Alkaline Electrolysers

The invention provides self-assembling and self-healing nanoparticle-based catalyst films that can be formed on the electrodes of alkaline electrolysers. Notably, provided is an electrocatalytic device comprising an electrolytic cell divided into anode and cathode chambers by an exchange membrane, said anode and cathode chambers being provided with electrodes immersed in electrolytic suspensions, wherein the electrolytic suspensions comprise surface charged catalytic particles forming self-healing films on the electrodes, and the use of the device as electrolyser, preferably for electrolysis of water.

Background of the Invention

The energy production from renewable sources faces a major challenge, its intermittency. New efficient energy conversion and storage pathways will pave the way for the ongoing change from fossil fuel-based to renewable energy production (Dincer, I.; Renew. Sust. Energ. Rev. 4: 157-175 (2000)). Storage of surplus electrical energy in the form of chemical bonds, especially hydrogen, is an attractive solution to level out fluctuations in energy production (Editorial. Hydrogen on the rise. Nat. Energy 1 : 16127 (2016)). The stored energy can then be recovered by oxidation of hydrogen in fuel cells when needed (Chu, S. & Majumdar, A. ; Nature 488: 294-303 (2012)). Electrochemical water splitting that consists of reduction and oxidation of water in separate compartments is a promising method for the conversion of electrical energy into hydrogen (Du, P. & Eisenberg, R. ; Energy Environ. Sci. 5 : 6012-6021 (2012)). The overpotential that is necessary to drive the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER) at cathode and anode, respectively, makes the use of efficient electrocatalysts indispensable. Therefore, enormous efforts are currently devoted to the development of highly active catalysts for both reactions (Galan- Mascaros, J. R. ; ChemElectroChem 2 : 37-50 (2015), Jiang, J. et al., J. Power Sources 278:445-451 (2015), McCrory, C. C. L. et al. ; J. Am. Chem. Soc. 135 : 16977-16987 (2013), Zou, X. et al. ; J. Am. Chem. Soc. 135 : 17242-17245 (2013), Hu, W.-H . et al . ; Carbon 100 : 236-242 (2016), Dai, X. et al . ; ACS Appl. Mater. Interfaces 8: 6439-6448 (2016), Masa, J. et al. ; Adv. Energy Mater. 6: 1502313(1 (2016), Konkena, B. et al.; Nat Com ms 7 : 12269 (2016), Ng, J. W. D. et al. ; Nat. Energy 1 : 16053 (2016)). Highly active materials often undergo structural changes during electrochemical reactions leading to instable forms of the initial material (Lutterman, D. A. et al .; J. Am. Chem. Soc. 131 : 3838-3839 (2009)). Hence, intrinsic stability of the catalyst as well as stability of the catalyst film (attachment of the catalytic material onto the current collector) are as important as catalytic activity for the envisaged technical application of catalysts, especially under the harsh electrolysis conditions. There are two general approaches of preparing catalyst films. I) Direct growth of catalysts on current collectors by electroless (Jiang, J. et al., J. Power Sources 278:445-451 (2015)) and electrochemical (Pintado, S. et al. ; J. Am. Chem. Soc. 135 : 13270- 13273 (2013), Mousty, C. & Walcarius, A.; J. Solid State Electrochem. 19 : 1905- 1931 (2015)) methods as well as vapor deposition methods (Peng, Q. et al. ; J. Vac. Sci. Technol. A 30 : 10803 (2012), Pickrahn, K. L. et al .; Adv. Energy Mater. 2 : 1269-1277 (2012)). And II) immobilization of a pre-synthesized catalyst powder (Geng, J. et al. ; ChemSusChem 8: 659-664 (2015), Di Tang et al. ; ACS Appl. Mater. Interfaces 6: 7918-7925 (2014)). Whereas the former may results in more stable catalyst films, the latter offers the possibility of exploring an immense number of powders of tuneable properties, particle sizes and shapes. Unfortunately, it is very challenging to immobilize catalyst powders, e.g. nanoparticles, on gas-evolving electrodes. Drop-coating is a widely used methodology for forming catalyst films from catalyst powder. This technique is based on the application of a suspension that contains catalytic particles and binder agent on the surface of a current collector. Evaporation of the solvent leads to formation of the catalyst film. However, films prepared by this method are not entirely stable under the harsh conditions of gas-evolving reactions resulting in catalyst detachment. In addition, the presence of binders in the catalyst film can interfere decreasing the initial activity of the catalyst (Wang, L. et al. ; ACS applied materials & interfaces 8: 16736-16743 (2016), Lu, X. & Zhao, C ; Nat. Commun. 6: 6616 (2015)). Thus, despite the increasing number of highly active catalyst powders reported for electrochemical water splitting, immobilization of powder onto the gas-evolving electrodes of electrolysers remains a challenge.

In CN-A-105714326 a suspension solution electrocatalytic device is described, the electrolytic cell of which is divided into anode and cathode chambers by a proton exchange matrix, said anode and cathode chambers comprising an anode sidewall and a cathode bottom, respectively, and said cathode chamber being provided with a microsphere suspension such as a molybdenum carbid gravity microspheres that intermittently touch the cathode material.

Short Description of the Invention

An innovative approach for the preparation of nanoparticle-based catalyst films for alkaline electrolysers was found . Instead of modifying the electrodes prior to assembly of the electrolyser, catalyst films are formed while the system is running . Pre-synthesised catalytic particles are directly added into the electrolyte reservoirs of an electrolyser forming catalyst/electrolyte suspensions. Pumping the suspensions through the running electrolyser leads to the formation of self- assembling catalyst films driven by electrostatics based immobilization of particles. The electrostatic interaction requires that the surfaces of electrode and particle are oppositely charged, e.g. particles with surfaces positively and negatively charged are immobilized at cathode and anode, respectively. Importantly, self-assembled catalyst films possess self-healing properties as long as particles are present in the electrolyte, which is a unique feature of this preparation technique. The proof-of-concept was demonstrated in a non-zero gap alkaline electrolyser using state-of-the-art catalyst powders, NiFe LDH and Ni _x B for anode and cathode, respectively. Catalyst films remained stable for at least three weeks at current densities ranging between 50 - 100 mA/cm ² . The invention thus provides

(1) an electrocatalytic device comprising an electrolytic cell divided into anode and cathode chambers by an exchange membrane, said anode and cathode chambers being provided with electrodes, wherein one or both of said electrodes are immersed in electrolytic suspensions, and wherein the electrolytic suspensions comprise surface-charged catalytic particles forming self-assembling and self-healing films on the electrodes;

(2) a method for preparing the device of aspect (1) comprising an electrolytic cell divided into anode and cathode chambers by an exchange membrane, said anode and cathode chambers being provided with electrodes, wherein the method comprises immersing said electrodes in electrolytic suspensions comprising surface-charged catalytic particles; and

(3) the use of the device of aspect (1) as electrolyser, including for electrolysis of water.

Short description of the Figures

Figure 1 : Electrochemical and visual evidence of the formation of self-assembling catalyst films. Drop in the cell voltage ( U _C eii) observed in a Ni//Ni electrolyser operating at I _app i= 50 mA/cm ² after addition of catalyst powder into the electrolyte (3 M KOH) for (a) NiFe LDH, (b) CoMn LDH, (c) Ni _x B and(d) Co _x P. Insets in the graphs are photographs of the corresponding electrode after disassembling the electrolyser. Concentration of the catalyst was about 1.3 mg/ml .

Figure 2 : Electrostatic nature of the formation of self-assembling catalyst films (a) Zeta potential of NiFe LDH, CoMn LDH, Ni _x B and Co _x P in 3 M KOH (N = 3 for each one) , (b) Drop in the voltage cell after addition of NiFe (anode) LDH, CoMn LDH (anode), Ni _x B (cathode) and Co _x P (cathode) into the electrolyte (3 M KOH) (N= 3 for each one). The applied current density and final concentration of catalyst were 50 mA/cm ² and about 1.3 mg/ml, respectively, (c) Schematic representation of an alkaline electrolyser with self-assembled catalyst films in the presence of suspended catalysts in the electrolyte.

Figure 3 : SEM images of the various catalyst films formed by self-assembling process. Cross-section SEM micrographs of (a) NiFe LDH (anode), (b) CoMn LDH (anode), (c) Ni _x B (cathode) and (d) Co _x P (cathode) films on Ni current collectors. Figure 4: Operation of a self-assembled NiFe LDH@Ni//Ni _x B@Ni electrolyser. (a) Evolution of the cell voltage over time (3 M KOH, room temperature) at current densities of 50 mA/cm ² and 0.1 A/cm ² in the presence (grey background) and absence (white background) of suspended particles. Mark I and II indicate the addition of NiFe LDH and Ni _x B in the anolyte and catholyte reservoirs, respectively. Mark III shows the replacement of suspension by fresh catalyst-free electrolyte. Mark IV and V shows current interruptions during operation (1 = 0 mA/cm ² ) and current steps in catalyst-free electrolyte, respectively. Mark VI and VII indicate the second addition of catalyst powders. And Mark VIII shows current steps in the presence of catalyst in the electrolyte (b) Cell voltage of the self-assembled NiFe LDH@Ni//Ni _x B@Ni electrolyser at 50 mA/cm ² (squares) and 0.1 A/cm ² (triangles) over a period of several days, (c) Current interruptions in a catalyst-free electrolyte, (d) Cell voltage in the presence (circles) and absence (squares) of suspended particles at various current densities (50 mA/cm ² to 0.1 A/cm ² in steps of 10 mA/cm ² ). (e) Comparison between the cell voltage of a N ianode//N icathode alkaline electrolyser (squares) and a self-assembled NiFe LDH @ Ni _an ode//Ni _x B @ Ni _cat hode electrolyser (circles) at 50 mA/cm ² and 0.1 A/cm ² (current steps of 10 mA/cm ² in between). The concentration of catalyst powder in the suspension was about 1.3 mg/ml. Figure 5 : Schematic representation of the different stages during the formation of self-assembled catalyst film and the corresponding cell voltage. (1) Catalyst- free electrolyte solution and corresponding cell voltage. (2) Addition of catalyst powder into the electrolyte leading to the formation of self-assembled catalyst film and a drop in cell voltage. (3) Self-healing process of catalyst film, and associated constant cell voltage. (4) Detachment of particles from the catalyst film in the absence of particle in the electrolyte, and resulting increase in cell voltage.

Figure 6: Schematic representation of the setup of the electrolyser used in the Examples.

Figure 7 : Photographs of the assembled electrolyser.

Figure 8: Evolution of the cell voltage over time after addition of 1.3 mg/ml (squares) and 2.67 mg/ml (circles) NiFe LDH into the anolyte (3 M KOH) of a Ni//Ni alkaline electrolyser (I _app i. = 50 mA/cm ² ).

Figure 9 : Evolution of the cell voltage over time for Ni//Ni electrolyser at 50 mA/cm ² . Two additions of Ni _x B in the anolyte reservoir are indicated with arrows. The cell voltage did not decrease upon addition of this catalyst in the anolyte which indicates that Ni _x B film does not self-assemble on the anode.

Figure 10 : Schematic representation of the bulk electrochemical cell with three- electrode configuration used for the immobilization of suspended particle by applying pzc _N i ± 200 mV.

Figure 11. Linear sweep voltammograms (3 M KOH, 5 mV/s, per surface area of 0.196 cm ² ) for (a) OER and (b) HER of electrodes modified with catalyst films immobilized using a three-electrode cell. Bare Ni foil (blue), NiFe LDH @ Ni (black), CoMn LDH @ Ni (red), Ni _x B @ Ni (green) and Co _x P @ Ni (pink) and the optical images of the corresponding electrodes.

Figure 12 : Linear sweep voltammograms (3 M KOH, 5 mV/s, per surface area of 0.126 cm ² ) for (a) OER and (b) HER of electrodes modified with catalyst films that were prepared using an electrolyser. (Blue) Bare Ni foil, (black) NiFe LDH@Ni, (red)CoMn LDH@Ni, (green) Ni _x B@Ni and (pink) Co _x P@Ni.

Figure 13 : Photographs of the self-assembled NiFe LDH and Ni _x B films before and after LSV measurements. The circle indicates the area where the electrochemical cell was placed for LSV. The catalyst films lost material during LSV measurements.

Figure 14: Theoretical corrosion rate of the anode Ni foil current collector at an applied current density of 50 mA/cm ² for various Faradaic efficiencies. The graph shows the corrosion depth over the time at which the current is applied. The vertical green lines indicate the time needed until the whole Ni foil (thickness = 100 prn) is corroded for Faradaic efficiencies of (squares) 98 %, (circles) 99 % and (triangles) 99.5 %.

Figure 15 : Comparison of cell voltage in (circle) Ni//Ni electrolyser (without suspension), (triangle) NiFe LDH @Ni//Ni _x B@Ni electrolyser (in the presence of particles in the electrolyte) and (square) NiFe LDH @Ni//Co _x P@Ni electrolyser (in the presence of particles in the electrolyte) operating at 0.1 A/cm ² at room temperature.

Figure 16: XRD pattern of the as-prepared NiFe LDH.

Figure 17 : XRD pattern of the as-prepared CoMn LDH.

Figure 18: XRD pattern of the as-prepared Ni _x B (Ni _x B-300).

Figure 19 : XRD pattern of the as-prepared Co _x P.

Detailed Description of the Invention

The aspect (1) of the invention provides for an electrocatalytic device comprising an electrolytic cell divided into anode and cathode chambers by an exchange membrane, said anode and cathode chambers being provided with electrodes, wherein one or both of said electrodes are immersed in electrolytic suspensions, and wherein the electrolytic suspensions comprise surface-charged catalytic particles forming self-assembling and self-healing films on the electrodes.

In a preferred embodiment of aspect (1) of the invention, the surface-charged catalytic particles are directly added into the electrolyte reservoirs of the device and form suspensions that are pumped through the cell. The surface-charged catalytic particles may be micro- or nanoparticles having a diameter from 1 nm to 1000 nm, and/or may have a shape of 0-dimensions (including, but not limited to quantum dots), 1-dimension (including, but not limited to nanowire), 2-dimensions (including, but not limited to nanosheet), or 3-dimensions (including, but not limited to nano- to micro-particles).

In a further preferred embodiment of aspect (1) of the invention, the negatively surface-charged particles are immobilized and forming a film on the surface of the positively charged electrode, such as an electrolyser anode, and the positively surface-charged particles are immobilized and forming a film on the surface of the negatively charged electrode, such as an electrolyser cathode, driven by electrostatic forces. In particular, a hydrogen evolution reaction (HER) catalysts that possess positively charged surfaces, including transition metal borides, phosphides, sulphides and selenides which contain at least one metal or a mixture of metals (including, but not limited to C0 ₂ B, CO _x P and Ni _x B), form films on the surface of the negatively charged electrode. Furthermore, oxygen evolution reaction (OER) catalysts that possess negatively charged surfaces, including transition metal oxides, hydroxides and oxyhydroxide which contain at least one metal or a mixture of metals, form films on the surface of the positively charged electrode. In a particular example, NiFe LDH particles and Ni _x B particles are immobilized and form films on the electrolyser anode and cathode, respectively. It is to be noted that the films formed on the electrodes possess self-healing properties and remained stable for at least three weeks at current densities ranging between 50 - 100 mA/cm ² .

In a further preferred embodiment of aspect (1) of the invention, the device is an electrolyser that possesses electrodes made from Ni, Ti, stainless steel and/or carbon, in the form of foil, foam, felt, cloth and mesh. It is preferred that in such electrolyser the cell further comprises gas outlets, such as oxygen and hydrogen outlets.

The aspect (2) of the invention provides for a method for preparing the the device of aspect (1) as defined hereinbefore. In a preferred embodiment of aspect (2) of the invention, the surface-charged catalytic particles are directly added into electrolyte reservoirs of the device and form suspensions that are pumped through the electrolytic cell.

The aspect (3) of the invention provides for the use of the device of aspect (1) as defined hereinbefore as electrolyser. In a preferred embodiment of aspect (3) of the invention, the device is used for the electrolysis of water. In a particularly preferred embodiment the device is used for the electrolysis of aqueous alkaline solutions.

The invention is explained in more detail in the following discussion .

Formation of self-assembling catalyst films: From electrochemistry of single nanoparticle collision events, it is known that catalytic particles suspended in solution can follow two types of behaviours: hit-and-run and hit-and-stand (Wang, W. & Tao, N. ; Anal. Chem. 86: 2-14 (2014)). While the former is the basis of recent electrochemical flowing devices such as semi solid flow batteries and electrochemical flow capacitors (Hatzell, K. B. et al. ; Chem. Soc. Rev. 44: 8664-8687 (2015), Hatzell, K. B. et al.; Electrochim. Acta 111 : 888-897 (2013)), the latter offers the possibility of modifying electrode surfaces. Liu et al. made use of the hit-and-stand behaviour to immobilize nanoparticles from a suspension onto a transparent electrode for electrochromic applications (Liu, L. et al. ; J. Mater. Chem. A 2 : 16224-16229 (2014)).

We propose the immobilization of catalytic particles suspended in electrolyte solution to form catalyst films on both electrodes of electrolysers. This concept can be easily tested by adding catalytic particles in the electrolyte reservoir of an already assembled non-zero gap electrolyser (see Methods and Fig 6 and 7). Ni was chosen as model electrode material for both cathode and anode because it is often used in commercial alkaline electrolyser systems (Holladay, J. D. et al. ; Catal. Today 139 : 244-260 (2009)). The formation of self-assembling catalyst films was separately evaluated for four highly active catalyst powders: NiFe layered double hydroxide (Song, F. & Hu, X. ; Nat. Commun. 5 :4477 (2014)) (LDH) and CoMn LDH (Song, F. & Hu, X. ; J. Am. Chem. Soc. 136: 16481-16484 (2014)) for the anode, and Ni _x B (Liang, Y. et al .; Nanotechnology 27 (2016)) and COxP (Masa, J. et al. ; ACS Energy Lett., 1192-1198 (2016)) for the cathode. The successful formation of catalyst films was confirmed (Figure 1) in real-time by the drop in cell voltage immediately after the addition of the catalytic particles in the electrolyte reservoir and ex-situ by direct visual inspection of the electrodes after disassembling the electrolyser.

It should be noted that a constant concentration of catalytic particles in suspension of 1.3 mg/ml was used in this work. Interestingly, higher concentrations, i.e. 2.6 mg/ml, did not result in more active catalyst films since the drop in cell voltage was the same (Fig . 8 and example 2). On the other hand, the steady state in cell voltage was reached in a shorter time for higher concentration, which indicates faster formation of the self-assembling catalyst film.

In the study of single nanoparticle collision events, the surface charge was suggested to play an important role in the nature of sticking interaction between suspended particle and electrode (Xiao, X. et al. ; J. Am. Chem. Soc. 130 : 16669- 16677 (2008)). The surface charge of the tested materials was determined by zeta potential measurements, which revealed that the surface of NiFe LDH and CoMn LDH was negatively charged, while the surface of Ni _x B and Co _x P was positively charged in the used electrolyte (3 M KOH) as shown in Figure 2a. Considering that the surface of anode and cathode is positively and negatively polarized, respectively, with respect to the potential of zero charge of the electrode (+80 mV vs. Ag/AgCI 3 M KCI in the case of a Ni electrode in alkaline media (McCafferty, E.; Electrochim. Acta 55 : 1630-1637 (2010), Kang, J. et al. ; Corros. Sci. 50 : 3576-3580 (2008)), the immobilization of NiFe LDH and CoMn LDH on the anode and Ni _x B and Co _x P on the cathode and the associated drop in cell voltage (Figure la-d) can be explained on the basis of electrostatic interaction between particle and electrode surface. To proof this hypothesis, Ni _x B that is an extremely active catalyst for HER and OER (Liang, Y. et al .; Nanotechnology 27 (2016)) was added to the catholyte and anolyte compartment. A clear drop in cell voltage was only observed when Ni _x B was added to the catholyte (Figures 2b), whereas no change in cell voltage was observed when Ni _x B was added to the anolyte (Figure 9). Direct visual inspection of the anode after disassembling the electrolyser confirmed the absence of Ni _x B films. Additionally, three-electrode configuration measurements (Example 4) support electrostatic nature of the formation of self-assembling catalysts films. The mechanism of the formation of self-assembling catalyst films on the electrodes of an electrolyser is illustrated in Figure 2c.

Characterisation of electrodes modified with self-assembled catalyst films:

After formation of the catalyst films by adding catalyst powder in the electrolyte reservoirs, the electrolysers were disassembled and the modified electrodes were characterized. As discussed above, direct visual inspection of the electrodes revealed the formation of a film along the area defined by the electrolyte flow channel (Figure 1). Scanning electron microscopy (SEM) images confirmed the formation of catalyst films (Figure 3). Cross-section SEM images showed that the films were composed of aggregates of particles and had a thickness of several micrometres.

The electrochemical activity of the catalyst films towards either HER or OER was investigated in a three-electrode configuration (Example 5, Fig 12). The activity of the different catalyst films, determined by linear sweep voltammetry (LSV), was in good accordance with the trend observed in the electrolysers: activity of NiFe LDH > CoMn LDH > Ni and Co _x P > Ni _x B > Ni. Interestingly, visual inspection of the electrodes after electrochemical evaluation in a three-electrode configuration revealed a clear loss of catalyst from the films (Fig 13). This indicates a weak particle - electrode interaction due to electrostatic forces rather than covalent bonds. Hence, the presence of particles in the electrolyte of the electrolyser might induce a self-healing process that would be highly beneficial for long term operation .

Operation of a self-assembled catalyst film based electrolyser: The proof-of- concept for an electrolyser based on the formation of self-assembling catalyst films was demonstrated using NiFe LDH and Ni _x B as catalysts for OER and HER, respectively. A home-made, non-zero gap, alkaline electrolyser operating at room-temperature and using Ni electrodes for both anode and cathode was employed as model system. As described above, catalyst powders were directly added to the electrolyte reservoirs during operation of the electrolyser. Figure 4a shows the evolution of the cell voltage ( U _C eii ) over 22 days. U _ce ii drops 450 mV and 120 mV, marked in Fig. 4a as I and II, upon addition of NiFe LDH and Ni _x B, respectively. After addition of both catalysts, U _ce ii reached a steady state at about 2.0 V over five days. The small fluctuations of U _ce ii around this value were due to changes in the pH of the catholyte, which was manually controlled (no maintenance during nights).

In the previous section, the presence of particles in the electrolyte after formation of the catalyst film was speculated to induce a self-healing process. Mark III in Figure 4a indicates the moment when the electrolyte suspension was replaced by fresh particle-free electrolyte. Immediately, U _ce ii increased about 70 mV and kept increasing slowly (a slope of 18 mV per day)) over a period of 7 days (between mark III and VI in Fig 4a). Note that the electrolyte was periodically replaced over this period to remove particles released from the film. This confirms that the presence of particles in the suspension induces a self- healing process: when some particles detached, others that were suspended in the electrolyte immobilized replacing the detached ones. On the other hand, the catalyst films were not washed away immediately in the absence of suspended particles. Even when the current was interrupted which is of high importance for practical applications, the catalyst films were not severely deteriorated since U _ce ii did not increase significantly (Figure 4 c and mark IV in Fig 4a). After operating for 7 days in the absence of suspended particles, catalyst powders were again added to the electrolyte reservoirs. Mark VI and VII indicate the drop in U _ce ii associated with this second addition of NiFe LDH and Ni _x B (130 and 60 mV, respectively). This decrease in U _ce ii confirms the self-healing process that occurs in the presence of catalyst in the electrolyte. Figure 4d shows the differences in Uceii in the presence and absence of suspended particles at various current densities showing the clear benefit of having particles in the electrolyte. In the last 7 days of test, the current density was maintained at 0.1 A cm ^"2 in the presence of particles and U _ce ii remained at about 2.25 V over the entire period of time. Overall, the electrolyser operated for 22 days over which time the catalyst films remained active (U _ce ii of an electrolyser without catalyst films is displayed in Figure 4e for comparison). The three weeks test allows us to rule out major corrosion issue. The calculated corrosion rate of the anode current collector revealed that the Faradaic efficiency of the anode was above 99.5 % (see Examples, Section 5). This value is not surprising since Ni current collectors are used in commercial electrolysers due to its high corrosion resistance in alkaline media (Holladay, J. D. et al. ; Catal. Today 139 : 244-260 (2009)). The cell voltage of the electrolyser was improved by employing a more active catalyst film for HER. A stable cell voltage of 2.05 V over 48 hours was obtained at 0.1 A/cm ² when using NiFe LDH and Co _x P on the anode and cathode, respectively (Fig 16). This value represents an improvement of 200 mV and 750 mV versus NiFe LDH@Ni//Ni _x B@Ni and Ni//Ni electrolysers, respectively. Further improvements are expected by optimizing the engineering of the system since the proof-of-concept was demonstrated using a home-made non-zero gap electrolyser operating at room temperature.

The different stages during the formation of self-assembled catalyst films are illustrated in Figure 5. In stage 1, U _ce ii is defined by the HER and OER activity of the initial N ianode// i _C athode electrolyser at a constant current density. In stage 2, the catalyst powder is added into the electrolyte forming a suspension that is pumped through the cell. Driven by electrostatic interaction, catalytic particles from suspension self-assemble on the electrode surface forming the catalyst film. Uceii drops as the catalyst films are been formed (stage 2) until it reaches a steady state. In the stage 3, U _ce ii remains at constant value as long as particles are present in the electrolyte enabled by the self-healing process of the catalyst film. When a particle detaches from the electrode surface, another that is suspended in the electrolyte instantaneously replaces the detached one. Stage 4 occurs when the suspension is replaced by particle-free electrolyte losing the self-healing feature. Catalytic particles are slowly removed from the catalyst film leading to an increase in U _C eii - As in stage 2, U _ce ii decreases when catalyst powder is again added to the electrolyte. And U _ce ii stabilizes remaining at a constant value as long as particles are present in the electrolyte (stage 3).

Conclusion : It is demonstrated an innovative method for the formation of catalyst powder-based films on the electrodes of electrolysers. The approach allows "operando" modification of electrodes while the electrolyser is operating due to its self-assembling character. The catalyst powders are simply added to the electrolyte reservoirs forming a suspension that is pumped through the electrochemical cell. Based on electrostatic interaction, particles with surface negatively and positively charged are immobilized on the anode and cathode, respective, forming the catalyst film. Importantly, the presence of particles in suspension play a role even after the formation of the film since it induces a self- healing effect on the catalyst film. The proof-of-concept was achieved in a homemade non-zero gap alkaline electrolyser using state-of-the-art catalyst powders, NiFe LDH and Ni _x B for anode and cathode, respectively. The catalyst films formed on the Ni electrodes of this electrolyser were stable for at least 22 days.

The invention is furthermore explained in the following Examples, which are however not to be construed as limiting the invention.

Examples

Material and Methods

All chemicals (analytical grade) were used without any further purification.

NiFe LDH synthesis: Ni(N0 ₃ ) ₂ »6H ₂ 0 (4.362 g, 15 mmol) and Fe(N0 ₃ ) ₃ »9H ₂ 0 (2.02 g, 5 mmol) were dissolved in decarbonated ultra pure water (100 ml) in the presence of triethanolamine (1.86 g, 12.5 mmol). Ammonium hydroxide solution (c = 25 wt.%) was added to the salts mixture under nitrogen atmosphere until a pH value of 8.5 (~ 4 ml NH ₄ OH) was reached . Afterwards the mixture was stirred for 2 h at room temperature and, then, kept in an oven at 60 °C for 14 h. The resulting mixture was filtrated and washed with decarbonated ultra pure water (500 ml). The precipitate was dried at 70 °C for 24 h. (For XRD see Figure 16).

CoMn LDH synthesis: CoMn LDH was synthesised according to the procedure proposed by Song and Hu (Song, F. & Hu, X. ; J. Am. Chem. Soc. 136: 16481- 16484 (2014)). Briefly, under N ₂ atmosphere Co(N0 ₃ ) ₂ »6H ₂ 0 (1.085 g, 3.75 mmol), Μη(Ν0 ₃ ) ₂ ·4Η ₂ 0 (0.4705 g, 1.875 mmol), NaN0 ₃ (0.765 g, 9 mmol) and NH ₄ F (0.925 g, 25 mmol) were dissolved in ultra pure nitrogen purged water (500 ml). After 30 min of vigorous stirring H ₂ 0 ₂ (30 wt%, 125 μΙ) was added. The pH of the solution was adjusted to 10 by dropwise addition of NaOH (0.1 M, 250 ml). The final suspension was aged at room temperature for a period of 20 h. The product was isolated and washed with water and ethanol, yielding a brownish solid (For XRD see Figure 17). NiyB synthesis: Aqueous NiCI ₂ »6H ₂ 0 (20 ml, 0.5 M) in a round-bottomed Schlenk flask was deaerated by means of a Schlenk vacuum technique, flushed with argon and maintained at 0 °C using an ice-bath . NaBH ₄ (1.0 M) in NaOH (0.1 M), separately deaerated and flushed with argon, was slowly added to the NiCI ₂ »6H ₂ 0 by means of a syringe. The reaction equation is given in equation (1) (Ganem, B. & Osby, J. O.; Chem. Rev. 86: 763-780 (1986)).

2NiCI ₂ + 4NaBH ₄ + 9H ₂ 0→ Ni _x B + 4NaCI + 12.5H ₂ + 3B(OH) ₃ (1)

As a precaution, vigorous frothing occurs if NaBH ₄ solution is added rapidly to NiCI ₂ »6H ₂ 0. Instantaneous formation of a dark precipitate was observed. The precipitate was collected by filtration and washed with large amounts of tri- d istil led water, followed by washing with ethanol. For XRD see Figure 18.

COvP synthesis: Cobalt phosphide (Co _x P) was prepared by thermal decomposition of Bis(triphenylphosphine)cobalt(II)dichloride [(C ₆ H ₅ ) ₃ P] ₂ CoCI ₂ , under a reductive atmosphere constituted of H ₂ (25%) in N ₂ at 400 °C for 2 h (Masa, J. et al. ; ACS Energy Lett., 1192-1198 (2016)). For XRD see Figure 19.

Electrochemical measurements: All electrochemical measurements were conducted in alkaline media (3 M KOH) at room temperature. For the electrolyser (a home-made, non-zero gap electrolyser, see Examples, Fig 6 and Fig 7), a BioLogic VMP3 multi-channel potentiostst/galvanostat (Biologic, France) was used. Ni foil (0.1 mm thick, 99.2 % Ni and a surface area exposed to electrolyte of 4.1 cm ² ) acted as anode and cathode. Anodic and cathodic compartments were separated by an anion exchange membrane (fumasep ^® FAA-3-PK-130, Fumatech). A two-channel Minipuls3 peristaltic pump (Gilson Inc., USA) was used to separately pump anolyte and catholyte through the system at a flow rate of 30 ml/min. Catalyst powders were directly added to either the anolyte or the catholyte. The concentration of catalyst in the electrolyte was about 1.3 mg/ml. The catalyst suspensions were stirred to prevent settling of the particles. The pH of the electrolyte was kept constant manually.

Linear sweep voltammetry measurements of the electrodes modified with self- assembled catalyst films were carried out in a three-electrode configuration in which the modified electrode, a commercial Ag/AgCI (3 M KCI) electrode and a piece of Ni foil acted as working electrode (WE), reference electrode (RE) and counter electrode (CE), respectively. Evaluation of activity in the three-electrode cell was conducted under static conditions at a scan rate of 5 mV/s (the uncompensated resistance was corrected afterwards) using an Autolab

PGSTAT128N potentiostat/galvanostat (Metrohm Autolab B.V., Netherlands). Scanning electron microscopy (SEM) : SEM images were taken using a Quanta 3D FEG scanning electron microscope (FEI, USA) operated at 20.0 kV.

Zeta potential determination : The zeta potentials of the different catalysts were measured in aqueous 3 M KOH medium using a Malvern Zetasizer Nano ZS (Malvern Instruments Ltd, UK), equipped with a He-Ne-laser with a wavelength of 633 nm, using the laser doppler micro electrophoresis.

X-ray photoelectron spectroscopy (XPS) : XPS spectra were recorded using a UHV set-up equipped with a Gammadata-Scienta SES 2002 analyzer. The base pressure in the measurement chamber was 5 ^χ 10 ^"10 mbar. Monochromatic Al Ka (1486.6 eV; 14.5 kV; 30 mA) was used as incident radiation, and a pass energy of 200 eV was chosen resulting in an effective instrument resolution higher than 0.6 eV. Charging effects were compensated using a flood gun, and binding energies were calibrated based on positioning the main C Is peak originating from carbon contaminations at 284.5 eV.

Example 1 : Electrolyser system.

The setup of the home-made, non-zero gap, alkaline electrolyser used in this work is shown in Figure 6. The electrochemical cell consists of two compartments separated by an anion exchange membrane. Each compartment was composed of a metallic end plate, a peek plate, a teflon gasket, the electrode (Ni foil, thickness = 100 prn) and a teflon sealed flow channel . Both compartments were stacked together and fixed with six screws (Figure 7). The electrolyte was pumped through the cell in two separate circuits, i.e. anolyte and catholyte circuit, to prevent mixing of particles and gases.

Example 2 : Dependence of the voltage drop on the particle concentration.

Addition of NiFe LDH into the anolyte (3 M KOH) of a Ni//Ni alkaline electrolyser leads to a significant drop in the measured cell voltage at an applied current density of 50 mA/cm ² as it can be seen in Figure 1. Figure 8 shows the evolution of the cell voltage over time upon addition of NiFe LDH for two different concentrations of particles in suspension (1.3 and 2.67 mg/ml). The total drop in cell voltage did not increase with increasing concentration of catalyst, which indicates that the activity of the catalyst film does not improve for higher concentration. What it does change is the time needed for the formation of a self-assembling film; the higher concentration, the faster formation.

Example 3 : Absence of formation of NiyB catalyst on the anode. The electrostatic nature of the formation of self-assembling catalyst film predicts that particles with surface positively charged will not stick onto the anode. Ni _x B is a state-of-the-art catalyst for OER that has positive zeta potential (surface positively charged). Ni _x B powder was added to the anolyte reservoir while the electrolyser was operating. The formation of self-assembling catalyst films leads to a significant decrease in the cell voltage of the electrolyser as shown in Figure 1. However, the addition of Ni _x B in the anolyte reservoir did not change the cell voltage (see Figure 9). This indicates that catalyst film was not formed when Ni _x B was added to the anodic compartment supporting the electrostatic nature of the film formation. In addition, a film was not seen on the anode after disassembling the electrolyser.

Example 4: Mechanism of the formation of self-assembling catalyst films.

The mechanism of the formation of self-assembling catalyst films is based on electrostatic interaction between the surfaces of the current collector and the suspended particle. The surface charge of the electrode is dependent on the applied polarization with respect to its potential of zero charge. The potential of zero charge of Ni electrodes (pzc _N i) in alkaline media was reported to be +80 mV vs. Ag/AgCI in 3 M KCI (Kang, J. et al ., Corros. Sci. 50 : 3576-3580 (2008), McCafferty, E., Electrochim. Acta 55 : 1630-1637 (2010)). Therefore, the surface of Ni is positively and negatively charged when the electrode is polarized at higher and lower potentials than +80 mV vs. Ag/AgCI (3 M KCI), respectively. As a consequence, the anode and cathode are positively and negatively charged, respectively, during electrolysis. The surface charge of suspended particles can be determined by measuring the zeta potential. Zeta potential values of -26.3 ± 3.5 mV (NiFe LDH), -55 ± 4.5 mV (CoMn LDH), + 104.3 ± 3.1 mV (Ni _x B) and + 110.7 ± 3.8 mV (Co _x P) were obtained in alkaline media (3 M KOH). According to the proposed immobilization mechanism, the polarisation of the electrode with respect to its potential of zero charge (pzc) enables tuning the surface from positively to negatively charged and vice versa. As a result, the immobilization of particles with surface positively or negatively charged can be controlled, to a certain extent, by the applied potential to the electrode. Bearing in mind that pzc _N i is + 80 mV vs. Ag/AgCI 3 M KCI, we attempted to immobilize catalytic material using a bulk electrochemical cell (Figure 10) at a constant potential of ±200 mV vs. pzc _Ni , i.e. -0.120 V and +0.280 V vs. Ag/AgCI 3 M KCI for materials with positive (Ni _x B and Co _x P) and negative (NiFe LDH and CoMnLDH) zeta potential, respectively. The immobilization was carried out using a three-electrode configuration where a piece of Ni foil (surface area of ca. 4 cm ² ), a piece of Ni foil in a separate compartment filled with particle free electrolyte and a Ag/AgCI (3 M KCI) electrode acted as working electrode (WE), counter electrode (CE) and reference electrode (RE), respectively. A suspension of particles (1.3 mg/ml) in 3 M KOH and particle-free 3 M KOH were used as electrolyte in the WE and CE compartments, respectively. The suspension was continuously stirred to provide a homogenous suspension during immobilization and to increase the collision frequency of particles at the electrode. The immobilization potential was applied to the WE for 60 min. After the immobilization procedure, the electrodes were taken from the suspension, and thoroughly rinsed with deionized water and then dried in air. Visual inspection of the electrodes revealed formation of films which covered the entire area immersed in the suspension (Figure 11). The catalytic activity of the modified electrodes was electrochemically investigated towards the specific reaction, either HER or OER (Figure 11). Linear sweep voltammograms (LSV) in 3 M KOH showed that the activity of all modified electrodes was higher than that of bare Ni electrode indicating the formation of a catalyst film. LSVs were recorded within the specific potential window for HER or OER at a scan rate of 5 mV/s. The potentials were corrected for uncompensated electrolyte resistance according to E _cor r = E - i · R. The uncompensated resistance was determined by means of electrochemical impedance spectroscopy (EIS) using an ac perturbation of 10 mV (rms) at open circuit potential in the frequency range from 100 kHz to 10 Hz.

Potentials are given against the used reference electrode (Ag/AgCI 3 M KCI) since the primary aim of LSVs is to compare the activity of catalyst films with that of bare Ni. However, the reversible hydrogen electrode (RHE) is the most commonly used reference system in electrocatalysis. Therefore, potentials were converted to the RHE scale for comparison reasons according to:

ERHE = Eappi + E(Ag/AgCI 3 M KCI) + 0.059 · pH.

The pH of the electrolyte was calculated according to the following formula pH = 14 + log(c»[OH ^" ]),

where c=activity coefficient and [OH ^" ] is the OH ^" concentration.

The activity coefficient has been reported to be 1.05 for 3 M KOH at 25 °C (Hamer, W. J. & Wu, Y.-C; J. Phys. Chem. Ref. Data 1:1047 (1972), Balej, J.; Collect. Czech. Chem. Commun. 61:1549-1562 (1996)). Electrodes modified with catalyst films that were prepared using a three- electrode cell showed similar catalytic activity than that of electrodes modified with catalyst films that were prepared using an electrolyser (Figure 12).

Example 5 : Characterisation of self-assembled catalyst films in three-electrode cell.

Electrodes modified using electrolysers (instead of three-electrode cell) were electrochemically characterized. The activity of these modified electrodes towards either HER or OER was studied using a three-electrode configuration. Figure 12 shows that the trend in the activity of modified electrodes was consistent with that observed in i) electrodes modified using a three-electrode cell and ii) the electrolyser (the corresponding cell voltage). Visual inspection of the modified electrode after evaluation in three-electrode configuration revealed the loss of catalyst material from the films (Figure 13), which indicates that interaction electrode - particle was not strong (covalent). The presence of particles suspended in the electrolyte that enables self-healing of the films is thus required for achieving long-term performance.

Example 6: Corrosion rate of the anode Ni current collector at 50 mA/cm ² .

In a galvanostatic measurement, the current (electrons) originate from the most favourable electrochemical reaction. In the case of the anode, the oxidation of transition metals in the collector takes place at lower potentials than OER. As a consequence, the corrosion of the current collector may simultaneously occur with OER in an electrolyser. The Faraday efficiency quantifies how much of the applied current goes to the oxidation of the current collector instead of evolution of oxygen. Faraday efficiency can thus help estimating the durability of an electrode. In our case, we have used 100 prn Ni foil as current collector. Metallic

Ni can be oxidized to Ni ³⁺ under the anodic condition of the anode. Using the

Faraday constant, molecular mass and density of Ni, we obtain that 0.835 mAh are necessary to oxidize a depth of 1 prn of an area of 1 cm (0.835 mAh/cm ² pm). We can now estimate the corrosion rate of the current collector at

50 mA/cm ² for different Faraday efficiencies (Figure 14). For a Faraday efficiency of 99.5 %, 0.25 mA/cm ² out of 50 mA/cm ² are consumed in the oxidation of the current collector leading to an oxidation rate of 3.3 h/pm. Therefore, the full depth of the current collector is oxidized within 14 days. However, our measurement lasted for 22 days, which means that the Faraday efficiency is higher than 99.5 %. This value is not surprising since Ni current collectors are used in commercial electrolysers due to its high corrosion resistance in alkaline media (Holladay, J. D. et al. ; Catal. Today 139 : 244-260 (2009)). It should be noted that the oxidation of the catalyst itself (20 mg of NiFe LDH) takes place in 1 min at 50 mA/cm ² (net current of 200 mA).

Example 7 : Operation of NiFe LDH@Ni//COvP@Ni alkaline electrolyser.

A comparison of cell voltage in Ni//Ni electrolyser (without suspension), NiFe LDH @Ni//Ni _x B@Ni electrolyser (in the presence of particles in the electrolyte) and NiFe LDH @Ni//Co _x P@Ni electrolyser (in the presence of particles in the electrolyte) operating at 0.1 A/cm ² at room temperature is shown in Figure 15.

Example 8: Structural characterisation of catalyst powders.

The XRD pattern of the as-prepared NiFe LDH, CoMn LDH, Ni _x B (Ni _x B-300)and Co _x P are shown in figures 16 to 19, respectively. For the sample used in the Ni _x B-300 study, broad and poorly defined peaks were obtained which is likely due to the small crystal size. The sample was annealed at higher temperature (i.e. 600 °C, Ni _x B-600) to increase the crystal size leading to better resolution of the peaks. The main components of Ni _x B were metallic Ni, β-Νϊ(ΟΗ) ₂ and a mixture of Ni ₂ B and Ni ₃ B.