Mechanism of the two-step alkaline water electrolysis
As shown in Fig. 1a, the H2 production (Step 1) involves a cathodic reduction of H2O on the HER electrode (H2O→H2) and a simultaneous anodic oxidization of the Ni(OH)2 electrode (Ni(OH)2→ NiOOH). The subsequent O2 production (Step 2) occurs on the OER electrode by an anodic oxidization of OH− (OH−→ O2), whereas the NiOOH cathode is reduced to Ni(OH)2. This approach leads to a device architecture for the alkaline electrolytic cell with several important advantages. First, the separate generation of O2 and H2 prevents the product gases from mixing over a range of current densities and simplifies the gas handling, which greatly increase the operation flexibility of alkaline electrolytic cells and make them suitable to be driven by sustainable energy (such as solar energy). Second, this device architecture can produce highly pure H2 and O2 with no membrane, which further reduces the cost of the alkaline water electrolysis technology. Third, the separate H2 and O2 productions require different driving voltages (or power inputs), which implies that we can flexibly use sustainable energy (such as solar or wind power) for H2 production or O2 production based on the output variation in these unstable power sources. Finally, the NiOOH that forms during the H2 production (that is, Step 1) can be coupled with a zinc anode to form a NiOOH-Zn battery for energy storage, and its discharge depends on the cathodic reduction of the NiOOH electrode (NiOOH→ Ni(OH)2) and the anodic oxidization of the zinc electrode (Zn→ ZnO22−). Herein, it should be noted that the cathodic reduction potential of NiOOH (0.45 V versus Hg/HgO) is significantly higher than the anodic oxidization potential of zinc (−1.15 V versus Hg/HgO) (ref. 48). Therefore, the NiOOH cathode and Zn anode can be coupled to form the NiOOH-Zn battery system that has been commercialized49,50,51.) Its discharge product (Ni(OH)2) can be used to produce H2 again, which provides an interesting rechargeable cycle that produces H2 with the charge (that is, electrolysis in Step 1) and delivers energy with the discharge of the NiOOH-Zn battery.
In this work, nickel hydroxide, which is the conventional electrode material for commercial rechargeable Ni-MH or Ni-Cd batteries, was used as a redox mediator to split the conventional alkaline water electrolysis process into two steps. Before the fabrication of this alkaline water electrolytic cell, the electrochemical profile of Ni(OH)2 in an alkaline electrolyte (1 M KOH) was investigated using a cyclic voltammogram (CV) with a typical three-electrode system, which used a Pt plate and a Hg/HgO electrode as the counter and reference electrodes, respectively. Carbon-nanotube-supported Ni(OH)2 particles were used as the active material to prepare the Ni(OH)2-based film electrode (see the Methods and Supplementary Fig. 2 for details) for the CV measurement, where the carbon nanotube support with high electronic conductivity was only used to alleviate the polarization that arose from the electrode impedance. The CV curve of Ni(OH)2 at a scan rate of 5 mV s−1 is shown in Fig. 1b (black line). The OER and HER potentials of the commercial RuO2/IrO2-coated Ti-mesh electrode and Pt-coated Ti-mesh electrode were also investigated using the three-electrode method for comparison (see the red and blue lines in Fig. 1b). As shown in Fig. 1b, a couple of redox peaks are clearly observed at 0.43 and 0.49 V (versus Hg/HgO) in the CV curve of the Ni(OH)2 electrode because of the reversible cycling between Ni(OH)2 and NiOOH. Obviously, the special potential window for the Ni(OH)2/NiOOH redox couple is located between the onset potential for the OER and the onset potential for the HER. The result indicates that Ni(OH)2 can be used as a redox mediator to split the conventional alkaline water electrolysis process into two steps according to Fig. 1a. The galvanostatic charge-discharge curve of the Ni(OH)2 electrode at a current density of 0.2 A g−1 is shown in Supplementary Fig. 3 to clarify the specific capacity of Ni(OH)2 (see the corresponding discussion about Supplementary Fig. 3).
Performance of the two-step alkaline water electrolysis
To test the hypothesis in Fig. 1a, an alkaline water electrolytic cell was constructed with a commercial Pt-coated Ti-mesh electrode (Supplementary Fig. 4) for the HER, a commercial RuO2/IrO2-coated Ti-mesh electrode for the OER (Supplementary Fig. 5) and a commercial Ni(OH)2 electrode of conventional Ni-MH or Ni-Cd batteries (Supplementary Fig. 6). The photo profile of the cell is shown in Supplementary Fig. 7, which shows that the Ni(OH)2 electrode (2.5 × 4 cm2) is located between the HER electrode (2.5 × 4 cm2) and the OER electrode (2.5 × 4 cm2). The water electrolysis of the cell was investigated by chronopotentiometry measurements with different applied currents of 100–500 mA. The chronopotentiometry curve (cell voltage versus time) of the electrolytic cell at a constant applied current of 200 mA is shown in Fig. 2a. The chronopotentiometry data of the anode (anodic potential versus time) and cathode (cathodic potential versus time) were also investigated during the electrolysis process and are provided in Fig. 2a. The electrolysis process includes two steps (Steps 1 and 2) with different cell voltages. As shown in Fig. 2a, Step 1 (that is, the H2-production process) exhibits a cell voltage of ∼1.6 V, which arises from the difference between the anodic potential of 0.5 V (versus Hg/HgO) of the Ni(OH)2 oxidation (Ni(OH)2→NiOOH) and the cathodic potential of −1.1 V (versus Hg/HgO) of the H2O reduction (H2O→H2). In Step 2 (that is, the O2-production process), the cell voltage is 0.4 V, which is equal to the potential difference (0.7–0.3 V versus Hg/HgO) between the anodic oxidation of OH− (OH−→O2) and the cathodic reduction of NiOOH (NiOOH→ Ni(OH)2). In Step 2, the cell voltage sharply increases sharply at the end of electrolysis (Fig. 2a), which indicates that all of the NiOOH has been reduced to Ni(OH)2. In other words, the electrolysis in Step 2 automatically finished after 600 s (=1,200–600 s), which is equal to the electrolysis time in Step 1 at the identical current of 200 mA. The equal electrolysis time clearly indicates a Coulombic efficiency of ∼100%. Photo profiles of the H2 generation in Step 1 and O2 generation in Step 2 are shown in Fig. 2b,c to further characterize the separated steps. In addition, the video evidence also clearly demonstrates the separate H2/O2 generation directly (Supplementary Movies 1 and 2). To clarify the operation flexibility of this electrolyser, the water electrolysis was also investigated at a lower current of 100 mA and a higher current of 500 mA (Supplementary Fig. 8). It should be noted that as a notably mature electrode material, Ni(OH)2 exhibits high efficiency and a long cycle life. These characteristics are also notably important to facilitate the cycle of H2 generation (Step 1) and O2 generation (Step 2). To demonstrate this point, the H2/O2 generation cycle performance was investigated with an applied current of 200 mA. As shown in Fig. 2d, this alkaline electrolytic cell exhibits stable H2 and O2 generation over 20 consecutive cycles. Furthermore, 100 consecutive cycles of H2/O2 generation are shown in Supplementary Fig. 9 to further demonstrate the stability. As shown in Fig. 2d (or Fig. 2a), the separate H2 production (Step 1) and O2 production (Step 2) require different driving voltages (or power inputs), which implies that we can flexibly use renewable energy, such as solar or wind power, to produce H2 or O2 based on the output variation in these unstable power sources. For example, we can use solar energy at noon to drive the H2-production step, which requires a high driving voltage (or power input), and solar energy at dusk to power the O2-production step, which requires a low driving voltage. The flexibility can increase the use of sustainable energy.
In the above investigation (Fig. 2), a step time of only 10 min (600 s) was used to characterize the separate H2 and O2 production. Such a short time was used to emphasize that we can flexibly change the operation of our system even within a notably short time. In fact, the electrolysis time in each step can be easily controlled by the applied current. As shown in Supplementary Fig. 10, the electrolysis time in each step can be increased to 12 h with a low current of 20 mA. In addition, the Ni(OH)2 electrode can be cycled with different charge depths (Supplementary Fig. 11). Thus, we can also control the electrolysis time in each step by adjusting the charge depths of the Ni(OH)2 electrode (Supplementary Fig. 12). However, as mentioned in the introduction section, alkaline water electrolysis can use non-precious electrodes for the H2/O2 production36,37. Therefore, non-precious electrodes (a Co3O4-based OER electrode and a metal-Ni-foam-based HER electrode) were used to further demonstrate the separate H2 and O2 production (see Supplementary Fig. 13 and Supplementary Movies 3 and 4 for details). Furthermore, according to the previous report by Cronin et al.39, the efficiency of two-step water electrolysis can be evaluated by comparing its total driving voltage (Step 1+Step 2) to the driving voltage of the corresponding one-step water electrolysis. Therefore, the efficiency of the two-step alkaline water electrolysis using precious or non-precious HER/OER electrodes was calculated according to the method described by Cronin et al. (see Supplementary Fig. 14 for detail). As shown in Supplementary Fig. 14a,b, the efficiency of the two-step water electrolysis using precious electrodes (a RuO2/IrO2-coated Ti-mesh electrode for the OER and a Pt-coated Ti-mesh electrode for the HER) is 92% (=1.829/1.985) compared with its corresponding one-step water electrolysis. According to the data shown in Supplementary Fig. 14c,d, the efficiency of the two-step water electrolysis using non-precious electrodes (a Co3O4-based electrode for the OER and a metal Ni-foam electrode for the HER) is also ∼92% (=1.973/2.137) compared with its corresponding one-step water electrolysis. The achieved efficiency is slightly higher than that (79%) of the two-step PEM water electrolysis reported by Cronin’s group39.
Purity of the generated H2/O2
To confirm the purity of the H2/O2 in the separate steps, in situ differential electrochemical mass spectrometry was used to measure the gas evolution of the total water electrolysis process at a constant applied current of 200 mA. In this experiment, a quadrupole mass spectrometer with a leak inlet was connected to the alkaline water electrolytic cell with two tubes as the purge/carrier gas inlet and outlet (see the Methods and Supplementary Fig. 15 for details). A pure Ar gas stream was used as the purge gas before electrolysis and the carrier gas during the electrolysis process. Before the online gas analysis, the system was purged with a pure Ar stream for 1 h. The system was further purged with a pure Ar stream for another 1 h, with an online analysis record (Fig. 3a) showing that both O2 and H2 reached a stable background line. Then, the H2-production step (Step 1) was started, and the H2 evolution is clearly observed in the online analysis record. The ion current intensity of O2 obviously remained at the background level in Step 1, which indicates that no O2 was generated in the H2 production process of 30 min (Fig. 3a,b). After the H2 production (Step 1) finished, a rest step of 130 min was performed with a pure Ar stream to eliminate remnant H2 in the system, and a hysteresis of H2 could be observed in the online analysis record. Afterward, the O2-production step (Step 2) was started. As shown in Fig. 3b, the O2 production automatically finished with a total electrolysis time of ∼30 min (1,700 s), which is close to the H2 production time (1,800 s). The minor difference of 100 s may be because of the slight self-discharge of NiOOH in the rest step. However, this description does not indicate that the self-discharge of the NiOOH electrode will be significantly aggravated with a longer rest time (see Supplementary Fig. 16 for an extended discussion about the self-discharge of the nickel hydroxide electrode). As shown in Fig. 3a, there is no H2 evolution in the O2 production process. Therefore, the results in Fig. 3a,b well demonstrate the purity of the H2/O2 in the separate steps. Herein, it should be noted that the online gas analysis in our experiment was only used to characterize the purity of the H2 and O2 in separate steps. A typical drainage method (Supplementary Fig. 17) was used to quantify the H2 generation over a specific time length. In this experiment, the H2 production rate (ml s−1) was measured with an applied current of 1,000 mA for 100 s (Fig. 3c,d). Figure 3c,d shows that ∼12 ml H2 was generated in the 100 s electrolysis, which is close to the theoretical value (12.67 ml). Therefore, the Faradaic efficiency is 94.7% (12/12.67). In theory, the Faradaic efficiency should be 100%, but the impedance and dissolution of H2 in the aqueous solution may slightly reduce the efficiency. This method was used to measure the generated O2 volume in Step 2 at the identical current of 1,000 mA. The obtained result indicates that ∼6 ml O2 was generated in Step 2. Therefore, the H2-to-O2 ratio is 2:1 in the consecutive cycle of Steps 1 and 2.
Combination between the H2-production and NiOOH-Zn battery
Interestingly, the aforementioned O2-production step (Step 2) can be replaced by the discharge step of the NiOOH-Zn battery (Step 2′), which will enable the coupling of H2 production with a discharge step of the NiOOH-Zn battery (Fig. 4a). As shown in Fig. 4a, the H2 production (Step 1) includes the cathodic reduction of H2O on the HER electrode (H2O→H2) and the anodic oxidization of the Ni(OH)2 electrode (Ni(OH)2→NiOOH). Next, the NiOOH electrode that is formed in Step 1 is coupled with a zinc anode to form a NiOOH-Zn battery. The subsequent discharge step (Step 2′) of the NiOOH-Zn battery is based on the cathodic reduction of the NiOOH electrode (NiOOH→Ni(OH)2) and the anodic oxidization (Zn→ZnO22−) of zinc48,49,50. In other words, the architecture in Fig. 4a provides an interesting rechargeable cycle that produces H2 with charge (that is, electrolysis in Step 1) and delivers energy with the discharge of the NiOOH-Zn battery (Step 2′). To confirm this hypothesis, the NiOOH electrode, which formed after the electrolysis for H2 production with an applied current of 200 mA for 600 s, was directly coupled with a zinc-plate electrode in the electrolysis cell to construct a NiOOH-Zn battery. The discharge profile of the NiOOH-Zn battery was investigated with a current of 200 mA. As shown in Fig. 4b, the NiOOH-Zn battery displays a discharge voltage of ∼1.6 V with a total discharge time of ∼600 s. Furthermore, the consecutive H2-production step (Step 1) and discharge step (Step 2′) of the NiOOH-Zn battery can be cycled exactly like a rechargeable battery (inset of Fig. 4b). Further cycle data of Step 1 (H2-production step) and Step 2′ (discharge of the NiOOH-Zn battery) are provided in Supplementary Fig. 18. Therefore, the architecture in Fig. 4a introduces a new energy storage/conversion approach, in which solar energy can be used to drive the electrolysis (charge process) of Step 1 to produce H2 during the daytime, and the discharge step (Step 2′) of the NiOOH-Zn battery can be used to deliver energy to power electronic devices overnight. To clarify this point for the layperson reader, the NiOOH-Zn battery that formed after the H2-production step was used to power a 1.6 V electric fan (Supplementary Movie 5). The cycle of Steps 1 and 2′ also increases the Zn(OH)2 concentration in the alkaline electrolyte. The Zn(OH)2 alkaline solution can be used to produce O2 and metallic Zn through electrolysis at the proper time with other energy inputs, such as at night-time with wind power, nuclear fission and so on. In addition, the rechargeable system based on the H2 production (Step 1) and discharge of the NiOOH-Zn battery (Step 2′) exhibits a theoretical energy density of 280 Wh kg−1 (see Supplementary Fig. 19 for details), which is close to the theoretical energy density of conventional Ni-MH batteries, Ni-Cd batteries or Ni-Zn batteries and higher than the theoretical energy density of lead-acid batteries and aqueous Li-ion batteries51.
Solar simulator calibration and Solar Cell characterization
For the 1 sun measurement, we used a solar simulator (ABET Technologies, Model Sun2000) equipped with a 550 W xenon lamp as a light source. The GaInP/GaAs/GaInNAsSb multi-junction solar cell (0.316 cm2 area) was manufactured by Solar Junction.
As the cells were operated under concentrated sunlight, the calibration of the solar simulator for 1 sun conditions (100 mW cm−2) was carried out using the AM 1.5 Direct spectrum (ASTM G173). Although reference solar cells can be used to adjust a simulator for appropriate total power output, spectral control is crucial for accurate multi-junction cell measurements. For this reason, the external quantum efficiency (EQE) of each sub-junction of the multi-junction stack was measured using a grating monochromator (Newport CS260) calibrated with silicon and germanium photodetectors (Newport 918D-UV, 918D-IR). All light sources and photodetectors were calibrated by the manufacturers before the experiment. These EQE measurements were integrated with the AM 1.5D spectrum to determine short-circuit current densities and to understand the current-limiting junction of the cells. The EQE of each sub-junction and the AM 1.5D spectrum are shown overlaid in Supplementary Fig. 1. The cells were all top-junction limited, which allowed the simulator to be tuned without luminescent coupling impacting the current27. The 1 sun I–V characteristics of the cell were measured using this condition. These data are shown in Fig. 2a.
For the PV-electrolysis measurement, we used a multi-sun solar simulator (Newport, Model 66921) with a 1,000 W xenon lamp as the white light source. A water filter was applied to the light beam, to remove the excessive infrared component from the lamp spectrum and to better match the AM 1.5D spectrum. The cell package was mounted onto a water-cooling stage such that a surface temperature of 25 °C was maintained. The distance between the cell and the lamp was adjusted to achieve the desired short circuit current (JSC). The intensity of the concentrated white light was determined from the ratio of the JSC under concentration to the JSC under 1 sun illumination, consistent with standard practices for characterizing concentrated PV cells32,33. No concentrator optics were placed between the simulator and the cell and therefore the possible effects of concentrator optics were not considered in efficiency calculation. As the cell was cut and installed in a standard CPV cell package, the whole cell area is active under illumination; therefore, no aperture or mask was used. Numerical modelling based on the junction ideality factor was later conducted to determine whether open-circuit voltage (VOC) inflation due to the mismatch between the solar simulator spectrum and the AM 1.5D spectrum resulted in STH inflation36,37,38. The results of these calculations show that the lamp spectrum mismatch did not cause significant STH inflation in our experiments (details provided in Supplementary Note 4 and Supplementary Table 1).
During the duration of the experiment the solar cell performance was very stable, as expected for III–V solar cells. No hysteresis or time-transient behaviour was observed during the I–V measurement. The I–V characteristics results in Fig. 2 were measured with a forward voltage sweep rate of 50 mV s−1 and a sampling period of 0.02 s. Changing the sweep rate, direction or sampling rate did not generate a noticeable difference in the results.
Electrolyser fabrication and characterization
Membrane electrode assemblies were fabricated using a conventional catalyst-coated membrane technique. Nafion 115 membranes purchased from FuelCellsEtc were cut into 3.5 cm × 3.5 cm2 pieces. These membranes were pretreated by soaking in 3% H2O2 at 80 °C for 1 h, then soaking in 0.5 M H2SO4 at 80 °C for 1 h and finally soaking in Millipore (18.2 MΩ cm) water at 80 °C for 1 h. The membranes were removed from the water and blotted dry before the catalyst was deposited. Next, a Pt black catalyst (ETEK) and Nafion 117 ionomer solution (Aldrich) were mixed in a 3:1 weight ratio. Separately, an Ir black catalyst (Premetek) and Nafion 117 ionomer solution (Aldrich) were also mixed in a 3:1 weight ratio. The Pt and Ir catalyst/ionomer mixtures were both dispersed in 4:1 volume ratio mixtures of isopropanol and water. The catalyst/ionomer solutions were sonicated for several minutes and then deposited onto opposite sides of the Nafion membranes by spray casting. The Pt catalyst was loaded on the cathode side at 0.5 mg cm−2 and the Ir catalyst was loaded on the anode side at 2 mg cm−2 over a 2.5 cm × 2.5 cm area for a total device active area of 6.25 cm2. This catalyst-coated membrane was pressed between carbon paper (Sigracet GDL 35BC, Ion Power) on the cathode side and Ti mesh (Dexmet) on the anode side. Two identical assemblies prepared in this manner were loaded into cell assemblies (5 cm2, Fuel Cell Technologies, Inc.), which were maintained at a temperature of 80 °C for all measurements. The two electrolysers were connected in series. Millipore water (18.2 MΩ cm) preheated to 80 °C was fed into the anode side of the first electrolyser; there was no input to the cathode side. The cathode and anode outputs of the first electrolyser were connected to the cathode and anode inputs of the second electrolyser and both outputs of the second electrolyser were collected; thus, the H2 and O2 products could be quantified using a volume displacement Faradaic efficiency measurement apparatus.
PV-electrolysis system operation
To construct the PV-electrolysis system, the triple junction PV cell, the two electrolysers and a potentiostat (BioLogic, VMP3) were connected in series as follows: the working electrode port of the potentiostat was connected to the bottom contact of the solar cell, the top contact of the solar cell was connected to the anode of the first electrolyser, the cathode of the first electrolyser was connected to the anode of the second electrolyser and the cathode of the second electrolyser was connected to the counter electrode port of the potentiostat. The potentiostat reference lead was connected to the counter electrode lead so that it could measure the current passing through the closed system; no additional potential was applied. All electrical connections were made with standard copper cables, which introduced negligible resistance compared with other components of the system.
Before the start of the operation, preheated Millipore water was purged with H2 and O2, and pumped into the two electrolysers with a Chem-tech Series 100 pump at a flow rate of 42 ml min−1. The solar cell was kept at 25 °C on a water cooler stage and positioned under the multi-sun solar simulator. The distance between the cell and the solar simulator was adjusted so that ∼42 suns of solar concentration was achieved. At this concentration, the solar cell output a short circuit photocurrent of 184 mA (∼583 mA cm−2) and aligned the solar cell I–V curve for an optimal operation point to match the electrode size and electrolyser capacity.
To begin operation, the shutter of the solar simulator was opened. The system current was recorded continuously by the potentiostat and these data were used to calculate the STH efficiency as a function of time as shown in Fig. 4. The system was run continuously for 48 h without interruption or modification. Periodically throughout the experiment, the gas products from the cathodes and anodes of the electrolysers were collected using volume displacement devices to calculate the Faradaic efficiency. At the end of the 48 h operation, the shutter of the solar simulator was closed.
The data that support the findings of this study are available from the authors on request.