Introduction

Water splitting for hydrogen production offers the advantage of producing clean and sustainable fuel without carbon emissions1,2. To date, proton exchange membrane (PEM) water electrolysis is one of the most established ways in the field of green hydrogen production3,4,5. However, the high overpotential typically associated with the anodic oxygen evolution reaction (OER) poses a significant challenge to enhancing hydrogen production efficiency from water electrolysis6. Furthermore, the advancement of OER catalysts designed for acidic medium poses a greater challenge compared to those intended for alkaline medium. Since highly OER-active electrocatalysts are mainly comprised of metal oxides or hydroxides, most of which exhibit poor stability under acidic conditions7,8,9,10. Therefore, the development of efficient OER catalysts functioning in acidic medium is critical.

Current, effective acidic OER catalysts include Ru, Ir, and Mn-based metal oxides11,12,13,14,15,16,17. In particular, SrIrO3 with an ABO3 perovskite structure (where A typically represents an alkaline earth metal and B represents a transition metal) shows high acidic OER catalytic performance18,19,20,21,22,23,24,25,26,27,28,29,30. In recent years, extensive researchers have been conducted to understand the OER catalytic mechanism of SrIrO3. Jaramillo et al. prepared the (001) plane SrIrO3 film via a laser epitaxy strategy, and observed an incremental improvement in its OER catalytic performance throughout the catalytic process22. Through combining theoretical calculations and experiments, they discovered that the high catalytic activity of SrIrO3 could be attributed to the exposed IrOx sites following Sr dissolution. In further comprehensive studies, researchers confirmed the structural modifications of SrIrO3 under acidic OER process using secondary ion mass spectrometry (SIMS), in situ atomic force microscopy (AFM), and X-ray absorption spectroscopy (XAS), respectively23,24,25. They proposed a correlation between the Sr dissolution process and the formed IrOx surface activities. Moreover, an investigation involving SrIr0.1Co0.9O3 further indicates that the OER activity also originates from the amorphous IrOx structure formed by the dissolution of Co26. These studies appear to identify the active component of SrIrO3-based perovskites as the amorphous IrOx structure. Nevertheless, most studies overlook the impact of B-site dissolution on surface oxygen stability and Ir-O coordination structure, the precise reasons contributing to the high OER catalytic performance of Co-doped SrIrO3 catalysts remain unclear. This issue primarily stems from the challenge faced by researchers in elucidating two key aspects of the OER catalytic process: (1) The key role of Co dissolution in catalytic processes and (2) the influence of surface and bulk Co on IrOx sites.

To address the two key challenges mentioned above, we designed a B-site Co-doped SrIrO3 system to discern the dissolution mechanism at catalyst sites and the origins of IrOx catalytic activity, as shown in Fig. 1a. In situ inductively coupled plasma mass spectrometry (ICP-MS) experiments revealed simultaneous ion dissolution of Co and Sr, caused by acid corrosion prior to OER process. At the OER potential (1.60 V vs. RHE), the dissolution phenomenon was found to be negligible. Along with theoretical calculations, a series of in situ experiments including in situ Raman mapping, in situ XAS, and differential electrochemical mass spectrometry (DEMS) were conducted. These results highlighted the role of Co in two critical ways: (1) Surface Co reduces the stability of the Co-O-Ir bridge oxygen in SrIrO3, leading to the rapid exposure of the low-coordination IrOx structure; (2) Bulk lattice Co optimizes the OOH binding energy of IrOx, consequently reducing the overpotential. Thus, the synthesized Co-doped SrIrO3 demonstrated high OER activity, markedly surpassing commercial IrO2 catalysts in PEM water electrolyzer. The insights obtained from this research would significantly enhance the understanding of high OER catalytic performance of SrIrO3-based perovskite catalysts, providing key insights for designing and preparing high-performance acidic OER catalysts.

Fig. 1: Characterizations of samples.
figure 1

a OER catalytic mechanism diagram of Co-doped SrIrO3 catalyst; b–d HRTEM images and corresponding FFT diagrams of SI, SI6C1 and SI1C1; e O K-edge EELS spectra of SI6C1; f HAADF image of SI6C1 and corresponding EDS mapping images; g, h XRD spectra of SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, and SI samples. i ICP–MS diagram of Co/Ir and Sr/Ir ratios of SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, and SI samples.

Full size image

Results

Fabrication and characterizations of Co-doped SrIrO3 samples

SrIrO3 doped with different amounts of Co were synthesized by using the sol-gel method, and were fully acid washed before use. The samples were denoted as follows: Sr2IrCoOx (SI1C1), Sr3Ir2CoOx (SI2C1), Sr5Ir4CoOx (SI4C1), Sr7Ir6CoOx (SI6C1), Sr9Ir8CoOx (SI8C1), and SrIrOx (SI). The crystal structures of these samples were analyzed through high-resolution transmission electron microscopy (HRTEM) and X-ray diffraction (XRD) in Fig. 1b–h, respectively. For HRTEM testing, three samples (SI, SI6C1, and SI1C1) were selected. Specifically, SI exhibited an orthorhombic crystal structure. Fast Fourier transform (FFT) showed that the primary exposed crystal planes included (112), (110), and (002). Despite its low surface crystallinity, SI6C1 still clearly exposed crystal planes such as (112), (204), and (020). Meanwhile, SI1C1 displayed poor surface crystallinity with only a few areas showing short-range ordered crystallinity and no clear diffraction spots in the FFT pattern. To investigate the reason for the diminished crystallinity observed in SI6C1 sample, we conducted electron energy loss spectroscopy (EELS) analyses, the results of which are presented in Fig. 1e and Supplementary Fig. 3. The peaks at 532, 542, and 564 eV can be ascribed to the Ir 5d, while the peak at 536 eV is attributable to the Sr 4d31,32. Comparative analysis reveals that the Sr 4d peak of the SI sample is generally higher than that of the SI6C1 sample, suggesting that the presence of Sr contributes to maintaining the crystalline structure. Moreover, we identified a significant discrepancy between the surface and bulk Sr concentrations within the SI and SI6C1 samples, with this effect being particularly pronounced in the SI6C1 sample. This indicates that Co dissolution also impacts the proportion of Sr presenting on the surface. Further investigations were undertaken through energy-dispersive X-ray spectroscopy (EDS) mapping (Fig. 1f). The results confirm that the dissolution of Sr/Co and the subsequent formation of IrOx.

XRD revealed that the diffraction peaks of SI were consistent with typical pseudocubic (Pnma) perovskites. Also, Co-doped SI also demonstrated similar diffraction peaks, confirming the successful synthesis of the perovskite-based catalyst21,26. As shown in Fig. 1h, Co doping significantly reduced the crystallinity of SI and resulted in the diffraction peak shifting to the high angle, indicating lattice contraction in the Co-doped SI. In particular, SI1C1 and SI2C1 displayed almost no diffraction peaks, possibly due to the substantial dissolution of surface Co and Sr by acid washing process. These findings suggest that excessive Co doping may disrupt the pseudocubic perovskite structure after acid washing, a conclusion in line with the HRTEM results. The overall composition of the samples was analyzed by using ICP-MS, as shown in Fig. 1i. Notably, there was a significant discrepancy between the Co and Ir initial ratio and the final composition ratio of the samples. Samples with poor crystallinity, specifically SI1C1 and SI2C1, exhibited the highest Co/Ir ratio, with SI2C1 slightly higher than SI1C1. This could possibly be attributed to the difficulty in maintaining the perovskite structure in SI1C1, which led to a large amount of Co dissolution during the acid washing process and, consequently, a reduced Co/Ir ratio. Despite possessing the highest Co/Ir ratios, these two samples were still significantly lower than the initial ratio, further confirming the instability of surface Co in acidic conditions. The Co/Ir ratios of SI4C1, SI6C1, and SI8C1 showed a decreasing trend from SI4C1, SI6C1 to SI8C1 as shown in. Figure 1i suggesting that lattice maintenance assists in stabilizing Co atoms in the bulk phase. Among various SI-based samples, the SI1C1 and SI2C1 displayed the lowest Sr/Ir ratio, and the Sr/Ir ratio significantly increased with the decrease of Co doping. The Sr/Ir ratio of the SI sample was slightly higher than 1, indicating a slightly higher Sr proportion compared to Ir in bulk structure, as suggested by the EELS results shown in Supplementary Fig. 3. These conclusions were further confirmed by additional characterization methods such as scanning electron microscope (SEM), X-ray photoelectron spectroscopy (XPS), and Raman spectroscopy, as shown in Supplementary Fig. 5a, b.

High OER catalytic performance of Co-doped SrIrO3

The OER catalytic performances of the SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, and SI samples were investigated in acidic medium. The linear sweep voltammetry (LSV) curves for SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, and SI samples were recorded after 10 cyclic voltammetry (CV) cycles, as shown in Fig. 2a. The results indicate that SI6C1 necessitates an overpotential of only 245 mV to reach a current density of 10 mA/cm2, which is approximately 5 times higher than that of SI at the same potential. Given that the OER catalytic process primarily occurs on the surface, the performance of SI1C1, SI2C1, SI4C1, SI6C1 and SI8C1 samples is generally superior and exhibits a similar trend. This aligns with previous characterization results, suggesting that despite varying Co doping ratios, the catalysts’ surface structures are comparable. As presented in Fig. 2b, c, the SI1C1, SI2C1, SI4C1, SI6C1 and SI8C1 samples show Tafel slopes within the range 51.5–53.6 mV/dec, significantly lower than that of SI (59.5 mV/dec). This suggests that Co participation can enhance the reaction kinetics of OER. As shown in Fig. 2d, a comparison of the mass activity of the SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, SI and IrO2 samples reveals that the Co-doped samples all demonstrated high activity, significantly surpassing those of SI and IrO2.

Fig. 2: Electrochemical measurements of samples.
figure 2

a OER polarization curves of SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, and SI samples with a mass loading of 0.025 mg/cm2 in 0.5 M H2SO4 and (b) corresponding Tafel slopes. The resistance values for SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, SI and IrO2 were 3.9, 3.7, 3.6, 4.7, 3.2, 4.6 and 4.4 Ω, respectively. c Comparison of overvoltages and Tafel slopes of SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, and SI samples. d Comparison of OER mass activity of SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, SI and IrO2 samples. e Schematic diagram of PEM water electrolysis device. f PEM water electrolysis performance of SI6C1, IrO2, and SI samples, in set: PEM water electrolysis device photograph. g PEM water electrolysis stability of SI6C1 sample.

Full size image

To study the potential applications of SI series catalysts in water electrolysis, we further examined the electrocatalytic performance of three samples (SI, SI6C1, and IrO2) for water electrolysis by PEM. The schematic diagram of the PEM water electrolyzer is presented in Fig. 2e. The performance results of the PEM water electrolysis, shown in Fig. 2f, reveal that the catalytic activity of SI6C1 is significantly higher than those of SI and IrO2. It can achieve a current density exceeding 1000 mA/cm2 at 2.0 V cell voltage at 85 °C, which is much higher than those of the SI catalyst (580 mA/cm2) and IrO2 catalyst (560 mA/cm2) under the same conditions. We further tested the long-term stability of SI6C1 catalyst for PEM water electrolysis as shown in Fig. 2g and Supplementary Fig. 6. The results indicate that the SI6C1 catalyst exhibits high stability, with a performance decay rate of 0.21 mV/h, which is comparable to that of the SI catalyst and commercial IrO2 catalyst. This exceptional high stability indicates that the SI6C1 catalyst exhibits the potential for practical application in PEM water electrolysis.

The roles of Co doping and dissolution in SrIrO3 studied by theoretical calculations

The impact of Co doping on SrIrO3 was investigated through a theoretical study. As numerous early studies had substantiated that SrIrO3-based catalysts undergo significant Sr dissolution during OER23,24,25. This was examined by analyzing the Pourbaix diagrams of SrIrO3 and Sr4Ir3CoO12, particularly focusing on their stability in acidic medium (Fig. 3a and Supplementary Fig. 7). The findings reveal a marked similarity in the stability properties of SrIrO3 and Sr4Ir3CoO12. Under acidic conditions, Sr displays a thermodynamic inclination towards dissolution, thereby exposing a multitude of IrOx sites on the surface. Additionally, Co in the Sr4Ir3CoO12 also exhibits instability in acidic medium, which could also result in substantial dissolution, consistent with the EDS mapping and ICP-MS results in Fig. 1.

Fig. 3: Theoretical calculations of SrIrO3-based perovskite catalysts.
figure 3

a Pourbaix diagram of SrIrO3. b Possible computational models of IrOx, IrCosurfOx (surface Co doping) and IrunsatCo2LOx (second layer Co doping with unsaturated IrOx) for DFT calculations. c Adsorbate evolution mechanism (AEM) diagram. d OER free energy diagrams of different models. e Overpotential of different computational models. f Density of states diagrams for IrOx and IrunsatCo2LOx, and (g, h) energy band center and volcano plot for different computational models.

Full size image

As indicated by the Pourbaix diagram, the structure of catalyst is significantly influenced by the applied voltage. However, the two catalysts do not exhibit notable thermodynamic differences near the oxygen stability curve, although they display pronounced differences in thermodynamic tendencies across varying pH values. This suggests that the effect of SrIrO3 dissolution at different OER potentials may be significantly less than the effect of electrolyte pH, particularly in strongly acidic medium. Previous research suggests that the Ir-O coordination number of surface IrOx on SrIrO3 during OER catalysis is ~4.525. While the presence Sr content at the trace amount has been detected in numerous studies, determining the exact state of Sr at the subnanometer scale remains a challenge23. By combining the results from the Pourbaix diagrams and above characterizations, seven models were constructed to investigate the theoretical OER catalytic sites and catalytic performance of SrIrO3 and Co-doped SrIrO3 (Fig. 3b and Supplementary Fig. 8).

The theoretical calculations were conducted on the models, as shown in Fig. 3c. The computed free energy diagrams (Fig. 3d) reveal that the rate-determining step for all models is the formation of *OOH. The Co doping, whether at the surface and bulk phase, which would not change the coordination structure of IrO6 octahedron, can only enhance OER catalytic activity to a certain extent. However, such activity enhancements are restricted, especially for bulk-doped Co, where the OER overpotential is reduced by only 15 ~ 42 mV. In contrast, the surface unsaturated IrOx, which forms following Co dissolution, exhibits a significant improvement in the catalytic activity for oxygen evolution, with the overpotential decreasing by 311 ~ 344 mV (Fig. 3e). These findings suggest that Co may not directly participate in the catalysis, but rather promote the surface reconstruction through site dissolution, leading to rapid exposure of more low-coordination IrOx active sites.

To further investigate the activity origin, density of states (DOS) analyses were performed on these models (Fig. 3f). The data suggest that Co doping significantly affects the Co-O-Ir bridge oxygen, shifting the O 2p band center closer to the Fermi level. According to the related studies, such a shift in the O 2p band center substantially affect the surface stability of the catalyst and may be associated with oxygen dissolution33,34. This indicates that the primary effect of Co doping is to alter the surface stability of SrIrO3, and promote the surface reconstruction and the formation of IrOx active sites. Additionally, the doping and dissolution of Co significantly influence the Ir 3d band center (Fig. 3g). The displacement of the metal 3d orbitals typically directly impacts the binding energy to the adsorbate. In this study, the negative shift of the Ir 3d band center led to a decrease in *OOH free energy, significantly enhancing the OER catalytic activity (Fig. 3g). By establishing the relationship between the Ir 3d and O 2p band centers and the theoretical overpotential, a new OER volcano plot was formulated (Fig. 3h). The data revealed that the moderate O 2p band center and the lower Ir 3d band center play crucial roles in the catalytic activity of SrIrO3-based catalysts.

In situ characterizations of surface structural changes of catalysts

To elucidate the catalytic mechanism, DEMS tests were first conducted, as shown in Fig. 4a. The surfaces of SI1C1, SI2C1, SI4C1, SI6C1, SI8C1, SI, and IrO2 samples were labeled with 18O, as shown in Supplementary Figs. 10–12. For each sample, m/z = 34 signals at different cyclic voltammetry cycles were collected, and the ratio of 18O16O to 16O2 was utilized to eliminate the natural abundance of 18O in the air. As shown in Supplementary Fig. 13, the lattice oxygen evolution reaction (LOER) trend of most catalysts across different cyclic voltammetry cycles is similar, and the Co doping ratio appears to facilitate the release of lattice oxygen. The results indicated that Co doping reduces the stability of the surface oxygen in SrIrO3, consistent with O 2p center calculation results. Notably, the LOER of SI1C1 and SI2C1 is the most prominent, with SI2C1 slightly surpassing SI1C1. This phenomenon has been previously explained by XRD and ICP-MS characterizations, as SI1C1 has difficulty in maintaining the perovskite lattice, the proportion of Co doping decreases. In contrast, SI and IrO2 exhibit the smallest LOER process, which is linked to the stability of the catalyst surface lattice oxygen. The results indicated that Co doping and dissolution can activate the catalyst lattice oxygen, thereby accelerating the formation of IrOx.

Fig. 4: In situ characterizations of samples.
figure 4

a The 18O16O percentage of the samples test by DEMS. b In situ ICP-MS diagram of SI and SI6C1 and (c) its differential transformation diagram, in set: enlarged diagram. d, e In situ Raman and Raman mapping of SI samples and SI6C1 samples.

Full size image

In addition to LOER, in situ ICP-MS experiments were performed to detect the phenomenon of ion dissolution during OER, as shown in Fig. 4b, c. We observed that both SI6C1 and SI undergo considerable dissolution when immersed in an acidic electrolyte. Coupling this observation with prior XRD and Raman results, it suggests that the dissolution process is triggered by the rapid dissolution of Sr/Co-related oxides or compound heterophases. Furthermore, the subsequent steady dissolution trend indicates that ion dissolution of the catalyst does not necessarily occur during the catalytic process. It is noteworthy that the Sr dissolution phenomenon of SI is more significant than that of SI6C1, which is attributed to the higher concentration of surface and bulk Sr in SI. Moreover, SI6C1 is also accompanied by a small amount of Co dissolution, thus further confirming that Co dissolution promotes the formation of unsaturated IrOx.

To confirm the structural information of the catalyst, an in situ Raman mapping study was conducted, as shown in Fig. 4d, e. The peak around 600 cm−1 can be attributed to the Ir-μ-oxo stretching vibration of IrOx (involving the unprotonated bridge oxygen, Ir3+), and the characteristic peaks around 550 cm−1 and 720 cm−1 can be attributed to the typical Eg and B2g vibrational peaks of IrO2. In addition, the characteristic peaks around 300 and 700 cm−1 can be attributed to Sr-related oxide or compound35,36,37,38. The figure shows that before and after the OER process, SI exhibits an obvious Ir-μ-oxo stretching vibration peak at 600 cm−1, confirming the existence of the perovskite IrO6 structure. However, for SI6C1, a significant decrease in the Ir-μ-oxo stretching vibration peak is observed. This may be attributed to the destruction of the IrO6 structure by Codissolution to form the low-coordination IrOx structure, leading to a significant reduction in peak intensity. Moreover, the characteristic peaks of Sr-related oxides/compounds near 300 and 700 cm−1 nearly disappear from the open circuit voltage (OC). Together with the results of in situ ICP–MS, it can be further confirmed to be caused by the dissolution of Sr-related oxides/compounds.

Key evidences of highly active low-coordination IrOx in Co-doped SrIrO3 for OER

To elucidate the formation mechanism of IrOx on the surface, comprehensive in situ EXAFS spectra was recorded at the Ir-LIII edge to monitor the evolution of local coordination of Ir, as shown in Fig. 5a–e. Initially, we examined the coordination changes of the SI sample, as demonstrated in Fig. 5a, b. The pristine state SI sample owns a Ir-O coordination number of 5.6, indicating the existence of substantial intact Ir-O octahedral structure within the catalyst. During the pre-OER stage (OC, 1.23 V, 1.43 V vs. RHE), the Ir-O coordination numbers of SI sample exhibited a pattern of initial increase followed by a decrease.

Fig. 5: In situ XAS characterization of samples.
figure 5

a R-space fitting diagram for SI sample. b Changes in Ir-O coordination and Ir-O bond length of SI sample. c Structure change diagram of SI sample; d R-space fitting diagram of SI6C1 sample. e Ir-O coordination and Ir-O bond length variation diagram of SI6C1 sample. f Structure change diagram of SI6C1 sample. g Ir–LIII Absorption edge diagram of SI sample, in set: enlarged diagram. h Ir–LIII absorption edge diagram of SI6C1 sample, in set: enlarged diagram.

Full size image

Previous reports showed a significant augmentation in the Ir-O coordination number of SrIrO3 thin film samples during OER, which was ascribed to the oxygen refilling process24,25. In this study, the dissolution of Sr led to alterations in the surface structure of SI. However, there was no significant LOER in SI sample. Therefore, during the OER oxidation process, the rate of H2O filling exceeded the rate of LOER, resulting in the observed trend in coordination numbers. During the OER stage (1.63 V vs. RHE), the Ir-O coordination number of the SI sample exhibited a minor increase, suggesting that at high potentials, the lattice oxygen of SI remained inactive, leading to the higher H2O filling rate than the LOER rate25. Additionally, at 1.23 V vs. RHE, a significant reduction in the Ir-O bond length of the SI sample was observed, potentially indicating a transformation of the octahedral corner-sharing structure during the H2O filling process26. In sum, as shown in Fig. 5c, the surface site mechanism of the SI sample undergoes a dynamic process from a relatively complete perovskite structure to site dissolution and adsorbate filling, ultimately transitioning into a saturated IrOx structure.

We next analyzed the coordination changes in the SI6C1 sample, as illustrated in Fig. 5d, e. The initial Ir-O coordination number of SI6C1 was 6.0, suggesting a more perfect Ir-O octahedral structure compared with SI. However, prior to the OER stage (OC, 1.23 V, 1.43 V vs. RHE), the Ir-O coordination number of the SI6C1 sample exhibited a notable decreasing trend. This is consistent with the formation of low-coordination IrOx as observed in in situ ICP-MS results. Furthermore, during the OER stage, the Ir-O coordination number of SI6C1 further decreased to 5.4, significantly lower than SI, confirming that the active lattice oxygen in SI6C1 further facilitated the formation of highly active low-coordination IrOx.

To validate the generation of the unsaturated IrOx structure, in situ XANES spectra were analyzed at the Ir-LIII edge to study the oxidation state of Ir, as shown in Fig. 5g, h. The Ir-LIII absorption edge of SI displayed an increasing white line intensity during the OER process, indicating an increase in the oxidation state of Ir. Differently, the Ir-LIII absorption edge of SI6C1 exhibited a slight decrease in white line intensity, signifying a reduction in the oxidation state of Ir. Combined with the valence state information of reference samples Ir and IrO2, the mechanism of Ir valence state change during OER can be inferred. During the OER process of SI, the dissolution of Sr and the filling of H2O led to the saturation of the Ir-O coordination, displaying a higher valence state. Conversely, for SI6C1, the dissolution of Co resulted in the formation of unsaturated Ir-O, leading to a decrease in the average Ir valence state. This observation further substantiates that Co can enhance the generation of unsaturated IrOx structures, aligning with the results from prior characterizations.

Discussion

OER catalytic mechanism on Co-doped SrIrO3 catalyst

Our study has elucidated the surface reconstruction process of SIC series catalysts by theoretical calculations and a comprehensive series of in situ characterizations. We now proceed to discuss and summarize the potential OER mechanisms inherent to SIC series catalysts. There are two widely recognized mechanisms for OER, specifically the adsorbate evolution mechanism (AEM) and the lattice oxygen mechanism (LOM), as shown in Fig. 6a, b39,40. Our theoretical calculations have demonstrated that the surface O 2p band center of the Co/Ir model is closer to the Fermi level compared to the Ir model, which suggests that the bridging oxygen of Co-O-Ir is thermodynamically predisposed to be oxidized40. In situ Raman spectroscopy revealed that the Ir-μ-oxo stretching vibration peak (indicative of bridging oxygen) in the Co/Ir system catalyst was notably reduced during the OC and OER processes when compared to the Ir system. Further, DEMS tests have found that a higher content of Co promotes the LOM.

Fig. 6: Proposed possible OER mechanism.
figure 6

a Adsorbate evolution mechanism. b Lattice oxygen mechanism. c Lattice oxygen promoted adsorbate evolution mechanism. d Catalytic mechanism of Co-doped SrIrO3 catalyst.

Full size image

Despite these findings, our DEMS results suggest that the AEM remains the dominant mechanism, while the LOM significantly contributes to the formation of unsaturated IrOx. Consequently, we propose a distinctive catalytic mechanism, the lattice oxygen promoted adsorbate evolution mechanism (LOPAEM), as illustrated in Fig. 6c. Unlike the conventional catalytic mechanisms of OER, LOPAEM involves the synergistic action of both mechanisms. Specifically, when the oxidation rate of lattice oxygen in certain catalysts becomes excessively fast, it will lead to the formation of a large number of surface oxygen vacancies (i.e., unsaturated metal sites), thereby altering the original catalytic sites of the catalyst. These unsaturated metal sites may exhibit more efficient AEM performance, thus achieving LOPAEM. Our findings suggest that the LOM in Co/Ir system catalysts is an integral step for catalyst activation. The formation of Ir-O-Covac substantially reduces the catalyst’s AEM overpotential, thereby making it thermodynamically more favorable. This is in agreement with our theoretical calculations and DEMS results. Moreover, the LOM process is unable to produce a lower Ir-O coordination structure as the dissolution of lattice oxygen shifts the O 2p center away from the Fermi level, precluding further oxidation (as detailed in the theoretical calculation section).

In light of the above findings, we have summarized the mechanism for the Co/Ir system catalyst as represented in Fig. 6d. Initially, Co on the catalyst’s surface dissolves under acidic conditions, removing a portion of the bridging lattice oxygen to form an IrOx structure with lower coordination. Subsequently, lattice oxygen dissolves during the OER process, forming an unsaturated IrOx structure that exhibits high AEM activity. Lastly, under the dynamic balance of adsorbate filling and LOM, the catalyst conducts AEM to efficiently facilitate the OER.

In summary, we designed a highly active catalyst through Co doping and dynamic dissolution of Co/Sr bimetallic ions to study the catalytic mechanism of SrIrO3-based perovskite. Theoretical calculations and in situ characterizations (DEMS, in situ Raman mapping, in situ XAS and in situ ICP-MS) show that dynamic dissolution of Co is crucial for forming highly active unsaturated IrOx. The as-synthesized catalysts exhibit higher OER reaction kinetics than SrIrO3 and commercial IrO2 catalyst in both electrolyzer and PEM water electrolyzer, revealing the mechanism of catalytic activity enhancement by tuning catalytic sites. This work is of great significance for understanding the high OER catalytic performance of Ir-based catalysts, and will provide an important basis for the design and preparation of high-performance acidic OER catalysts.

Methods

Chemicals

The chemical reagents utilized in this study were all received from the manufacturer. Potassium hexachloroiridate (IV) [K2IrCl6, AR, Macklin], cobalt(III) nitrate hexahydrate [Co(NO3)3·6H2O, AR, Sigma-Aldrich], strontium(II) nitrate [Sr(NO3)2, AR, Guangdong chemical reagent)], citric acid monohydrate (AR, LookChem.) were utilized as precursors.

Preparation of SI Catalyst

Solution A was prepared by dissolving Sr(NO3)2 (280 mg) and citric acid (840 mg) in 5.0 mL of deionized water. Solution B was prepared by dissolving K2IrCl6 (80 mg) in 4.0 mL of ethylene glycol. Solution A was then added dropwise with stirring to solution B. The resulting mixture was dried at 150 °C for 12 h to obtain a brown solid product as a precursor. Subsequently, the precursor was calcined in air at 200 °C for 6 h, 300 °C for 6 h, 500 °C for 3 h, and 700 °C for 6 h with a heating rate of 2 °C/min. Afterward, the excess SrCO3 impurities were removed by reacting with a 1.0 M HCl solution for 12 h to obtain SrIrO3 (SI) catalyst.

Preparation of SI1C1 catalyst

Solution A was prepared by dissolving Sr(NO3)2 (280 mg) and citric acid (840 mg) in 5.0 mL of deionized water. Solution B was prepared by dissolving K2IrCl6 (40 mg) and Co(NO3)2 (24 mg) in 4.0 mL of ethylene glycol. The subsequent steps were identical to the preparation of the SI catalyst described above.

Preparation of SI2C1 catalyst

Solution A was prepared by dissolving Sr(NO3)2 (280 mg) and citric acid (840 mg) in 5.0 mL of deionized water. Solution B was prepared by dissolving K2IrCl6 (53 mg) and Co(NO3)2 (16 mg) in 4.0 mL of ethylene glycol. The subsequent steps were identical to the preparation of the SI catalyst described above.

Preparation of SI4C1 catalyst

Solution A was prepared by dissolving Sr(NO3)2 (280 mg) and citric acid (840 mg) in 5.0 mL of deionized water. Solution B was prepared by dissolving K2IrCl6 (64 mg) and Co(NO3)2 (10 mg) in 4.0 mL of ethylene glycol. The subsequent steps were identical to the preparation of the SI catalyst described above.

Preparation of SI6C1 catalyst

Solution A was prepared by dissolving Sr(NO3)2 (280 mg) and citric acid (840 mg) in 5.0 mL of deionized water. Solution B was prepared by dissolving K2IrCl6 (68 mg) and Co(NO3)2 (7 mg) in 4.0 mL of ethylene glycol. The subsequent steps were identical to the preparation of the SI catalyst described above.

Preparation of SI8C1 catalyst

Solution A was prepared by dissolving Sr(NO3)2 (280 mg) and citric acid (840 mg) in 5.0 mL of deionized water. Solution B was prepared by dissolving K2IrCl6 (71 mg) and Co(NO3)2 (5 mg) in 4.0 mL of ethylene glycol. The subsequent steps were identical to the preparation of the SI catalyst described above.

Materials characterizations

Characterization of the atomic-level crystal structure was performed using an aberration-corrected scanning transmission electron microscope (JEM-ARM200P, JAPAN) operated at 300 kV. Energy-dispersive X-ray (EDX) analysis was used to measure the relative elemental content. X-ray diffraction (XRD) patterns of SrIrO3 were recorded on an X-ray diffractometer (Smart lab) using Cu-Kα radiation (λ = 1.5418 Å) with a step size of 0.02° and a step time of 0.2 s in the 20°–80° range. X-ray photoelectron spectroscopy (XPS) was performed using a Thermo Scientific K-Alpha X-ray photoelectron spectrometer, and all XPS spectra were calibrated using the C 1s line at 284.8 eV. The surface morphology of SrIrO3 was characterized using a scanning electron microscope (TESCAN MIRA LMS). Considering the high acid resistance of SrIrO3, anti aqua regia was prepared by mixing hydrochloric acid and nitric acid in a 1:3 ratio for the experiment. In this solution, 1.0 mg of SrIrO3 powder was dissolved in 10 mL of aqua regia and left to stand for 1–3 week after thorough ultrasonic treatment. Finally, the proportions of each element in SrIrO3 were determined by ICP-MS (iCAP RQ) analysis.

In situ characterizations and LOER confirmation

X-ray absorption spectra of Ir L-edge were obtained at the BL17B and BL20U1 beamline of the Shanghai Synchrotron Radiation Facility (SSRF). The spectra were collected either in transmission mode or fluorescence mode using a Lytle detector. The corresponding reference samples were collected in transmission mode. The samples were ground and uniformly applied to special adhesive tape. In situ XANES characterizations were performed in fluorescence mode at the same beamline. The samples were sprayed onto carbon paper at a loading of 1.0 mg cm−2 as the working electrode. The measurements were conducted under the same conditions as the OER measurements in a self-designed cell. The in situ Raman spectroscopy characterizations were carried out using an inVia confocal Raman microscope from Renishaw. The laser power was set at 532 nm with 1% power at a grating of 1800 mm/1, the silicon peak was calibrated before testing. During the OER process, differential electrochemical mass spectrometry (DEMS) measurements were conducted using the QAS 100 apparatus from Shanghai Linglu Instruments to determine the volatile reaction products of the SI series catalysts and IrO2 catalyst labeled with 18O. Saturated Ag/AgCl and Pt wires served as the reference electrode (RE) and counter electrode (CE), respectively. The working electrode (WE) was prepared by sputtering Au onto a 50 μm-thick porous PTFE membrane, followed by depositing 10 μL of catalyst ink (1.0 mg/mL) onto the Au surface. The catalyst isotopic labeling was achieved by cycling the electrode in H218O for eight cycles using cyclic voltammetry (CV) with a scan rate of 5 mV/s in the range of 0–0.5 V vs. Ag/AgCl. Subsequently, the 18O-labeled electrode was rinsed with H216O to remove residual H218O. Finally, the electrode was electrochemically tested against Ag/AgCl in a 1.0 M H2SO4 solution at different potentials with a scan rate of 5 mV/s. The DEMS signal was normalized by current density (A/g). Simultaneously, real-time measurements of gas products with different molecular weights generated during the OER process were conducted using mass spectrometry. In situ ICP-MS experiments were performed using a Thermo Scientific iCAP RQ instrument. The experimental setup consisted of a standard three-electrode cell with a 3 mm glassy carbon working electrode, consistent with the electrochemical testing. The reference electrode used was a saturated calomel electrode (Hg/HgCl2), and the counter electrode was a platinum foil electrode. To ensure accurate detection of ion distribution, a stirrer was employed to prevent leaching and dissolution errors, with data sampling occurring every 15 s.

Electrochemical characterizations

To prepare the catalyst ink, a 0.5 mg amount of catalyst was mixed with 1.0 ml of a 0.05 wt% Nafion solution and neutralized. Subsequently, a 10 μl volume of the prepared ink was deposited onto a glassy carbon electrode (GCE) with a diameter of 5 mm and dried using an infrared lamp. Prior to use, the GCE was polished with 0.05 μm alumina powder and rinsed three times with a mixture of high purity water and ethanol. Electrochemical measurements were conducted in a three-electrode system using an electrochemical workstation (CHI 760E). The reference electrode used was an Hg/HgCl2 electrode in a 0.5 M H2SO4 electrolyte, while a carbon rod served as the counter electrode. The working electrode was the GCE with the catalyst. LSV and CV was performed in 0.5 M H2SO4 solution at a scan rate of 10 mV/s. A home-made PEM water electrolysis cell with a proton exchange membrane was used to evaluate the performance of SI series catalyst. The preparation step of catalyst inks is the same as above method. The total catalyst loading on the electrode was 1.0 mg and all the catalyst inks were deposited on carbon paper (1 cm × 1 cm). The cell temperature (25, 65, and 85 °C) was maintained by an electric heating plate and measured by a temperature probe in electrolyte.

Computational methods

Spin-polarized density functional theory (DFT) calculations were performed in the plane wave and ultrasoft pseudopotential (USPP) with Perdew-Burke-Ernzerhof (PBE) exchange functional correction as implemented in Quantum ESPRESSO41,42. An energy cutoff of 25 Ry was employed for the plane wave expansion of the electronic wavefunction. The atomic structures of the models were fully relaxed until self-consistency was achieved with a convergence criteria of 10−6 Ry for the energy and 10−3 Ry/Bohr for the atomic coordinates. To prevent interaction between layers, a vacuum slab of 12 Å was used to isolate the surface. For bulk geometry optimization, a 3 × 3 × 1 Monkhorst-Pack k-point set was used, while a 5 × 5 × 1 set was used for electronic structure calculations. The correction for every adsorbate and surface, with typical values of +0.35 eV, +0.05 eV, +0.35 eV for *OH, *O and *OOH respectively. To simulate the complex unsaturated IrOx structure, we start by constructing and optimizing the slab structure of the original SrIrO3. Then, the surface Sr atoms were removed from the SrIrO3 slab for further optimization. According to a previous report25, the Ir-O coordination number on the SrIrO3 surface is approximately 4.5, achieved by removing 4 Sr atoms and their corresponding 3 neighboring O atoms. The same procedure can be applied to the Co-doped system, where removing a Co atom also removes 2 neighboring O atoms.

XAS analysis

The acquired extended X-ray absorption fine structure (EXAFS) data were processed following standard procedures using the ATHENA module of the Demeter software package43. The EXAFS spectra were obtained by subtracting the post-edge background from the overall absorption and then normalized with respect to the edge-jump step. The χ(k) data were Fourier transformed to real (R) space using a Hanning window (dk = 1.0 Å−1) to separate the contributions from different coordination shells. To determine the quantitative structural parameters around the central atoms, least-squares curve parameter fitting was performed using the ARTEMIS module of the Demeter software package.