Porous membrane, its method of production, and an alkaline electrolyzer with such membrane

Abstract

Porous membrane, its method of production, and an alkaline electrolyzer with such membrane A porous membrane for alkaline water electrolysis is produced by a mix of a polymer, an alkoxide of an inorganic metal as a precursor for conversion into hydrophilic metal oxide or metal hydroxide particles, and a stabilizing agent for suppressing agglomera- tion of metal oxide or metal hydroxide particles during conversion of the precursor. The mix is cast as a layer on a support and exposed to nonsolvent-induced phase separation, NIPS, for converting the precursor in the layer into metal oxide particles or metal hy- droxide particles by hydrolyzing the precursor. The resulting membranes performed well in alkaline electrolysis.

Description

[0001] Porous membrane, its method of production, and an alkaline electrolyzer with such membrane

[0002] FIELD OF THE INVENTION

[0003] The present invention relates to a porous membrane that contains hydrophilic inorganic particles and a method of producing it, as well as its use in an alkaline electrolyzer for alkaline water electrolysis, for example for hydrogen production.

[0004] BACKGROUND OF THE INVENTION

[0005] Alkaline electrolysis for production of hydrogen is constantly subject to improvement, especially with respect to the microporous membranes between the electrodes. Such membranes should have high ionic conductivity and good gas separation capabilities. One of the problems encountered for such membranes is attachment of gas bubbles on the surface of the pores of the membrane, closing the pores in the membrane surface and causing an increase of voltage loss due to reduction of ion transport through the pores that are closed by gas bubbles. The problem with gas bubbles in the membrane increases when the hydrophilicity of the membrane decreases. In order to maintain the membrane hydrophilic, the material of the membrane includes large concentrations of hydrophilic inorganic particles in the porous polymer matrix. However, a problem encountered in such membranes is aggregations of such particles and detachment of the particles, which leads to reduction of hydrophilicity, as the polymer is hydrophobic. Accordingly, substantial efforts are put into production of useful membranes for alkaline electrolysis by optimizing the concentration and fixation of the hydrophilic particles in the membrane.

[0006] Various production methods of membranes are disclosed in WO2017 / 176599A1, US2014 / 158612A1 CN105013356A, JP2021-090909A. Production of ultrafiltration membranes are described in the article “Ultrafiltration membranes based on hybrids of an amphiphilic graft copolymer and titanium isopropoxide” by Park et al. and published in J. APPL. POLYM. SCI. (2018) DOI: 10.1002 / APP.45932. An early approach of production of membranes as electrolytic separators for alkaline electrolysis is disclosed in an article “Aromatic Polymers for Advanced Alkaline Water Electrolysis - III Polysulphone-TiCh Films” published by Modica et al. in Int. J. Hydrogen Energy, Vol. 11, No. 5, pp. 307-308, 1986. In this article, a production method is disclosed in which polysulphone (PSU) was dissolved in N-methyl-2-pyrrolidone (NMP) together with Carbowax 6000. For providing hydrophilic inorganic particles in a porous PSU matrix, titanium-butoxide Ti(OBu)4 was added. The resulting dope was poured onto a doctor’s blade and immersed in water, which resulted in Ti(OBu)4 being converted into BuOH +TiO2, where the TiCh was maintained as hydrophilic inorganic additive in the final membrane. The final porosity of the membrane was achieved by immersion in boiling water, washing out the Carbowax and leaving pores. Without Carbowax, no pores were formed. As explained in this article, the concentration of TiCh particles was limited to 19%, as higher concentrations resulted in agglomeration of the particles that could be shaken off the membrane. Additionally, it was found that the mechanical strength decreased greatly with increasing concentration of the TiCh.

[0007] As a high concentration of TiCE is desired for optimized hydrophilicity, this approach by Modica et al. of using Ti(OBu)4 , limiting the titanium concentration to less than 19%, did not appear promising for membranes in alkaline electrolysis and was not used as basis for later developments of commercially available membranes. Instead, the development of useful membranes for alkaline electrolysis took a different direction, as explained in the following.

[0008] Basis for later large-scale commercialization became a membrane production principle disclosed by Vermeiren et al. in WO93 / 15529 and in the follow-up article “Zirfon®: A New Separator for Alkaline Batteries and Fuel Cells” published by Vermeiren et al. in Int. J. Hydrogen Energy Vol. 21, No. 8. pp. 679 684, 1996, as well as the article “Evaluation of the Zirfon® Separator for use in Alkaline Water Electrolysis and Ni-H2 Batteries” published by Vermeiren et al. in Int. J. Hydrogen Energy, Vol. 23. No. 5, pp. 321-324, 1998. In this article, the production method by Vermeiren et al. was reported to provide Zirfon® separator membranes containing as much as 85% ZrCh and 15% PSU and being chemically stable for 10,000 h in 30% KOH at 120°C. It was explained that the structure was dependent on speed of removal of solvent and type of nonsolvent, and the properties were dependent on the amount of ZrO2, which had to be high. Based on these early principles, a series of improvements were made and are still ongoing towards commercially available membranes. Manufacturing methods were discussed by Vermeiren et al in the article “The influence of manufacturing parameters on the properties of macroporous Zirfon® separators”, published in Porous Mater (2008) 15:259-264 [DOI 10.1007 / s 10934-006-9084-0]. It emphasizes the important finding of the pore size of the membrane becoming smaller with increasing ZrO2 amounts. However, as already discussed in earlier prior art, proper production results with increasing particle content is a challenge. As an alternative option to high particle concentration, this disclosure also reports the finding that porosity is inversely proportional to polymer concentration, and polymer concentration has same effect on pore size as the amount of ZrCh. This points towards increased polymer concentration as an alternative to high particle concentration and advices to pre-evaporate a portion of solvent (NMP) prior to phase inversion.

[0009] With refined production methods since the 1990’ies, the Zirfon® type membranes are current state-of the-art and commercially available by the international company AGFA® under the initial trademark Zirfon®. The production principles are to a large extend based on Vermeiren’ s initial works of dissolving PSU in a solvent, such as NMP, adding ZrCh powder, spreading the solution onto a support and then using a nonsolvent for phase inversion.

[0010] Not only have the Zirfon® membranes become state-of-the-art membranes but they are also used as a benchmark in connection with tests, for example as disclosed in the articles by Lee et al. “Advanced Zirfon®-type porous separator for a high-rate alkaline electrolyser operating in a dynamic mode”, published in Journal of Membrane Science 616 (2020) 118541 [https: / / doi.Org / 10.1016 / j.memsci.2020.118541] and “The synthesis of a Zirfon®-type porous separator with reduced gas crossover for alkaline electrolyzer” published in Int. J. Energy Res. Volume 44, Issue 3, 10 March 2020, Pages 1875-1885 [http s : / / doi . org / 10.1002 / er .5038].

[0011] Generally, in the technical field, this technique is termed nonsolvent-induced phase separation (NIPS). Different phase inversion variants have been reported, using liquid or vapor, and a corresponding terminology with acronyms LIPS and VIPS. Later publications based on these early publications develop this principle into suitability for commercialization and further optimization. Examples are found in W02006 / 015462, W02009 / 147084, W02009 / 147086, WO2016 / 203701,

[0012] KR102328727B1, W02021 / 04811, W02023 / 280760, W02023 / 280600,

[0013] WO2023 / 280598, WO2022 / 063584, W02022 / 002904, EP3272908.

[0014] Further production methods and products are disclosed in CN115677269A, CN115693017A, JP2022176792A, CN115181996A, CN115896863A,

[0015] CN115895014A, CN114432906A, JP2020007574A, KR102436919B1,

[0016] WO2016 / 148302A1, and W02019 / 011844A1.

[0017] As the commercially available Zirfon®-type membranes are relatively costly, especially for large alkaline electrolysis ^-production plants, attempts have been going on to provide alternatives, which is also reflected in some of the above-mentioned references. Thus, there is not only a need for improvements but also a desire for alternative cost-effective membranes on the market.

[0018] However, as it appears from the disclosures above, the general principle of addition of hydrophilic inorganic particles, typically ZrCh or TiCh, to the dissolved polymer has been maintained in further developments and appears as the most promising way forward. None of the above disclosures describe methods according to principles of the above-mentioned work by Modica et al. with respect to Ti(OBu)4 being converted into BuOH +TiO2.

[0019] DESCRIPTION / SUMMARY OF THE INVENTION

[0020] It is therefore an objective of the invention to provide an improvement in the art. In particular, it is an objective to provide an alternative to the Zirfon®-type membrane production method. This objective and further advantages are achieved with a membrane that contains hydrophilic inorganic additive and a method of producing it as well as its use for alkaline water electrolysis as described below and in the claims.

[0021] In short, a porous membrane for alkaline water electrolysis is produced by a mix of a polymer, an alkoxide of an inorganic metal as a precursor for conversion into hydrophilic metal oxide or metal hydroxide particles, and a stabilizing agent for suppressing agglomeration of metal oxide or metal hydroxide particles during and after conversion of the precursor. The mix is cast as a layer on a support and exposed to nonsolvent-induced phase separation, NIPS, for converting the precursor in the layer into metal oxide particles or metal hydroxide particles by hydrolyzing the precursor. The resulting membranes performed well in alkaline electrolysis experiments. Details are described below.

[0022] In the following, the following abbreviations have been used for simplicity:

[0023] NIPS - nonsolvent-induced phase separation

[0024] LIPS - nonsolvent liquid induced phase separation

[0025] VIPS - nonsolvent vapor induced phase separation

[0026] CTFE - Chlorotrifluoroethylene

[0027] Cyrene - Dihydrolevogluco senone

[0028] DMAc - Dimethylacetamide

[0029] DMF - Dimethylformamide

[0030] DMI - Dimethylisosorbide

[0031] DMSO - Dimethylsulfoxide

[0032] ECTFE - Ethylene Chlorotrifluoroethylene

[0033] EDS - Energy Dispersive X-ray Spectroscopy

[0034] ETFE - Ethene-co-tetrafluoroethene

[0035] GVL - Gamma- valerolactone

[0036] NBP - N-Butyl-2-Pyrrolidone

[0037] NEP - N-Ethyl-2-Pyrrolidone

[0038] NMP - N-Methyl-2-Pyrrolidone

[0039] PA - Polyamide

[0040] PAI - Polyamide-Imide

[0041] PAN - Polyacrylonitrile

[0042] PE - Polyethylene

[0043] PEEK - Polyether ether ketone

[0044] PEKK - Poly etherketoneketone

[0045] PEG - Polyethyleneoxide

[0046] PES - Polyethersulfone

[0047] PET - Polyethylene terephthalate PI - Polyimide

[0048] PMP - Polymethylpentene

[0049] Polarclean™ - Methyl 5-(dimethylamino)-2-methyl-5-oxopentanoate

[0050] PP - Polypropylene

[0051] PPS - Polyphenylene sulfide

[0052] PPSU - Polyphenylsulfone

[0053] PS - Polystyrene

[0054] PSU - Polysulfone / Polysulphone

[0055] PTFE - Polytetrafluoroethylene

[0056] PVA - Polyvinylalcohol,

[0057] PVP - Polyvinylpyrrolidone

[0058] SEB - Styrene-Ethylene-Butylene

[0059] SEBS - Styrene-Ethylene-Butylene-Styrene

[0060] SEM - Scanning Electron Microscope

[0061] SPEEK - Sulfonated polyether ether ketone

[0062] TTBO - Titanium-butoxide

[0063] TTIP - Titanium-isopropoxide

[0064] All percentages for substances given herein are by weight unless indicated otherwise. Ther term w% is used to indicate it.

[0065] The term nonsolvent is used for the substance in which the polymer for the membrane, for example PSU, is not soluble. This term is used, irrespective of the nonsolvent, for example water, being a solvent for other chemical compounds.

[0066] In the method described herein, a porous membrane is produced. Its use is, advantageously, for alkaline water electrolysis. In this method, a polymer solution is provided by dissolving a polymer resin, for example PSU, for the porous membrane in a solvent, for example NMP. Furthermore, an alkoxide of an inorganic metal is selected, for example titanium-butoxide Ti(0Bu)4 , as a precursor for conversion into hydrophilic metal oxide or metal hydroxide particles, for example TiCK The term alcoholate is often used as synonym for alkoxide. In addition, the conversion of the alkoxide into hydrophilic metal oxide or metal hydroxide particles in situ generates an alcohol, which is a non- solvent to the polymer resin and further inducing phase inversion as well as generating an in situ concentration gradient of non-solvents.

[0067] Despite some similarity with the method disclosed in the aforementioned article of 1986 by Modica et al, the problems that Modica encountered with the agglomeration of the TiC at concentration of 19w% and above has been overcome with the current approach, so that the invention as described herein is a successful further development. This has been achieved by adding a stabilizing agent, for example PVP. The stabilizing agent is different from solvents for the polymer of the matrix, as it must not dissolve the polymer.

[0068] It is further pointed out that the pore formation, as described in the following, is done by NIPS simultaneously with the hydrolyzation of the precursor or, alternatively, in a NIPS step before the hydrolyzation of the precursor. It is also possible to perform the pore formation, for example by VIPS, and obtain a partial hydrolyzation simultaneously with the pore formation in a first step and, then, finalizing the hydrolyzation in a subsequent second step. However, in any case, this is in contrast to the method by Modica et al., in which the pore formation was done after the hydrolyzation in a water bath, namely by a 24h immersion in boiling water in order to get the Carbowax leached out of the polymer.

[0069] Additionally, there is no need of Carbowax in the method described herein, in contrast to the method in Modica’ s article.

[0070] Clearly differentiated from the Modica’ s article, the method herein comprises the steps of selecting a stabilizing agent among agents that are suppressing agglomeration of such metal oxide or metal hydroxide particles in the solution. The stabilizing agent is mixed with the precursor and the polymer solution to obtain a mix. A layer of the mix is then provided on a rigid or flexible supporting substrate, typically by casting the mix onto the substrate. Potentially, a slot-die dispenser is used for a continuous process.

[0071] Then, the layer is exposed to nonsolvent- induced phase separation, NIPS, for providing pores in the polymer matrix. Furthermore, the precursor in the layer is converted into metal oxide or metal hydroxide, typically sub-micrometer particles, by hydrolyzing the precursor.

[0072] By using a stabilizing agent, it has surprisingly turned out that the concentration of the metal oxide or metal hydroxide additive obtained by hydrolyzing the precursor can be increased to more than 20w% in the porous membrane as compared to the total weight of the porous membrane. In experiments, concentrations up to 33w% were reached. This is a major improvement of the method as disclosed by Modica et al., where the limit was below 19%.

[0073] Examples of the stabilizing agents are PVA, PEO, Cellulose Acetate, Glycerol, and PVP.

[0074] Good results were obtained with PVP as stabilizing agent. Advantageously, the molecular weight of PVP is in the range 5.000-4.000.000 g / mol, where the mid-range of 100.000 to 1.000.000 is regarded as particularly useful. In experiments, PVP of the type K90 was used with 360.000 g / mol.

[0075] Advantageously, the method is performed without completely removing the stabilizing agent from the polymer matrix.

[0076] Using a precursor has a number of advantages. In particular, agglomeration of the hydrophilic inorganic particles is avoided. The latter is a specific problem when such metal oxide particles are dispersed in the solution as per the prior art. Furthermore, the invention has an advantage over the prior art in that there is no problem with inhomogeneous dispersions of particles, which occurs especially at higher particle concentrations and is detrimental to good membrane performance. From this perspective, the invention also solves additional problems encountered in the prior art.

[0077] Even further, it should be pointed out that the small hydrophilic inorganic particles that are used in production processes in the prior art imply a potential health hazard for the workers, which can also be advantageously avoided by the invention. As a consequence, safety installations for providing protection of the workers against the fine powder can be avoided, reducing overall costs and making the production method highly suitable for large-scale production of membranes. Accordingly, the invention solves multiple problems in a surprisingly simple manner.

[0078] For example, candidates for the metal of the precursor is selected from the group of Zr, Ti, Ce, Hf, Ba, Al, Cu, Li, Na, K, Si, V, Sc, Y, Ca, Mg, Sr, optionally from the group of Zr, Ti, Ce, Hf, Ba, Al.

[0079] As an example, the alkoxide / alcoholate for the precursor is selected among the group of methoxides, ethoxides, propoxides, iso-propoxides, butoxides, tert-butoxides, pentoxides, hexoxides and mixtures thereof.

[0080] Typically, the concentration of the precursor in the mix is in the range of 8w% to 25w% of the mix.

[0081] Typically, the ratio between precursor and polymer resin in the dope is in the range of 0.4 (w / w) to 2.5 (w / w), for example 0.6 (w / w) to 2.2 (w / w), optionally 0.9 to 1.7 (w / w). Experimentally, stable membranes were produced in the ratio range of 0.66 (w / w) to 2.13 (w / w).

[0082] Examples of the polymer resin are found among PSU, PES, PPS, PPSU, PEEK, PEKK, PP, PE, PMP, PI, PAI, PA, PAN, PS, SEB, SEBS, and mixtures thereof.

[0083] In concrete embodiments, the mix contains 6w% to 26w% polymer resin, for example 8w% to 21w%, optionally 10w% to 17w%.

[0084] Typically, the content of such stabilizing agent in the mix is in the range of 0.5w% to 15w%, for example 1.5w% to 12w%, optionally 2.0w% and 9w%, or even 2.5w% to 7w%, relatively to the total weight of the mix.

[0085] Typically, the ratio of the stabilizing agent to the precursor in the mix solution is in the range of 0.05 (w / w) to 2.00 (w / w), for example 0.10 (w / w) to 1.00 (w / w), optionally 0.12 (w / w) to 0.75 (w / w). The content of hydrophilic metal oxide or hydroxide particles can be calculated as following. One mol of alkoxide achieves one mol of hydrophilic metal oxide or hydroxide particle. By taking the molar mass of the hydrophilic metal oxide or hydroxide divided by the molar mass of the alkoxide, a ratio (M) defines the mass of the in situ produced hydrophilic metal oxide or hydroxide particle per mass of alkoxide. In general terms, the ratio M is defined as the molar amount of hydrophilic metal oxide or hydroxide particles per molar amount of precursor considering the stochiometric balance between hydrophilic metal oxide or hydroxide particles and precursor, which is typically one for alkoxides. When further considering the ratio (R) between alkoxide and the polymer resin, the amount of hydrophilic metal oxide or hydroxide particles in the final membrane (T) is achieved according to the following equation. T[w%] = 100

[0086] For example, M for Titanium Butoxide (Molar mass 340.32 g / mol) and TiCh (Molar mass 79.87 g / mol) is M = 0.23469. An exemplified ratio between Titanium Butoxide and Poly sulfone is 1.25. Then, the content of TiCh in the final membrane is T = 22.7w%.

[0087] In some embodiments, all metal oxide and / or metal hydroxide particles in the membrane have been obtained only by hydrolyzation of a precursor. Optionally, the metal oxide and / or metal hydroxide particles in the membrane obtained by hydrolyzation of a precursor are the only hydrophilic inorganic particles in the membrane.

[0088] However, although, the method is particularly useful for avoiding use of a powder for providing hydrophilic inorganic particles, the method with the hydrolyzation of the precursor can also be combined with existing methods. The advantage is that the final concentration of the hydrophilic inorganic particles can be increased while avoiding many of the problems that occur when high concentrations are achieved by only adding powder, as the high concentration of powder implies risks of inhomogeneous distributions and agglomeration of the particles in the mix even prior to NIPS. For example, the concentration of the metal oxide, sulfate, or hydroxide in the final membrane can be increased, so that it reaches 40w% or higher, for example up to 75w% or even up to 90w%. In such cases, depending on the final concentration, at least 10w%, for example at least 18w%, and optionally up to 35w% of the total content of the metal oxide, sulfate or hydroxide in the final membrane is achieved by hydrolysis of a precursor.

[0089] Accordingly, in some embodiments, prior to the NIPS, sub-micron particles of a hydrophilic metal oxide, sulfate or hydroxide, for example from the group of Zr, Ti, Ce, Hf, Ba, Al, are added for increasing the concentration of hydrophilic inorganic particles in the porous membrane as compared to concentration of the metal oxide and / or metal hydroxide particles in the membrane obtained by hydrolyzation of a precursor. For example, at least 10w% of the total content of the metal oxide, metal sulfate, and metal hydroxide in the membrane is achieved by hydrolysis of the precursor. Average size of the added particles as powder is less than 1 micrometer, for example less than 0.5 micrometer, optionally less than 0.2 micrometer. A possible way of measuring average particle sizes is laser diffraction. However, other methods known to the skilled person can be used.

[0090] In some embodiments, the NIPS is performed by VIPS, in which at least one side of the mix on the substrate is exposed to a vapor of a nonsolvent for vapor-induced phase separation. For example, the mix on the supporting substrate is exposed to humidified air having at least 65% relative humidity. Typical VIPS durations are 5 seconds to 5 minutes.

[0091] Alternatively or additionally, NIPS is performed in which the mix on substrate is immersed in a bath comprising a liquid nonsolvent for liquid induced phase separation LIPS. This method of LIPS is faster than VIPS and useful for a fast continuous production process, for example when a belt is guided through various stations in a continuous process.

[0092] Advantageously, when using LIPS, the temperature of the coagulation bath is at maximum 14°C or even better at maximum 10°C. It was observed that the gas crossover increases with increasing temperature in the coagulation bath.

[0093] The phase inversion itself can be combined with the hydrolyzation of the precursor, for example during immersion in the NIPS nonsolvent bath. For the hydrolyzation with water, this coagulation bath would require a certain content of water, for example at least 10w%. In some useful embodiments, however, the portion of nonsolvent, for example water, is advantageously in the range of 40w% to 100w%.

[0094] Alternatively, the NIPS coagulation bath is not used for the hydrolyzation, but a further bath is used for the hydrolyzation, for example, containing at least 10w% water. A hydrolyzation bath specifically for hydrolyzing the precursor is also advantageous, if the NIPS is done by exposure to vapor, VIPS. Such bath may remove the stabilizing agent, at least to a large extent. However, it is pointed out that a bath is not strictly necessary for the hydrolyzation, as other methods can be used, for example spray treatment, where the spray contains a certain concentration of water, for example at least 10w%.

[0095] The NIPS is done with a NIPS-agent, be it in liquid form for LIPS or in vapor form for VIPS. The NIPS-agent comprises a NIPS -nonsolvent selected among water, methanol, ethanol, propanol, isopropanol, butanol, iso-butanol, hexanol, acetone, methyl-ethyl ketone, formic acid, acetic acid, propionic acid, butyric acid, acetylacetone, ethyl acetate, ammonium acetate, ethylenediaminetetraacetic acid, potassium hydroxide and mixtures thereof.

[0096] In some instances, it has been found advantageous, if the NIPS-agent, in addition to the nonsolvent, comprises a minor portion of solvent. For example, such NIPS-solvent is selected among NMP, DMSO, NEP, NBP, DMAc, DMF, DMI, GVL, Cyrene, methyl 5-(dimethylamino)-2-methyl-5-oxopentanoate, the latter marketed as Polarclean™, or mixtures thereof. This portion of NIPS-solvent is less than NIPS -nonsolvent, and typically in the order of 1-50 w%. In experiments, NMP was mixed with an aqueous solution of acetic acid (5w% acetic acid, 95w% water) in weight ratios between 5:95 and 50:50.

[0097] The mix is optionally applied onto a rigid substrate. However, for large-scale production, the casting onto a flexible substrate is advantageous, for example a belt that is guided through various stations, for example comprising a slot-die station and one or more baths.

[0098] Advantageously, the casting is further conducted onto a reinforcing porous substrate, which becomes part of the final membrane, placed adjacent to the belt. Typically, the thickness of the final membrane on the supporting substrate is in the range 100-1200 micrometer, for example 150-1000 micrometer, optionally 200-700 micrometer

[0099] Typically, the membrane is separated from the substrate after production, although, this is not strictly necessary, as the substrate can form part of the final product as a composite membrane with a support and a membrane produced on one or both sides of the support, and where such composite membrane is used for the final purpose, for example in an electrolyzer. In case that the substrate is part of the final product, it can be covered on both sides with the membrane material, which would, then, have a symmetrical pore structure if treated equally on opposite sides.

[0100] To obtain the favorable pore structure, it is advantageous to cover the porous membrane support on one side with a non-porous support which does not become part of the final membrane. The one-sided coverage ensures that the NIPS treatment is one-sided.

[0101] Useful substrate materials include PPS, PEEK, SPEEK, PS, PET, ETFE, PTFE, CTFE, ECTFE, PP, PE, PMP, PI, PAI, PA, or glass.

[0102] By the method as described above, membranes are produced having a high level of porosities above 40% and even up to 95%.

[0103] A membrane produced according to the method indicated above is useful for electro- lysers in which hydrogen is produced by splitting of water. However, the membrane may also find use in other electrochemical cells, for example fuel cells and batteries.

[0104] As will be explained in more detail below, the membranes resulting from a production as described herein have good gas impermeability and low ionic resistance when soaked with liquid electrolyte, require low production costs, and exhibit increased cell efficiency due to the stabilized hydrolysis of the precursor. This has been proven by increased cell efficiency, as seen by the low resistance and the low hydrogen in oxygen gas crossover and low cell voltage, which is expected to stem from well-distributed sub- micrometer particles by the hydrolysis of the precursor which has been stabilized by a stabilizing agent within the membrane.

[0105] SHORT DESCRIPTION OF THE DRAWINGS

[0106] The invention will be explained in more detail with reference to the drawing, where FIG. 1 is a sketch illustrating the production process;

[0107] FIG. 2 is a scanning electron micrograph of a first side of the membrane;

[0108] FIG. 3 is a scanning electron micrograph of a second side of the membrane;

[0109] FIG. 4 is an EDS graph for a portion of the bulk of the membrane;

[0110] FIG. 5 A-D illustrate element- specific EDS mapping scans

[0111] FIG. 6 are experimental results from measurements when the membranes were used in alkaline electrolysis, using KOH liquid electrolyte;

[0112] FIG. 7 illustrate impedance measurements with A) fixed current density and B) fixed voltage.

[0113] FIG. 8 illustrates hydrogen in oxygen gas crossover at fixed current density;

[0114] FIG. 9 shows a polarization curve after 25 h of electrolysis of a sample DC263, in comparison to a Zirfon® reference sample;

[0115] FIG. 10 shows gas crossover during steady-state electrolysis in the 30w% KOH based alkaline electrolysis bath at 80°C with a fixed current density of 200 mA / cm2;

[0116] FIG. 11 shows a photo of the sample DC170**, which was made at 17°C.

[0117] DETAIEED DESCRIPTION / PREFERRED EMBODIMENT

[0118] Experimentally, porous membranes were produced according to the following method, which is illustrated in FIG. 1.

[0119] For the initial mix, one part PSU was dissolved in 50 to 100 parts by weight pure NMP. Pure PVP was added as a stabilizer. The PVP of the type K90 was used with 360.000 g / mol.

[0120] The weight ratio between PVP and PSU was varied in the range of 0.16 to 0.99 for different samples, and stable membranes were produced for mixes throughout this entire range. Into this solution, pure titanium butoxide Ti(0Bu)4 liquid or, alternatively, titanium tetra-iso-propoxide liquid as a precursor was added to provide a mix for subsequent phase inversion, NIPS.

[0121] The mix was poured onto a glass plate as supporting substrate and distributed to obtain well defined a layer. Stable membranes were produced for mixes throughout the entire range of thicknesses of 0.12 mm and up to 1.20 mm. From the results, it is reasonable to assume that stable membranes with thicknesses down to 0.2 mm can be produced, or even down to 0.15 or 0.1 mm.

[0122] The glass plate with the mix was immersed into a nonsolvent bath.

[0123] Various nonsolvents were tested. Stable membranes were successfully produced with the following nonsolvents:

[0124] - pure water,

[0125] - 30w% KOH, pH 14.6,

[0126] - aqueous acetic acid at pH within the range of 2-5, in particular at 2.5 and 3.5,

[0127] - aqueous acetic acid at pH within the range of 2-5, in particular at 2.5 and 3.5 + NMP up to 50w%,

[0128] - aqueous acetic acid at pH within the range of 2-5, in particular at 2.5 and 3.5 + acetone up to 67w%,

[0129] - aqueous NH4-acetate at pH of 4.75 or 9.25,

[0130] - aqueous acetic acid at pH within the range of 2-5, in particular at 2.5 and 3.5 + isopropanol up to 50w%.

[0131] It turned out to be useful with a bath at a temperature not higher than 14°C, and even better if not higher than 10°C.

[0132] Stable membranes were produced at minus degrees down to -8°C for aqueous NH4- acetate, down to -7 °C for acetic acid + acetone, down to -12°C for acetic acid + NMP, and down to -12°C for KOH.

[0133] This bath created the pores by NIPS, which due to the one-sided contact with the nonsolvent result in pores larger at the side towards the glass plate and smaller pores towards the nonsolvent. The water in the bath converted the precursor into TiCh and BuOH in the hydrolysis process.

[0134] The TiCh tend to form sub-micron particles of various sizes but no agglomerations were observed, which is due to the stabilizing agent.

[0135] It has not been able to verify whether a portion of the TiCh is additionally embedded in non-particular form in the polymer matrix. However, as the metal oxide is converted from a liquid form, this cannot be excluded and may contribute to some extent to the good performance of the membranes, which, however, remains speculative at this stage.

[0136] It has turned out that the bath for the NIPS did not remove all of the PVP, which remains in the NIPS-treated membrane and suppresses agglomeration of the TiCh particles during the hydrolyzation of the Ti(0Bu)4.

[0137] For sake of clarification, the NIPS and the hydrolyzation can be achieved in a single bath or in two baths, or alternatively by VIPS followed by a bath containing water. The method of using two steps has the advantage of being able to tune the pore size of the membrane. This may be achieved by choosing different amounts of water in the two steps, which allows a control of the rate of hydrolysis of the precursor and thereby the amount of hydrophilic particles, specifically on the side facing the nonsolvent. As described above, it was previously found that the amount of hydrophilic particles influences the pore size. In particular, higher amounts of particles lead to smaller pore sizes.

[0138] In such processes, it is possible to achieve a partial hydrolyzation of the precursor during the pore formation. The hydrolyzation can then be finalized in a subsequent step, for example in a water bath. Hydrolyzation would typically require at least 5w%, or better at least 10w%, water in the nonsolvent bath, for example for the combined treatment of NIPS and hydrolyzation.

[0139] The finally obtained membrane in the experiments floated off the glass plate, and could easily be collected. The following FIG. 2-8 are related to an experimental sample internally denoted as DC173, which was produced according to the above method and had a thickness of 620 micrometer. This porous membrane was produced in which 1 weight unit of PVP was used for 2 weight units of PSU, yielding a weight ratio of 0.5. The content of TiCh was 23.8w% relatively to the total weight of the final membrane. NMP was used as solvent, and water was used as nonsolvent. NIPS was done in a bath into which the mix on the supporting glass plate was immersed. The bath had a temperature of 4°C.

[0140] FIG. 2 and 3 are scanning electron micrographs made by a ZEISS-SEM of opposite sides of the membrane. The side of the membrane that was towards the NIPS agent is shown in FIG. 2 at 50k times magnification, and the opposite side in contact with the supporting glass plate is shown in FIG. 3 at 10k times magnification. By comparison, it is observed that the pores on the side with the NIPS treatment and opposite to the side covered by the glass are much smaller. The pores at the covered side have an averaged size up to 3 micrometer, whereas the pores on the NIPS-exposed side are an order of magnitude smaller and in average size smaller than 0.2 micrometer. Also observed in FIG. 3 are particles in the porous polymer matrix of the membrane.

[0141] Porosities were found to be in the range of 45%-95%.

[0142] FIG. 4 shows an Energy Dispersive X-ray Spectroscopy, EDS, graph taken on the crosssection of the sample of FIG. 2, with the image of the area from which the measurements were taken is inserted into the graph as an overlay. The graph has indications of the elements identified on the surface. Nitrogen was measured in the bulk, which was investigated further. When comparing spectra from the surface, as in FIG. 2, with spectra inside the porous structure, as in FIG. 3, it was found that nitrogen was measured in the inner porous structure. This nitrogen was believed to be due to a portion of the PVP not having been completely removed from the membrane.

[0143] FIG. 5A illustrates a scanning electron micrograph of a selected region of the sample of FIG. 2 and 3. Clearly visible is a particle in the porous matrix. For identification of the chemical elements, filtered mapping for specific elements was applied in FIG. 5B, 5C, and 5D. The results for carbon are illustrated in FIG. 5B, not giving any specific signal from the particle. However, as seen in FIG. 4C, titanium is clearly visible, and as observed in FIG. 5D, also oxygen is enhanced in the location of the particle, indicating the particle as being TiCh, especially, when the signals strengths of titanium and oxygen were compared with the carbon background level.

[0144] FIG. 6 illustrates performance of the 0.62 mm thick membrane as produced according to the method and with the parameters above and as illustrated in FIG. 2-4. It was tested in 30w% KOH based alkaline electrolysis at 80°C with nickel (Ni) electrodes. The experimental results are illustrated in comparison to a Zirfon® membrane as a benchmark. It is seen that the performance of the produced membrane was improved compared to a Zirfon® membrane. This means that such membranes are useful alternatives to Zirfon® membranes in alkaline electrolysis.

[0145] FIG. 7A illustrates measurements of the membrane in the 30w% KOH based alkaline electrolysis bath at 80°C with a fixed current density of 100 mA / cm2, and FIG. 7B illustrates results at a fixed voltage of 1.3 V. It is observed that the ionic resistance is very low in the order of 0.15 Ohm- cm2.

[0146] FIG. 8 illustrates measurements of the membrane in the 30w% KOH based alkaline electrolysis bath at 80°C with a fixed current density of 220 mA / cm2during steady-state operation at a runtime between 81h and 89h. It is observed that the hydrogen in oxygen gas crossover is very low, around or less than 0.2vol% hydrogen in oxygen.

[0147] FIG. 9 and FIG 10 are related to a membrane, internally denoted as DC263, in which the non-solvent bath temperature was increase to 14°C. This porous membrane was produced in which 0.9 weight units of PVP was used for 2 weight units of PSU, yielding a weight ratio of 0.45. The content of TiO2 was 23.8w% relatively to the total weight of the final membrane. NMP was used as solvent, and a mixture of water and acetic acid at a pH of 2.5 was used as nonsolvent. NIPS was done in a bath tempered to 14°C into which the mix on the supporting glass plate was immersed. The final thickness of the membrane was 440 micrometer.

[0148] FIG. 9 shows the cell voltage as a function of the current density after 25 hours of electrolysis of the membrane DC263, in comparison to a Zirfon® reference sample. It was tested in 30w% KOH based alkaline electrolysis at 80°C with nickel (Ni) electrodes. It is seen that the performance of the produced membrane DC263 was better than a Zir- fon® membrane due to the DC263 membrane delivering identical currents at lower voltage.

[0149] FIG. 10 shows the gas crossover for the DC263 membrane during steady-state electrolysis in the 30w% KOH based alkaline electrolysis bath at 80°C with a fixed current density of 200 mA / cm2. The obtained hydrogen in oxygen crossover is around l.lvol% hydrogen in oxygen. The increased level of crossover of this sample produced in a nonsolvent bath of 14°C compared to a sample produced in a non-solvent bath with a temperature of 4°C is believed to stem from the increased hydrolysis rate of the alkoxide during the phase inversion process leading to larger pores.

[0150] To further investigate the effect of temperature, two samples were prepared at different temperatures. These samples, internally denoted as DC170 and DC170**, were produced as follows. One weight unit of PVP was used for two weight units of PSU, yielding a weight ratio of 0.5. The content of TiO2 was 33.3w% relatively to the total weight of the final membrane. NMP was used as solvent, and water was used as nonsolvent. NIPS was done as LIPS in a bath into which the mix on the supporting glass plate was immersed. The bath had a temperature of 4°C for DC 170, and a temperature of 17°C for DC170**. The obtained sample of DC170 had a thickness of 350 micrometer and showed a conductivity of 229 mOhm*cnT2.

[0151] FIG. 11 shows a picture of the sample DC 170**, which was made at 17°C. It is clearly visible, by the highlighting marks, that the sample is inhomogeneous, leading to the formation of hole-like structures. The sample DC170** was therefore not investigated further.

[0152] In conglomeration, the increasing gas crossover levels by increasing the non-solvent temperature from 4°C to 14°C,. and the non-successful membrane formation at 17°C clearly shows that the maximum allowable temperature for this process is 14°C.

[0153] These results are remarkable, as this means that an alternative to Zirfon® membranes for alkaline electrolysis has been found, where the production method is not only simple but in particular suitable for low-cost, large-scale production in a simpler way, in particular by avoiding hydrophilic metal oxide or hydroxide powders.

[0154] In the following, some further test samples and their characteristics are presented in the tables below.

[0155] For the various mixes in the production, the following materials were used:

[0156] PSU: Udel P-3500-NT-LCD from Solvay®

[0157] Precursor: Titanium-butoxide (TTBO), Titanium-isopropoxide (TTIP) both via SigmaAldrich®

[0158] PVP: K90 via SigmaAldrich®0

[0159] NMP: >99w% via SigmaAldrich®

[0160] Zr®2 particles: <100 nm (TEM) via SigmaAldrich®

[0161] The following parameters were evaluated as follows:

[0162] The resistance was evaluated in a zero gap alkaline water electrolyzer, with 2 perforated Ni plates at 4 cm2, as electrodes, and the membrane squeezed between these electrodes.

[0163] The electrolyte was 30 w% KOH and the measurement conducted at room temperature of 21 °C ± 2°C.

[0164] The resistance was evaluated from the high frequency resistance in Electrochemical Impedance Spectroscopy at DC=1 mA with 0.1 mA amplitude. An example of such is shown in FIG 7.

[0165] The thickness was evaluated with a thickness gauge, Electronic Micrometer from RS Pro®, RS Stock No. 705-1229, at 3-5 places of the membrane, and an average was defined.

[0166] The above mentioned test sample DC173 was identical to DC162 in the tables below, apart from the fact that DC173 had a thickness of 620 micrometer and DC162 700 micrometer. All percentages in weight percentages and ratios in weight-ratios. From these results, it can be understood that the production method can deliver a variety of membranes with low resistance at thicknesses as low as 120 micrometer and up to 1200 micrometer. Furthermore, it shows that the addition of hydrophilic metal oxide or hydroxide particles to the mix is not required to decrease the resistance of the mem- branes. Although the invention is not limited to it, it demonstrates that the invention does not require the addition of hydrophilic metal oxide or hydroxide particles, and it may optionally be chosen to add particles to the mix.

Claims

CLAIMS1. Method of producing a porous membrane for alkaline water electrolysis, comprising- providing a polymer solution in which a polymer resin for the porous membrane is dissolved in a solvent,- selecting an alkoxide of an inorganic metal as a precursor for conversion into hydrophilic metal oxide or metal hydroxide particles,- in a mixing step, mixing the precursor and the polymer solution to obtain a mix;- in a conversion step, converting the precursor in the layer into metal oxide particles or metal hydroxide particles by hydrolyzing the precursor, characterized in that the method further comprises- selecting a stabilizing agent among agents that are suppressing agglomeration of metal oxide or metal hydroxide particles during conversion of the precursor, wherein the stabilizing agent is not a solvent for the polymer resin;- in the mixing step, mixing also the stabilizing agent with the polymer solution and the precursor to obtain the mix; and exposing the layer to nonsolvent-induced phase separation, NIPS, for pore formation.

2. The method according to claim 1, wherein the method comprises carrying out the pore formation by NIPS simultaneously with hydrolyzation of a portion or all of the precursor or carrying out the pore formation in a NIPS step prior to hydrolyzation of the precursor but not after the hydrolyzation.

3. The method according to claim 2, wherein the method is free of using Carbowax.

4. The method according to anyone of the claims 1-3, wherein the method comprises adjusting the content of metal oxide particles or metal hydroxide particles by hydrolyzing the precursor to at least 20w% in the porous membrane as compared to the total weight of the porous membrane.

5. The method according to anyone of the claims 1-3, wherein the method comprises adding a powder of sub-micrometer particles of a hydrophilic metal oxide, sulfate orhydroxide, for example from the group of Zr, Ti, Ce, Hf, Ba, Al, in the mixing step for obtaining the mix with increased concentration of hydrophilic inorganic particles in the porous membrane, wherein at least 10w% of the total content of the metal oxide, metal sulfate, and metal hydroxide in the membrane is achieved by hydrolysis of the precursor.

6. The method according to any preceding claim, wherein the method comprising adjusting concentrations to at least one of the following:- the amount of precursor in the mix is in the range of 8w% to 25w%.- the content of the stabilizing agent in the mix is in the range of 0.5w% to 15w% relative to the weight of the mix,- the weight-ratio between the precursor and the polymer resin in the mix is in the range of 0.4 (w / w) to 2.5 (w / w);- the weight-ratio of stabilizing agent to the precursor in the mix is in the range of 0.05 (w / w) to 2.00 (w / w), optionally 0.12 (w / w) to 0.75 (w / w).

7. The method according to any preceding claim, wherein the method comprising adjusting concentrations to all of the following:- the amount of precursor in the mix is in the range of 8w% to 25w%.- the content of the stabilizing agent in the mix is in the range of 0.5w% to 15w% relative to the weight of the mix,- the weight-ratio between the precursor and the polymer resin in the mix is in the range of 0.4 (w / w) to 2.5 (w / w);- the weight-ratio of stabilizing agent to the precursor in the mix is in the range of 0.05 (w / w) to 2.00 (w / w), optionally 0.12 (w / w) to 0.75 (w / w).

8. The method according to any preceding claim, wherein the stabilizing agent is selected from the group of Polyvinyl alcohol, Polyethylene oxide, Cellulose Acetate, Glycerol, and Polyvinylpyrrolidone.

9. The method according to any preceding claim, wherein the alkoxide is selected from the group of a methoxide, an ethoxide, a propoxide, an iso-propoxide, a butoxide, a tert- butoxide, a pentoxide, a hexoxide and mixtures thereof.

10. The method according to anyone of the claims claim 1-9, wherein the method comprises providing a layer of the mix on a non-porous substrate and exposing only one side of the layer to the nonsolvent for one-sided NIPS and obtaining the porous membrane on the substrate by the NIPS, and only after the NIPS separating the porous membrane from the substrate.

11. The method according to claim 10, wherein the NIPS is done by liquid induced phase separation, LIPS, in a coagulation bath at a temperature not higher than 14°C.

12. The method according to anyone of the claims claim 1-9, wherein the NIPS is done by liquid induced phase separation, LIPS, in a coagulation bath at a temperature not higher than 14°C.

13. The method according to claim 11 or 12, wherein the method comprises providing the bath with 40w% to 100w% nonsolvent, for example water.

14. The method according to anyone of the claims 1-9, wherein the NIPS comprises exposure to vapor of the nonsolvent for vapor- induced phase separation, VIPS.

15. The method according to any one of the claims 1-9, wherein the method comprises a step with NIPS and a subsequent additional step comprising immersion into a bath that contains at least 10w% water to hydrolyze the precursor.

16. The method according to any preceding claim, wherein the polymer resin is PSU, PES, PPS, PPSU, PEEK, PEKK, PP, PE, PMP, PI, PAI, PA, PAN, PS, SEB, SEBS.

17. The method according to any preceding claim, wherein the metal for the precursor is from the group of Zr, Ti, Ce, Hf, Ba, Al, Cu, Li, Na, K, Si, V, Sc, Y, Ca, Mg, Sr, optionally from the group of Zr, Ti, Ce, Hf, Ba, Al.

18. The method according to any preceding claim, wherein the NIPS is done with a NIPS -agent comprising a NIPS -nonsolvent, which is a nonsolvent to the polymer, selected among methanol, ethanol, propanol, isopropanol, butanol, iso-butanol, hexanol, water, acetone, methyl-ethyl ketone, formic acid, acetic acid, propionic acid, butyric acid, acetylacetone, ethyl acetate, ammonium acetate, ethylenediaminetetraacetic acid, potassium hydroxide and mixtures thereof.

19. The method according to claim 18, wherein the NIPS-agent also comprises a NIPS-solvent which is a solvent to the polymer and wherein the relative content of the NIPS-solvent in the NIPS-agent is in the range of l-50w%.

20. The method according to claim 19, wherein the NIPS-solvent is selected among NMP, DMSO, NEP, NBP, DMAc, DMF, DMI, GVL, Cyrene, methyl 5-(dime- thylamino)-2-methyl-5-oxopentanoate, the latter marketed as Polarclean™, or mixtures thereof.

21. The method according to any preceding claim, wherein- the metal for the precursor is selected from the group of Zr, Ti, Ce, Hf, Ba, Al, Cu, Li, Na, K, Si, V, Sc, Y, Ca, Mg, Sr, optionally from the group of Zr, Ti, Ce, Hf, Ba, Al- the alkoxide for the precursor is selected from the group of a methoxide, an ethoxide, a propoxide, an iso-propoxide, a butoxide, a tert-butoxide, a pentoxide, and a hexoxide.- the amount of precursor in the mix is in the range of 8w% to 25w%,- the stabilizing agent is selected from the group of Polyvinyl alcohol, Polyethylene oxide, Cellulose Acetate, Glycerol, and Polyvinylpyrrolidone;- the content of the stabilizing agent in the mix is in the range of 0.5w% to 15w% relative to the weight of the mix,- the weight-ratio between the precursor and the polymer resin in the mix is in the range of 0.4 (w / w) to 2.5 (w / w);- the weight-ratio of stabilizing agent to the precursor in the mix is in the range of 0.05 (w / w) to 2.00 (w / w).

22. A porous membrane produced by a method according to anyone of the claims1-21.

23. The membrane according to claim 22, wherein the average pore size of the pores on a first side of the membrane is less than 0.3 micrometer and the average pore size on a second, opposite side of the membrane is larger but less than 3 micrometer, and wherein the membrane has a porosity in the range of 40-95%.

24. An alkaline electrolyzer for production of hydrogen gas comprising a membrane according to claim 22 or 23.

Images