WO2025231331A1 - Water electrolyzer - Google Patents

Abstract

A direct impure water electrolysis (DIWE) approach generates green hydrogen in a modified proton-exchange membrane pure water electrolyzer (PEM-PWE), that avoids fouling, corrosion, deactivation, and side reactions normally caused by the ions in impure or saline waters. Conventional electrolyzers require ultrapure deionized (DI) water as feed because: 1) the proton-exchange membrane (PEM) and electrocatalysts are readily poisoned by the anions, e.g., chloride, and cations, e.g., sodium, calcium, and magnesium that are present in seawater or brackish water; and 2) the chloride anions readily form chlorine at the PEM-electrolyzer anode, which is toxic and corrosive. This adds substantially to the cost and complexity of the electrolyzer plant due to the water treatment plant needed for producing ultrapure DI water. The tolerance of impure water as described herein avoids reverse osmosis and deionization requirements steps which is beneficial for use in semi-arid regions with a paucity of fresh water.

Description

WATER ELECTROLYZER

Inventors: Ravindra Datta, Patrick Emerick,

Srivatsava V. Puranam and Erik P. Kinstler

Attorney Docket No.: VLT24-01PCT

BACKGROUND

Worldwide electricity generation is expected to double by 2050, in large part due to the rapidly increasing demand from economically emerging nations. It is further expected that by 2050, two-thirds of the power generation will be from variable renewable electricity I'VRE) generators, e.g., solar and wind, replacing the conventional power plants based on fossil fuels, i.e., coal, and natural gas. Temporally responsive electrical energy storage, including batteries and water electrolyzers, would be a key enabler in this transition, since solar and wind power are unpredictable and, unlike fossil-fueled power plants, cannot be turned on-or-off to meet demand. Thus, there is a need for storing excess renewable power generated during periods of low demand.

Green hydrogen (H₂) from electrolysis of water, is expected to serve as the central link between a variable renewable electric (VRE) grid and hard to abate energy sectors, storing as hydrogen the excess energy from VRE generators, solar and wind, when the electricity demand is low, and using it to generate electricity when the demand is high. Green hydrogen may also be employed as a fuel in transportation and/or as a feedstock in the chemical industry, where when combined with captured or recycled C0₂, it can replace the conventional fossil feedstocks, or can provide green ammonia when combined with N₂ as a H₂ carrier or as a fertilizer. Further, it can help to decarbonize other large-scale industries including steel and cement manufacture. The low-temperature electrolysis (LTE) candidate technologies poised to be the most widely employed in green H₂ generation include: 1) the liquid alkaline -water electrolyzer (L-AWE) involving a porous diaphragm (PD) and an alkaline water electrolyte feed; and 2) the proton-exchange membrane pure water electrolyzer (PEM-PWE) with a pure water feed. The L-AWE is a mature, century-old technology, that however, is not suitable for direct integration with VRE. For this, the PEM- PWE is the leading candidate. However, the PEM-PWE requires ultrapure deionized (DI) water as feed to avoid fouling of membranes and catalysts by the ionic impurities in water, which adds to cost and complexity of the balance-of-plant (BOP). This invention modifies the conventional PEM-PWE such that it can tolerate a direct impure water feed, thus reducing the cost, complexity, and bulk of the BOP.

SUMMARY

A water electrolysis device includes a membrane-electrode assembly (MEA) incorporating a catalyzed proton-exchange membrane disposed between an anode consisting of a porous-transport layer (PTL) flanked by an anode flow field connected to an aqueous input and for removing oxygen produced at the anode, and a cathode comprising of another PTL flanked by a cathode flow field for removing the hydrogen gas produced at the cathode. A voltage source connects between the anode (positive electrode) flow field and the cathode (negative electrode) flow field for imparting a voltage differential across the MEA to cause water electrolysis. A selective layer disposed between the anode and/or cathode flow field and the aqueous input prevents the passage of ions and other impurities to the MEA, such that the water need not undergo rigorous deionization and/or filtration pretreatment as in conventional PEM-PWE. The selective layer is a porous hydrophobic (water-fearing) layer, such as, Teflon®, that is further aerophilic (gas-loving) and thus allows for the passage of water vapor as well as the evolved gases while preventing feed liquid-water, along with its impurities, to pass through.

Configurations herein describe a direct impure water electrolyzer (DIWE), thus adapted from a conventional proton-exchange membrane -pure water electrolyzer (PEM- PWE), that allows direct impure, brackish, or saline water electrolysis, based on the use of one or more in-situ porous hydrophobic layers (PHLs), or membranes, for retaining dissolved salts and any other impurities in impure water feeds, while allowing only molecularly pure water vapor to permeate its hydrophobic pores to reach the MEA, and removing the gases produced. In the disclosed approach, the water vaporization necessary for transport through the PHL occurs in-situ within the cell, and does not need any external energy. Thus, the required heat of vaporization is supplied largely by the heat of the oxygen evolution reaction (OER) at the anode, combined with any heat of vapor condensation within the MEA. Further, the electrolyzer is designed to retain the condensed liquid water within the MEA so that the PEM remains well-hydrated as does the ionomer in the catalyst layers, providing good performance.

In a particular configuration, the porous hydrophobic layer (PHL) for in situ water purification is interposed, along with necessary modifications to the flow-fields and gaskets, between the impure liquid-water feed flow-field channels at the anode and a conventional PEM-electrolyzer membrane-electrode assemble (MEA). The latter, for instance, typically includes a titanium (Ti) anode flow-field, a porous Au- or Pt-plated Ti felt porous transport layer (PTL) for the oxygen electrode, an IrCE catalyst for the oxygen evolution reaction (OER), a proton-exchange membrane (PEM) such as Nafion®, or a nanocomposite-Nafion®, the latter being more suited to a low humidity, higher-temperature cell environment, a Pt/C catalyst for a Hydrogen Evolution Reaction (HER), a carbon gas-diffusion layer (C-GDL) with a microporous layer (MPL) for the negative electrode (cathode), and a graphite cathode flow field.

A PEM water electrolyzer can be modified as disclosed herein for use with a direct impure or saline water feed with little modification to the cell architecture, facilitating adaptation of existing PEM-PWE manufacturing/ assembly processes with small changes, along with a smaller balance of plant (BOP). Furthermore, the direct removal of evolved gases via the PHL without significant bubble formation at the electrode which otherwise cover the catalyst limiting its effectiveness and resulting in an overpotential, results in efficient operation of the overall electrolyzer system. Such an electrolyzer would be especially useful in terrestrial or extra-terrestrial applications with a paucity of fresh water sources, and/or an absence of gravity or presence of micro-gravity, where the absence of buoyancy forces makes the removal of evolved gas bubbles from the catalyst as well as gasliquid separation substantially more challenging.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing and other objects, features and advantages of the invention will be apparent from the following description of particular embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.

Fig. 1 is a context diagram of an electrolyzer for producing hydrogen as an energy vector and as a raw material for various industries;

Fig. 1 A shows a block diagram of a PEM water electrolyzer in the context of Fig. 1 suitable for use with configurations herein;

Fig. 2 shows a schematic diagram of a conventional PEM electrolyzer plant showing a PEM-PWE stack and the balance-of-plant (BOP) including water purification steps; Fig. 3 shows a schematic of available processes that can occur during the electrolysis of saline water;

Figs. 4 and 4A show a particular configuration of a direct impure water electrolyzer (DIWE) cell with a single porous hydrophobic layer (PHL);

Fig. 5 shows a DIWE cell with a vapor knock-out pot and condenser to recycle the pure water for better cathode humidification and improved performance combined with generation of a desalinated pure water byproduct;

Fig. 6 shows a single PHL embodiment of the invention with impure water feed at both electrodes;

Fig. 7 shows a DIWE structure with twin porous hydrophobic layers (PHLs) for improved PEM and ionomer hydration and reduced resistance;

Fig. 8 shows a DIWE structure with twin PHLs with feed of the direct impure water to the cathode while the hydrophobic membrane at the anode helps to improve PEM and ionomer hydration and cell performance;

Fig. 9 depicts a context of a simplified BOP for the DIWE system;

Fig. 10 shows water-vapor diffusion toward the catalyst layer (CL), with counterdiffusion of O2, through the porous transport layer (PTL), microporous layer (MPL), the CL, and the dense ionomer (I) film encapsulating the catalyst. Also shown schematically is the formation of salt concentration polarization and temperature polarization phenomena at the porous hydrophobic layer-impure feed interface that lower transmembrane flux of pure water;

Fig. 11 is a graph of predicted anode limiting current density for water-vapor fed electrolyzer as a function of cell temperature, assuming AT = T — T_m = 5 °C;

Fig. 12 is a graph of predicted polarization plots for liquid-water versus water-vapor fed PEM electrolyzer;

Fig. 13 is a graph of conductivity of Nation® versus relative humidity (RH) of water vapor;

Fig. 14 shows experimental results in the form of polarization plots for liquid pure- water electrolyzer (PWE) versus DIWE cell with an anode PHL for a DI water feed at 60 °C and 80 °C with a flow rate of 175 mL/min;

Fig. 15 shows the effect of a switch to a saline water feed on performance of a PHL- DIWE cell and a conventional PEM-electrolyzer at 60 °C;

Fig. 16 shows experimental polarization plots showing the effect of salt concentration in the feed water to DIWE at 60°C; and Fig. 17 provides experimental DIWE cell voltage over a period of 60 h operating at 600 mA/cm² current density on raw tap water in the town of Spencer, Massachusetts (USA), with a flow rate of 175 mL/min.

DETAILED DESCRIPTION

The description and drawings below depict example configurations of electrolysis using the disclosed water electrolyzer suitable for use with impure (non-deionized, nondesalinized) water sources in several configuration incorporating a selective membrane for controlling and shifting the electrolysis reaction for accommodating various input streams and output products. Fig. 1 is a context diagram of an electrolyzer 100 for producing hydrogen as an energy vector and as a raw material for various industries. Referring to Fig.l, green hydrogen (H₂) from electrolysis of water is expected to serve as the central link between a variable renewable electric (VRE) grid and hard to abate energy sectors, storing excess energy from VRE generators, solar and wind, when the electricity demand is low, and employed as a fuel in transportation and/or as a feedstock in the chemical industry, where when combined with captured or recycled C0₂, it can replace the conventional fossil feedstocks, petroleum and natural gas, or can provide green ammonia when combined with N₂ as a hydrogen carrier or as a fertilizer. It can also potentially help to decarbonize other large-scale industries including steel and cement manufacture. It should be noted that the notion of “green” hydrogen production refers to a reduction or absence of combustion, high temperatures and carbon emissions associated with conventional hydrogen production, often relying on fossil fuels, as from methane steam reforming.

One technology likely to be most widely employed in such green H₂ generation is low-temperature electrolysis (LTE) of water, because of its technological maturity and high efficiency, and its ability to potentially directly use DC from solar and wind generators. The two commercially mature LTE technologies are: 1) the liquid alkaline -water electrolyzer (L- AWE) involving a porous diaphragm (PD) and an alkaline water electrolyte feed; and 2) the proton-exchange membrane pure water electrolyzer (PEM-PWE) with a pure deionized (Dl)-water feed. The L-AWE is an inexpensive, scalable, and commercially available century-old technology, but suffers from low current density, inefficiency, and slow-start and limited operational flexibility needed for coupling to the solar or wind generators, which can instantaneously go from zero to full power or vice versa, or for using grid electricity during off-peak hours. The replacement of the porous diaphragm by an anion- exchange membrane (AEM) in an alkaline water electrolyzer (AEM-AWE) could potentially provide lower Ohmic losses and better gas separation and flexibility. Unfortunately, the durability of AEMs in alkaline electrolytes is low, although AEMs with better stability are being actively researched, so that it is not yet commercially deployed. However, the AEM-AWE technology can benefit from additional development and refinement to promote it as viable for commercial deployment.

In contrast to the L-AWE, the commercially available PEM-PWE is responsive and amenable to start-stop operation and thus direct VRE integration, further allowing differential operating pressure, high current and power densities, and high efficiencies. The main issues with PEM-PWE are that: 1) it requires ultra-high purity water, 2) its cost is high because of the requirement of prccious-group metal (PGM) catalysts along with an anode porous transport layer (PTL) that is based on Pt- or Au-plated titanium, and 3) its life is limited in large part because the MEA eventually gets poisoned by ionic impurities in even ultra-high purity DI water. In fact, the realization of the envisioned green hydrogen economy depends in large part on the underlying expectations that: 1) the cost of the electrolyzers can be reduced dramatically; and that 2) there is an adequate supply of potable water collocated with abundant sunshine and/or wind resources across the globe. However, many arid areas that have a high VRE potential have only a limited supply of fresh water.

Fig. 1 A shows a block diagram of the proposed water electrolyzer in the context of Fig. 1 suitable for use with configurations herein. Referring to Figs. 1 and 1A, the water electrolyzer 100 is generally defined by a containment 101 for encapsulating a layered or stacked structure for fluid and gaseous exchange as provided by the respective layers. The containment 101 includes an anode 120 in an anode flow field 125 in fluid communication with an aqueous input 122, and a cathode 140 in a cathode flow field 145 configured for passing a gaseous product 142. Both the aqueous input 122 and the gaseous product 142 are in fluid communication with a respective impure water feed 123 and a gaseous output 143, and form a generally contained fluid volume.

A proton-exchange membrane (PEM) 150 is disposed between the anode flow field 125 and the cathode flow field 145 and provides for proton migration for hydrogen gas (Hz) generation, based on appropriate catalysts as described below. A voltage source 105 connects between the anode 120 and the cathode 140 for imparting a voltage differential across the proton-exchange membrane 150. The layered structure of the containment 101 provides appropriate electrical communication between the electrode (anode 120 and cathode 140) layers, the voltage source 105 and the flow fields 125, 145. A selective membrane 155’ is disposed between the anode 120 and the aqueous input 122 preventing passage of contaminants in the aqueous input 122 and relieves the need for ultrapure water found in conventional electrolysis operations.

The selective membrane 155’ is typically a porous hydrophobic layer (PHL) 155 permeable to water vapor for allowing only passage of gaseous water and a return of oxygen gas (O2) to be expelled. It is expected, therefore that the selective membrane is hydrophobic and while permeable to water vapor and oxygen gas is impermeable to liquid water and the contaminants therein. In general, the PHL 155 is a hydrophobic membrane resistive to the transport of dissolved ions, while other solid contaminants and particles will likely have been filtered from the impure water feed 123.

Fig. 2 shows a schematic diagram of a conventional PEM electrolyzer 10 plant showing a system 200 employing conventional PEM-PWE stack and the balance-of-plant (BOP) including the various water purification steps. The attainment of low-cost targets would require steep reductions in both operating and capital expenses of the electrolyzer stack, as well as those of the balance-of-pant (BOP), as shown in Fig, 2. Referring to Figs. 1 and 2, a significant factor in operating expenses is the cost of renewable electricity, which declines substantially during off-hours. Thus, electrolyzers should be able to operate dynamically to access the low-cost VRE, with frequent starts and stops.

For reducing the capital expenses, cheaper and Earth-abundant materials that are further not supply constrained must be utilized, and the BOP, i.e., the DI water purification and the power conditioning plants, needs to be further simplified and provide a smaller footprint. Conventional low-temperature electrolyzers require ultra-pure DI water as feed so as not to contaminate the membranes and the electrodes, and to not produce any chlorine containing side products. This adds substantially to the cost and footprint of the overall electrolyzer plant, which is roughly divided as a third for the power conditioning plant, a third for water purification, gas-separation, and pure water recirculation plant, and a third for the electrolyzer stack. In short, to meet the cost targets, 1) the electrolyzers must be capable of being directly coupled to the variable DC generators to reduce power conditioning costs, and 2) should preferably be able to work directly with impure water feeds containing ions to reduce water treatment and recycle costs.

There is, however, a paucity of fresh water in many arid parts of the world that are otherwise endowed with abundant sunshine and wind resources. Saline water, ranging from brackish to sea water, is characterized by a high dissolved salt content, mainly NaCl, present as Na⁺ and Cl- ions, ranging from >1,000 (0.1%) for brackish water to 30,000 ppm (3.0 %) for sea water, but also containing hard-water components such as Ca²⁺, Mg^{2 +}, and C0|“ ions. Other impurities can include organics, microplastics, bacteria, particulates, and gases, which can foul the membrane, electrodes, and catalysts. The process from raw water to the ultrapure water required for electrolysis can be divided into two steps (Fig. 2): 1) pretreatment of raw water such as desalination at 202, and 2) polishing 204 to the required ultrapure standard needed for the electrolyzer stack. The goal is to produce water with the preferred quality of ASTM Type I DI water, i.e., with a resistivity of >10 Mfl cm.

The pretreatment steps, depending on the source of raw water, include, e.g., aeration and sand filtration to remove dissolved redox-active species involving iron and manganese, and filtration and other steps including ultrafiltration and UV to remove particles, organics, and microorganisms. Beyond the pretreatment steps, production of ultrapure water suitable for electrolyzers include removal of hardness due to multivalent cations such as Ca²⁺ and Mg²⁺, which are exchanged with soluble monovalent cation Na⁺ via water softening, and finally removal of most of the ionic load via reverse osmosis (RO). The permeate from the first RO step is passed through a secondary RO system for further reduction of salts. The permeate from the first RO process is filtered again in a secondary RO system. Thereafter, to produce ultrapure water with very low conductivity required for electrolyzers, a final deionization (DI) or an electrodeionization (EDI) step is used, wherein any remaining ions are exchanged with for H⁺ and OH-. All these steps can add substantially to the BOP footprint and cost, as typically 10 L of ultrapure water is needed per kg of H₂ produced. Clearly, if impure/saline water, after the pretreatment steps, but containing most of the ions present in raw water such as sea water could be directly used in a PEM electrolyzer 100, it could provide a significant reduction in size and cost of the plant. This is the objective of configurations described herein.

However, there are significant challenges for using direct impure/saline/sea water feeds to PEM or alkaline water electrolyzers, including:

1. The fouling and blockage of the ion-exchange membrane by the ions such as Ca^{2 +}, Mg²⁺, and CO^{2 -}, causing a significant increase in membrane resistance, and a reduction in lifetime.

2. The fouling and blockage of the electrocatalysts by these same ions including via precipitation as alkaline-earth hydroxides and carbonates.

3. The chlorine evolution chloride reaction (CER) under low pH conditions at the anode from the chloride (Cl-) anion in feedwater that competes with the oxygen evolution reaction (OER), and the formation of hypochlorite ion (0C1- ) at higher pH.

4. The corrosion of cell components and catalysts caused by the Cl- and 0C1- anions. Fig. 3 shows a schematic of available processes that can occur during the electrolysis of saline water. Referring to Fig. 3, if there is salt (NaCl) in saline/sea water, it results in chloride (Cl-) anion in it via ionization, which can lead to the undesirable chlorine evolution reaction (CER) at the anode, as follows:

2NaC 2Na⁺ + 2C1-

2C1- Cl₂ + 2e- l₂ (5) + 2Na⁺ + 2e- ; 0° ₀ = +1.36 V (pH = 0) in addition to the oxygen evolution reaction (OER), since CER has an electrode standard Nernst potential (O° ₀ ⁼ +1-36 V vs SHE at pH = 0) not much higher than that of the OER (OQ = + 1.223 V vs SHE at pH = 0). Here, SHE is the standard hydrogen electrode potential. The electrode operating potential is typically significantly higher to overcome the large overpotential of the OER under acidic conditions, so that there can be a significant overpotential promoting CER as well, as the CER is typically a more facile reaction than the OER. The generation of chlorine in a PEM water electrolyzer is undesirable because of its toxic and corrosive nature. Further, the resulting sodium cations (Na⁺) cross over to the cathode and result in the generation of NaOH and H₂: ; O° ₀ = -0.828 V (pH = 14) where the electrode standard Nernst potential 02 ₀ = —0.828 V at a pH = 14.

The overall process of these undesirable side reactions in the cell may hence be represented by the electrolysis of saline/sea water:

Side Reaction: 2H₂O + 2NaCl 2Na0H + H₂ + Cl₂ ; T_o° = O°,₀ - 02, ₀ = 2.1875 V which, in fact, represents the commercial chlor-alkali process for producing NaOH and Cl₂ from brine, the H₂ being a side product in the production of chlorine and caustic soda. The consumption of protons and the generation of NaOH at the cathode further results in a local pH increase at the electrode-electrolyte interface. This can further cause precipitation at the electrode due to the formation of sparingly soluble species such as Ca(OH)₂, Mg(OH)₂, and Ca(CO)₃, from the Ca²⁺, Mg²⁺, and C0|“ ions present in saline water besides NaCl.

Under conditions when the anode is not strongly acidic but rather neutral or alkaline, chloride ion can undergo alternate electrode reactions, e.g., with the hydroxyl ion to produce the hypochlorite ion:

NaCl Na⁺ + Cl" cr + 20H“ H₂0 + ocr + 2e“

Anode NaCl + 20H“ H₂0 + Na⁺ + 0C1" + 2e" ; 0>° ₀ = +0.89 V with the electrode standard Nernst potential +" ₀ = +0.89 V at a pH = 14. In comparison, the ORR at a pH = 14 has a standard Nernst potential <J>° ₀ = +0.401 V, and an operating voltage, including OER overpotential, of around 0.7 to 0.8 V, i.e., below the Nernst potential of 0.89 V, so that conditions are typically not conducive to the formation of hypochlorite ion (0 Cl “ ). In other words, interference from chloride ion chemistry at the OER electrode is thermodynamically unlikely under neutral or alkaline conditions (pH > 7.5). Thus, the maximum thermodynamic potential difference between the chloride chemistry and the water oxidation is 480 mV for pH > 7.5, for chloride to be oxidized to hypochlorous acid, HC1O, or to the hypochlorite ion, 0C1-. Thus, liquid-alkaline water electrolyzers (L-AWE) do not suffer from chloride-based side reactions, and can hence tolerate some salinity in feed water. On the other hand, the PEM-PWE are extremely sensitive to saline-water feeds.

Thus, the conventional approach to avoid chlorine generation and other side reactions, membrane fouling, and electrode precipitation in PEM-PWEs is to rigorously desalinate a saline water feed via reverse osmosis followed by deionization (DI) or electrodeionization (EDI) to reduce the salt and other ion content down to virtually zero to provide an ultra-pure water feed, as discussed above. However, this is expensive, and the water treatment plant has a large footprint. This is especially an issue where space is at a premium. For instance, there is a particular interest in offshore direct seawater electrolysis systems for their compactness, since plant footprint dominates the installation costs. However, because of microbial content and other particulate contaminants that are not dissolved, a simple pretreatment including filtration of the saline or natural water feeds is essential even for direct water feeds.

It is noteworthy that conventional approaches aimed at developing direct seawater electrolysis have been based on alkaline water electrolyzers, which operate at low current densities and are less sensitive to ionic impurities in feed water and can largely avoid the unwanted chloride reactions, as discussed in the preceding. However, L-AWEs are not suitable for direct integration with VRE. On the other hand, PEM-PWE operate at high current densities providing high performance and dynamic operation suitable for VRE integration, and are especially attractive where space is at a premium, e.g., on an offshore platforms. However, there have been no reports so far of efforts to develop PEM-watcr electrolyzers for direct saline water electrolysis. Thus, we believe that the invention described here wherein a PEM-PWE is equipped with an in-situ porous hydrophobic layer (PHL) that allows only pure water vapor to access its MEA and operates at higher current densities than the L-AWE is without precedence . However, the conceptual basis of this invention is in part supported by a report wherein the researchers proposed to utilize an externally raised vapor feed from saline water to the anode of a PEM-PWE, thus avoiding the CER and the attendant corrosion issues, while the saline water feed is sent to the cathode chamber, where chloride electrode chemistry is not a concern, and the Nafion® PEM used is able to largely reject the chloride ions diffusion from the cathode to the anode.

Turning now to the beneficial approach disclosed herein, configurations below adapt a PEM-electrolyzer for direct impure water electrolysis by equipping it with one or two in-situ porous hydrophobic layers (PHLs), that are highly effective, by retaining all dissolved salts and any other impurities in impure liquid water feeds, while allowing only molecularly pure water vapor to permeate its hydrophobic pores and access the MEA. In our approach, the membrane-based water vaporization occurs within the cell, and does not need any input of external energy. Rather, the required heat of vaporization is supplied in-situ by the heat of the oxygen evolution reaction (OER) at the anode, combined with any heat of vapor condensation within the ionomer layer in catalyst layer. Further, the electrolyzer is designed to retain liquid water within the cell so that the PEM remains well-hydrated as does the ionomer in the catalyst layers, providing good performance. In fact, especially in the two PHL configuration, since the PHL forbids any liquid to enter or exit, the PEM can also support non-volatile liquid or gel acids (e.g., phosphoric acid), or solid acids (e.g., heteropolyacids), or functionalized acidic materials (e.g., sulfated zirconia), providing higher proton conductivities and allowing higher operating temperatures. Thus disclosed are electrochemical cells and methods that allow the direct use of impure water for electrolysis without resulting in any fouling of the PEM electrolyzer cell membrane or catalysts by the ions present in saline water, or the corrosion caused by the generation of chlorine gas or hypochlorite ion within the cell. Further, the aerophilic nature of the PHL that is in close proximity to the catalyst layer where the gases are evolved, draws the formed gases directly into the PHL pores, without significant nucleation and bubble formation on the catalyst surface that can cover the catalyst surface. This improves the catalyst utilization as well as avoids the associated overpotential.

Figs. 4 and 4A show a particular configuration of a direct impure water electrolyzer (DIWE) cell with a single porous hydrophobic layer (PHL) 155. Referring to Figs. 1 A and 4, a porous hydrophobic layer (PHL) 155 for water purification is interposed, with appropriate modifications to the flow-fields and gaskets for maintaining the layered structure of the containment 101, between the impure liquid-water feed channels and a PEM electrolyzer membrane-electrode assemble (MEA) 104. The latter typically comprises of a titanium (Ti) anode flow-field 425, a porous Au- or Pt-plated Ti felt porous transport layer (PTL) for the oxygen electrode, or cathode 120, an IrCL catalyst 121 for the OER, the proton-exchange membrane (PEM) 150 such as Nafion®, or a nanocomposite-Nafion®, the latter being more suited to a low humidity, higher-temperature cell, a Pt/C catalyst 141 for the HER, a carbon gas-diffusion layer (GDL) with a microporous layer (MPL) for the negative electrode (cathode), and a graphite cathode flow field 445. In short, it is noteworthy that the disclosed PEM water electrolyzer can be used with a direct impure water feed 123 with feasible modification to the cell or containment 101 architecture, facilitating adaptation of existing PEM-PWE manufacturing/ assembly processes with small changes.

In the configuration of Fig. 4, the anode flow field 425 includes the anode layer 120 and an anode catalyst layer 121, such that the anode layer 120 is electrically coupled to the aqueous flow and the anode catalyst layer 121 in communication with the proton-exchange membrane 150. The anode layer 120 includes titanium and the anode catalyst layer 121 includes iridium, where the iridium contributes to an oxygen evolution reaction separating protons and electrons from oxygen. Similarly, the cathode flow field 445 includes the cathode layer 140 and a cathode catalyst layer 141, where the cathode layer 140 is electrically coupled with the gaseous product 143 in the cathode flow field 445. The cathode layer 140 includes graphite and the cathode catalyst layer 141 includes platinum supported on carbon particles, such that the platinum contributes to the hydrogen evolution reaction forming hydrogen gas (Hi). As shown further in Fig. 4, the raw tap/brackish/saline/sea water feed 123 containing the salts and other impurities, after the pretreatment steps, e.g., of flocculation and filtration, but without the usual RO/DI/EDI polishing steps, does not come in direct contact with the MEA. Rather, driven by its chemical potential difference, water 420 evaporates on the PHL feed side 423, permeates through its pores 424 as shown schematically in Fig. 4A, and thence through the Ti PTL, finally condensing on the ionomer in the catalyst layer (CL), and diffusing through it to the catalyst site where it undergoes OER to produce H ⁺ , e“and 0₂, the resulting 0 422 counter-diffusing back. The protons formed diffuse through the proton-exchange membrane carrying with them roughly 3 molecules of water of hydration each from the anode to the cathode via electroosmosis.

Also produced in the CL is heat due to the OER irreversibility and its large overpotential, along with that due to water vapor condensation in the CL, which is conducted to the water feed side of the PHL via a temperature gradient, resulting in in situ thermal integration and a high energy efficiency. It is worth stressing that no external energy is needed in the process of in-situ water evaporation in the membrane.

The proposed MEA 104 structure (Fig. 4) thus ensures that the water coming in contact with the MEA is molecularly pure and completely devoid of any ions and any other nonvolatile impurities, including any remaining macromolecules or colloids, which are left behind in the feed water on the feed side of the PHL. The GDL at the hydrogen electrode may be further hydrophobicized with PTFE treatment as shown in Fig. 4, much as is done in the conventional PEM fuel cell, to reduce the loss of water vapor in the H₂ effluent stream and the resulting drying of the Nafion® membrane and hence an increased membrane resistance.

Fig. 5 shows a DIWE cell with a vapor knock-out pot and condenser to recycle the pure liquid water for better cathode humidification and improved performance combined with generation of a fully desalinated pure water byproduct as a result of electroosmotic transport of water from the anode to the cathode via the membrane. In the alternate configuration as shown in Fig. 5, to further avoid drying of the vapor-equilibrated Nafion® membrane and the resulting increase in PEM 150 resistance further exacerbated by electroosmosis, the water vapor 503 issuing from the cathode, along with the hydrogen gas product 143, can be condensed in a knock-out (KO) pot 501, and returned as a molecularly pure liquid back to the cathode chamber 505 to facilitate PEM humidification and improved performance.

The excess pure liquid water 510 hence produced can also be utilized as ultrapure desalinated water byproduct from the saline water feed 123. Assuming that the liquid water transported from anode 120 to cathode 140 across the PEM 150 is via electro-osmosis, roughly 2-3 molecules of H₂0 per H⁺, or about 4-6 molecules of H₂0 per molecule of H₂ produced, it amounts to roughly 30-50 kg of ultrapure water per kg of H₂ produced, which is a substantial amount of desalinated pure water, that does not require input of external energy. In comparison, about 10 kg of water is consumed per kg of H₂ produced.

Fig. 6 shows a single PHL 155 cell configuration with impure water feed at both electrodes. In this configuration, the impure water feed 123, 123’ can in fact be introduced at both the anode 120 and the cathode 140, and/or cycled via 523, 523’ as shown in Fig. 6. It may be recalled that the chloride ion in saline water causes unwanted side reactions including the CER mainly at the anode rather than at the cathode, which is at a lower potential on the HER scale. These are avoided in this embodiment by the in-situ vaporization of impure water in the PHL. Since the cathode is not greatly affected by ions, it can also be fed with saline water even in the absence of a PHL. This results in little chlorine evolution, since the chloride ion is largely rejected by the PEM 150. On the other hand, the PEM can still be deleteriously affected by the cations in impure water if present in large concentrations.

Fig. 7 shows a DIWE structure with twin porous hydrophobic layers (PHLs) 155- L.155-2 for improved PEM and ionomer hydration and reduced resistance. Referring to Figs. 6 and 7, in still another configuration, the DIWE cell may be configured with two porous hydrophobic layers 155-1..155-2, with the second layer in the cathode chamber 505 in an essentially mirror configuration to the first layer in the anode chamber 705. The purpose of the second porous hydrophobic layer (PHL) 155-2 is to help retain liquid water within the cathode and thus provide improved hydration of the ionomer in the CLs and the PEM 150.

In fact, the configuration in Fig. 7 is also suitable for retaining any aqueous liquid electrolyte, e.g., an acidic (e.g., phosphoric acid) electrolyte, within the MEA to assist in ion transport and allow higher operating temperatures while avoiding recirculation or pumping of such electrolyte as well as the attendant corrosion issues normally caused by a liquid electrolyte because of its encapsulation within the PHL 150 “sandwich,” or layered structure.

Fig. 8 shows a DIWE structure with twin PHLs 155 with feed of the direct impure water 823 to the cathode 140. In yet another configuration of the twin-layer (155-1..155-2) configuration shown in Fig. 8, the impure water 823 feed is introduced at the cathode 140, rather than at the anode 120. The hydrophobic layer 155-1 at the anode 120 helps to improve PEM 150 and ionomer hydration and cell performance. Although, the second layer may still lose vapor because of the elevated cell temperature, it will still serve to retain liquid water and maintain higher hydration levels within the MEA.

Fig. 9 depicts a context of a simplified BOP for the DIWE system 900, in contrast to the conventional system of Fig. 2. Referring to Figs. 109, the overall DIWE system 900 including the electrolyzer 100 and the attendant simplified BOP is shown schematically in Figure 9. While the pretreatment-steps 902 for raw impure water are still needed, the energy and capital intensive steps of RO, DI, or EDI are avoided, thus greatly simplifying the BOP for water treatment and purification system normally needed for electrolyzer grade ultrapure water as shown in Fig. 2 at 202. Furthermore, the recycling system of water after separation from oxygen is not needed, as indeed this stream would have a higher salinity or impurities than the raw water feed. Thus, this may be rejected or used as a brine feed for a chlor-alkali plant. However, pure water recovered from hydrogen separation may be recycled to the cathode for improved cell hydration, as shown in Fig. 9. Finally, the excess water stream from the hydrogen KO drum is a pure desalinated water side-product that is suitable for drinking or other purposes, and does not require any input of external energy for its generation.

It is beneficial to validate and explore the efficacy of the proposed DIWE cell 100 with the help of a robust theoretical model. Thus, we utilize a theoretical description of the DIWE cell based on a comprehensive modeling approach for conventional PEM water electrolyzers. The model is suitable for analysis of a single electrolyzer cell as well as an electrolyzer stack. As discussed above, an issue in the proposed DIWE design is the limitation on the cell performance imposed by: 1) a lower limiting current density for vaporphase water, as a result of the additional mass transfer resistance posed by the PHE; and 2 ) a higher resistance because of lower PEM and ionomer hydration when in contact with the vapor-phase water. It is noteworthy, that mass transfer limitations with a liquid feed are largely absent in a conventional PEM electrolyzer, except as a result of bubble coverage of the catalyst layer.

Fig. 10 schematically shows transport of water vapor as water-vapor diffusion toward the catalyst layer (CL), with counter-diffusion of O2, through the porous transport layer (PTL), microporous layer (MPL), the CL, and the dense ionomer (I) film encapsulating the catalyst, along with profiles of temperature and the partial pressures p, of H 0 and 0 . Also shown schematically is the formation of salt concentration polarization and temperature polarization phenomena at the porous hydrophobic layer-impure feed interface that lower transmembrane flux of pure water. Thus, as shown in Fig. 4A, the hydrophobic nature of the PHL forbids the passage of liquid water through its pores. Driven by a water chemical potential difference across the membrane, thus, the water must evaporate and diffuse through the PHL pores while the produced 0₂ gas diffuses in the reverse direction.

The permeation mechanisms include ordinary diffusion, Knudsen diffusion, and D’Arcy flow. The resulting permeance of the water through the PHL, P_PHL, determines in a large part the limiting current density at the anode, namely, i_{+ L}. The water vapor then diffuses across the porous transport layer (PTL) and any microporous layer (MPL), and finally the dense ionomer (I) film encapsulating the catalyst to reach the active sites, where it undergoes OER to produce H⁺, e“and 0₂, the resulting 0₂ counter-diffusing back. Also produced in the CL is heat due to the OER irreversibility and its large overpotential, along with that due to any water vapor condensation in the CL, which is conducted to the water feed side of the hydrophobic layer via a temperature gradient. The coupled heat and mass transfer as shown schematically in the inset on the right in Fig. 10, thus, determine the transport of pure water vapor to the catalyst.

Besides the additional transport resistance of the porous hydrophobic layer, thus, we may expect higher proton transport resistance in the PEM as well as in the ionomer film, as vapor-equilibrated Nafion® has a lower water content than that in equilibrium with liquid water. Further, we may expect the catalyst roughness, y_M, to be reduced as a result. Means to counter the effect of this is to optimize the equivalent weight of Nafion® membrane and gel, and also to incorporate inorganic nanoparticles in these to allow better humidification at the lower RH expected at the anode. Further, of course, such nanocomposite PEMs allow higher temperature operation, which would be beneficial in DIWE.

For the DIWE, the actual voltage V at a given current density i differs from the cell Nernst potential F_o because of internal overpotentials Dj, i.e., V = F_o + .ry Thus: where the second and third terms on the right represent overpotentials of the positive (+) and negative (— ) electrodes, r}₊ and r _, respectively, i_{+ 0} and i_ ₀ are the exchange-current densities, i'_{+ L} and i_ _L are the corresponding limiting current densities, reflecting masstransfer limitations to the electrode surface, while E_r is the effectiveness factor accounting bulk-diffusion limitations across the porous electrode thickness. This can also be construed as the effective roughness, or the effective electrochemical active surface area (ECSA) of the catalyst layer as a ratio of the MEA area. The last term on the right includes the Ohmic potential drop in the PEM, where a is its ionic conductivity, which varies with the relative humidity (RH), and L is the PEM thickness. The anode limiting current density because of diffusion limitations of water vapor with a feed vapor pressure p^₂o where 2 represents the stoichiometric coefficients of electrons per mol of H₂0, while its total permeance PH₂O i^s due t° contributions from the different layers including from PHL, PpHM ^ar|d YM i^s the catalyst roughness factor and P is the permeance of the ionomer layer in the CL, shown in the inset on the left in Fig. 10.

Fig. 11 is thus a graph of predicted anode limiting current density for water-vapor fed electrolyzer as a function of cell temperature, assuming AT = T — T_m = 5 °C. Of course, water vapor pressure p^₂o i^s strong function of temperature, and as a result so is the limiting current density for a vapor-fed WE. Thus, predicted limiting current density as a function of temperature is plotted in Fig. 11, using a permeance PH₂O ~ 0-3 cm s ’ is estimated based on literature data, i.e., roughly a third of that for the GDL of 1.96 cm s^-1 when including the PHL in DIWE cell design. The hence predicted limiting current density for the anode is shown in Fig. 11 as a function of operating temperature.

In other words, we may conclude that feed of water to the CL as a vapor is unlikely to substantially impede the performance of the proposed DIWE electrolyzer at temperatures of 80 °C or above, although the performance may be expected to be lower because of the PHL resistance, an increase in PEM resistance owing to lower hydration and a reduction in effective catalyst roughness. The use of nanocomposite-Nafion® membranes should allow higher temperatures, especially for pressurized electrolyzer operation. Thus, the optimal current density for DIWE is not as high as for PEM-PWE with a pure liquid DI water feed. However, DIWE would have a lower PGM catalyst loading due to the lower current density and the absence of impurities even in DI water that deactivate the catalysts and membranes, and reduce electrolyzer life. Fig. 12 is a graph of predicted polarization plots for the performance of a pure water electrolyzer (PWE) vs. DIWE. The results are reminiscent of those of Fornaciari et al. who compared performance of a PEM-PWE with liquid vs. externally raised vapor feed, and found that the high-frequency resistance increased by an order of magnitude to roughly 1 ft cm² at 80 °C. Thus, while with a liquid feed the HFR is quite low (~ 0.1 fl cm²) and independent of the current density, that with vapor feed is higher (~ 0.5 fl cm²), and varies with current density, being high at lower and higher current densities, in part because PEM resistance increases at higher currents due to additional membrane dehydration via electroosmosis. However, this may be alleviated in part by recycling water recovered from hydrogen knock-out drum (Fig. 5).

In short, there are the following factors that increase the slope of the polarization plot for DIWE versus that of PWE: 1) the hydration level 2 (defined as moles of water per mole of sulfonic acid site) of the PEM reduces substantially from - 20 for liquid-water equilibrated Nafion® to -13 for vapor-equilibrated Nation® at low current densities, and to - 6 at high current densities; 2) reduction of the effective catalyst roughness factor due to shrinking of the ionomer particles in the catalyst layer; and 3) mass transfer limitations of the vapor across the microporous hydrophobic layer which reduces the anode limiting current density. Keeping these factors in mind, the ionomer loading at the DIWE electrodes should be higher for a vapor feed. Further, in the absence of any leaching liquid water in the MEA in some of the proposed configurations, e.g., those with twin PHLs 155, the PFSA membrane, or an alternate such as a PBI membrane, could be doped with another liquid or gel acid electrolyte such as phosphoric acid to enhance the PEM 150 conductivity.

As described above, the proposed MEA structure for the subject invention, the direct impure water electrolyzer (DIWE) cell, comprises of a water purification membrane, e.g., a microporous hydrophobic layer, that is simply interposed, along with necessary modifications to the current collectors and gaskets, between the liquid-water feed channels and a conventional state-of-the-art PEM electrolyzer MEA. The latter comprises of a porous Au- or Pt-plated Ti felt PTL for the oxygen electrode, an IrO catalyst (0.4-1 mg cm^-2) for OER, a proton-exchange membrane (PEM) such as Nafion® (e.g., NR212, N115, N117), a Pt/C catalyst (0.1-0.4 mg cm^-2) for HER, and a carbon GDL with an MPL for the negative electrode (cathode). The catalyst loadings mentioned here are somewhat lower than the typical precious group metal (PGM) loadings used in current PEM electrolyzers, but reflect the desired future loadings, and used in a benchmark MEA by the US Department of Energy. It is, in short, noteworthy that the commercial PEM water electrolyzer can be used with a direct impure water feed with small modifications to its cell architecture, i.e., a simple incorporation of one or more water purification membranes (i.e., PHL) within the cell along with the necessary attendant changes such as in gaskets and in the bipolar plate to accommodate the water-purification layer. The remaining cell components remain virtually unchanged from those in the commercial PEM-PWE, although the optimized characteristics of these components may in general be somewhat different in a DIWE operating on impure water than in a PEM-PWE with a pure DI water feed. These optimized characteristics may include different catalyst loadings, different ionomer loadings, different PEM equivalent weight and thickness, as well as inclusion of non-volatile liquid, gel, or solid acidic materials/oxide nanoparticles for higher-temperature/lower RH operation, and different flow field architecture, wherein the current collection is in a picture-frame configuration around the PHL. Thus, the main choice to be made in DIWE design is the selection of the in-situ water purification membrane. Thereupon, the design and fabrication of other standard PEM- PWE components can be fine-tuned to optimize the DIWE cell. The discussion below is thus limited to the selection of proton-exchange membrane and the microporous hydrophobic layer for water purification, as the nature of the other components including the catalysts, electrodes, and the bipolar plates is virtually identical to those in a state-of-the-art PEM-PWE.

Fig. 13 is a graph of conductivity of Nafion® versus relative humidity (RH) of water vapor. The desired proton-exchange membranes (PEMs), e.g., Nafion®, or more generally, cation-exchange membranes (CEMs) used in the proposed cell are selective cation (A⁺) conducting and electronically insulating membranes. The CEMs generally constitute a neutral hydrophobic polymer backbone bonded to repeating hydrophilic pendant ionomer units. The pendant ionomer units include anionic functional groups including carboxylates, sulfonates, phosphonates, and the like. The strength of the pendant ionomer units, e.g., — (COO“)H⁺ versus — (SC>3 )H⁺, is used to accomplish a desired pH within the membrane. In CEMs, the pendant ionomer groups are charge -balanced by exchangeable and mobile cations such as mono-, di-, or higher-valent cations A⁺, such as protons, H⁺, or alkaline earth metals, such as Li⁺, Na⁺, and K⁺, that act as charge carriers for the membrane.

When hydrated, the membrane swells, and the protons attached to the sulfonic acid groups dissociate forming hydronium ions within aqueous clusters away from the hydrophobic backbone that are responsible for proton transport via Grotthuss and ordinary diffusion mechanisms providing high conductivity. The conductivity of the membrane is strongly dependent on the extent of hydration characterized by , the number of water molecules per ionomer group, e.g., the sulfonic acid group. The parameter A is not only dependent on the nature of the polymer and its equivalent weight (EW), but also on whether the membrane is equilibrated with liquid or vaporous water, the relative humidity (RH), as well as with the operating temperature. Thus, the conductivity a increases over two orders of magnitude with membrane water content which rises with RH, as shown in Fig. 13. for a Nafion® membrane equilibrated with water vapor, e.g., from 1 mS/cm at 20% RH to roughly 100 mS/cm when fully hydrated (100% RH) with A « 13. On the other hand, the A « 20, when the membrane is in contact with liquid water. This is a part of the reason that the membrane resistance is higher in a vapor equilibrated Nafion® membrane, as is the case in the present invention, resulting in a higher slope of the polarization plot for DIWE as compared with conventional PEM-PWE, e.g., as predicted in Fig. 12.

In principle, there are a number of possible membranes available for in situ water purification within an electrolyzer. These include those for: 1) reverse osmosis (RO): 2) forward osmosis (FO); 3) electrodialysis (ED); 4) membrane distillation (MD), and 5) pervaporation (PV). The selection of suitable membrane may be made based on cost, desired water purity, and energy consumption. Thus, while RO remains the desalination technology of choice for potable water because of its low energy consumption and commercial maturity, it is unable to produce DI water of the quality needed for electrolysis, and is thus followed by DI or EDI steps in the BOP, as shown in Fig. 2. The membranes used in MD/PV technologies, on the other hand, are potentially suitable for our application, as they can provide water with the needed purity combined with low relatively energy consumption, especially in our DIWE invention where the energy required for vaporization on feed side of the membrane is supplied in part by the heat generated by the electrode reaction in the proximate catalyst layer, as schematically illustrated in Fig. 10.

The preference for the porous hydrophobic layer (PHL) for use in DIWE from among those employed in MD/PV applications is based on the mechanism of selective water transport and the corresponding flux. In a porous hydrophobic layer (PHL) as used, e.g., in membrane distillation (MD) process, the transport mechanism includes vaporization at the hydrophobic membrane -liquid water interface, followed by vapor transport via combined Knudsen diffusion, molecular diffusion, and Poiseuille flow, providing a higher flux and thus a higher limiting current density than the alternative PV membranes operating on the mechanism of solution-diffusion, size exclusion, or charge exclusion, through a hydrophilic dense or molecular sieving thin layer on a supporting underlayer.

Fig. 14 shows experimental results in the form of polarization plots for liquid pure- water electrolyzer (PWE) versus DIWE cell with an anode PHL for a DI water feed at 60 °C and 80 °C with a flow rate of 175 mL/min. The polarization plot depicts a comparison of a conventional PEM-PWE with that of a DIWE cell equipped with a porous hydrophobic layer (PHL) at the anode, both fed with a DI water feed. Other experimental conditions are described below.

Thus, two 5 cm² electrolyzers, a traditional PEM-PWE, termed here as pure water electrolyzer (PWE), and another identical one but equipped with a porous hydrophobic layer (PHL), termed here as direct impure water electrolyzer (DIWE), were fabricated and operated in parallel with the same feed and identical operating conditions. The catalyst- coated membrane (CCM) included a Nafion 115 membrane, coated on one side with a 3 mg/cm² PtRuO. anode catalyst, and on the other with a 3 mg/cm² Pt black cathode catalyst. This was sandwiched between an anode porous transport layer (PTL) was a 0.4 mm thick, Grade 2 Ti felt, with - 55% porosity, and plated with a very thin layer of Pt -50 mg/cm², and a cathode gas-diffusion layer (GDL) that was 0.45mm thick, carbon cloth, - 63% porosity, and with a wet-proofed microporous layer (MPL). The DIWE additionally included an e-PTFE porous hydrophobic layer with a 1 mm pore size and a thickness of 150 mm.

The results in Fig. 14 show that as expected: 1) the slope of the DIWE polarization plot is higher due to lower hydration and higher resistance of the PEM and the catalyst ionomer (Figure 13); and 2) although not evident in Fig. 14, the limiting current density of DIWE is reduced from that of the conventional PEM-PWE because of the additional resistance to water vapor transport offered by the porous hydrophobic layer. It is to be noted that the contribution of the temperature polarization effect is included in the additional transport resistance of the PHL, although the contribution of the concentration polarization effect is not present for these experiments with a DI water feed.

Fig. 15 shows the comparative behavior of a conventional PWE and a DIWE cell when the feed is switched from DI water to a saline water feed (2 g/L NaCl) at 4.5 h into a run with DI water operating at a 60 °C and a current density of 500 mA cm^-2. It is seen in Fig. 15 that while the conventional PEM electrolyzer died almost immediately, the DIWE cell continued to operate with a barely perceptible increase in cell voltage, for this salt concentration, which is quite saline, but still an order of magnitude lower than seawater.

Fig. 16 shows the effect of increasing salt concentration in feed water on the polarization behavior of a PEM-DIWE with an anode PHL operating at 60 °C. As expected, the slope of the polarization curve beyond the kinetic region (low current densities) increases with increasing salt concentration. This could be ascribed to the additional effect of concentration polarization on the PHL surface, as shown schematically in the inset on the right in Fig. 10.

Thus, the higher salt concentration at the membrane surface has two conceivable effects: 1) it reduces the vapor pressure at the membrane surface and hence the flux driving force, i.e., N_w = PPHM ~ Pw,a)’ where p_w f is the partial pressure of water at the feedside, and p_{w d} is that on the downstream side, and 2) the overall PHL permeability, PPHM, including the membrane and the salty liquid-phase film, reduces by virtue of the salt buildup on the membrane surface. This is evident in the polarization curve at the highest salt concentration in Fig. 16, where it curves upward at higher current density, an indication of transport limitations. It is clear from these results, nonetheless, that the DIWE cell behaves largely as expected

Fig. 17 shows the stability of the DIWE cell operating on domestic raw tap water in the town of Spencer, Massachusetts (USA). The sharp initial rise in voltage is as the PEM, which is initially saturated with water, dries out to within a few minutes to reach a steady state with the water vapor coming across the PHL. This is thus the total overpotential, including that due to increased PEM resistance plus that due to the mass transport resistance across the PHL, which may be ascribed to the inclusion of the in situ anode PHL within the DIWE, as compared with the conventional PEM-PWE. The subsequent slow rise may be ascribed to the fouling of the PHL by the non-ionic impurities in the unfiltered raw tap water used in this experiment.

While the system and methods defined herein have been particularly shown and described with references to embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.

Claims

CLAIMS What is claimed is:

1. An electrolysis device, comprising: an anode in an anode flow field in fluid communication with an aqueous input; a cathode in a cathode flow field configured for passing a gaseous product; a proton-exchange membrane disposed between the anode flow field and the cathode flow field; a voltage source connected between the anode and the cathode for imparting a voltage differential across the proton-exchange membrane; and a selective membrane between the anode and the aqueous input for preventing passage of contaminants in the aqueous input.

2. The device of claim 1 wherein the selective membrane is a porous hydrophobic layer permeable to water vapor.

3. The device of claim 2 wherein the selective membrane is a hydrophobic membrane resistive to the passage of liquid water and ions and contaminants therein.

4. The device of claim 1 wherein the selective membrane is hydrophobic and permeable to water vapor and gaseous products.

5. The device of claim 1 wherein the anode flow field includes an anode layer and an anode catalyst layer, the anode layer electrically coupled to the aqueous flow and the anode catalyst layer in communication with the proton-exchange membrane.

6. The device of claim 5 wherein the anode layer includes titanium and the anode catalyst layer includes iridium, the iridium contributing to an oxygen evolution reaction separating protons and electrons from oxygen.

7. The device of claim 1 wherein the cathode flow field includes a cathode layer and a cathode catalyst layer, the cathode layer electrically coupled with the gaseous product in the cathode flow field.

8. The device of claim 7 wherein the cathode layer includes graphite and the cathode catalyst layer includes platinum, the platinum contributing to a hydrogen evolution reaction forming hydrogen gas (H2).

9. The device of claim 2 wherein the porous hydrophobic layer includes at least one of polymeric, ceramic, or carbon materials based on a hydrophobicity and pore size for passing water vapor and hydrogen and oxygen gas while retaining liquid water and contaminants therein.

10. The device of claim 7 wherein a heat of vaporization at the selective membrane is generated by heat resulting from an oxygen evolution reaction (OER) occurring at the anode.

11. The device of claim 10 wherein the heat of vaporization at the selective membrane is received from a heat of vapor condensation within an ionomer layer in the cathode catalyst layer in communication with the cathode flow field.

12. A proton-exchange membrane water electrolyzer flow cell device for impure water comprising: a cathode configured to be in fluid communication with a hydrogen recovery system configured for gas dehydration and water recycling; an anode configured to be in fluid communication with a direct impure water source and an oxygen recovery system; a hydrogen evolution catalyst layer in contact with the cathode; a proton-exchange membrane (PEM) in contact with the hydrogen evolution catalyst layer; an oxygen evolution catalyst layer in contact with an opposed side of the PEM and an anode; and a porous hydrophobic layer (PHL) in contact with the anode and the direct impure water feed.

13. The flow electrolyzer cell device of claim 12, wherein the porous hydrophobic layer (PHL) provides only pure water vapor to pass to the anode, an anode flow field allowing oxygen to egress, the porous hydrophobic layer configured to retain dissolved ions, nonvolatile and particulate water impurities on an impure water feed side of the PHL; the impure water feed side configured to expel the impure water feed and the egressed oxygen.

14. The flow electrolyzer cell device of claim 12, further comprising a PHL- water interface defined by contact of the impure water feed with the PHL, a water vaporization process at the PHL-water interface utilizes the heat produced within the electrolyzer flow cell device.

15. The flow electrolyzer cell device of claim 12, wherein the cathode is in fluid communication with the water recovered from the hydrogen recovery system.

16. The flow electrolyzer cell device of claim 12, wherein a second porous hydrophobic layer is disposed between the cathode and a hydrogen flow channel.

17. The flow electrolyzer cell device of claim 16, further comprising a second direct impure water feed to the cathode.

18. The flow electrolyzer cell device of claim 12, wherein the porous hydrophobic layer includes at least one of polymeric, ceramic and carbon materials for providing a predetermined hydrophobicity.

19. The flow electrolyzer cell device of claim 12, wherein the PEM further comprises a liquid electrolyte supported on a polymeric or a ceramic support and encapsulated by the PHL.

20. A system comprising: a source of water that has been pretreated to remove particulate and microbial impurities but without the reverse osmosis (RO) or the deionization (DI) steps and thus retains salts and soluble minerals normally present in impure water sources such as city water, saline water, and sea water; a source of electrical power that has been conditioned appropriately to provide DC power to the electrolyzer cells; and a balance of plant to recover oxygen from the anode exhaust stream and discard or reuse the remaining impure water; a balance of plant to recover hydrogen from the cathode exhaust stream and recycle the recovered water to the electrolyzer cells or for other use as pure desalinated water; a plurality of flow electrolyzer cells in fluid communication with the sources of impure water, the flow electrochemical cells comprising: an anode comprising of an oxygen evolution catalyst, a cathode comprising of a hydrogen evolution catalyst, a proton-exchange membrane in contact with the anode and the cathode, respectively, a first porous hydrophobic layer between the anode and the flow plate, and/or a second porous hydrophobic layer between the cathode and the flow plate, wherein the flow electrolyzer cells are configured for hydrogen and oxygen generation and electrical energy storage.