Yolk–shell structured microspheres consisting of CoO/CoP hetero-interfaced nanocomposites as highly active hydrogen evolution reaction electrocatalysts for AEM electrolyzer stacks

Abstract

An anion exchange membrane water electrolyzer (AEM electrolyzer) is an advanced technology for converting electrical energy into hydrogen energy with high efficiency and low cost. However, its advantages are diminished by significant voltage losses due to insufficient catalytic activity. To address this issue, yolk–shell structured microspheres consisting of CoO/CoP hetero-interfaced nanocomposites (YS-CoO/CoP) have been developed as highly active hydrogen evolution reaction (HER) electrocatalysts, which exhibited and an over potential −126 mV at −10 mA cm⁻² and was applied in an AEM electrolyzer. The yolk–shell structure improves ion and gas transport, reducing mass transport losses, while enhancing HER activity resulting from electron redistribution at the CoO/CoP interface. The AEM electrolyzer equipped with YS-CoO/CoP for the HER and NiFe-LDH for the OER shows reduced activation and mass transport losses, achieving a high current density of 0.6 A cm⁻² at 1.8 V_cell. Additionally, a 3-cell AEM electrolyzer stack equipped with non-platinum group metal (non-PGM) electrocatalysts for both the OER and HER demonstrates high performance.

---

1. Introduction

Global warming has accelerated the shift from fossil fuels to renewable energy sources.^1,2 Hydrogen energy, in particular, is gaining recognition as an eco-friendly alternative due to its zero carbon emissions and high energy density positioning it as a strong candidate to replace fossil fuels.^3–5 One promising technology is water electrolysis, which converts electrical energy into hydrogen gas by splitting water molecules.⁶ While 1.23 V is the theoretical voltage required for water electrolysis, practical conditions demand higher voltages due to overpotentials in the oxygen evolution reaction (OER) and the hydrogen evolution reaction (HER).^7–9 Minimizing these overpotentials is critical for achieving high efficiency in water electrolysis. Consequently, development of highly efficient electrocatalysts is essential for enhancing hydrogen production.

Among the various water electrolysis technologies, an anion exchange membrane water electrolyzer (AEM electrolyzer) is considered a next-generation water electrolysis technology, as it combines the advantages of both an alkaline water electrolyzer (AWE) and proton exchange membrane electrolyzer (PEM electrolyzer).^10,11 Operating in an alkaline environment similar to that of AWE, the AEM electrolyzer enables the use of abundant and inexpensive non-platinum group metals (non-PGMs) as electrocatalysts. Additionally, like the PEM electrolyzer, it features a zero-gap configuration, which reduces ohmic losses and enhances high energy density.^12,13 However, the commercialization of AEM electrolyzers remains challenging due to the lower catalytic activity of non-PGMs compared to platinum group metals (PGMs). As a result, the technology is still in its early stages and requires further extensive research to achieve maturity.

Several strategies have been proposed to improve catalyst activity of non-PGM based electrocatalysts, such as nanostructuring,^14–16 alloying,^17–19 defect engineering,^18,20,21 heterostructuring,^19,22,23 and doping.^24–26 For efficient improvement of catalytic activity, a strategy that simultaneously enhances both the qualitative and quantitative activity of the electrocatalyst is essential. Among these approaches, using heterostructures that create interfaces between different phases are a promising strategy for enhancing the qualitative activity. The lattice mismatch between these phases induces defect formation,²⁷ while the interactions between them lead to electron redistribution and modifications in the electronic structure.²⁸ This can also regulate surface energy and improve the kinetics of catalytic reactions. Nanostructuring, such as dendrites, nanosheets, nanocages, and other similar shapes, improves the quantitative activity of the electrocatalyst by increasing its porosity or surface area, allowing for more active sites and enhancing overall catalytic performance.²⁹ Particularly in AEM electrolyzers, where operation at high current densities above 0.5 A cm⁻² (under conditions of vigorous bubble generation) is sustained for extended periods, having a structure that efficiently removes gas is critically important.

In this study, we developed yolk–shell structured microspheres consisting of a CoO/CoP hetero-interfaced nanocomposite (YS-CoO/CoP) as highly active HER electrocatalysts for AEM electrolyzer stacks. The yolk–shell structure facilitates efficient ion and gas transport, minimizing mass transport losses during high current density operation of the AEM electrolyzer. Furthermore, the HER activity is significantly enhanced by improving the Volmer step, driven by electron redistribution at the interface between CoO and CoP. The AEM electrolyzer equipped with YS-CoO/CoP as the HER electrocatalyst and NiFe-LDH as the OER electrocatalyst was successfully implemented, demonstrating a significant reduction in both activation losses related to catalytic activity and mass transport losses associated with bubble transport. Furthermore, a 3-cell AEM electrolyzer stack equipped with non-PGM electrocatalysts for both the OER and HER, was successfully realized, delivering excellent performance.

2. Results and discussion

The formation mechanism of yolk–shell structured microspheres consisting of a CoO/CoP hetero-interfaced nanocomposite is illustrated in Scheme 1. To prepare yolk–shell structured microspheres as a starting frame, a one-pot spray pyrolysis process was carried out to obtain cobalt oxide yolk–shell microsphere powder as shown in Scheme 1a. The droplet was formed by using a spray nebulizer from the spray solution, which contained soluble cobalt nitrate and sucrose. Sucrose as a carbon source was the core material for producing yolk–shell microspheres by a one-pot spray pyrolysis process. When the droplet was transported into a hot furnace, it dried and decomposed to a dense structured cobalt oxide–carbon composite intermediate. The excess sucrose in the spray solution facilitated dense structured cobalt oxide–carbon composite microspheres. Using air as a carrier gas, cobalt oxide–carbon@void@cobalt oxide microspheres could be formed by carbon combustion, which was described in previous literature studies.^30,31 As it continued to pass through the furnace, the cobalt oxide@void@cobalt oxide yolk–shell configuration was finally obtained, and this entire process took place within seconds by a one-pot process. The synthesized cobalt oxide yolk–shell microspheres were heat-treated with sodium hypophosphite sources at various temperatures as shown in Scheme 1b. Under the post-treatment condition of 300 °C, Co₃O₄ sprayed precursor yolk–shell microspheres were converted to CoO yolk–shell microspheres, not cobalt phosphide related phase microspheres. This revealed that phosphidation did not proceed under 300 °C conditions. On increasing post-treatment temperature to 400 °C, CoO/CoP nanocomposite yolk–shell microspheres were produced, and hetero-interfaced CoO/CoP nanocrystals were uniformly dispersed over the yolk–shell microspheres. The primary CoO@voids@CoP core–shell structured nanocrystals were formed based on the nanoscale Kirkendall diffusion effect from CoO during the phosphidation process. When the temperature was increased to 500 °C, the cobalt oxide phase was entirely converted to the cobalt phosphide (CoP/CoP₂) nanocomposite. Hetero-interfaced CoO@CoP nanocrystals were transformed into CoP@CoP₂ nanocrystals while maintaining the heterostructure.

Scheme 1 Schematic of the synthetic procedure. (a) Spray pyrolysis process for YS-Co₃O₄. (b) Formation mechanism of YS-CoO, YS-CoO/CoP and YS-CoP/CoP₂.

The morphological properties of one-pot sprayed precursor cobalt oxide yolk–shell microspheres (denoted as YS-Co₃O₄) are presented in Fig. S1.† The SEM images in Fig. S1a and b† show the morphology of yolk–shell structured microspheres with a mean size of 0.9 μm. The little difference in particle size is due to differences in the size of the droplets formed from the spray generator. The XRD data indicated that one-pot sprayed precursor yolk–shell microspheres had the Co₃O₄ phase (PDF#80-1536) as shown in Fig. S1c.† The post-treated cobalt oxide yolk–shell microspheres at 300 °C under a reducing atmosphere with sodium hypophosphite sources showed almost the same yolk–shell structure as shown in Fig. S2.†

The morphological characteristics of yolk–shell structured microspheres consisting of a CoO/CoP hetero-interfaced nanocomposite (denoted as YS-CoO/CoP), which were prepared by post-treatment at 400 °C, are exhibited in Fig. 1. The SEM images in Fig. 1a and b reveal that the yolk–shell structure was well maintained even after post-treatment of phosphidation at a high temperature of 400 °C. The TEM images in Fig. 1c and d verify that multi-shelled microspheres were formed and the yolk part had a porous structure, which could contribute to providing high reactive surface area, effectively improving the acceleration of the bubble release rate and electrolyte penetration, and avoiding the random particle aggregation during the HER process. The multi-shell structure could be synthesized by multi-step carbon combustion mechanisms as reported in previous papers.^32,33 The TEM image in Fig. 1e reveals that core–shell structured CoO@CoP nanocrystals were formed by the phosphidation process from the sprayed Co₃O₄ yolk–shell microspheres, which consisted of dense CoO nanocrystals. Furthermore, some voids in core–shell CoO@CoP nanocrystals were exhibited as indicated by arrows in Fig. 1e. This was proved by renowned nanoscale Kirkendall diffusion as reported in previous literature studies.^34,35 The diffusion rates of the Co and O components were faster than that of phosphorus gas during the thermal phosphidation process resulting in the formation of hollow cavities. The high-resolution (HR)-TEM image in Fig. 1f shows that the lattice fringe spacings were measured to be 0.21 and 0.19 nm, which correspond to planes of CoO (200) and CoP (211), respectively. Moreover, a heterointerface (denoted by a white dotted line) between CoO (200) and CoP (211) planes was revealed, indicating the existence of the CoO/CoP heterojunction. It is noteworthy that lattice disorder, local defects, and the formation of the heterostructure occurred simultaneously, mainly due to the oxygen loss and the grain growth in the process of solid-state phosphating. This heterostructure could improve kinetics of electrochemical reactions and deliver a platform for exposing active sites. The selected area electron diffraction (SAED) pattern in Fig. 1g also verifies the formation of the cubic-CoO and orthorhombic-CoP composite. The elemental mapping images in Fig. 1h show the uniform and overall well dispersed CoO/CoP hetero-interfaced nanocomposites in the yolk–shell structured microspheres.

Fig. 1 Morphological characteristics of YS-CoO/CoP. (a and b) SEM images of YS-CoO/CoP. (c–e) TEM images of YS-CoO/CoP. (f) HR-TEM image of YS-CoO/CoP. (g) SAED patterns of YS-CoO/CoP. (h) TEM-EDS mapping images of YS-CoO/CoP.

The morphological properties of yolk–shell structured microspheres consisting of a CoP/CoP₂ hetero-interfaced nanocomposite (denoted as YS-CoP/CoP₂), which was prepared by post-treatment at 500 °C, are shown in Fig. 2. The SEM images in Fig. 2a and b show similar yolk–shell structured microspheres as compared to those of YS-CoO/CoP in Fig. 1a and b. On the other hand, the TEM images in Fig. 2c and d indicate that nanocrystals, which were composed of yolk–shell microspheres, were considerably grown and sintered resulting in a decrease in the porosity of the entire yolk–shell microsphere. This could make the electrolyte penetration and the bubble release difficult during the HER process. Moreover, no internal voids are observed in the grown nanocrystals in Fig. 2e compared to Fig. 1e. It could be comprehended that internal voids were filled by crystal growth during phosphidation at higher temperatures. The HR-TEM image in Fig. 2f shows that the lattice fringe spacings were measured to be 0.23 and 0.25 nm, which correspond to the CoP (201) and CoP₂ (200) planes, respectively. Similarly, an interface between the CoP (201) and CoP₂ (200) planes was found compared to CoO/CoP in Fig. 1f. It could be interpreted that CoO was converted to CoP, and CoP was converted to the P-rich CoP₂ phase in YS-CoP/CoP₂ at a higher temperature while maintaining hetero-interfaces. The selected area electron diffraction (SAED) pattern in Fig. 2g also shows that the two phases of orthorhombic-CoP and monoclinic-CoP₂ coexisted in YS-CoP/CoP₂. The elemental mapping images in Fig. 2h show that phosphidation progressed sufficiently at a temperature of 500 °C as indicated by weakening of the color mapping signal of oxygen; cobalt oxide was well converted to cobalt phosphide, and CoP/CoP₂ nanocrystals well formed a yolk–shell structured microsphere.

Fig. 2 Morphological characteristics of YS-CoP/CoP₂. (a and b) SEM images of YS-CoO/CoP₂. (c–e) TEM images of YS-CoP/CoP₂. (f) HR-TEM images of YS-CoP/CoP₂. (g) SAED patterns of CoP/CoP₂. (h) TEM-EDS mapping images of CoP/CoP₂.

The crystal structures of post-treated yolk–shell microspheres at 300, 400, and 500 °C of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂, respectively, from one-pot sprayed precursor cobalt oxide yolk–shell microspheres were analyzed by X-ray diffraction (XRD) as shown in Fig. 3a. The XRD pattern of post-treated YS-CoO at 300 °C under a reducing atmosphere with sodium hypophosphite sources represented transformation of the Co₃O₄ phase sprayed precursor into the CoO phase not containing cobalt phosphide phases. This revealed that Co₃O₄ was not sufficiently reduced to metallic Co for phosphidation but was reduced to the CoO phase even though hydrogen phosphide gas was generated from sodium hypophosphite with a decomposition temperature of 200 °C. The yolk–shell structured microspheres consisting of YS-CoO/CoP, which were prepared by post-treatment at 400 °C showed mixed CoO and CoP phases in the XRD pattern. This verified that metallic Co sufficiently reduced on the surface of the nanocrystals consisting of the yolk–shell microspheres and was phosphidized to the CoP phase at 400 °C, and that the CoO phase was formed inside the nanocrystals, which was confirmed by hetero-interfaced CoO/CoP nanocrystals as shown in Fig. 1f. On the other hand, the cobalt oxide phase of YS-Co₃O₄ was completely transformed into cobalt phosphides (CoP and CoP₂ composite) at 500 °C as shown in the XRD pattern. This showed that the CoO phase, which was not sufficiently reduced at 400 °C, was converted to CoP at 500 °C, and the existing CoP phase was converted to the P-rich CoP₂ phase by the continuous hydrogen phosphate gas reaction at a higher temperature of 500 °C. Fig. 3b displays the Raman spectra of hetero-interfaced YS-CoO/CoP and YS-CoP/CoP₂, which were prepared by post-treatment at 400 and 500 °C, respectively. YS-CoO/CoP contained five peaks at wavenumbers 184, 464, 508, 604, and 667 cm⁻¹, which were associated with the CoO phase.³⁶ Peaks corresponding to CoP are not detected due to the weak vibration of the Co–P bond, which is in good agreement with previous literature.³⁷ On the other hand, YS-CoP/CoP₂ exhibited several peaks at 151, 196, 258, and 354 cm⁻¹, which correspond to the vibrational modes of the Co–P bond in the CoP₂ phase while peaks at 274, 318, and 443 cm⁻¹ correspond to those of the CoP phase.^38–41 This verified the formation of the CoP/CoP₂ hetero-interfaced composite as confirmed by XRD data. As a result of the phosphidation process, new weak peaks at 258 cm⁻¹ appear in the CoO/CoP spectrum, and peaks at 151 cm⁻¹ and 258 cm⁻¹ overlap in the CoP/CoP₂ spectrum, assigned to phosphate (P–O bonds).³⁷ The chemical states for YS-CoO/CoP and YS-CoP/CoP₂ were further investigated by X-ray photoelectron spectroscopy (XPS) analysis as shown in Fig. 3c–e and S3.† The wide scans for YS-CoO/CoP and YS-CoP/CoP₂ suggested the presence of Co, P, and O elements as shown in Fig. S3.† The high-resolution Co 2p spectra in Fig. 3c of the two samples were mostly similar, but showed a slightly shifted pattern. In YS-CoO/CoP, which was prepared at 400 °C, the peaks were fitted at 778.5 and 793.4 eV, which could be assigned to Co 2p_3/2 and Co 2p_1/2, respectively, and corresponded to Co³⁺ in the Co–P phase.^42,43 The peaks of high binding energies at 781.6 eV (Co 2p_3/2) and 798 eV (Co 2p_1/2) were associated with Co²⁺ in the Co–O phase.^42,43 Additionally, Tang's group confirmed the electronic distribution phenomena induced by the CoO/CoP heterointerface through density of states (DOS) calculations. The DOS intensity of CoO/CoP is significantly higher than that of CoP and CoO on the right side of the Fermi level, indicating enhanced electrical conductivity and improved charge transfer efficiency due to heterointerface formation.⁴⁴ Interestingly, the fitted peak at 778.9 eV (Co 2p_3/2) from YS-CoP/CoP₂, which was prepared at 500 °C, was shifted in a positive direction compared to that (778.5 eV) of YS-CoO/CoP. On the other hand, in the P 2p spectra in Fig. 3d, peaks of YS-CoO/Co were found at 129.2 and 131.1 eV, which could be ascribed to P 2p_3/2 and P 2p_1/2, respectively.⁴² Contrary to P 2p, the fitted peak at 128.9 (P 2p_3/2) and 129.8 eV (P 2p_1/2) from YS-CoP/CoP₂, which was prepared at 500 °C, was shifted in a negative direction compared to that of YS-CoO/CoP.⁴⁵ Both shifts strongly suggest that the Co species have a partial positive charge (δ⁺) and the P species possess a partial negative charge (δ⁻), indicating the resulting formation of CoP and CoP₂ as a result of a transfer of electron density from Co to P.^46,47 Other peaks in P 2p spectra were identified at 133.5 (YS-CoO/CoP) and 133.8 eV (CoP/CoP₂), which were attributable to oxidized P due to sample preparation in air.⁴⁶ As shown in Fig. 3e, the peaks at 531.5 eV and 533.1 eV correspond to the Co–O bond and P–O bond in CoO and CoP phases in the O 1s XPS spectrum of YS-CoO/CoP, respectively.⁴² Compared with YS-CoO/CoP, negative shifts of O 1s peaks (shifts to 531.4 and 532.8 eV) are observed for YS-CoP/CoP₂ implying that the local charge density of CoO has changed after modification with CoP.^48–50 The Brunauer–Emmett–Teller (BET) surface areas and pore structures of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂ were analyzed through N₂ gas adsorption and desorption isotherms, as well as Barrett–Joyner–Halenda (BJH) measurement as shown in Fig. S4.† The BET surface areas of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂ were 15.6, 14.7, and 6.39 m² g⁻¹, respectively. As the post-heat treatment temperature increases from 300 to 500 °C, it could be confirmed that the specific surface area gradually decreases due to crystal growth. However, YS-CoO and YS-CoO/CoP did not show a significant difference even when the synthesis temperature was increased. This could be understood as the voids inside the primary nanocrystals, which consist of yolk–shell microspheres, formed by the nanoscale Kirkendall effect affecting the specific surface area. The specific surface area was drastically reduced because the internal void disappeared due to YS-CoP/CoP₂ crystal growth at high temperature of 500 °C. This was demonstrated by TEM images in Fig. 1e and 2e. Moreover, pore distributions of the three samples showed that YS-CoO and YS-CoO/CoP, which were prepared at 300 and 400 °C, respectively, had micropores (<2 nm) and mesopores (4 nm and above 10 nm). On the other hand, YS-CoP/CoP₂ prepared at 500 °C showed only mesopores of around 4 nm and micropores and mesopores larger than 10 nm were not observed. This could also be attributed to the effect of drastic crystal growth at 500 °C. As shown in Table S1,† the molar ratio of YS-CoO, YS-CoO/CoP and YS-CoP/CoP₂ was analyzed by using inductively coupled plasma optical emission spectroscopy (ICP-OES). The results indicate that the phosphorus content increased as the phosphidation temperature increased.

Fig. 3 Crystal structures and chemical states of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂. (a) XRD patterns of YS-CoO, YS-CoP/CoP, and YS-CoP/CoP₂. (b) Raman spectra of YS-CoO/CoP, and YS-CoP/CoP₂. High-resolution XPS spectra of (c) Co 2p, (d) P 2p, and (e) O 1s.

To evaluate the electrocatalytic activity for the HER, LSV curves were measured in a 1 M KOH solution, as shown in Fig. 4a. The overpotentials at a current density of −10 mA cm⁻² were compared to assess the HER activity. YS-Co₃O₄ exhibited poor catalytic activity, while YS-CoO showed enhanced HER activity, with an overpotential of approximately −218 mV. Notably, YS-CoO/CoP demonstrated significantly enhanced HER activity, with an overpotential of approximately −126 mV. CoP exhibits optimal adsorption energy for the HER, attributed to modifications in its electronic structure induced by Co–P bonding.⁵¹ Additionally, the interface between CoO and CoP promotes electron redistribution and defect formation, further enhancing HER performance.⁵² Furthermore, this interface accelerates the Volmer step in alkaline electrolyte by improving the water dissociation ability, thereby boosting HER activity.⁴⁴ Despite being a ceramic, CoP exhibits metallic behavior and possesses high electrical conductivity.⁵³ However, YS-CoP/CoP₂ showed lower catalytic activity compared to YS-CoO/CoP. CoP₂ has low electrical conductivity due to the strong covalent bonds between Co and P, and its intrinsic catalytic activity for the HER is lower than that of CoP.⁵⁴ Therefore, YS-CoO/CoP demonstrated the highest HER activity. Previous studies have shown that the formation of the CoP/CoO heterointerface significantly enhances water affinity, thereby accelerating the Volmer step in alkaline HER. This also facilitates hydrogen adsorption and desorption, ultimately improving HER kinetics.⁴⁴

Fig. 4 Electrochemical analysis of YS-Co₃O₄, YS-CoO, YS-CoO/CoP and YS-CoP/CoP₂. (a) polarization curves for the HER. (b) Tafel plots. (c) Nyquist plots at −0.45 V_RHE. (d) Comparison of the overpotential at −10 mA cm⁻² and Tafel slope. (e) Durability test of YS-CoO/CoP at −100 mA cm⁻² for 300 h.

The Tafel slope, which indicates the kinetics of the electrochemical reaction, was calculated from the LSV curves, as shown in Fig. 4b. The Tafel slopes of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂ were measured to be 96 mV dec⁻¹, 65 mV dec⁻¹, and 65 mV dec⁻¹, respectively. These results indicate that YS-CoO exhibited the slowest HER kinetics, whereas YS-CoO/CoP and YS-CoP/CoP₂ demonstrated significantly enhanced HER kinetics. In alkaline HER, there are two reaction mechanisms: the (1) Volmer–Heyrovsky mechanism and (2) Volmer–Tafel mechanism. These reaction mechanisms are identified through the Tafel slope for the HER. Tafel slopes of 120 mV dec⁻¹, 40 mV dec⁻¹, and 30 mV dec⁻¹ correspond to cases where the Volmer, Heyrovsky, and Tafel reactions, respectively, are the rate determining steps in alkaline HER. The Tafel slope of CoO was 96 mV dec⁻¹, indicating that the reaction mechanism for the HER follows the Volmer–Heyrovsky pathway. Similarly, the Tafel slope of CoO/CoP was 65 mV dec⁻¹, also indicating the Volmer–Heyrovsky mechanism. A lower Tafel slope indicates improved HER kinetics, meaning that the rate-determining step is facilitated. According to a previous study, the heterointerface of CoO and CoP has been reported to enhance water adsorption capability, which aligns with our experimental results.⁴⁴ A Nyquist plot was obtained to measure the charge transfer resistance using the equivalent circuit shown in Fig. 4c, where CPE, R_cl, and R_ct represent the constant phase element, catalyst layer resistance, and charge transfer resistance, respectively. The R_ct value for YS-CoO/CoP was 1.7 Ω, which is lower than that of YS-CoO (3.5 Ω) and YS-CoP/CoP₂ (2.0 Ω), demonstrating that the CoO/CoP heterostructure is a more efficient HER electrocatalyst. Fig. 4d provides a summary of the overpotential at a current density of −10 mA cm⁻² and the Tafel slope. The electrochemical surface area (ECSA) is another parameter used to evaluate catalytic performance. Fig. S5† shows cyclic voltammetry curves of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂ at various scan rates, with the C value directly related to the ECSA. As shown in Fig. S5,† the C_dl value decreased as the phosphidation temperature increased. This suggests that high-temperature annealing drove atom diffusion,⁵⁵ thereby reducing the microscopic roughness and surface area of the electrocatalyst.^56–58 To assess durability, a galvanostatic test was conducted at a constant current density of −100 mA cm⁻² for 300 hours, as shown in Fig. 4e. Over 300 hours, the catalyst exhibited a degradation rate of only 0.3 mV h⁻¹, indicating exceptional durability.

To verify the applicability of YS-CoO/CoP in a full-cell system, the synthesized catalysts were applied and evaluated in an AEM electrolyzer. Fig. 5a presents a simplified schematic diagram of the AEM electrolyzer. As shown in Fig. 5b, the AEM electrolyzer equipped with YS-CoO/CoP achieved a current density of 0.6 A cm⁻² at 1.8 V, outperforming those equipped with YS-CoO (0.1 A cm⁻²) and YS-CoP/CoP₂ (0.3 A cm⁻²). Notably, YS-CoP/CoP₂ exhibited a gradually increasing slope at higher current densities, indicating a decline in performance. This decline is attributed to its relatively small specific surface area, leading to reduced mass transport characteristics. To further analyze this, the overvoltages of the AEM electrolyzer were separated: ohmic losses (η_ohm), activation losses (η_act), and mass transport losses (η_mass). The ohmic losses of the AEM electrolyzer equipped with YS-CoO/CoP were the lowest, indicating that YS-CoO/CoP had the highest conductivity (Fig. S6†). Furthermore, the activation losses decreased in the order of YS-CoO, YS-CoP/CoP₂, and YS-CoO/CoP, with YS-CoO/CoP exhibiting the least voltage loss (Fig. S7†). This trend is consistent with the half-cell test results, confirming that YS-CoO/CoP had the highest activity, followed by YS-CoP/CoP₂ and YS-CoO. However, in terms of mass transport loss, YS-CoO, YS-CoO/CoP and YS-CoP/CoP₂ showed progressively lower losses (Fig. S8†). This is attributed to the collapse of the yolk–shell pore structure during phosphidation. YS-CoP/CoP₂, lacking mesopores, showed particularly high mass transport loss. Fig. S9† presents the total voltage losses of the AEM electrolyzer equipped with YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂ across all current densities. To further validate this, EIS analysis was conducted (Fig. S10†). High-frequency resistance (HFR) values for the AEM electrolyzer equipped with YS-CoO, YS-CoO/CoP and YS-CoP/CoP₂ were approximately 0.1 Ω cm², 0.09 Ω cm², and 0.11 Ω cm², respectively, supporting the reasonableness of the ohmic loss results. The radius of the semicircle represents the polarization resistance (R_pol) of the AEM electrolyzer, with YS-CoO/CoP exhibiting the smallest semicircle, as shown in Fig. S10.† Fig. S11† shows the performance of the AEM electrolyzer tested in previous studies, incorporated into the polarization curve of an AEM electrolyzer equipped with YS-CoO/CoP. The performance of the AEM electrolyzer equipped with YS-CoO/CoP, while inferior to that of the AEM electrolyzer utilizing PGM based catalysts (grey symbols), was superior to that employing non-PGM based catalysts (blue symbols) in previous studies. Table S2† summarizes the performance in the referenced previous studies. To evaluate the conversion rate of the electrocatalysts, the turnover frequency (TOF) was calculated (Fig. 5c). The AEM electrolyzer equipped with YS-CoO/CoP achieved a hydrogen conversion rate of 0.47 H₂ s⁻¹. This rate was three times and 1.5 times faster than that of the electrolyzer equipped with YS-CoO (0.15 H₂ s⁻¹) and YS-CoP/CoP₂ (0.29 H₂ s⁻¹), respectively, indicating that YS-CoO/CoP exhibited a significantly enhanced reaction rate. When operating at a current density of 1 A cm⁻², it reached an efficiency of 66%, with an electrical energy consumption of 50 kW h kg⁻¹ (Fig. 5d). To assess durability, a constant current test was conducted at a current density of 0.5 A cm⁻² for 135 hours. The degradation rate was 1 mV h⁻¹ and cell efficiency decreased by only 7%, indicating stable durability.

Fig. 5 1-Cell AEM electrolyzer equipped with YS-CoO, YS-CoO/CoP and YS-CoP/CoP₂. (a) Schematic of a 1-cell AEM electrolyzer. (b) Polarization curves of a 1-cell AEM electrolyzer. (c) Turnover frequency (TOF) of a 1-cell AEM electrolyzer. (d) Electrical energy to generate 1 kg of hydrogen and efficiency at 1.0 A cm⁻². (e) Durability test and 1-cell efficiency of YS-CoO/CoP at 0.5 A cm⁻² for 135 h.

Finally, YS-CoO/CoP was applied to a 3-cell AEM electrolyzer stack to evaluate its performance and durability. Fig. 6a presents a schematic diagram of the 3-cell stack. The stack equipped with YS-CoO/CoP achieved a current density of 0.3 A cm⁻² at 5.4 V_stack (equivalent to 1.8 V_cell), which was lower than the single-cell performance (Fig. 6b). To investigate this performance degradation, EIS analysis was performed. As shown in Fig. 6c, the HFR of the 3-cell stack was approximately 0.29 Ω cm². The HFR of a single cell was approximately 0.09 Ω cm², and since three cells are connected in series, the ideal resistance would be around 0.27 Ω cm². However, the HFR of the 3-cell stack was confirmed to be slightly higher. This increase in resistance, attributed to the contact between stack components, contributed to the observed performance decline. In addition, connecting the three cells in series increased the complexity of fluid flow, restricting ion and gas transport, and further complicating the electrochemical resistance elements within the stack.⁵⁹ At a current density of 1.0 A cm⁻², the stack efficiency and electrical energy consumption were calculated to be 64.2% and 51.9 kW h kg⁻¹, respectively (Fig. 6d). To assess the durability of the 3-cell stack, a constant current test was conducted at a current density of 0.5 A cm⁻² for 158 hours (Fig. 6e). During this period, the degradation rate of the 3-cell stack was approximately 3 mV_stack h⁻¹, indicating a faster degradation rate than that of the single cell. Similar to the performance degradation, this is attributed to the increased complexity of the electrochemical resistance elements as the number of cells in the stack increases.⁵⁹ To evaluate the stability of the catalysts, we performed SEM, XRD, and XPS analyses on the YS-CoO/CoP materials after the durability test. The SEM images in Fig. S12a and b† reveal that YS-CoO/CoP maintains its yolk–shell structure. However, the partial growth of nanoplates on the yolk–shell structure is noticeable. Previous study has reported significant reconstruction of catalysts after durability tests, which reduces their contact area.⁶⁰ Likewise, YS-CoO/CoP also undergoes reconstruction, but its yolk–shell structure does not collapse, and it can sustain contact area between the catalysts. To further assess the stability of the crystal structure, we conducted XRD analysis after the durability tests. The XRD patterns in Fig. S12c† confirm that YS-CoO/CoP retains both CoO and CoP crystal structures. Additional peaks corresponding to the carbon cloth and Co(OH)₂ peaks were detected. Moreover, a small shift in the peaks caused by interlayer ions was observed, along with the appearance of new Co(OH)₂ peaks.^61–63 In addition, XPS spectra of YS-CoO/CoP were analyzed to investigate the surface chemical states of elements in detail. As shown in Fig. S13a,† the XPS survey spectrum after durability tests detected signals from Co, P, O, C, and F, the latter of which is derived from the Nafion residue on catalyst elements.⁶⁴ The high-resolution Co 2p spectra revealed new characteristic peaks attributed to Co(OH)₂ located at 783.1 eV as shown in Fig. S13b.⁶⁵† Notably, the disappearance of Co³⁺ (Co–P) peaks suggests changes in the bonding structure of the transition metal phosphides. As is well known, transition metal phosphide consists of covalent and ionic bonding, with ionic bonds having much longer bond lengths.⁶⁶ During the electrochemical reaction, surface Co–P covalent bonds are rapidly converted into ionic bonds. Therefore, YS-CoO/CoP catalysts are predominantly in an ionic state rather than a covalent site. This characteristic can enhance their catalytic activity.^66–68 The P 2p spectrum showed that surface phosphorus remains unchanged after durability tests⁶⁴ (Fig. S13c†). However, in the O 1s spectrum, changes in the surface oxygen chemical states were observed (Fig. S13d†). Specifically, the intensity of the Co–O bond decreased, while the intensity of the P–O bonds increased, and O_abs peaks appeared. These results indicated that during the HER process, H⁺ can preferentially absorb on the surface of CoP and tends to bond with neighboring CoP sites.⁶⁹

Fig. 6 3-Cell stack equipped with YS-CoO/CoP. (a) Schematic of a 3-cell stack. (b) Polarization curves of the 3-cell stack. (c) Nyquist plot of the 3-cell stack at 0.1 A cm⁻². (d) Electrical energy to generate 1 kg of hydrogen and efficiency at 1.0 A cm⁻². (e) Durability test and stack efficiency at 0.5 A cm⁻² for 158 h.

Similarly, the stack efficiency was found to decrease more rapidly than that in the single cell. Although the performance and durability of the 3-cell stack were inferior to that of the single cell, our study suggests that the non-PGM HER electrocatalysts developed in our laboratory have the potential for application in stacks. Our unique approach, which combines the spray pyrolysis method, Co–O–P heterointerface formation, and the nanovoid structure derived from the Kirkendall effect, enhances HER activity and offers a novel pathway with potential for commercial application. To compare our strategy with related studies, we summarized previous studies on yolk–shell structures in Table S3.†

3. Conclusion

In summary, we developed yolk–shell structured microspheres consisting of a CoO/CoP hetero-interfaced nanocomposite (YS-CoO/CoP) as highly active HER electrocatalysts for AEM electrolyzer stacks. The yolk–shell structure promotes efficient ion and gas transport, minimizing mass transport losses during high current density operation of the AEM electrolyzer. The CoO/CoP heterostructure significantly enhanced the HER activity by improving the Volmer step, driven by electron redistribution at the interface between CoO and CoP. The AEM electrolyzer equipped with YS-CoO/CoP for the HER and NiFe-LDH for the OER electrocatalyst was successfully implemented, showing a marked reduction in both activation and mass transport losses. Furthermore, a 3-cell AEM electrolyzer stack utilizing non-PGM electrocatalysts for both the OER and HER was successfully realized, demonstrating high performance.

4. Experimental section/methods

4.1. Preparation of the sprayed precursor (YS-Co₃O₄) and phosphidized samples (YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂)

Yolk–shell structured Co₃O₄ microspheres (YS-Co₃O₄) were synthesized through a one-step spray pyrolysis process. The aqueous spray solution was prepared by dissolving sucrose (0.7 M, C₁₂H₂₂O₁₁, Samchun Chemical Co., Ltd) and Co (NO₃)₂·6H₂O (0.2 M, Samchun Chemical Co., Ltd) in distilled water (1 L). During the spray pyrolysis, a 1.7 MHz ultrasonic nebulizer equipped with six oscillator generated droplets. The quartz reactor, with a length of 1200 mm and diameter of 50 mm, was operated at 700 °C. Air is utilized as the carrier gas is utilized at a flow rate of 10 L min⁻¹. The generated droplets were delivered into a quartz reactor, where they were dried and decomposed into yolk–shell structured Co₃O₄ microspheres through a carbon combustion mechanism. The synthesized yolk–shell structured Co₃O₄ microspheres were collected in a Teflon bag filter. Prepared yolk–shell structured Co₃O₄ microspheres, along with NaH₂PO₂ as the phosphorus source, were subjected to a chemical vapor deposition (CVD) process in a tube furnace. NaH₂PO₂ was placed upstream and the Co₃O₄ microspheres were positioned downstream in the tube furnace. Subsequently, the reaction was carried out at different temperatures (300 °C, 400 °C, and 500 °C) for 3 h at a rate of 2 °C min⁻¹ under a N₂ atmosphere, resulting in the formation of YS-CoO, YS-CoO/CoP, and YS-CoP/CoP₂, respectively.

4.2. Characterization

The morphological characteristics of the samples were investigated via field emission scanning electron microscopy (FE-SEM, S-4700, Hitachi) and field emission-transmission electron microscopy (FE-TEM, JEM-2100F) at the Korea Basic Science Institute (Daegu). The crystal structures of the samples were analyzed using X-ray diffraction spectroscopy (XRD, X'Pert PRO with Cu K_α radiation, λ = 1.5418 Å). X-ray photoelectron spectroscopy (XPS) of the samples was performed using an ESCALAB-250 with Al K_α radiation (1486.6 eV). The structural characteristics of the sample were investigated via Raman spectroscopy (Jobin Yvon LabRam HR800, excited by using a 632.8 nm He/Ne laser) at room temperature.

4.3. Electrochemical characterization

All electrochemical measurements were conducted by using a potentiostat (BP2, WonAtech) in a three-electrode system and were iR-corrected. A platinum mesh, Hg/HgO (1 M KOH), and a glassy carbon rotating disk electrode (RDE, 0.198 cm²) were used as the counter electrode, reference electrode, and working electrode, respectively. A 1 M KOH aqueous solution was used as an electrolyte, and the electrolyte was purged with N₂ gas. The ink was prepared by mixing the electrocatalyst (20 mg), Vulcan carbon (1 mg), ethanol (900 μL), and 5 wt% Nafion solution (100 μL). The mixture was sonicated for 15 minutes. Then, the prepared ink (5 μL) was dropped onto the RDE and then dried in a convection oven for 10 minutes. Linear sweep voltammetry (LSV) was conducted at a scan rate of 5 mV s⁻¹ in 1 M KOH solution. The Tafel slope was calculated using the Tafel formula: η = b log j + a, where η indicates an overpotential, j is the current density, and b is the Tafel slope.⁷⁰ Electrochemical impedance spectroscopy (EIS) analyses were conducted over a frequency range of 100 kHz to 1 Hz, with an applied potential of −0.45 V vs. RHE. The Nyquist plots were fitted using the Z-view program. Electrochemical surface area (ECSA) was calculated using the equation ECSA = C_dl/C_s, where C_dl is the double layer capacitance, and C_s is the specific capacitance value (40 μF cm⁻²).^71,72 The double layer capacitance (C_dl) was obtained by fitting cyclic voltammetry data at various scan rates of 5, 10, 20, 30, 40, and 80 mV s⁻¹. The durability test was conducted by applying a constant current of −100 mA cm⁻² for 300 hours.

4.4. AEM electrolyzer test

The AEM electrolyzer consists of an anion exchange membrane (AEM), an anode (OER electrode), a cathode (HER electrode), a gold-plated current collector, a titanium flow channel, and a stainless steel-end plate. A Sustainion®X37-50 grade RT from Dioxide Materials was used as the AEM. The cathode was prepared by spraying the ink onto carbon cloth. The ink for the cathode was prepared by mixing the HER electrocatalyst, isopropyl alcohol (IPA), deionized water (DI water), Vulcan carbon, and an ionomer (Fumion® FAA-3-solut-10), followed by 25 minutes of sonication to create a homogeneous solution. The loading mass of HER electrocatalysts was 2 mg cm⁻² ±5%. The anode was prepared by spraying the ink onto commercial nickel foam. The ink for the anode was prepared by mixing the OER electrocatalysts, IPA, DI water, and the ionomer (Fumion® FAA-3-solut-10), followed by 15 minutes of sonication to create a homogeneous solution. NiFe-LDH was employed as the OER electrocatalyst. It was synthesized using a co-precipitation method, following synthesis procedures from previous studies.⁷³ The loading mass of OER electrocatalysts was 5.0 mg cm⁻² ±5%. The active area of the AEM electrolyzer was 4.9 cm², and the operating temperature was approximately 50 °C. The electrolyte was the 1 M KOH aqueous solution with a flow rate of 100 mL min⁻¹. The electrochemical analysis of the AEM electrolyzer was conducted using a DC power supply (MK-W102, MKPOWER). The Nyquist plots were measured at a current density of 0.1 A cm⁻². The galvanostatic test was conducted at a current density of 0.5 A cm⁻² for 135 hours. The cell efficiency was calculated using the following equation:
where ΔH_H2,LHV is a lower heating value reaction enthalpy for water electrolysis (241.8 kJ mol⁻¹), n_H2,measured is hydrogen production (mol s⁻¹) measured by using a DC power supply, I is the applied current (A) and V is the applied voltage (V).^74–76 The turnover frequency (TOF) was calculated using the formula TOF = (a × i)/(2 × n × F).⁷⁷ In the formula, a represents the active area, i is the current density, 2 is the number of electrons involved in the hydrogen evolution reaction, n is the number of metal atoms and F is the Faraday constant (96 485C mol⁻¹). The electrical energy to produce 1 kg of hydrogen was calculated using the following equation:
where N_A represents Avogadro’s number, M (H₂) is the molecular mass of hydrogen, and η_F is the faradaic efficiency, which was assumed to be 100% for the calculation.⁷⁸ The three-cell stack (3-stack) configuration consists of three MEAs connected in series, and the active area was the same as that of the single cell. The durability test was conducted at 0.5 A cm⁻² for 158 hours.

Data availability

The data supporting this article have been included as part of the ESI.† Additional data are available on reasonable request to the authors.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) (Grant No. RS-2023-00217581 and RS-2023-00236572).

References

1. A. Kalair, N. Abas, M. S. Saleem, A. R. Kalair and N. Khan, Energy Storage, 2021, 3, e135.
2. Y.-N. Zhou, Y.-W. Dong, Y. Wu, B. Dong, H.-J. Liu, X.-J. Zhai, G.-Q. Han, D.-P. Liu and Y.-M. Chai, Chem. Eng. J., 2023, 463, 142380.
3. S. Evro, B. A. Oni and O. S. Tomomewo, Int. J. Hydrogen Energy, 2024, 78, 1449–1467.
4. I. Rolo, V. A. F. Costa and F. P. Brito, Energies, 2024, 17, 180.
5. Y.-N. Zhou, F.-T. Li, B. Dong and Y.-M. Chai, Energy Environ. Sci., 2024, 17, 1468–1481.
6. M. A. Khan, H. Zhao, W. Zou, Z. Chen, W. Cao, J. Fang, J. Xu, L. Zhang and J. Zhang, Electrochem. Energy Rev., 2018, 1, 483–530.
7. H. Wu, C. Feng, L. Zhang, J. Zhang and D. P. Wilkinson, Electrochem. Energy Rev., 2021, 4, 473–507.
8. X. Zhang, C. Feng, B. Dong, C. Liu and Y. Chai, Adv. Mater., 2023, 35, 2207066.
9. R.-Y. Fan, Y.-N. Zhou, M.-X. Li, J.-Y. Xie, W.-L. Yu, J.-Q. Chi, L. Wang, J.-F. Yu, Y.-M. Chai and B. Dong, Chem. Eng. J., 2021, 426, 131943.
10. S. Palmas, J. Rodriguez, L. Mais, M. Mascia, M. C. Herrando and A. Vacca, Curr. Opin. Electrochem., 2023, 37, 101178.
11. G. A. Lindquist, PhD, University of Oregon, 2023.
12. R. Vinodh, S. S. Kalanur, S. K. Natarajan and B. G. Pollet, Polymers, 2023, 15, 2144.
13. M. Moreno-González, P. Mardle, S. Zhu, B. Gholamkhass, S. Jones, N. Chen, B. Britton and S. Holdcroft, J. Power Sources Adv., 2023, 19, 100109.
14. Y. Aykut and A. Bayrakçeken Yurtcan, Electrochim. Acta, 2023, 471, 143335.
15. M. A. Gaikwad, V. V. Burungale, D. B. Malavekar, U. V. Ghorpade, U. P. Suryawanshi, S. Jang, X. Guo, S. W. Shin, J.-S. Ha, M. P. Suryawanshi and J. H. Kim, Adv. Energy Mater., 2024, 14, 2303730.
16. M. Mottakin, M. Sukor Su'ait, V. Selvanathan, M. A. Ibrahim, H. Abdullah and M. Akhtaruzzaman, J. Alloys Compd., 2024, 1002, 175351.
17. Y. Wang, N. Gong, G. Niu, J. Ge, X. Tan, M. Zhang, H. Liu, H. Wu, T. L. Meng, H. Xie, K. Hippalgaonkar, Z. Liu and Y. Huang, J. Alloys Compd., 2023, 960, 171039.
18. H. Wang, W. Li, J. Wu, D. Wang, D. Hou, M. Sheng and R. Guo, Energy Fuels, 2023, 37, 19103–19112.
19. X. Lu, M. Du, T. Wang, W. Cheng, J. Li, C. He, Z. Li and L. Tian, Int. J. Hydrogen Energy, 2023, 48, 34009–34017.
20. J. Yang, L. An, S. Wang, C. Zhang, G. Luo, Y. Chen, H. Yang and D. Wang, Chin. J. Catal., 2023, 55, 116–136.
21. J. Liang, S. Li, F. Li, L. Zhang, Y. Jiang, H. Ma, K. Cheng and L. Qing, J. Colloid Interface Sci., 2024, 655, 296–306.
22. L. Tian, Y. Liu, C. He, S. Tang, J. Li and Z. Li, Chem. Rec., 2023, 23, e202200213.
23. S. Wang, X. Ning, Y. Cao, R. Chen, Z. Lu, J. Hu, J. Xie and A. Hao, Inorg. Chem., 2023, 62, 6428–6438.
24. J. Wu, Z.-F. Wang, T. Guan, G. Zhang, J. Zhang, J. Han, S. Guan, N. Wang, J. Wang and K. Li, Carbon Energy, 2023, 5, e268.
25. G. Ma, J. Ye, M. Qin, T. Sun, W. Tan, Z. Fan, L. Huang and X. Xin, Nano Energy, 2023, 115, 108679.
26. F. Saeed, M. Ahmad, Samia, A. Zada, D. Qi and Y. Wang, Int. J. Hydrogen Energy, 2024, 59, 1196–1204.
27. G. H. Olsen, J. Cryst. Growth, 1975, 31, 223–239.
28. Q. Wang, R. He, F. Yang, X. Tian, H. Sui and L. Feng, Chem. Eng. J., 2023, 456, 141056.
29. I. Hussain, C. Lamiel, M. Sufyan Javed, M. Ahmad, X. Chen, S. Sahoo, X. Ma, M. A. Bajaber, M. Zahid Ansari and K. Zhang, Chem. Eng. J., 2023, 454, 140313.
30. M. H. Kim, Y. J. Hong and Y. C. Kang, RSC Adv., 2013, 3, 13110–13114.
31. Y. B. Kim, C. Kim, S.-H. Kim, Y. C. Kang, D. Lee and G. D. Park, Rare Met., 2024, 1–14.
32. S. K. Patel, S. H. Choi, Y. C. Kang and J.-K. Lee, Nanoscale, 2016, 8, 6728–6738.
33. S. H. Choi, S. K. Park, J.-K. Lee and Y. C. Kang, J. Power Sources, 2015, 284, 481–488.
34. H. Wang, X. Shi, S. Ji, X. Wang, L. Zhang, H. Liang, D. Brett, X. Wang and R. Wang, Nanotechnology, 2020, 31, 425404.
35. Y. Wu, H. Wang, S. Ji, B. G. Pollet, X. Wang and R. Wang, Nano Res., 2020, 13, 2098–2105.
36. X. He, X. Song, W. Qiao, Z. Li, X. Zhang, S. Yan, W. Zhong and Y. Du, J. Phys. Chem. C, 2015, 119, 9550–9559.
37. Y. Pan, H. Ren, R. Chen, Y. Wu and D. Chu, Chem. Eng. J., 2020, 398, 125660.
38. I. Pathak, B. Dahal, D. Acharya, K. Chhetri, A. Muthurasu, Y. R. Rosyara, T. Kim, S. Saidin, T. H. Ko and H. Y. Kim, Chem. Eng. J., 2023, 475, 146351.
39. L. Su, X. Cui, T. He, L. Zeng, H. Tian, Y. Song, K. Qi and B. Y. Xia, Chem. Sci., 2019, 10, 2019–2024.
40. Y. Wang, Y. Jiao, H. Yan, G. Yang, C. Tian, A. Wu, Y. Liu and H. Fu, Angew. Chem., 2022, 134, e202116233.
41. M. Ali, M. Wahid and K. Majid, J. Appl. Electrochem., 2023, 53, 95–108.
42. Q. Zhou, R. Sun, Y. Ren, R. Tian, J. Yang, H. Pang, K. Huang, X. Tian, L. Xu and Y. Tang, Carbon Energy, 2023, 5, e273.
43. Q. Li, Y. Wang, J. Zeng, Q. Wu, Q. Wang, L. Sun, L. Xu, T. Ye, X. Zhao and L. Chen, Chin. Chem. Lett., 2021, 32, 3355–3358.
44. Q. Zhou, R. Sun, Y. Ren, R. Tian, J. Yang, H. Pang, K. Huang, X. Tian, L. Xu and Y. Tang, Carbon Energy, 2023, 5, e273.
45. X. Zhang, L. Zhang, G. Xu, A. Zhao, S. Zhang, T. Zhao and D. Jia, Inorg. Chem., 2020, 59, 12232–12239.
46. K. Chen, Y. Cao, W. Wang, J. Diao, J. Park, V. Dao, G.-C. Kim, Y. Qu and I.-H. Lee, J. Mater. Chem. A, 2023, 11, 3136–3147.
47. S. Li, G. Zhang, X. Tu and J. Li, ChemElectroChem, 2018, 5, 701–707.
48. F. Guo, X. Huang, Z. Chen, H. Sun and L. Chen, Chin. J. Chem. Eng., 2021, 40, 114–123.
49. D. Duan, C. Xing, K. Chen, X. Zhou and S. Liu, J. Electroanal. Chem., 2022, 920, 116644.
50. R. Ren, Z. Zhao, Z. Meng and X. Wang, J. Colloid Interface Sci., 2022, 608, 1576–1584.
51. G. Zhou, M. Li, Y. Li, H. Dong, D. Sun, X. Liu, L. Xu, Z. Tian and Y. Tang, Adv. Funct. Mater., 2020, 30, 1905252.
52. H. Wu, Y. Cheng, B. Wang, Y. Wang, M. Wu, W. Li, B. Liu and S. Lu, J. Energy Chem., 2021, 57, 198–205.
53. Y. Hu, M. Liu, Q. Yang, L. Kong and L. Kang, J. Energy Chem., 2017, 26, 49–55.
54. Z. Yang, L. Liu, X. Wang, S. Yang and X. Su, J. Alloys Compd., 2011, 509, 165–171.
55. R.-Y. Shao, X.-C. Xu, Z.-H. Zhou, W.-J. Zeng, T.-W. Song, P. Yin, A. Li, C.-S. Ma, L. Tong, Y. Kong and H.-W. Liang, Nat. Commun., 2023, 14, 5896.
56. S. Kogularasu, Y.-Y. Lee, B. Sriram, S.-F. Wang, M. George, G.-P. Chang-Chien and J.-K. Sheu, Angew. Chem., Int. Ed., 2024, 63, e202311806.
57. J. Qi, B. Jin, W. Liu, W. Zhang and L. Xu, Fuel, 2021, 285, 119163.
58. Y. Guo, Z. Wang, S. Wang, H. Chen, J. Kong, X. Liu, Q. Yang, Y. Zhang and Z. Lü, J. Power Sources, 2024, 592, 233900.
59. M. J. Jang, S. H. Yang, M. G. Park, J. Jeong, M. S. Cha, S.-H. Shin, K. H. Lee, Z. Bai, Z. Chen, J. Y. Lee and S. M. Choi, ACS Energy Lett., 2022, 7, 2576–2583.
60. C. Lei, K. Yang, G. Wang, G. Wang, J. Lu, L. Xiao and L. Zhuang, ACS Sustain. Chem. Eng., 2022, 10, 16725–16733.
61. Q. Zhou, Q. Bian, L. Liao, F. Yu, D. Li, D. Tang and H. Zhou, Chin. Chem. Lett., 2023, 34, 107248.
62. H. Yoon, H. J. Song, B. Ju and D.-W. Kim, Nano Res., 2020, 13, 2469–2477.
63. G. Li, Y. Zhang, W. Ma, F. Liu and H. Li, Microporous Mesoporous Mater., 2024, 379, 113302.
64. X. Guo, M. Duan, J. Zhang, B. Xi, M. Li, R. Yin, X. Zheng, Y. Liu, F. Cao and X. An, Adv. Funct. Mater., 2022, 32, 2209397.
65. Y. Xu, Y. Zhao, M. Sun, W. Xie, Y. Wu, G. Cheng, Y. Zhong, S. Han and L. Yu, Chem. Eng. J., 2024, 490, 151697.
66. W. Li, Y. Jiang, Y. Li, Q. Gao, W. Shen, Y. Jiang, R. He and M. Li, Chem. Eng. J., 2021, 425, 130651.
67. Y. Dong, Z. Deng, Z. Xu, G. Liu and X. Wang, Small Methods, 2023, 7, 2300071.
68. Y. Shi, M. Li, Y. Yu and B. Zhang, Energy Environ. Sci., 2020, 13, 4564–4582.
69. F. Nie, Z. Yang, X. Dai, Z. Ren, X. Yin, Y. Gan, B. Wu, Y. Cao, R. Cai and X. Zhang, J. Colloid Interface Sci., 2022, 621, 213–221.
70. I. T. Kim, S.-H. Kim, J. S. Ha, T. H. Kim, J. Cho, G. D. Park and Y. S. Park, J. Mater. Chem. A, 2023, 11, 16578–16585.
71. J. S. Ha, Y. Park, J.-Y. Jeong, S. H. Lee, S. J. Lee, I. T. Kim, S. H. Park, H. Jin, S. M. Kim, S. Choi, C. Kim, S. M. Choi, B. K. Kang, H. M. Lee and Y. S. Park, Adv. Sci., 2024, 11, 2401782.
72. S. J. Lee, S. H. Lee, J. S. Ha, I. T. Kim, S. H. Park, H. K. Park, H. J. Park, B. K. Kang and Y. S. Park, ACS Sustain. Chem. Eng., 2024, 12, 275–282.
73. Z.-Y. Liu, Q.-Y. Wang and J.-M. Hu, Catal. Sci. Technol., 2024, 14, 110–118.
74. K. In Tae, Y. Jeong-Mi, C. Sun-Yong and P. Yoo Sei, J. Mar. Eng. Technol., 2024, 48, 133–137.
75. S. H. Park, S. H. Lee, J.-Y. Jeong, H. Jin, J. S. Ha, S. J. Lee, I. T. Kim, C. Kim, S. Kim, M. Bae, H. Lee, S. M. Choi, k. Yangdo and Y. S. Park, Int. J. Hydrogen Energy, 2023, 48, 29877–29886.
76. I. T. Kim, S. H. Lee, S. J. Lee, J. S. Ha, S. H. Park, H. Jin, W. J. Lee, B. K. Kang, H. Lee, Y. Kim and Y. S. Park, Energy Fuels, 2024, 38, 23034–23042.
77. H. Zhang, F. Wan, X. Li, X. Chen, S. Xiong and B. Xi, Adv. Funct. Mater., 2023, 33, 2306340.
78. S. Yang, Z. Zhang, A. M. Oliveira, S. Xi, M. Zhiani, J. Zhang, Z. Tu, F. Xiao, S. Wang, Y. Yan and J. Xiao, Adv. Funct. Mater., 2024, 34, 2313275.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta07211a

‡ In Tae Kim and Tae Ha Kim contributed equally to this work.

---