Open AccessFeature PaperArticle

3D-Printed Reactor for Coupling Photoelectrochemical (Sea)Water Splitting with Solid-State H₂ Storage

Paweł Wyżga

^1,2,*

Joanna Macyk

Yuan-Chih Lin

Emil Høj Jensen

Matylda N. Guzik

Krzysztof Bieńkowski

⁴

Renata Solarska

⁴

and

Wojciech Macyk

^1,2,*

InPhoCat—Innovative Photocatalytic Solutions Sp. z o.o., ul. Brzask 49, 30-381 Krakow, Poland

Faculty of Chemistry, Jagiellonian University, ul. Gronostajowa 2, 30-387 Krakow, Poland

Department of Technology Systems, University of Oslo, NO-2027 Kjeller, Norway

⁴

Centre of New Technologies, Laboratory of Molecular Research for Solar Energy Innovations, University of Warsaw, ul. Banacha 2C, 02-097 Warsaw, Poland

Authors to whom correspondence should be addressed.

Catalysts 2024, 14(12), 941; https://doi.org/10.3390/catal14120941

Submission received: 18 October 2024 / Revised: 5 December 2024 / Accepted: 12 December 2024 / Published: 20 December 2024

(This article belongs to the Special Issue Environmental Catalysis in Advanced Oxidation Processes, 2nd Edition)

Download

Browse Figures

Graphical abstract
"> Figure 1
Linear sweep voltammetry (LSV, scan rate: 10 mV·s−1) experiments with Fe2O3 photoanodes, conducted in 1M NaOH under the illumination of a solar simulator: (a) Power density of simulated solar light for each photoelectrode. Electrodes are highlighted with bold lines and labels. The power density was measured at three points per electrode, and, then, the average value was taken for further analysis. (b) The photocurrent for individual Fe2O3 electrodes and all together connected in parallel (bold line). Measurements were collected under constant and modulated illumination. Reference electrode (RE): Hg/HgO; Counter electrode (CE): Ni-helix. (c) Photocurrent density normalized by light power density. (d) Applied bias photon-to-current efficiency (ABPE) derived from data in (b). "> Figure 2
Photo-assisted charging/discharging (C/D) experiments of (a,c) LaNi4.7Al0.3 and (b,d) LaNi4.3Co0.4Al0.3 cathodes. The absolute values of charging/discharging currents and potential changes upon illumination are given in red. "> Figure 3
Specific discharge capacities for LaNi5-based cathodes derived from C/D measurements. The discharge current densities are given in red. "> Figure 4
Experimental (dot), calculated (red line), and differential (black line) profiles from Rietveld refinement of electrochemically charged LaNi4.7Al0.3 (a) and LaNi4.3Co0.4Al0.3 (b) samples. Bragg peaks indicated with * appeared after C/D experiment (see <a href="#app1-catalysts-14-00941" class="html-app">Figure S5</a>) in both samples but could not be indexed with any well-known secondary phase. "> Figure 5
LSV (scan rate: 20 mV·s−1) experiments with WO3 photoanodes, conducted in acidified synthetic seawater (pH = 1.55) under the illumination of the solar simulator: (a) Light power density per electrode (average of three points). The average power density equals 52.4 mW∙cm−2. (b) The photocurrent for individual electrodes and all-together connected electrically in parallel (bold line). (c) Photocurrent density normalized by light power density. (d) ABPE efficiency derived from (b). The noisy signal for W1 and W5 samples (top row of electrodes) is probably connected to poor contact with the current collector. The color code is the same in subfigures (b–d). RE: Ag/AgCl (sat. KCl), CE: Pt-coil. "> Figure 6
LSV (scan rate: 20 mV·s−1) experiments with WO3 photoanodes, conducted in acidified 0.5M Na2SO4 (pH = 1.55) under the illumination of the solar simulator: (a) Photocurrent density normalized by light power density. (b) Applied bias photon-to-current efficiency derived from the photocurrent. The noisy signal for W1 and W5 samples (top row of electrodes) is probably connected with poor contact with the current collector. RE: Ag/AgCl (sat. KCl), CE: Pt-coil. "> Figure 7
(a) The two-electrode chronoamperometric (CA) measurement in acidified synthetic seawater (pH = 1.55) under approx. 0.5 Sun illumination at the potential of 1.23 V vs. CE. WE: 8 WO3 photoanodes connected electrically in parallel, CE: stainless steel foil. (b) The potential of WE vs. RE (Ag/AgCl electrode). It was measured independently from the CA setup using auxiliary potentiostat. "> Figure 8
(a) Photo-assisted C/D experiments of the LaNi4.3Co0.4Al0.3 cathode. Data of potential vs. RHE during 4–7. cycles are missing because the RE was temporarily above the electrolyte. (b) The first twelve hours of the experiment. The absolute values of charging/discharging currents and potential changes upon illumination are given in red. "> Figure 9
Short photo-assisted C/D experiments with different charging rates. The specific current densities of 20, 50, 60, and 200 mA·g−1 correspond to absolute current values of 2.02, 5.05, 6.06, and 20.18 mA, respectively. Modulated light was applied to photoanodes during experiments. "> Figure 10
A schematic assembly of the PEC-MH reactor: (1) an anodic compartment, (2) a cathodic compartment, (3) lids with ports for electrodes, (4) fixing plates with windows, (5) silicone gaskets, (6) an electric contact (the other seven are not shown), (7) a photoanode, (8) a reference electrode (optional), (9) cathode, (10) an ion-exchange membrane/diaphragm, and (11) the PMMA/glass plate. For clarity, some elements of the assembly are not shown, e.g., the cathode holder, additional counter electrode, hydraulic fittings, electrolyte tubing, screws, and nuts. ">

Versions Notes

Abstract

The modular photoelectrochemical (PEC) reactor accommodating eight photoelectrodes with a total active area of up to 46 cm² has been designed and manufactured using the fused deposition modeling method. The device was equipped with an electrolyte flow system, a relay module for the photoelectrode connection, and a feedback-loop module for switching between counter electrodes. The performance and durability of the system were tested within three case study experiments. The water splitting process was successfully combined with an in situ hydrogen storage in the form of metal hydride phases (confirmed by powder X-ray diffraction) using Fe₂O₃- or WO₃-based photoanodes and LaNi₅-based cathodes. The PEC water oxidation at the anodes was realized either in a strongly alkaline electrolyte (pH > 13.5) or in acidified synthetic seawater (pH < 2) for Fe₂O₃ and WO₃ electrodes, respectively. In the latter case, the photoresponse of the anodes decreased the cell charging voltage by 1.7 V at the current density of 60 mA∙g⁻¹. When the seawater was used as an anolyte, the oxygen evolution reaction was accompanied by the chlorine evolution reaction. The manufactured PEC-metal hydride reactor revealed mechanical and chemical stability during a prolonged operation over 300 h and in the broad range of pH values.

Keywords:

PEC reactor; 3D-printing; photoelectrochemical water splitting; Fe₂O₃; WO₃; metal hydride; hydrogen storage; solar battery; chlorine evolution

Graphical Abstract

1. Introduction

Solar radiation is a virtually inexhaustible source of energy with 120 PW of light power constantly crossing the Earth’s atmosphere [1]. However, the amount of available solar energy is governed by seasonal and hourly sunlight fluctuations, often incompatible with the highest energy demand periods. Consequently, temporal storage of the harvested energy, e.g., in the form of chemical bonds, and its subsequent release on-demand are required to fully exploit the potential of solar energy. The storage can be based on the formation of stable fuel, for instance, hydrogen, that can be later reused, e.g., in fuel cells for electricity generation. One of the promising solar-to-chemical energy conversion routes is based on photoelectrochemical (PEC) water splitting, in which a hydrogen evolution reaction (HER) and an oxygen evolution reaction (OER) take place at the cathode and anode, respectively, and at least one of these electrodes is photoactive. Although the state-of-the-art efficiency of the photoelectrolysis of water is below 10% (without additional bias from photovoltaic/dye-sensitized cells), it could replace the more complex PV-electrolyzer unit, which often requires the use of noble metal HER and OER catalysts, as well as other power control devices like inverters [2,3]. Nevertheless, to store and reuse produced hydrogen, both systems must be coupled with external units, e.g., gas cylinders with compressors, and fuel cells. These additional components can, in principle, be avoided when a standard cathode in a PEC cell is replaced by a hydride-forming material that functions as a hydrogen-evolving/storing electrode [4]. This can be a single element like Pd or Zr [5,6,7] or an intermetallic compound like LaNi₅ [8,9]. Hydrogen is reversibly introduced into the crystal structure of cathode material in an electrochemical manner, and it can be released in the same way (in the PEC cell) or through thermal desorption (outside the PEC cell) [4]. The working principle of such a PEC-metal hydride (MH) hybrid system, when equipped with an additional counter electrode (CE) for the discharge process, can be considered a photo-assisted rechargeable battery [10,11].

In the first approach, the PEC-MH device can be designed as a standard PEC cell operating with photoanodes, in which the cathode/counter electrode is replaced with a hydride-forming electrode [4]. For the operation of photoelectrodes, at least one face of the cell must be transparent, or a photoelectrode port should be provided. The reactor should consist of individual anodic and cathodic compartments separated by a porous diaphragm or an ion-exchange membrane (IEM) to minimize the crossover of evolved gases, an additional counter electrode (for electrochemical discharge) or a straightforward system for an introduction/removal of the cathode (for thermal discharge), a degassing option in at least the anodic compartment, and an electrolyte feeding system. Moreover, an efficient charging rate of the MH-based cathode requires a minimum electric current provided by the photoactive anode. Taking into account typical values of current densities needed for charging electrodes based on different hydride-forming compounds [12], photoactive areas larger (i.e., ≫1 cm²) than those usually used in lab-scale investigations may be required. Consequently, scaling up the whole reactor may be considered. The number of reported large PEC reactors is, however, rather low. A. Mendes et al. provided important contributions to the field by designing and investigating PEC reactors called Porto cell and CoolPEC [13,14,15]. The devices were characterized by a modular design and equipped with photoelectrodes with an active area of 20–100 cm². The CoolPEC reactor revealed a stable operation over 42 days with a single photoelectrode having an active area of 50 cm². The authors stressed two detrimental effects in the operation of such large photoelectrodes: (i) notable ohmic losses caused by the resistance of the transparent conducting oxide (TCO) substrate layer, and (ii) an increased recombination rate within the photoactive material (in that case, Fe₂O₃). To address these challenges, they proposed a modified version of the reactor, in which a single large photoelectrode was replaced by eight smaller ones with a total active area of about 28 cm². As prompted by the simulation of electric potential and current vectors, they introduced internal separators that minimize overpotential losses and parasitic ionic paths. The multi-electrode approach allowed not only for the minimization of problems reported for a previous setup version but also increased the efficiency of the whole cell due to the higher homogeneity of the smaller electrodes. A similar-sized reactor (the BiVO₄ photoanode with an active area of 64 cm²) coupled with a Si PV cell was proposed in the ArtipHyction project [16,17]. The authors recognized the problem of efficiency reduction (down to 33% of the initial value) during the long-term operation (about 1000 h) due to mass transfer limitations and bubble formation at the photoanode. To minimize the latter, the cell was equipped with separators that formed parallel channels. The follow-up simulation study discussed the pros and cons of this modified design [18]. Another example of a larger reactor (active area of 50 cm²) was reported by Becker et al. [19]. The cell body was made of poly(methyl methacrylate) (PMMA) and accommodated circular electrodes/windows as well as a compartment separator. The results of the study supported the conclusions from [18] that modular design provides a straightforward assembly and flexibility necessary for system optimization (e.g., an interelectrode distance, a type of intercompartment separator, etc.).

The manufacturing of compartments/bodies in the standard electrochemical and PEC reactors made of polymers (including already presented examples) usually follows two steps: (i) the injection molding or extrusion of a simple polymer block, and (ii) the fabrication of details (windows, ports, threads, and grooves) using subtractive manufacturing methods (e.g., milling, turning, and grinding). The commonly utilized materials include PMMA, polytetrafluoroethylene (PTFE), polycarbonate (PC), or polyether ether ketone (PEEK). Such a fabrication route is highly robust and used routinely in industry; however, it might not be suitable for fabricating complex details (e.g., internal flow channels). In the previous decade, additive manufacturing methods, also known as 3D printing, emerged as an interesting alternative for producing not only early-stage prototypes but also specialized end products [20]. Among several polymer-based 3D printing techniques, fused deposition modeling (FDM) is the most common. During operation, a thermoplastic filament is extruded through the heated nozzle onto a build plate. The pre-programmed movement of the nozzle and build plate in X/Y− and Z−directions, respectively, results in the fabrication of the object in a layer-by-layer manner. Although the printing resolution of FDM is limited to ~250 μm, there is a broad selection of compatible polymer filaments (including electrically conductive ones), and FDM desktop printers are nowadays readily available. Reported examples of 3D-printed PEC cells or their components include simple open cells with electrodes fully immersed in the electrolyte [21,22,23] as well as small modular reactors [24,25]. Their designs were adjusted to the investigated photoelectrodes and sometimes equipped with additional ports for electric contacts, electrolyte circulation, or gas collection. Nevertheless, none of the reported reactors has reached the size and complexity of demonstrators fabricated with traditional subtractive manufacturing.

In this work, we designed a photoelectrochemical reactor accommodating eight photoelectrodes with up to 46 cm² in total active area and operating under the flow of water-based electrolytes. The reactor was manufactured by the FDM technique, followed by necessary post-processing. The constructed system was used in several (photo)electrochemical experiments (presented as three case studies). This report demonstrates that additive manufacturing can be successfully utilized for producing fully functional and flexible scaled-up reactors. The presented results provide a proof-of-concept example of the PEC-MH hybrid device with LaNi₅-based hydrogen-storage cathodes, supplemented by simultaneous seawater splitting. The manufactured system can be utilized for various photocatalytic and photoelectrochemical reactions and processes.

2. Results and Discussion

2.1. Case Study 1—PEC Water Splitting Coupled with MH-Based Hydrogen Storage in Alkaline Electrolyte

The concept of combining PEC water splitting with the in situ storage of generated hydrogen by hydride-forming intermetallic compound was thoroughly discussed by Lin et al. [4]. When such a device operates in an aqueous solution, water oxidation takes place at the anode, whereas water reduction at the cathode is completed by the formation of a hydride phase. The hydrogen absorption is reversible and thus allows for the recovery of stored hydrogen, in either thermal or electrochemical way. The combination of a PEC system with an MH-forming cathode extends the functionality of the former so that it follows the concept of a solar battery. The photocurrent generated by photoelectrode contributes to the charging current and effectively decreases the absolute value of the applied potential required for charging. For these experiments, Fe₂O₃ photoanodes and the LaNi₅-based cathodes were chosen and operated in strongly alkaline electrolytes.

The performance of the Fe₂O₃ photoelectrodes was studied by linear sweep voltammetry (LSV). The scans were measured for each electrode separately and for all of them connected electrically in parallel (bold line in Figure 1b–d and Figure S2 in Supplementary Materials). Photocurrents generated by individual photoanodes (during constant illumination) did not exceed 1.3 mA in the analyzed potential range. Recombination spikes were visible below 1.4 V vs. RHE during chopped illumination. The observed values of photocurrent densities were much lower [at least three times at 1.5 V vs. RHE for the C3 sample (Figure S2), taking into account lower light power used in this study] than those reported for analogous but smaller Ti-doped Fe₂O₃ electrodes [26], which indicates the problem with the scaling-up process and homogeneity of the samples. Interestingly, the photocurrent generated by the combination of eight electrodes was around 0.97 mA at 1.23 V vs. RHE, approximately 33% of the sum of photocurrents measured individually for each photoanode (2.92 mA). These discrepancies were decreasing along with the applied potential, i.e., 2.8 mA was observed at 1.5 V, which was 95% of the sum of individual photocurrents. Large losses at low potentials were observed previously in the multi-electrode system with one electrode shaded [15], and they were analogous to mismatch losses between different solar cells connected in parallel or series [27]. In such cases, the worst-performing electrode (i.e., the one generating the lowest photocurrent) limits the overall photocurrent of the combination of electrodes. The presented variation in photocurrents for the individual electrodes (Figure 1b) comes not only from differences in the PEC efficiency but also from the high spatial inhomogeneity of incoming light. These two factors must be improved to generate a higher overall photocurrent. To exclude the effect of light variation, a comparison of photocurrent density normalized by light power density can be made (Figure 1c). There are clearly two groups of electrodes—the one (electrodes C2–C4) with higher and the second (electrodes C1, C5, C6, and C8) with lower photoresponse. A similar trend was observed during measurements in a standard lab cell with constant power of light, and, thus, the influence of light power on electrode photoactivity plays a minor role in this case. The above-discussed phenomena and a high onset potential for the Fe₂O₃ electrodes (>900 mV) resulted in negligibly low applied bias photon-to-current efficiency (ABPE, three-electrode system).

The photo-assisted charging of hydride-forming cathodes (LaNi_4.7Al_0.3 and LaNi_4.3Co_0.4Al_0.3) vs. eight Fe₂O₃ photoanodes and dark discharging vs. Ni-helix electrode are presented in Figure 2. The average light power illuminating the reactor was in the range of 50–54.5 mW·cm⁻². To highlight the effect of generated photocurrents on the charging potential, the modulated illumination (ON/OFF every 15 min) was applied to photoelectrodes. The protocol of charging/discharging (C/D) experiments comprised (i) a single C/D cycle at the current density of 60 mA·g⁻¹ up to a 60% state of charge (SoC), (ii) a single C/D cycle at 60 mA·g⁻¹ up to a 90% SoC, and (iii) ten C/D cycles at 200 mA∙g⁻¹ up to a 90% SoC. Such a sequence was found to be optimal for reaching a high and stable specific capacity of the cathodes. Values of specific current densities were chosen based on the peak current value determined for the hydride formation by chronovoltammetry scans (CV, Figure S3) and previous literature reports. Steps (i) and (ii) of the experimental protocol are magnified in Figure 2c,d. The dark cell voltage required to charge cathodes at the current density of 60 mA∙g⁻¹ varied between −1.82 to −1.77 V and −1.85 to −1.78 V for LaNi_4.7Al_0.3 and LaNi_4.3Co_0.4Al_0.3, respectively. The higher absolute values, as expected from CV measurements, were required for charging at 200 mA·g⁻¹: −2.08 to −2.01 V and −2.08 to 2.00 V, respectively. These numbers are comparable for both electrodes. The cathode potential was between −0.05 and −0.01 V vs. RHE for 60 mA·g⁻¹ and between −0.09 and −0.16 V vs. RHE for 200 mA∙g⁻¹. When photoanodes were irradiated, the charging potential increased (its absolute value decreased) by 70–80 mV at 60 mA∙g⁻¹ for both samples. After increasing the current density to 200 mA∙g⁻¹, the changes in potential dropped to 20–35 mV. The dark potential of CE vs. RHE (calculated as the difference between E_WE-CE and E_WE-RHE) was approx. 1.7–1.8 V and 1.9 V for 60 mA∙g⁻¹ and 200 mA∙g⁻¹, respectively, which was higher than the dark current onset potential for Fe₂O₃ photoanodes (Figure S4). The small changes in the cell voltage upon illumination indicated that the water oxidation reaction at the photoanodes was driven mainly by the dark current. Increasing the specific current density (i.e., magnitude of current flowing in the system) forces photoelectrodes to operate under conditions in which their photoresponse plays a minor role. Photo-induced changes in cell potential for particular current density values remained unchanged in the subsequent charging steps, indicating a stable photoanode operation over a prolonged period.

The specific discharge capacity derived from discharge time is presented in Figure 3. Capacity for both samples reached saturation after 2–3 cycles at 200 mA∙g⁻¹ with the values of 160 and 190 mAh·g⁻¹ for LaNi_4.7Al_0.3 and LaNi_4.3Co_0.4Al_0.3, respectively. In general, determined capacities fall into the range of values reported in the literature for analogous active materials [28,29,30,31]. However, it is difficult to directly compare measurements performed in small laboratory cells (usually with Ar purging and the more concentrated KOH electrolyte) with our experiments conducted in a scaled-up reactor due to different compositions of the cathode mixture, a larger distance between the WE and the CE or the two-electrode mode of operation, to mention just a few reasons.

To directly prove the formation of metal hydrides and, thus, hydrogen sorption by the studied cathodes, C/D experiments with an analogous protocol were performed in a small laboratory cell (with the three-electrode setup) using cathodes with a larger powder mixture (300 mg). After the last charging cycle, the powders were extracted from the cathodes and analyzed by powder X-ray diffraction (PXRD). The results of the Rietveld refinements for the collected diffraction patterns are presented in Figure 4 with calculated values of unit-cell parameters given in Table S1. The collected data clearly indicated the presence of characteristic Bragg reflections associated with the formation of the pristine LaNi₅-based intermetallic phases (space group P6/mmm). However, the occurrence of the second set of diffraction peaks, shifted towards lower 2θ angles, was also apparent and indicated the formation of the second phase with the same hexagonal symmetry but increased unit-cell volume. The sample phase analysis confirmed that the latter was associated with the formation of the hydrogenated LaNi₅-based compound. After electrochemical charging (hydrogenation), the unit-cell volumes increased to 13.59(1) Å³ and 10.82(1) Å³ for LaNi_4.7Al_0.3 and LaNi_4.3Co_0.4Al_0.3, respectively, which correspond to the following chemical compositions of hydride phases: LaNi_4.7Al_0.3H_4.218(3) and LaNi_4.3Co_0.4Al_0.3H_3.371(3). Hydrogen concentrations were derived from the estimated volume expansion of 3.21 Å per hydrogen atom in the LaNi₅-based hydrides [32].

2.2. Case Study 2—PEC (Sea)Water Splitting

Performing water splitting in seawater attracts more and more attention because of the naturally large abundance and low cost of this electrolyte [33]. From the (photo)anode perspective, the utilization of seawater may introduce a chlorine evolution reaction (ClER, E° = 1.36 V [34]) taking place alongside the oxygen evolution reaction (E° = 1.23 V). Depending on the current density range, either the ClER or OER dominates the electrode reaction [35,36]. Moreover, dissolved chlorine undergoes disproportionation (E° = 1.59 V [34]), or chloride ions may hydrolyze (E° = 1.49 V [34]), leading to the formation of hypochlorite HClO acid and further acidification of the solution. Although OER reveals the lowest standard potential among the other mentioned reactions, the chlorine-involving ones require lower overpotentials because of their two-electron character (OER is the four-electron process). As suggested by Jadwiszczak et al. [5], such a chlorine-containing solution could be reused as a disinfecting agent, which would be a supplementary benefit of water splitting in addition to hydrogen generation. Case study 2 comprises short LSV (three-electrode setup) and prolonged chronoamperometry (CA, two-electrode setup) measurements using eight WO₃ photoanodes as the working electrode, steel foil as the counter electrode (CE), and Ag/AgCl (sat. KCl) as the reference electrode (RE). The experiments were performed in acidified synthetic seawater (pH < 2) in both compartments separated by Nafion 117 membrane. A pH below four is required for the stable operation of the WO₃ electrodes [37].

WO₃ electrodes revealed a relatively high photocurrent, ranging from 0.4 to 3.1 mA at 1230 mV vs. RHE, which almost saturated at 1500 mV vs. RHE (Figure 5). The photocurrent onset potential was observed approx. at 500 mV vs. RHE. Recombination peaks were not observed, which indicated the long lifetime of charge carriers in the studied electrodes. Interestingly, the photocurrent generated by all electrodes connected equaled 62% of the sum of individual photocurrents at 500 mV vs. RHE (onset potential) and exceeded 99% at 1125 mV vs. RHE. Smaller differences between both values for WO₃ in comparison to Fe₂O₃ could be associated with the higher electrical conductivity of the former [38] and the larger uniformity within a single film and between different WO₃ electrodes. Nevertheless, the variation in normalized photocurrent densities among studied electrodes (Figure 5c) suggests that the reproducibility of films could be improved, and their activity may depend on the power of incident light. The individual WO₃ electrodes revealed maximum ABPE efficiency in the range of 780–880 mV. The highest value of 0.23 % at 835 mV was determined for the W6 electrode, which was also illuminated with the strongest light. In general, higher photocurrent density and lower (by approx. 400 mV) onset potential for WO₃ anodes enable reaching higher by two orders of magnitude ABPE efficiency (three electrode setup) in comparison to Fe₂O₃.

The PEC activity of WO₃ photoanodes was additionally evaluated in the acidified Na₂SO₄ solution (pH = 1.55), in which only the OER takes place. Although the onset potential measured in Na₂SO₄ was almost the same as in seawater (approx. 500 mV vs. RHE), the increase in the photocurrent below 1000 mV is much slower in Na₂SO₄ electrolyte. For some samples (e.g., W2, W3, W4), values of the photocurrent at potentials > 800 mV vs. RHE overcame those measured in seawater (Figure 6a and Figure S6). Consequently, maximum ABPE values were shifted towards higher potentials > 900 mV, and, thus, they were slightly lower. Similar behavior was also found for nanorod-TiO₂ electrodes (measurements performed in 0.5M of NaCl and 0.5M Na₂SO₄) [39], suggesting that the ClER (and possibly other reactions involving Cl-species) is more efficient than the OER at low overpotentials (<1000 mV), independent of the type of photoanode.

The stability of the system in seawater splitting was evaluated during a prolonged CA experiment (two-electrode configuration) with eight photoanodes connected as the WE and stainless steel foil acting as the CE (Figure 7a). The measurement was performed at 1230 mV (500 mV in the first 60 s) vs. CE with anolyte circulating between the anodic half-cell and an external reservoir. Additionally, the potential of the WE vs. the Ag/AgCl electrode (WE-RE potential) was independently measured (Figure 7b). Imposing the bias of 1230 mV between the WE and the CE resulted in WE-RHE potential of 1168 mV (3. min), which indicated that the overpotential connected with resistances of electrolyte, membrane, etc., equaled 62 mV at these conditions (inset in Figure 7b). The dark current density during the first 3 min was below 0.01 mA·cm⁻² and rapidly increased to 0.3 mA·cm⁻² after switching on the light, which was associated with the drop of WE-RHE potential by 141 mV. On the other hand, the start of electrolyte circulation caused the increase in both the WE-RHE potential by 20 mV and current density by 7 µA·cm⁻². During 17.7 h of experiment, current density decreased to 0.177 mA·cm⁻² (drop of 33%), whereas the WE-RHE potential decreased to 829 mV (drop of 339 mV). The plausible reasons for such a behavior are (i) a decrease in light power during the measurement, (ii) a degradation of photoelectrodes, or (iii) an increase of overpotentials. The two first hallmarks seem to play only a minor role in the observed changes (see discussion in Supplementary Materials, Figures S7 and S8). On the other hand, the color of the stainless steel electrode was changed after the CA experiment and likely indicates a corrosion process occurring at the surface and, thus, the possible increase in the cell overpotential. This type of electrode is clearly not suitable for prolonged operation in acidified seawater, even under cathodic bias.

A characteristic odor of chlorine was readily noticeable, and the anolyte was turning more and more yellow during the CA experiment. The pH of anolyte and catholyte increased from 1.55 to 1.61 and from 1.52 to 1.69, respectively. Such an observation can be expected in the cathodic half-cell because of the hydrogen evolution reaction (HER). Concerning anodic reactions, the ClER does not influence the pH, while the OER and other reactions involving Cl-species result in acidification of the solution. An increase in pH seems to result only from the transfer of H₃O⁺ ions across the membrane to minimize the pH gradient between both compartments. Such behavior suggests that the efficiency of the OER was lower than the HER because of competition with the ClER. The upper limit of the total amount of produced chlorine (calculated from the total charge, ~486 C) reached almost 450 mg·L⁻¹, which translates into the concentration of approx. 440 ppm after the conclusion of the CA experiment (assuming 100% Faradaic efficiency of the ClER). This value is rather highly overestimated due to the preference for the OER process at current densities below 1 mA·cm⁻² [35]. The simple estimation of chlorine concentration in the anolyte was conducted using test strips. The intensive purple color of the upper pad (Figure S9) indicated the presence of at least 10 ppm of chlorine in the solution after the CA experiment. Moreover, no chlorine was detected in the catholyte, suggesting no crossover of Cl-containing ions between compartments. Further analysis of the chlorine formation was performed within case study 3 (see text below).

2.3. Case Study 3—PEC Seawater Splitting Coupled with MH-Based Hydrogen Storage

The last case study combines two former, i.e., photo-assisted seawater splitting in the anodic half-cell was coupled with in situ hydrogen storage in the cathodic compartment. The use of WO₃ photoanodes (in an acidic electrolyte) and the LaNi_4.3Co_0.4Al_0.3 cathode (in an alkaline electrolyte) implicated a significant overpotential (~0.7 V according to the Nernst equation) due to the pH gradient between both compartments. Although the Na-substituted Nafion membrane was again used as the separator of compartments, the proton transfer seemed to be the dominating conduction mechanism through the membrane, in agreement with a general order of cation permeability in cation-exchange membranes [40]. This process led to a decrease in the inter-compartment pH gradient. Consequently, to maintain pH < 4 in the anolyte solution (required for WO₃ photoanodes), the larger volume of electrolyte (2 L) was circulating between the anodic half-cell and the external reservoir.

The photo-assisted galvanostatic C/D sequence, analogous to the one from case study 1 (CS-1), is presented in Figure 8. The experiment was performed in a two-electrode mode with charging of the cathode vs. eight WO₃ photoanodes and discharging vs. the Ni-helix electrode. Moreover, cathode potential was independently measured vs. the Hg/HgO electrode placed in the cathodic half-cell. Due to a different setup (two electrolytes, membrane separation, different photoanodes) compared with CS-1, cell potentials during charging in the dark at current densities of 60 mA∙g⁻¹ and 200 mA∙g⁻¹ were much lower (i.e., higher absolute values), reaching −3.4 V and −3.8 V, respectively. Moreover, WE potentials also decreased to −0.16 V and −0.29 V vs. RHE. Upon irradiation (~50 mW·cm⁻²) the cell potential increased by 1.65–1.75 V at 60 mA∙g⁻¹ (Figure 8b), which was more than 20 times higher than in the case of Fe₂O₃ photoelectrodes (Figure 2). Such a voltage drop at the given current density (absolute current of 6.06 mA) decreased the power withdrawn from the galvanostat by 10–11 mW. Taking into account the total active area of the studied photoelectrodes (28.5 cm²), the light-to-electricity efficiency of the system (calculated as the ratio of electrical and light power) reached 0.7%. Moreover, the absolute cell potential under illumination was also lower than in the CS-1, even with a large pH gradient. The CE potential at 60 mA∙g⁻¹ under illumination was equal to 0.8 V vs. RHE, at which the maximum ABPE efficiency of the WO₃ electrodes was found (Figure 5d). The latter observation explains a high cell voltage reduction due to the photoresponse of the anodes. WE-RE potential remained unaffected during light modulation. After increasing the current density to 200 mA∙g⁻¹, the photoanodes operated at a potential >2.7 V vs. RHE, which was much higher than the onset potential of the dark current. Consequently, the dark current dominated the reaction at the anodes; thus, applied illumination had a minor impact on the cell potential (reduction of around 40 mV), similar to the observations in CS-1 (Figure 2). The changes in cell potential remained constant for over 17 h (for 200 mA∙g⁻¹), indicating operational stability of the system in dual electrolyte setup.

The influence of the specific current density on the magnitude of the cell voltage change was verified during short charging experiments (Figure 9), in which the light was switched ON/OFF every 60 s (15 min was used for 200 mA∙g⁻¹). The dark charging potential decreased (absolute value increased) along with the growth of the charging current, as expected from the cyclic voltammetry curves (Figure S3). On the other hand, cell voltage change was gradually suppressed for larger currents, which is coherent with the fact that photovoltage decreases for higher (photo)currents. The highest voltage gain was approx. 1.8 V at 20 mA∙g⁻¹ (2.02 mA), which is almost 57% of the dark charging voltage. This suggests that using a more intense illumination or multiplying photoelectrodes could bring the device closer to bias-free charging for the given current density. Nevertheless, to facilitate hydride formation at the cathode (i.e., to yield hydrogen storage), a minimum charging current of approx. 5 mA for LaNi₅-based cathodes (Figure S3) is required, which limits the maximum available voltage gain from photoelectrodes operating in the galvanostatic mode.

The specific discharge capacity was significantly lower (7–35 mAh∙g⁻¹, Figure S10) than values determined in CS-1 (140–240 mAh∙g⁻¹, Figure 3), which may be connected with the lower ionic strength of the catholyte (0.5M NaOH in CS-3 vs. 1M KOH in CS-1), differences in the experimental setup, and/or the possible degradation/oxidation of cathode active materials between CS-1 and CS-3 (a period of four months).

The chlorine evolution in the anodic compartment was again confirmed by test strips (Figure S11) and by the formation of the intense characteristic odor and a color change in the electrolyte. No chlorine was detected in the catholyte, as in experiments within CS-2. Two electrolytes were compared: (i) the first one (Figure S12a) after shorter experiments partly presented in Figure 9, and (ii) the second (Figure S12b) after the complete C/D experiment shown in Figure 8. Interestingly, the yellowish color was significantly more intense in the former case despite the fact that a much smaller electric charge transferred during the process (288 C vs. 1420 C). This suggests a higher concentration of the colorless ionic chlorine-based species in the second case. The analysis of a free chlorine (Cl₂)-based species with the spectrophotometric DPD (N,N-diethyl-p-phenylenediamine) method [41] revealed the concentration between 20 and 30 mg·L⁻¹ (approx. 19–29 ppm) for the second electrolyte. These values indicate that no more than 11% of the total charge transferred during the C/D experiment was involved in the ClER, which is in line with the preference towards the OER for currents in the mA-range [35]. The observed chlorine concentration was much lower than 5%, the concentration commonly found in commercial antifungal/antimold products. Thus, such solutions might be used as a disinfection agent or oxidating reagent with a lowered pH (final value of around 2).

2.4. Stability of the Reactor

As discussed in the introduction, the 3D-printed reactors reported in the literature are typically early-stage prototypes operating for up to 30 h in a specified electrolyte [21,22,23,25]. In contrast, the PEC cell manufactured in this study was used for several prolonged experiments, each lasting between 17 and 45 h (e.g., four experiments analogous to those in CS-1 were not presented), giving more than 300 h in total. The designed and tested setup revealed mechanical and chemical stability, as well as tightness in various chemical environments (with the pH ranging from 1 to 14) and under the flow of diverse electrolytes. After the C/D experiments in seawater, the initially transparent impregnation layer inside the anodic half-cell, the outer layer of electrolyte tubing (only inside the external reservoir, where tubes were immersed in the electrolyte), and the multi-electrode gasket changed color (Figures S13 and S14) due to the contact with chlorine species (in the form of Cl₂, HClO, ClO⁻, etc.), which are known to be strong oxidative agents. Nevertheless, these changes did not alter the durability of these components. The tailor-made gaskets provided sufficient water tightness; however, they were single-use because of the substantial deformation, which remained after the disassembly of the reactor (Figure S14). Moreover, the noisy signal observed for the top-row electrodes suggests the need for improving the design of the current collectors and the uniformity of pressure applied to the set of electrodes (the type of gasket also affects this factor).

3. Materials and Methods

3.1. Manufacturing of the Reactor

The reactor and additional components, including compartment bodies, top lids, the cathode holder, the diaphragm, closing plates, stoppers, adapters, caps for electrolyte bottles, and additional boxes were designed in the Freecad v. 0.19 parametric-modeler [42]. The slicing of the 3D models, setting the printing conditions, and the generation of .gcode files were performed in Ultimaker Cura v. 4.9.1 software [43]. Fused deposition modeling was conducted with the Ender 5 Plus printer (Creality, Shenzhen, China) equipped with an enclosure, which minimizes temperature gradients between the printer’s build plate and the environment and reduces the cooling rate of the printout. Several printing parameters were optimized to obtain objects with the best properties, e.g., wall thickness, single-layer height, infill type and density (the fraction of the object’s internal volume filled with filament), nozzle and build-plate temperatures, printing speed, and print orientation. They affect the quality of the printouts, the water tightness of certain faces, the printing time, and the amount of building material. The exemplary parameters for different objects were compared in Table 1. The ASA filament (acrylonitrile-styrene-acrylate filament, Polar White, Spectrum Filaments, Pecice, Poland) was used to print all objects. This polymer was chosen due to its chemical stability in acidic and alkaline aqueous solutions, and its resistance to UV radiation. To increase the adhesion of the printout to the build plate, the latter was manually coated with a thin ASA layer. After printing, all supporting elements were removed, and rough faces were smoothed with acetone (if applicable and needed). The interior of cell compartments was additionally impregnated with transparent epoxy resin 049 (AG TermoPasty, Sokoly, Poland) to ensure water tightness. Gaskets for electrode/lid ports (methyl vinyl silicone, RubberPro, Liszki, Poland) were glued to the cell body/lid using a cyanoacrylate adhesive (Technicqll, Trzebinia, Poland). Softer gaskets in contact with photoanodes and a membrane/diaphragm were cast in-house from mold-forming silicone (SiliForm 25, AgBet, Dobrzelin, Poland) using 3D-printed tailor-made molds.

3.2. Design of the Reactor and Peripheral Modules

The manufactured reactor was based on a modular design, which allows for various configurations of electrodes and processes. The multi-electrode half-cell (no. 1 in Figure 10) accommodated up to eight (photo)electrodes with dimensions of 50 mm × 22mm (an active area up to 5.8 cm² and 46 cm² for a single electrode and the whole set, respectively). On the contrary, the single-electrode compartment (no. 2 in Figure 10) was suitable for a larger electrode of size 100 mm × 100 mm (an active area up to 71 cm², not presented in this report). In both variants, electrode(s) were pressed against the half-cell equipped with an intermediate gasket and Cu foil acting as an electrical connection (no. 6 in Figure 10). Depending on the type of experiment/process studied, the electrode window in each half-cell could be alternatively closed by the PMMA plate (no. 11 in Figure 10). Compartments were separated by an IEM (e.g., Nafion 117, Sigma Aldrich, Burlington, MA, USA) or a porous diaphragm, which provided an ionic connection between both parts. Top lids enabled the introduction of additional electrodes or tubing to the cell. The removal of evolved gases out of the compartments was assumed to take place along with a separated circulation of anolyte and catholyte between the cell and external reservoirs. For this reason, each compartment contained at least ten standard 1/4”- hydraulic ports for the connection of the electrolyte flow system. Figure 10 illustrates the PEC-MH configuration of the reactor, with eight photoanodes and the cathode introduced through the lid, whereas the disassembled reactor is presented in Figure S15. The multi- and single-electrode half-cells are characterized by internal volumes of approx. 170 and 240 mL, respectively. The electrolyte circulation system was based on a set of membrane pumps with a maximum flow of up to 900 mL∙min⁻¹, inert tubing (Versilon SE-200, Saint-Gobain, Courbevoie, France), and plastic reservoirs. The individual switching of each electrode was realized by a computer-controlled relay module.

3.3. Preparation of Cathode Assembly

Hydride-forming intermetallic compounds (LaNi_4.7Al_0.3 and LaNi_4.3Co_0.4Al_0.3) were synthesized by arc melting. Stoichiometric amounts of La (99.9%, Alfa Aesar, Ward Hill, MA, USA), Ni (99.98%, Alfa Aesar), Co (99.95%, Alfa Aesar), and Al (99.99%, Alfa Aesar) were melted together in a pure argon (oxygen-free) atmosphere, in an arc furnace (MAM-1, Edmund Bühler GmbH, Bodelshausen, Germany) utilizing an electric arc formed between tungsten and copper electrodes. The arc temperature was adjusted for each composition (up to max. 3500 °C) by varying the applied electric current. To reach a high chemical and structural homogeneity, the obtained ingots were repeatedly melted and flipped upside down multiple times. In the final stage of the synthesis process, the as-cast button-shaped samples were once again melted and cast into rod-shaped samples. To further increase the homogeneity of the synthesized materials, the rod-shaped ingots were subsequently enclosed in stainless steel tubes under the Ar atmosphere and annealed at 1200 K for 96 h. In the next step, the annealed samples were pulverized and then mixed with binding and conducting materials to form a cathode. The studied cathodes comprised a powder mixture of an annealed hydrogen-storing intermetallic compound (the active material, 90 wt.%), the binder (PTFE, 5 wt.%), and carbon black (5 wt.%, to increase electrical conductivity) of approx. 120 mg total mass, which was cold-pressed between two sheets of Ni-foam. To minimize powder losses and to increase the mechanical stability of such a sandwich-type electrode, the cathode was enclosed in a 3D-printed cathode holder supplemented with two silicone gaskets and a Ni-frame acting as an electrical connection (Figure S16).

3.4. Preparation of Photoanodes

Thin films of Ti-doped Fe₂O₃ were deposited on FTO-glass substrates using iron(III) acetylacetonate (FeAcac, 99.9 %, Sigma Aldrich) as the precursor following the procedure reported before [26]. Shortly, an appropriate amount of TiCl₄ (99.9 %, Sigma Aldrich) and FeAcac were sequentially dissolved in ethanol and homogenized using sonication. FTO substrates were cleaned (detergent solution, de-ionized water, and isopropyl alcohol), masked at one edge (for electrical contact), and placed onto a hot plate (450 °C). The precursor solution was then sprayed onto FTO glass with the Ar carrier gas pressure of 100 psi. After 10 min, the samples were calcinated at 650 °C for 1 h.

Approximately 2.5 µm thick WO₃ films were obtained by using a sol–gel method based on a tungstic acid/polyethylene glycol (PEG) 300 precursor following a previously described procedure [37]. FTO glass (resistance 8–10 ohm·sq⁻¹, Sigma Aldrich) was used as substrate. The WO₃ films were formed during six consecutive applications of the colloidal precursor onto the substrates with a surface area of 5.8 cm² using the doctor-blade technique. Each application was followed by heating in flowing oxygen at 500 °C for 30 min. The annealing provides highly crystalline WO₃ films, which crystallize in a monoclinic structure.

3.5. PEC Characterization

Linear sweep voltammetry, chronoamperometry, and chronopotentiometric charge/discharge measurements were controlled by potentiostat/galvanostat OrigaStat 100 (OrigaLys, Lyon, France), whereas additional voltage measurements were performed with the potentiostat 2.0 (Instytut Fotonowy, Krakow, Poland). The source of solar radiation was the solar simulator consisting of the xenon arc lamp (XBO, 150 W, Instytut Fotonowy) and the AM1.5G filter, coupled with a water filter. The light spectrum (Figure S17) was measured using a Black-Comet-SR spectrometer (StellarNet Inc., Tampa, FL, USA) equipped with a cosine receptor and fiber optic cable. The used solar simulator was characterized by a small light power variation over time (less than 5 %). However, it produced a highly focused beam (spot smaller than 1 cm²) because of the installed elliptic reflector. Consequently, light power over 100 cm² area (size of the photoanode port) was highly inhomogeneous, and each photoanode was illuminated differently. To account for this effect, the spatial distribution of light power was measured before each experiment at three points per electrode using a light calibrator (Instytut Fotonowy), fitted with a reflective neutral density filter (required to avoid saturation of sensor, optical density 2, Edmund Optics, Barrington, NJ, USA), and a tailor-made holder resembling the geometry of the multi-electrode side (three points per electrode, Figure S18). The average power for each electrode was used to calculate ABPE efficiency. The photoanode active areas in multi-electrode compartments were determined graphically (Figure S19) using ImageJ v. 2 software [44]. Electrolytes were freshly prepared before experiments. The pH values of electrolytes were measured with the CX-461 multifunction meter (Elmetron, Zabrze, Poland) coupled with EPS-1/EPX-4 electrodes and a temperature probe.

3.6. Case Study 1 (CS-1)

The photo-assisted charging of LaNi_4.7Al_0.3- and LaNi_4.3Co_0.4Al_0.3-based cathodes (working electrode WE) was performed under simulated solar radiation with eight Fe₂O₃ photoanodes connected electrically in parallel (CE). The discharging step was performed in dark vs. the Ni-helix electrode (CE) with a cut-off potential of 0.7 and 0.79 V vs. CE (0.52 V and 0.58 V vs. RHE) for LaNi_4.7Al_0.3 and LaNi_4.3Co_0.4Al_0.3, respectively. Although C/D experiments were performed in two-electrode mode (to simulate real operation), the potential of WE vs. Hg/HgO was additionally measured by an independent potentiostat (Figure S20a). The automatic switching of two types of counter electrodes (photoanodes or Ni-helix electrodes) between charging and discharging was realized by a feedback-loop module based on a microcontroller (Seeeduino Nano, Seeed Studio, Shenzhen, China) coupled with a relay. Both compartments of the reactor were separated by a porous membrane made of an ASA polymer, allowing the mixing of electrolytes (1M KOH, pH = 13.5–13.7, prepared from KOH pellets, 88%, VWR, Radnor, PA, USA) but limiting the crossover of evolved gases. The scheme with distances between electrodes is given in Figure S21a, and pictures of anodic and cathodic compartments are presented in Figure S1. The total volume of the electrolyte in the system amounted to 1.2 L. Powder X- ray diffraction experiments were performed with the Rigaku Miniflex diffractometer (Cu K_α1 radiation, 2θ-range = 5–120°, Δ2θ = 0.02°).

3.7. Case Study 2 (CS-2)

Three-electrode LSV and two-electrode CA measurements were performed with eight WO₃ photoanodes connected electrically in parallel (WE) and Pt-coil or stainless steel foil (CE), respectively. Ag/AgCl (sat. KCl) was used as the RE. The experiments were performed in synthetic seawater (AF Perfect Water, Aquaforest, Brzesko, Poland) and 0.5M Na₂SO₄ (from powder, 99%, Alfa Aesar), acidified with HCl (35–38%, Chempur, Piekary Slaskie, Poland) and H₂SO₄ (95%, POCH, Gliwice, Poland) to pH = 1.55 and 1.7, respectively. The measurements were performed in a two-compartment configuration with a Na-substituted Nafion 117 membrane (ion exchange in 0.5–1M NaCl lasted at least 5 days). Both compartments were initially filled with the same starting electrolyte. The set of membrane pumps maintained the electrolyte flow. Total volumes of circulating anolyte and catholyte were approx. 0.4 L each. The presence of chlorine in each compartment was verified qualitatively using test strips (Aqua Star).

3.8. Case Study 3 (CS-3)

The photo-assisted C/D experiment for LaNi_4.3Co_0.4Al_0.3-based cathode coupled with simultaneous seawater oxidation was performed in a two-electrode configuration. The setup was analogous to CS-1, except for the Na-substituted Nafion 117 membrane separating both compartments (Figure S20b and Figure S21b). For the anolyte and catholyte, acidified synthetic seawater (pH = 1.7, 2 L) and 0.5M NaOH (pH = 13, 0.8 L, from NaOH pellets, 98 % Honeywell, Charlotte, NC, USA) were used, respectively.

4. Conclusions

The objective of this work was the design and production of a photoelectrochemical reactor for coupling PEC (sea)water splitting with in situ hydrogen storage in hydride-forming intermetallic compounds. The large-scale modular PEC-MH cell was manufactured using the 3D-printing method supplemented by appropriate impregnation. The reactor size was larger than a standard laboratory cell and comparable with the upscaled devices manufactured by traditional subtractive methods. It can operate with up to eight photoactive electrodes with a total active area of up to 46 cm². The device was equipped with an electrolyte flow system, a relay module for the photoelectrode connection, and a feedback-loop module for switching between counter electrodes.

The fabricated system was tested in three case-study experiments. Photo-assisted galvanostatic charging/discharging of LaNi₅-based cathodes with the Fe₂O₃ photoanodes resulted in the formation of the metal hydride phases (confirmed by powder X-ray diffraction) and proved a successful operation of the upscaled PEC-MH device. Owing to the low efficiency of the tested photoanodes (the total photocurrent density for eight electrodes of 0.03 mA·cm⁻²) and a potential–current mismatch between anodes and the cathode, the reduction in the cell charging voltage due to the photoresponse from the photoanodes was only ~4% (charging at 6 mA). Significantly higher photoactivity was observed for WO₃ photoelectrodes (the total photocurrent density of 0.34 mA·cm⁻²) towards oxygen evolution reaction (performed in Na₂SO₄ solution) and simultaneous chlorine and oxygen evolution reactions (in synthetic seawater). The higher PEC efficiency of WO₃ electrodes and the optimal potential–current conditions for both the anodes and the cathode were substantiated by the significant reduction in the cell charging voltage by 50% (charging at 6 mA). The experiment was performed in a dual electrolyte system (the catholyte: 0.5M NaOH, the anolyte: acidified seawater), with the intrinsic overpotential of approx. 0.7 V caused by the pH gradient. It was proposed that increasing the power density of light would bring the system closer to bias-free operation. The experiments involving seawater as an electrolyte were associated with the simultaneous presence of the ClER and the OER. The estimated concentrations of dissolved Cl₂ were in the range of 20–30 ppm. These values indicate that the OER is the dominating process when operating the device at the mA-current range.

The performed experiments confirmed the versatility of the manufactured device, which can be operated in a broad range of pH and the presence of aggressive chemical species. The reactor revealed stability over several measurements, lasting between 17 and 43 h each. It was recognized that a single set of tailor-made gaskets is suitable only for experiment(s) that do not require cell disassembly. Moreover, the quality of electric contacts and the uniformity of the pressure applied on photoelectrodes by the fixing plate should be improved. The suitability of the constructed cell is not limited to the few examples of water splitting presented in this work as it could serve as a flow reactor for (photo)electrochemical synthesis or water purification.

5. Patents

The manufactured two-compartment photoelectrochemical flow cell and cathode holder were registered as industrial designs in the European EUIPO database with IPR numbers: 015055914-0001 and 015055830-0001, respectively.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/catal14120941/s1, Figure S1: The PEC-MH setup used in charge/discharge experiments. (a) Multi-electrode anodic compartment, (b) cathodic compartment with transparent PMMA window. 1—Fe₂O₃ photoanode, 2—Cathode holder with the cathode, 3—Ni-helix electrode, and 4—ASA porous diaphragm; Figure S2: Photocurrent density derived from linear sweep voltammetry (LSV) measurements (presented in Figure 2 in the main text); Figure S3: Ten chronovoltammetry cycles for LaNi_4.3Co_0.4Al_0.3 cathode measured in ~1M of KOH solution (pH = 13.5). RE: Hg/HgO, CE: Ni-helix, scan rate: 10 mV·s⁻¹. Peaks corresponding to H₂-absorption/desorption are highlighted; Figure S4: Dark LSV (scan rate: 10 mV·s⁻¹) experiment with Fe₂O₃ photoanodes connected in parallel, conducted in 1M of NaOH; Table S1: Unit-cell parameters for the pristine and hydrogenated LaNi₅-based phases derived from the Rietveld refinements; Figure S5: Comparison of the PXRD patterns for active material and powder mixture embedded in LaNi_4.7Al_0.3-based cathode. Bragg peak indicated with * could not be indexed with any well-known secondary phase; Figure S6: LSV (scan rate: 20 mV·s⁻¹) experiments with WO₃ photoanodes, conducted in acidified 0.5M of Na₂SO₄ (pH = 1.55) under the illumination of the solar simulator. (a) Light power density per electrode (average of three points). (b) The photocurrent for each electrode and all-together connected in parallel (bold line). (c) Photocurrent density. The noisy signal for W1 and W5 samples (top row of electrodes) is probably connected to poor contact with the current collector. The color code is the same in subfigures (b,c). RE: Ag/AgCl (sat. KCl), CE: Pt-coil; Figure S7: Changes in light power density during experiments of case study 2 for individual electrodes. The change in average total power was smaller than 1.5%. The first light measurement was performed 3.5 h before the start, while the second was 2.5 h after the end of CA experiments (period of approx. 24 h); Figure S8: LSV curves in the form of current density normalized by light power density measured before and after the CA experiment (case study 2). Data were collected for eight WO₃ electrodes connected in parallel. The largest deviation at 1500 mV is lower than 5.7 %. RE: Ag/AgCl (sat. KCl); Figure S9: Semi-qualitative chlorine detection using test strips. Although the Cl₂-pad turns yellow in acidic electrolytes, the intensive purple color clearly indicates the presence of free chlorine in the solution after the CA experiment; Figure S10: Discharge specific capacity of the LaNi_4.3Co_0.4Al_0.3-based cathode derived from charging/discharging measurements in CS-3; Figure S11: Semi-qualitative chlorine detection using test strips. Although the Cl₂-pad turns yellow in acidic electrolytes, the intensive purple color clearly indicates the presence of free chlorine in the anolyte after the C/D experiment in CS-3. No trace of free chlorine was detected in the catholyte; Figure S12: (a) Anolyte after short charging experiments with different C-rates (Figure 9 in the main text) and two charging steps at 20 mA·g⁻¹ (total charge transferred through photoanodes ~288 C). (b) Anolyte after a prolonged C/D experiment (Figure 8 in the main text, total charge transferred ~1420 C); Figure S13: Elements of PEC-MH device after experiments with seawater oxidation. (a) Anodic half-cell. The internal impregnation layer was colored yellow. (b) Catholyte tubing. (c) Anolyte tubing—only the external layer of tubing was oxidized; Figure S14: Multi-electrode gasket after experiment with seawater in CS-2. Deformations and discoloration are clearly visible; Figure S15: Components of the PEC-MH device. 1—Multi-electrode half-cell accommodating eight transparent photoanodes with electric contacts to photoanodes (Cu-foil) and hydraulic ports for electrolyte circulation. 2—Cathodic half-cell. 3—Backplate for fixing PMMA window. 4—Front plate for fixing photoelectrodes. 5, 6—Back and front thermal shields. 7—PMMA back window. 8—Fe₂O₃ photoanodes. 9—Ni-helix electrode. 10—Top lids with electrode ports. 11, 12—Photoelectrode gaskets. 13—Fixing bolts. The compartment separator (membrane or diaphragm) and its gaskets are not shown; Figure S16: The subsequent steps (a–d) of assembly of the cathode holder. The holder was 3D-printed with the ASA polymer, with screws and nuts made of polypropylene. The Ni-contact is combined with an adapter fitting the electrode port in the lid of the reactor; Figure S17: Spectral irradiance of the solar simulator (xenon arc lamp + AM1.5G filter) coupled with water filter. The spectrum was measured at an arbitrary distance from the lamp; Figure S18: A holder with a photodiode for calibration of light power over 100 cm². The holder resembles the geometry of a multi-electrode compartment (three points per electrode); Figure S19: Active areas of electrodes in cm²: (a) case study 1, (b) experiments in seawater in case study 2, (c) experiments in 0.5M of Na₂SO₄ in case study 2 and case study 3; Figure S20: Schemes of electrical circuits during (a) charging and (b) discharging experiments. The potential of WE vs. RE was measured with an auxiliary potentiostat (labeled as a voltmeter in the schemes); Figure S21: Schemes of experimental setups used in case study 1 (a) and 3 (b). Horizontal distances are given with an accuracy of ±3 mm due to the irreproducible shrinkage of gaskets.

Author Contributions

Conceptualization, P.W., J.M., Y.-C.L., M.N.G., K.B. and W.M.; methodology, P.W., Y.-C.L., E.H.J. and K.B.; validation, P.W., J.M., Y.-C.L. and E.H.J.; investigation, P.W., E.H.J. and K.B.; formal analysis, P.W. and E.H.J.; data curation, P.W. and E.H.J.; writing—original draft preparation, P.W.; writing—review and editing, P.W., J.M, Y.-C.L., E.H.J., M.N.G., R.S. and W.M.; visualization, P.W.; supervision, M.N.G., R.S. and W.M.; project administration, J.M. and M.N.G.; funding acquisition, M.N.G., R.S. and W.M. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Center for Research and Development (NCBR) within the HERA (Hydrogen Energy Rechargeable Architectures) project (NOR\POLNOR\HERA\0043\2019).

Data Availability Statement

The original contributions presented in this study are included in this article/Supplementary Materials, and further inquiries can be directed to the corresponding authors.

Acknowledgments

The authors gratefully acknowledge Alicja Macyk for chlorine analysis and Marcin Kobielusz for the fruitful discussions.

Conflicts of Interest

Authors Paweł Wyżga, Joanna Macyk and Wojciech Macyk are employed by the InPhoCat – Innovative Photocatalytic Solutions Sp. z o.o. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

References

Chen, Z.; Dinh, H.N.; Miller, E. Photoelectrochemical Water Splitting. In Standards, Experimental Methods, and Protocols; Springer: Berlin/Heidelberg, Germany, 2013; ISBN 9781461482987. [Google Scholar]
Kim, J.H.; Hansora, D.; Sharma, P.; Jang, J.W.; Lee, J.S. Toward Practical Solar Hydrogen Production-an Artificial Photosynthetic Leaf-to-Farm Challenge. Chem. Soc. Rev. 2019, 48, 1908–1971. [Google Scholar] [CrossRef]
Hamdani, I.R.; Bhaskarwar, A.N. Recent Progress in Material Selection and Device Designs for Photoelectrochemical Water-Splitting. Renew. Sustain. Energy Rev. 2021, 138, 110503. [Google Scholar] [CrossRef]
Lin, Y.-C.; Wyżga, P.; Macyk, J.; Macyk, W.; Guzik, M.N. Solar-Driven (Photo)Electrochemical Devices for Green Hydrogen Production and Storage: Working Principles and Design. J. Energy Storage 2024, 82, 110484. [Google Scholar] [CrossRef]
Jadwiszczak, M.; Jakubow-Piotrowska, K.; Kedzierzawski, P.; Bienkowski, K.; Augustynski, J. Highly Efficient Sunlight-Driven Seawater Splitting in a Photoelectrochemical Cell with Chlorine Evolved at Nanostructured WO3 Photoanode and Hydrogen Stored as Hydride within Metallic Cathode. Adv. Energy Mater. 2020, 10, 1903213. [Google Scholar] [CrossRef]
Schoonman, J.; Kiliaan, H.S. In-Situ Storage of Solar Hydrogen in a PEC Cell. Solid. State Ion. 1983, 9–10, 1087–1092. [Google Scholar] [CrossRef]
Schoonman, J. Electrochemical Conversion and Storage of Solar Energy. Berichte Bunsenges. Phys. Chem. Chem. Phys. 1982, 86, 660–664. [Google Scholar] [CrossRef]
Danko, D.B.; Sylenko, P.M.; Shlapak, A.M.; Khyzhun, O.Y.; Shcherbakova, L.G.; Ershova, O.G.; Solonin, Y.M. Photoelectrochemical Cell for Water Decomposition with a Hybrid Photoanode and a Metal-Hydride Cathode. Sol. Energy Mater. Sol. Cells 2013, 114, 172–178. [Google Scholar] [CrossRef]
Kongsawatvoragul, K.; Chomkhuntod, P.; Sawangphruk, M. Solar-Driven Energy Storage Enhancement of Nickel Hydroxide Nanomaterials. Electrochim. Acta 2021, 388, 138654. [Google Scholar] [CrossRef]
Akuto, K.; Sakurai, Y. A Photorechargeable Metal Hydride/Air Battery. J. Electrochem. Soc. 2001, 148, A121. [Google Scholar] [CrossRef]
Lei, B.; Li, G.R.; Chen, P.; Gao, X.P. A Solar Rechargeable Battery Based on Hydrogen Storage Mechanism in Dual-Phase Electrolyte. Nano Energy 2017, 38, 257–262. [Google Scholar] [CrossRef]
Gholami, T.; Pirsaheb, M. Review on Effective Parameters in Electrochemical Hydrogen Storage. Int. J. Hydrogen Energy 2021, 46, 783–795. [Google Scholar] [CrossRef]
Lopes, T.; Dias, P.; Andrade, L.; Mendes, A. An Innovative Photoelectrochemical Lab Device for Solar Water Splitting. Sol. Energy Mater. Sol. Cells 2014, 128, 399–410. [Google Scholar] [CrossRef]
Vilanova, A.; Lopes, T.; Spenke, C.; Wullenkord, M.; Mendes, A. Optimized Photoelectrochemical Tandem Cell for Solar Water Splitting. Energy Storage Mater. 2018, 13, 175–188. [Google Scholar] [CrossRef]
Vilanova, A.; Lopes, T.; Mendes, A. Large-Area Photoelectrochemical Water Splitting Using a Multi-Photoelectrode Approach. J. Power Sources 2018, 398, 224–232. [Google Scholar] [CrossRef]
Tolod, K.R.; Hernández, S.; Russo, N. Recent Advances in the BiVO4 Photocatalyst for Sun-Driven Water Oxidation: Top-Performing Photoanodes and Scale-up Challenges. Catalysts 2017, 7, 13. [Google Scholar] [CrossRef]
ArtipHyction Project: “Fully Artificial Photo-Electrochemical Device for Low Temperature Hydrogen Production”. Available online: https://www.artiphyction.org/ (accessed on 22 November 2023).
Torras, C.; Lorente, E.; Hernández, S.; Russo, N.; Salvadó, J. Hydrodynamics and Oxygen Bubble Characterization of Catalytic Cells Used in Artificial Photosynthesis by Means of CFD. Fluids 2017, 2, 25. [Google Scholar] [CrossRef]
Becker, J.P.; Turan, B.; Smirnov, V.; Welter, K.; Urbain, F.; Wolff, J.; Haas, S.; Finger, F. A Modular Device for Large Area Integrated Photoelectrochemical Water-Splitting as a Versatile Tool to Evaluate Photoabsorbers and Catalysts. J. Mater. Chem. A Mater. 2017, 5, 4818–4826. [Google Scholar] [CrossRef]
Ambrosi, A.; Pumera, M. 3D-Printing Technologies for Electrochemical Applications. Chem. Soc. Rev. 2016, 45, 2740–2755. [Google Scholar] [CrossRef]
Andrei, V.; Hoye, R.L.Z.; Crespo-Quesada, M.; Bajada, M.; Ahmad, S.; De Volder, M.; Friend, R.; Reisner, E. Scalable Triple Cation Mixed Halide Perovskite–BiVO₄ Tandems for Bias-Free Water Splitting. Adv. Energy Mater. 2018, 8, 1801403. [Google Scholar] [CrossRef]
Taştan, Ü.; Ziegenbalg, D. Getting the Most out of Solar Irradiation: Efficient Use of Polychromatic Light for Water Splitting. Chem. A Eur. J. 2016, 22, 18824–18832. [Google Scholar] [CrossRef]
Mousset, E.; Huang Weiqi, V.; Foong Yang Kai, B.; Koh, J.S.; Tng, J.W.; Wang, Z.; Lefebvre, O. A New 3D-Printed Photoelectrocatalytic Reactor Combining the Benefits of a Transparent Electrode and the Fenton Reaction for Advanced Wastewater Treatment. J. Mater. Chem. A Mater. 2017, 5, 24951–24964. [Google Scholar] [CrossRef]
Pornrungroj, C.; Andrei, V.; Reisner, E. Thermoelectric–Photoelectrochemical Water Splitting under Concentrated Solar Irradiation. J. Am. Chem. Soc. 2023, 145, 13709–13714. [Google Scholar] [CrossRef]
Dinesh, G.K.; Dessì, P.; Tong, W.; González-Gómez, R.; Farràs, P. 3D-Printed Photoelectrochemical Cell and Its Application in Evaluation of Bismuth Vanadate Photoanodes: Synthesis, Cell Design and Testing. Johns. Matthey Technol. Rev. 2022, 66, 386–392. [Google Scholar] [CrossRef]
Ramalingam, K.; Bieńkowski, K.; Nawaz, N.; Solarska, R. Enhanced PEC Generation of Hydrogen from Seawater Driven by Efficient and Stable Ti-Fe2O3 Photoanode. Appl. Surf. Sci. 2025, 679, 161246. [Google Scholar] [CrossRef]
Tobías, I.; del Cañizo, C.; Alonso, J. Crystalline Silicon Solar Cells and Modules. In Handbook of Photovoltaic Science and Engineering; Wiley: Hoboken, NJ, USA, 2003; pp. 255–306. [Google Scholar]
Todorova, S.; Abrashev, B.; Rangelova, V.; Mihaylov, L.; Vassileva, E.; Petrov, K.; Spassov, T. Hydrogen Gas Phase and Electrochemical Hydriding of LaNi5−xMx (M = Sn, Co, Al) Alloys. Materials 2020, 14, 14. [Google Scholar] [CrossRef] [PubMed]
Giza, K.; Bala, H.; Drulis, H.; Hackemer, A.; Folcik, L. Gaseous and Electrochemical Hydrogenation Properties of LaNi4.3(Co, Al)0.7 -XInx Alloys. Int. J. Electrochem. Sci. 2012, 7, 9881–9891. [Google Scholar] [CrossRef]
Dymek, M.; Bala, H.; Hackemer, A.; Drulis, H. Hydrogenation and Electrochemical Corrosion Properties of LaNi4.5Co0.5 Alloy Doped with Aluminum. Solid. State Ion. 2015, 271, 116–120. [Google Scholar] [CrossRef]
Feng, F. Electrochemical Behaviour of Intermetallic-Based Metal Hydrides Used in Ni/Metal Hydride (MH) Batteries: A Review. Int. J. Hydrogen Energy 2001, 26, 725–734. [Google Scholar] [CrossRef]
Fukai, Y. The Metal-Hydrogen System; Springer: Berlin/Heidelberg, Germany, 2005; Volume 21, ISBN 978-3-540-00494-3. [Google Scholar]
Dresp, S.; Dionigi, F.; Klingenhof, M.; Strasser, P. Direct Electrolytic Splitting of Seawater: Opportunities and Challenges. ACS Energy Lett. 2019, 4, 933–942. [Google Scholar] [CrossRef]
Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions; National Association of Corrosion Engineers: Houston, TX, USA, 1974. [Google Scholar]
Bennett, J.E. Electrodes for Generation of Hydrogen and Oxygen from Seawater. Int. J. Hydrogen Energy 1980, 5, 401–408. [Google Scholar] [CrossRef]
Dionigi, F.; Reier, T.; Pawolek, Z.; Gliech, M.; Strasser, P. Design Criteria, Operating Conditions, and Nickel-Iron Hydroxide Catalyst Materials for Selective Seawater Electrolysis. ChemSusChem 2016, 9, 962–972. [Google Scholar] [CrossRef] [PubMed]
Santato, C.; Ulmann, M.; Augustynski, J. Photoelectrochemical Properties of Nanostructured Tungsten Trioxide Films. J. Phys. Chem. B 2001, 105, 936–940. [Google Scholar] [CrossRef]
Kafizas, A.; Godin, R.; Durrant, J.R. Charge Carrier Dynamics in Metal Oxide Photoelectrodes for Water Oxidation. In Semiconductors and Semimetals; Elsevier Inc.: Amsterdam, The Netherlands, 2017; Volume 97, pp. 3–46. [Google Scholar]
Wyżga, P.; Tabari, T.; Trochowski, M.; Macyk, W. Optimizing the Morphology of Titania Nanorods for Enhanced Solar Seawater Splitting. Results Eng. 2023, 17, 100921. [Google Scholar] [CrossRef]
Luo, T. Porous Ion Exchange Membranes for Vanadium Redox Flow Battery; Rheinisch-Westfälischen Technischen Hochschule Aachen: Aachen, Germany, 2017. [Google Scholar]
Harp, D.L. Current Technology of Chlorine Analysis for Water and Wastewater; HACH Company: Loveland, CO, USA, 2002. [Google Scholar]
Riegel, J.; Mayer, W.; van Havre, Y. FreeCad 0.19 [Software]. Available online: https://www.freecad.org (accessed on 1 June 2021).
Ultimaker B.V. Ultimaker Cura 4.9.1 [Software]; Ultimaker B.V.: Utrecht, The Netherlands, 2021. [Google Scholar]
Schneider, C.A.; Rasband, W.S.; Eliceiri, K.W. NIH Image to ImageJ: 25 Years of Image Analysis. Nat. Methods 2012, 9, 671–675. [Google Scholar] [CrossRef] [PubMed]

Figure 1. Linear sweep voltammetry (LSV, scan rate: 10 mV·s⁻¹) experiments with Fe₂O₃ photoanodes, conducted in 1M NaOH under the illumination of a solar simulator: (a) Power density of simulated solar light for each photoelectrode. Electrodes are highlighted with bold lines and labels. The power density was measured at three points per electrode, and, then, the average value was taken for further analysis. (b) The photocurrent for individual Fe2O3 electrodes and all together connected in parallel (bold line). Measurements were collected under constant and modulated illumination. Reference electrode (RE): Hg/HgO; Counter electrode (CE): Ni-helix. (c) Photocurrent density normalized by light power density. (d) Applied bias photon-to-current efficiency (ABPE) derived from data in (b).

Figure 2. Photo-assisted charging/discharging (C/D) experiments of (a,c) LaNi_4.7Al_0.3 and (b,d) LaNi_4.3Co_0.4Al_0.3 cathodes. The absolute values of charging/discharging currents and potential changes upon illumination are given in red.

Figure 3. Specific discharge capacities for LaNi₅-based cathodes derived from C/D measurements. The discharge current densities are given in red.

Figure 4. Experimental (dot), calculated (red line), and differential (black line) profiles from Rietveld refinement of electrochemically charged LaNi_4.7Al_0.3 (a) and LaNi_4.3Co_0.4Al_0.3 (b) samples. Bragg peaks indicated with * appeared after C/D experiment (see Figure S5) in both samples but could not be indexed with any well-known secondary phase.

Figure 5. LSV (scan rate: 20 mV·s⁻¹) experiments with WO₃ photoanodes, conducted in acidified synthetic seawater (pH = 1.55) under the illumination of the solar simulator: (a) Light power density per electrode (average of three points). The average power density equals 52.4 mW∙cm⁻². (b) The photocurrent for individual electrodes and all-together connected electrically in parallel (bold line). (c) Photocurrent density normalized by light power density. (d) ABPE efficiency derived from (b). The noisy signal for W1 and W5 samples (top row of electrodes) is probably connected to poor contact with the current collector. The color code is the same in subfigures (b–d). RE: Ag/AgCl (sat. KCl), CE: Pt-coil.

Figure 6. LSV (scan rate: 20 mV·s⁻¹) experiments with WO₃ photoanodes, conducted in acidified 0.5M Na₂SO₄ (pH = 1.55) under the illumination of the solar simulator: (a) Photocurrent density normalized by light power density. (b) Applied bias photon-to-current efficiency derived from the photocurrent. The noisy signal for W1 and W5 samples (top row of electrodes) is probably connected with poor contact with the current collector. RE: Ag/AgCl (sat. KCl), CE: Pt-coil.

Figure 7. (a) The two-electrode chronoamperometric (CA) measurement in acidified synthetic seawater (pH = 1.55) under approx. 0.5 Sun illumination at the potential of 1.23 V vs. CE. WE: 8 WO₃ photoanodes connected electrically in parallel, CE: stainless steel foil. (b) The potential of WE vs. RE (Ag/AgCl electrode). It was measured independently from the CA setup using auxiliary potentiostat.

Figure 8. (a) Photo-assisted C/D experiments of the LaNi_4.3Co_0.4Al_0.3 cathode. Data of potential vs. RHE during 4–7. cycles are missing because the RE was temporarily above the electrolyte. (b) The first twelve hours of the experiment. The absolute values of charging/discharging currents and potential changes upon illumination are given in red.

Figure 9. Short photo-assisted C/D experiments with different charging rates. The specific current densities of 20, 50, 60, and 200 mA·g⁻¹ correspond to absolute current values of 2.02, 5.05, 6.06, and 20.18 mA, respectively. Modulated light was applied to photoanodes during experiments.

Figure 10. A schematic assembly of the PEC-MH reactor: (1) an anodic compartment, (2) a cathodic compartment, (3) lids with ports for electrodes, (4) fixing plates with windows, (5) silicone gaskets, (6) an electric contact (the other seven are not shown), (7) a photoanode, (8) a reference electrode (optional), (9) cathode, (10) an ion-exchange membrane/diaphragm, and (11) the PMMA/glass plate. For clarity, some elements of the assembly are not shown, e.g., the cathode holder, additional counter electrode, hydraulic fittings, electrolyte tubing, screws, and nuts.

Table 1. Examples of parameters used for printing different objects.

Parameter	Compartment Body	Cathode Holder	Box for Pumps	Gasket Mold
Filament material	ASA
Nozzle diameter/mm	0.4	0.1	0.4	0.4
Nozzle/bed temperature/°C	245/60
Layer height/mm	0.1	0.08	0.2	0.2
Infill/%	20			15
Wall thickness/mm	4	4	1.2	1.2
Top/bottom thickness/mm	4	4	0.6	0.6
Top/bottom printing speed/mm∙s⁻¹	20

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Wyżga, P.; Macyk, J.; Lin, Y.-C.; Jensen, E.H.; Guzik, M.N.; Bieńkowski, K.; Solarska, R.; Macyk, W. 3D-Printed Reactor for Coupling Photoelectrochemical (Sea)Water Splitting with Solid-State H₂ Storage. Catalysts 2024, 14, 941. https://doi.org/10.3390/catal14120941

AMA Style

Wyżga P, Macyk J, Lin Y-C, Jensen EH, Guzik MN, Bieńkowski K, Solarska R, Macyk W. 3D-Printed Reactor for Coupling Photoelectrochemical (Sea)Water Splitting with Solid-State H₂ Storage. Catalysts. 2024; 14(12):941. https://doi.org/10.3390/catal14120941

Chicago/Turabian Style

Wyżga, Paweł, Joanna Macyk, Yuan-Chih Lin, Emil Høj Jensen, Matylda N. Guzik, Krzysztof Bieńkowski, Renata Solarska, and Wojciech Macyk. 2024. "3D-Printed Reactor for Coupling Photoelectrochemical (Sea)Water Splitting with Solid-State H₂ Storage" Catalysts 14, no. 12: 941. https://doi.org/10.3390/catal14120941

APA Style

Wyżga, P., Macyk, J., Lin, Y. -C., Jensen, E. H., Guzik, M. N., Bieńkowski, K., Solarska, R., & Macyk, W. (2024). 3D-Printed Reactor for Coupling Photoelectrochemical (Sea)Water Splitting with Solid-State H₂ Storage. Catalysts, 14(12), 941. https://doi.org/10.3390/catal14120941

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

3D-Printed Reactor for Coupling Photoelectrochemical (Sea)Water Splitting with Solid-State H₂ Storage

Abstract

1. Introduction