture is from 1 : 1 to 1 : 10;

(ii) heating the mixture at a temperature above 80 °C to effect a precipitation and produce a precipitate;

(iii) isolating the precipitate;

(iv) calcining the precipitate; and

(v) washing the precipitate with deionised water.

In step (i) a mixture is prepared comprising an iridium salt and a lithium salt. The mixture may be a solution or a suspension depending on the concentration of the iridium salt and the lithium salt. Due to the low solubility of lithium carbonate in water, the mixture is generally a suspension.

The iridium salt may be an iridium (III) salt (e.g. IrCh) or an iridium (IV) salt (e.g. ^IrCle). The iridium salt is preferably an iridium (III) salt such as iridium (III) chloride.

The lithium salt is preferably lithium carbonate.

The order of addition of components in step (i) is not thought to be particularly important. An exemplary embodiment involves adding an aqueous solution of iridium trichloride to an aqueous slurry of lithium carbonate.

A variety of different species may be formed following calcination depending on the content of Li present in the precipitated material. The present inventors have observed peaks in the XRD pattern which can be attributed to Ir, lrC>2 (rutile) and LisIrCU. Without wishing to be bound by theory, LisIrOds thought to be a highly active species for the oxygen evolution reaction. This phase appears to be present in the XRD pattern of the material described in Esquius, although it is mistakenly characterised in the reference as “Li _x lrO2-hollandite”.

The molar ratio of Ir : Li in step (i) is from 1 : 1 to 1 : 10. Using these lower loadings of lithium salt allows easier filtration of the precipitate formed in step (ii), particularly when lithium carbonate is used as the lithium salt. It is preferred that the molar ratio of Ir : Li in step (i) is from 1 : 1 to 1 : 8, preferably 1 : 1 to 1 : 6, such as 1 : 2 to 1 : 6.

The concentration of iridium in the mixture in step (i) is typically 10 to 500 mmol/L. When manufacturing the material at scale it is preferred that the concentration of Ir is as high as is possible without negatively impacting the properties of the end product. Using a high concentration of iridium may help to reduce the amount of effluent produced from filtration and washing, which offers benefits to metal recovery. In a preferred embodiment the concentration of iridium in the mixture in step (i) is 50 to 500 mmol/L.

In the procedure reported by Esquius, the mixture of iridium trichloride and lithium carbonate is stirred for 16 h at room temperature before heating to reflux. The present inventors have found that extended mixing before the heating step (ii) is not necessary.

In step (ii) the mixture from step (i) is heated to effect a precipitation reaction. During this step a blue/black precipitate forms. The mixture is preferably heated to a temperature above 80 °C, preferably to a temperature of 80-120 °C, ideally at reflux. The mixture is preferably heated at this temperature for a period of from 15 mins to 6 hours, such as from 15 mins to 4 hours. Typically, the mixture is then allowed to cool to room temperature before commencing step (iii).

In step (iii) the precipitate is isolated. This may be achieved by filtration techniques, such as suction filtration.

It is preferred that no washing step is carried out between steps (iii) and (iv). Carrying out a washing step after step (iii) removes both lithium and chloride ions from the structure. It is desired to keep the lithium ions in the structure during calcination in order to produce the desired LisIrCU phase.

In step (iv) the precipitate is calcined to produce a more crystalline material. The calcination should ideally be carried out at a temperature of 400 to 600 °C. Outside of this temperature range the catalyst performance of the calcined catalyst rapidly drops off. A temperature of approximately 500 °C is preferred. A duration of 30 to 90 minutes is generally appropriate. Longer calcination times are not thought to provide any benefit and have the downside of higher energy costs. In step (v) the precipitate is washed with deionised water, ideally until the filtrate has a conductivity of < 100 pS. Conductivity can be measured by any suitable means, for example by a conductivity probe. The role of this step is to remove chloride residues from the precipitate, and to remove excess U2CO3. The solubility of lithium carbonate in water increases with decreasing temperature, and in some embodiments the deionised water used for washing is cooled to below room temperature. However, particularly when manufacturing at scale, it may be preferable and less costly to use water at room temperature.

The product of step (v) is an active catalyst for the oxygen evolution reaction and has an excellent balance between activity and stability (low Li and Ir losses).

The product of step (v) may be grinded to remove large agglomerates. For example, using a pestle and mortar or by milling.

Catalyst

In one aspect the invention relates to an oxygen evolution catalyst. The catalyst may be obtained or obtainable by the method of the present invention.

The catalysts of the present invention show peaks in the X-ray diffraction (XRD) pattern associated with both the LisIrCU phase and the I rC>2 (rutile) phase. The XRD pattern of LisIrCU has been described previously in the article “Approaching the limits of cationic and anionic electrochemical activity with the Li-rich layered rocksalt LisIrCU” Nature Energy, 2, 954-962 (2017). The presence of the LisIrCU phase can be confirmed by the presence of reflections at 20 = 18° and 43°. These peaks are intense for LisIrCU and are distanced from the peaks associated with metallic iridium and iridium oxide (rutile). The presence of the lrC>2 (rutile) phase can be confirmed by the presence of reflections at 20 = 28° and 54°. These peaks are intense for Irel and are distanced from the peaks associated with LisIrCU.

The catalyst has a N2-BET surface area of > 50 m ² /g. As used herein “N2-BET” means that the surface area is measured by the BET method using N2 as the adsorption gas. Surprisingly, reducing the ratio of Ir : Li in step (i) produces a catalyst having a higher BET surface area. For comparison, a catalyst made following the procedure in Esquius has a BET surface area of c. 34 m ² /g (see examples). It is thought that the increased surface area may be at least partly responsible for the better performance of catalysts according to the present invention.

In preferred embodiments the catalyst has a N2-BET surface area of > 50 m ² /g, preferably > 70 m ² /g, such as > 80 m ² /g or > 90 m ² /g.

In preferred embodiments the catalyst has a N2-BET surface area of < 150 m ² /g, preferably < 130 m ² /g, such as < 120 m ² /g or < 110 m ² /g.

In preferred embodiments the catalyst has a N2-BET surface area of 50 to 150 m ² /g, preferably 70 to 130 m ² /g, such as 80 to 120 m ² /g or 90 to 110 m ² /g.

The catalyst may be used as an oxygen evolution reaction catalyst in a water electrolyser, especially in the anode of a water electrolyser.

The catalyst may also be used in a fuel cell, especially in a fuel cell anode for the purposes of cell reversal tolerance. For a discussion of the use of I rC>2 materials in fuel cells for the purpose of cell reversal tolerance see WO2012/107738 (Johnson Matthey PLC).

Catalyst coated membrane

A catalyst coated membrane (CCM), also referred to as a polymer electrolyte membrane (PEM), comprises a membrane having an anode catalyst layer on a first face thereof and a cathode catalyst layer on a second face thereof.

The catalyst may be formulated as an ink, typically by dissolving or dispersing the catalyst in a mixture of a fluoropolymer (such as Nation®) and water. The inks may be applied on an ion exchange membrane to produce the anode catalyst layer of a catalyst coated membrane. The CCM may include additional components (e.g. recombination catalysts, reinforcements, multiple layers) as will be known to those skilled in the art.

Examples

Measurement of BET surface area

The sample was degassed to remove any adventitiously adsorbed species from the sample surface. A nitrogen adsorption isotherm was obtained by measuring the quantity of N2 gas adsorbed as a function of gas pressure at a constant temperature (using liquid nitrogen at its boiling point at one atmosphere pressure). The data is used to plot 1 I [V _a ((Po/P)-1)] vs P/Po for P/P _o values in the range 0.05 to 0.3, where V _a is the quantity of gas adsorbed at pressure P and Po is the saturation pressure of the gas. A straight line is fitted to the plot to yield the monolayer volume (V _m ), from the intercept 1/V _m C and slope (C-1)/V _m C, where C is a constant. The surface area of the sample is determined from the monolayer volume by taking the cross-sectional area of a nitrogen molecule at 16.2 A ² .

Comparative Example 1 (CE1)

CE1 was prepared following the procedure in Chapter 5 page 135 of Esquius.

IrCh (4.1 mmol, 1.5 g), U2CO3 (32.7 mmol, 2.42 g) and 41 mL deionised (DI) water were added to a 100 ml round bottom flask. The molar ratio of Ir : Li was approximately 1 : 16. The solution was stirred overnight (17 h) at room temperature. pH ~9.5, solution remained yellow/brown after stirring overnight. Another 41 ml of DI water was added and the solution heated to reflux for 3 h. During reflux, the solution turned dark blue almost immediately. The solution was allowed to cool to room temperature. The dark blue/black precipitate was recovered via vacuum filtration. The filtrate pH was 10.4. The residue was dried at 105 °C for 3.5 h. Yield was 2.09 g. A portion of this material was calcined at 500 °C for 3 h and then washed with ~2 L of DI water at room temperature.

The XRD pattern of catalyst CE1 is shown in Figure 1a. This corresponds closely to the XRD pattern reported for “Li _x lrO2-hollandite” in Figure 5.5 of Esquius.

Catalyst CE1 had a N2 BET surface area of 34 m ² /g.

Comparative Example 2 (CE2)

CE2 follows a similar method to CE1 but on a larger scale and with a shorter calcination time.

U2CO3 (305.4 g, 4.13 mol) was suspended in 5 L of DI. IrCh (191.0 g, 522 mmol) dissolved in 2.5 L of DI was added slowly. The molar ratio of Ir : Li was approximately 1 : 16. The suspension was then heated to reflux for ~ 3.5 hours. The suspension was then allowed to cool to ~ 35 °C then filtered. Filtration was slow and was allowed to continue overnight under gravity. The residue was allowed to dry in an oven at 105 °C for 96 h. The mass of recovered material was 278.5 g. The dried sample was transferred to an alumina crucible and then calcined (ramp from room temperature to 500 °C at 5 °C/min, then hold at 500 °C for 1 h). After cooling to room temperature, the catalyst was washed by vacuum filtration using room temperature DI until the filtrate had a conductivity below 100 pS. The material was then dried in an oven at 105 °C for 18 hours.

The XRD pattern of catalyst CE2 is compared against CE1 in Figure 1a. Both catalysts include a mixture of phases including LisIrC and I rC>2 (rutile).

Catalyst CE2 had a N2 BET surface area of 37.4 m ² /g.

Example 1 (E1)

U2CO3 (7.76 g, 0.105 mol) was suspended in 0.350 L of DI. IrC (19.4 g, 0.053 mol) dissolved in 0.125 L of DI was added slowly. The molar ratio of Ir : Li was approximately 1 : 4. The suspension was then heated to reflux for 4 hours. The suspension was then allowed to cool to ~ 35 °C then recovered by vacuum filtration. Filtration was slow. The residue was allowed to dry in an oven at 103 °C for 96 h. The mass of recovered material was 17.8 g. The dried sample was transferred to an alumina crucible and then calcined (ramp from room temperature to 500 °C at 5 °C/min, then hold at 500 °C for 1 h). After cooling to room temperature, the catalyst was washed by vacuum filtration using room temperature DI until the filtrate had a conductivity below 100 pS. The material was then dried in an oven at 105 °C for 18 hours.

The XRD pattern of catalyst E1 is compared against CE2 in Figure 1b. The XRD pattern of E1 also shows the presence of a mixture of phases including LisIrCU and lrC>2 (rutile), but the proportion of LisIrCU is reduced compared to CE2.

The catalyst had a N2 BET surface area of 96.0 m ² /g.

Figure 2 is a plot of jmass / A g ^-1 i _r vs E / V for catalysts CE1 , CE2 and E1. Catalyst E1 has a lower Tafel slope than CE1 , which means that E1 has improved mass activity than CE1. This is expected to be accentuated at the higher potentials (> 2V) relevant for water electrolysis at scale. Figure 3 compares the cyclic voltammogram of catalysts CE1 , CE2 and E1. The greater double-layer capacitance of E1 compared to CE1 and CE2 shows that E1 has a higher electrochemically accessible surface area, contributing to its improved activity >1.5 V.