1 Introduction

In the recent decade, electrochemical water splitting has been recognized as a practical strategy for large-scale hydrogen production in a green and sustainable way by using renewable energy, such as solar, wind, and hydropower [1,2,3,4,5]. The proton exchange membrane water electrolysis (PEMWE) exhibits high current densities, high gas purity, and high durability and thus has an extensive application prospect [6,7,8,9]. However, due to the corrosion condition of the acidic electrolyte, the anode oxygen evolution reaction (OER) involves the processes of four electron metastasis and the adsorption/undertaking process of the intermediate, which results in high potential and slow dynamics [10,11,12]. The OER catalysts must have high electrochemical stability because of the anode-rich and high potential environment during the hydrolysis process. At present, only precious metals, such as Ir and Ru, can maintain good OER activity and stability under acidic and high potential conditions [13,14,15,16]. Specifically, Ir shows better stability, so it is considered the most ideal OER catalyst. However, the noble Ir is expensive, and the utilization and stability of Ir-based catalysts have become the focus of research [17,18,19,20]. To date, several strategies—(i) exposing more active sites by structural engineering [21,22,23] and (ii) tailoring the electronic configuration by the element doping—are developed to improve the OER performance of metallic Ir electrocatalysts [15, 17, 24]. The OER catalytic activity would be substantially improved by combining these two strategies. Accordingly, Ir-based materials with various configurations, such as single-atoms [25,26,27], nanoribibbons [16, 28, 29], and metallene [12, 18, 30], have displayed excellent catalytic properties. Especially, the core–shell structure has attracted extensive attention with large amount of active sites depending on the electronic structure. The different physicochemical characteristics of the core–shell composition regulate the electronic structure of the atoms at heterogeneous interfaces, thereby changing the binding energy of the reaction intermediates [31,32,33,34]. In addition, the element dropping between the inner and interface enhances the stability and OER performance. Hence, the design of the core–shell structure catalyst, which is the candidate material for metallic Ir or IrO2 catalytic in PEMWE, faces challenges. Au is widely used to modify the electronic configuration of other active elements, such as Ir and Pt [35, 36]. Just recently, Huiming Wang [37] and coauthors reported a novel core–shell structure with Au core and AuIr2 alloy shell (Au@AuIr2). Thus, Ir with Au is sought to improve the OER catalysis performance. Therefore, the design of advanced AuIr configuration can modify the electric structures, thereby realizing the regulation of the reaction intermediates and improving the acidic OER catalytic activity and stability.

In the present work, we investigate a novel core–shell structure catalysis with rich Ir core and AuIr alloy shell (Ir@AuIr). The core–shell structure exposed numerous active sites and improved the utilization efficiency of electrocatalyst. The strong electronic effect between two different species in the alloy electrocatalysts could effectively improve catalytic activity and stability owing to the tailored electronic configuration of active sites. The X-ray photoelectron spectroscopy (XPS) test showed that the electrons transformed from Ir to Au, leading to the upshifting of the d-band center of Ir. In turn, a mild adsorption of intermediate products on the catalyst surface and the reaction energy barrier decreation occur. The Irx@Au0.25Ir0.75−x catalyst exhibited merely 235 mV at the current density of 10 mA cm−2, which was 55 mV and 73 mV lower than those of Ir and commercial Ir black, respectively. Practically, the Irx@Au0.25Ir0.75−x catalysis exhibits 4.6 (5.6) times larger intrinsic (mass) activity than the commercial Ir black catalyst during the OER reaction. Moreover, Irx@Au0.25Ir0.75−x catalysis demonstrated excellent stability. The optimized Irx@Au0.25Ir0.75−x catalyst exhibited outstanding performance in PEMWE devices and confirmed the application prospect.

2 Experiment section

2.1 Materials and chemicals

The chloroiridic acid (H2IrCl6, AR) and sodium tetrachloroaurate (Na2AuCl6, AR) were purchased from Kunming Boren Precious Metals Co. Ltd. Disodium pyrophosphate (Na2H2P2O7, AR), commercial iridium black, and Pt/C were obtained from Johnson Matthey. Ethylene glycol (CH2OH)2 was provided by Aladdin Technology. A total of 5 wt% Nafion solution was acquired from the Chemours Company.

2.2 Synthesis of Ir@AuIr

First, 0.5606 mg Na2H2P2O7 was dissolved in 15 mL H2O and ethylene glycol under stirring and heating at 130 °C. Second, a certain proportion of H2IrCl6 and Na2AuCl6 solution (1:3, 1:1 and 3:1) was added into the mixed solution. The mixture was further stirred at 130 °C for 6 h. Third, the product was collected by centrifugation and washed with ethyl alcohol and water three times. Finally, the resulting AuIr sample was dried by a freeze drier, and denoted as Irx@Au0.75Ir0.25−x, Irx@Au0.5Ir0.5−x, Irx@Au0.25Ir0.75−x.

2.3 Catalyst characterization

X-ray diffraction (EMPYREAN, PANalytical) was performed to characterize the crystal structure of the catalysts. X-ray photoelectron spectroscopy (XPS, ESCALab250Xi, Thermo Scientific) was characterized by the electronic structure and their changes by the binding energy shift. The distribution and morphology of elements were observed using transmission electron microscopy (TEM, JSM-2100 F, JOEL). The in situ Raman (HNG-ISRaman) illustrated the reaction mechanism. Elemental content analysis was determined by inductively coupled plasma optical emission spectrometry (NexION 300, PerkinElmer).

2.4 Electrochemical measurements

The electrocatalyst (5 mg) was well dispersed in 1 mL solution (1 mL = 950 µL of isopropyl + 50 µL of Nafion) and sonicated to form “ink.” The glassy carbon electrode was coated with 10 µL “ink,” and Hg/HgCl (SCE) and Pt plate electrode were used as the reference and counter electrodes, respectively. Prior to the operation of the electrochemical activity tests, the Hg/HgCl reference electrode was calibrated by measuring the open circuit voltage in a hydrogen-saturated 0.5 M H2SO4 electrolyte that uses a Pt mesh as the working electrode (i.e., the homemade reversible hydrogen electrode (RHE)). The electrochemical tests were performed in 0.5 M H2SO4 electrolyte at 26.8 °C. The overpotential, Tafel slope, and mass-specific activity were determined by linear sweep voltammetry (LSV) at the scan rate of 5 mV s−1. The electrochemical surface area was performed using cyclic voltammetry (CV) test with the potential of the double-layer charging region, between 0.55 and 0.60 V vs. RHE, at different scan rates (40, 60, 80, 100, and 200 mV s−1). The electrochemical impedance spectroscopy was used to evaluate the impedance under oxygen evolution conditions at 1.26 V vs. RHE with the frequency range of 10 kHz to 1 Hz.

3 Results and discussion

The Ir@AuIr electrocatalysts were synthesized through ethylene glycol reduction approach. Typically, the aqueous solutions of all precursors, including metal precursors (Na2AuCl6 and H2IrCl6) and reducing agent (H2O and ethylene glycol) and surfactant (Na2H2P2O7), were mixed together at room temperature to obtain a dark brown solution as the reaction solvent by oil bath heating path. Methodologically, Na2H2P2O7 was selected as a surfactant to prevent the apparent agglomeration of metal nanocrystals.

Fig. 1
figure 1

a XRD patterns of various electrocatalysts. TEM of b Ir NPs, c Irx@Au0.75Ir0.25−x, d Irx@Au0.5Ir0.5−x and e Irx@Au0.25Ir0.75−x. f and related EDS mappings of Irx@Au0.25Ir0.75−x.

Full size image

The room-temperature powder X-Ray diffraction patterns of all prepared AuIr catalysts are shown in Fig. 1a. The peak location of the obtained AuIr component was consistent with the bulk Au (PDF#04-0784) that had face-centered-cubic (fcc) structure [38, 39]. Additionally, the diffraction peaks of Ir@AuIr catalysts were positively shifted compared with those of pure Au, indicating the strain effect formed in the AuIr structure. No diffraction peaks were attributed to Ir atom because of the small particle size. Figure 1b shows the low-magnification dark-field TEM image of Ir nanoparticles (Ir NPs). The outer-shell Ir NPs were homogeneously dispersed with an average diameter of 2 nm, which is in line with the results measured by XRD spectra. As revealed in Fig. 1c, the separated crystalline phases present two phases in an alloy that contains the superficial Ir and fcc AuIr alloy of Irx@Au0.75Ir0.25−x. The core–shell structure with lattice spacing of 0.22 nm, representing AuIr (111) phase [40], was observed from the HR-TEM image shown in Fig. 1e. In the structure, the ultrasmall Ir nanoparticles (2 nm) without aggregation were uniformly dispersed on the surface of AuIr alloy. The relevant fast Fourier transform pattern also exhibits the characteristics of fcc structure viewed from the direction (inset in Fig. 1f). Also, these two phases can be demonstrated by the contrast difference in their high-angle annular diffraction field STEM (HAADF-STEM) image (Fig. 1d), whereas the elemental mapping could further reveal the different compositions of the two-phase separation of the produced Au–Ir alloy. The elemental mapping of Irx@Au0.75Ir0.25−x and Irx@Au0.5Ir0.5−x is shown in Fig. S1.

Fig. 2
figure 2

a Ir 4f and b Au 4f XPS spectra of Ir NPs, Au NPs, Irx@Au0.75Ir0.25−x, Irx@Au0.5Ir0.5−x and Irx@Au0.25Ir0.75−x, respectively

Full size image

XPS test was carried out to determine the valence states of Ir@AuIr catalysts, as shown in Fig. 2. The XPS survey spectra showed the coexistence of Ir and Au in the Fig. S2. The fitting of the high-resolution XPS spectra of Ir 4f reveals up to four different Ir species of the Irx@Au0.25Ir0.75−x catalyst. Peaks at 60.9 eV and 64.0 eV were identified as the metallic Ir species, which are upshifted compared with the Irx@Au0.5Ir0.5−x and Irx@Au0.75Ir0.25−x, indicating that the confined Ir species had a higher oxidation state possibly ascribed to the formation of an Au–Ir coordination structure [41]. Irx@Au0.5Ir0.5−x and Irx@Au0.75Ir0.25−x were compared and found that the Ir/Ir4+ atomic ratio of Irx@Au0.25Ir0.75−x decreased from 4:1 to 0.6:1, indicating an increase in the oxidation state of Ir. Such finding suggests that a redox reaction should occur between Ir and Au alloy. The peaks at 61.2 eV and 64.3 eV are ascribed to Ir4+ and related satellites, respectively [42]. Moreover, for the Au 4f XPS spectra of Irx@Au0.25Ir0.75−x, the fitted peaks at 84.1 eV/87.8 eV could be attributed to metallic Au (Fig. 2b) [43]. According to the contact angle measurement, the Irx@Au0.25Ir0.75−x catalysts showed good hydrophilicity (Fig. S3).

Fig. 3
figure 3

Electrochemical testing of Irx@Au0.75Ir0.25−x, Irx@Au0.5Ir0.5−x, Irx@Au0.25Ir0.75−x, Ir NPs, and commercial Ir black in N2 saturated 0.5 M H2SO4 (aq) at 25 °C. a CV curves a scan rate of 10 mV s−1. b OER polarization curves obtained under a scan rate of 5 mV s−1. c–e overpotential, Tafel plot, and mass activity obtained from (b). f The electrochemical surface area obtained from the CV curves between 0.55 and 0.60 V

Full size image

OER catalytic activity and stability were tested using 0.5 M H2SO4 electrolyte, and commercial Ir black was also tested for comparison. As shown in Fig. 3a, the CV curves show the typical redox peaks of Ir with different valence states. The redox peaks in the range of 0.55–1.1 VRHE correspond to the valance transition process that is higher than that of Ir3+/4+, and the redox peaks in the range of 1.1–1.42 VRHE correspond to the valance transition process higher than that of Ir4+. From iR-corrected corresponding LSV curves in Fig. 3b, the overpotential of the Irx@Au0.25Ir0.75−x catalyst at a current density of 10 mA cm−2 is only 235 mV, indicating that it has higher OER catalytic activity than the Ir. The reason is that the core–shell structure can expose more active sites, thereby improving the utilization of the catalyst. The electronic interaction between the Au atom and the Ir atom at the interface can also reduce the d-band electron center of the Ir atom, thereby achieving mild adsorption of the intermediate product on the catalyst surface, reducing the reaction energy barrier, and improving the intrinsic catalytic activity of the catalyst. The Au catalyst has almost no OER catalytic activity, indicating that the excellent catalytic activity of AuIr is derived from the Ir layer. The overpotentials of Irx@Au0.25Ir0.75−x and Irx@Au0.5Ir0.5−x catalysts at a current density of 10 mA cm−2 were 390 mV and 313 mV, respectively. Such values indicate that the prepared catalysts considerably improved the utilization of Ir. The loading amount was also positively correlated with the catalytic activity within a certain range. The Tafel slope of Irx@Au0.25Ir0.75−x is 56.1 mV dec−1, which is much lower than that of commercial Ir black (68.3 mV dec−1), self-made Ir (60.2 mV dec−1), Irx@Au0.75Ir0.25−x (98.3 mV dec−1), and Irx@Au0.5Ir0.5−x (76.4 mV dec−1), indicating that it has the fastest catalytic reaction kinetics. Compared with the self-made Ir, the core–shell structure of Irx@Au0.25Ir0.75−x improves the utilization of the catalyst and the interaction of interface electrons, and the Irx@Au0.25Ir0.75−x catalyst exhibits remarkable OER kinetic performance. The catalytic activity of OER was normalized by mass normalization. The diagram illustrates that the activity of Irx@Au0.25Ir0.75−x catalyst is as high as 248 mA mgIr−1 at an overpotential of 370 mV, which is nearly three times that of commercial Ir black, indicating that the core–shell structure can provide more active sites and reduce the use of Ir. CV scanning was performed in the non-faradaic region, and the electric double-layer capacitance of the catalyst was calculated by fitting Fig. S4. The results showed that the active site of Irx@Au0.25Ir0.75−x was 1.14 mF cm−2, which was 4 times that of commercial Ir black, and 8.76, 4.75, and 1.32 times those of Irx@Au0.75Ir0.25−x, Irx@Au0.5Ir0.5−x and self-made Ir, respectively. Its high intrinsic catalytic activity was mainly due to the strong electronic structure between Au and Ir atoms, which changed the electronic structure of Ir atoms and affected its OER intrinsic catalytic activity.

Fig. 4
figure 4

a Chronopotentiometry curve of f Irx@Au0.25Ir0.75−x catalyst at a sequential current density of 10, 30, 50, and 100 mA cm-2. b Ir, c Au and d O XPS spectra before and after the stability operation

Full size image

The stability test of the electrode was conducted by multi-step chronopotentiometry at current densities of 10, 30, 50, and 100 mA cm−2, respectively. Fig. 4a shows the change in the electrode potential with time. The stability of Irx@Au0.25Ir0.75−x catalyst is low, which can only maintain the basic activity at 10 mA cm−2, and the electrode potential increases from 1.8 to 1.9 VRHE. When the test current density is increased to 10 mA cm−2, the electrode potential rises rapidly to more than 2.0 VRHE, and the electrode decays seriously. However, the electrode potential of Irx@Au0.75Ir0.25−x catalyst did not increase within 50 h, indicating that the catalyst demonstrated good stability.

The valence state of the catalysts before and after the stability operation was analyzed by XPS, and the results are shown in Fig. 4b–d. At 84.4 eV and 87.7 eV, the valence state corresponds to 4f7/2 and 4f5/2 of Au0, indicating that Au exists at 0 valence before and after stability [44]. Compared with the oxidation reaction of Ir before stability, this phenomenon is mainly due to the oxidation of Ir atoms on the surface of AuIr alloy. The oxidized Ir is mainly composed of the satellite peaks of Ir4+ binding energy at 61.2 eV (64.3 eV) and Ir3+ binding energy at 62.2 eV (65.2 eV) [45]. The formation of Ir3+ is mainly derived from two processes, one is the direct oxidation of metallic Ir to Ir3+; and the second is the conversion of IrO2 to Ir3+. The decomposition of Ir4+O2OH to release O2 and intermediate HIr3+O2. Ir3+ with electronic defects and low coordination number are beneficial in improving the catalytic activity of the alloy catalyst [10]. Metallic Ir occurs at 60.8 eV (Ir4f7/2) and 64.0 eV (Ir4f7/2), which is mainly considered to be more stable in AuIr alloy. However, the binding energy position of metallic Ir is positively shifted by 0.1–0.2 eV due to the electron transfer caused by the difference in the electronegativity between Au and Ir atoms. After OER, the XPS spectrum shows that the Ir species with high binding energy can still be retained, indicating the excellent stability of the crystal structure of the AuIr catalyst. The XPS spectrum of O 1s is shown in Fig. 4d. The O 1s signal can be fitted to two peaks, namely, 630.1 eV and 631.6 eV, which corresponds to lattice oxygen and hydroxyl oxygen species, respectively [46, 47]. The proportion of hydroxyl oxygen species increased greatly after the stability test. Therefore, we speculate that the Ir3+ that formed on the surface after achieving stability may be the hydroxylated Ir (Ir–OH) produced near the defect.

Fig. 5
figure 5

In situ Raman spectra of Irx@Au0.25Ir0.75−x at different potentials in 0.5 M H2SO4 solution

Full size image

In situ electrochemical Raman spectroscopy was further performed to explain the reconstruction of Ir toward OER (Fig. 5). The stretching vibration of Ir–O (Ir3+–O) appears at 460 cm−1, and the negative shift of the stretching vibration peak of Ir–O appears near 653 cm−1, which is mainly attributed to the oxidation of lattice oxygen at the oxidation potential [48]. As the voltage increases from 1.0 to 1.7 V, the stretching vibration peak of Ir–O at 660 cm−1 is weakened, and the vibration peak of Ir–O bond appears at 620 cm−1. At 850–950 cm−1, with the increase in voltage, an asymmetric and broadened peak corresponds to the O–O vibration in Ir–OOH, which is attributed to the non-planar vibration of Ir–O [49]. The stretching vibration peak of Ir4+–O appeared near 630 cm−1, and the stretching vibration peak (γ peak ) attributed to –OH in Ir4+–OH appeared at 450 cm−1. When the applied voltage was 1.3–1.8 V, the position of the γ peak showed a red shift, indicating that the surface deprotonation of the AuIr catalyst occurred before the OER reaction. The γ peak position returns to the initial state after the OER reaction, indicating that the OER process starts with surface deprotonation and ends with surface hydroxyl reconstruction.

Fig. 6
figure 6

Performance of PEM electrolyzer using Irx@Au0.25Ir0.75−x and Ir black as anode electrocatalyst. a Polarization curve, b CV curves and c EIS Nyquist plots of the PEM electrolyzer obtained at 80 °C with Nafion 115 membrane. d Chronopotentiometric curves of the PEM electrolyzer using Irx@Au0.25Ir0.75−x at the current density of 0.5 A cm−2. e CV curves and f EIS Nyquist plots obtained before and after 100 h at a current density of 0.5 A cm−2

Full size image

To evaluate the performance of the electrolytic cell of the composite catalyst, the Irx@Au0.25Ir0.75−x catalyst with the highest mass-specific activity during the battery test was selected as the anode catalytic layer to prepare the membrane electrode and assemble the electrolytic cell for performance test. The membrane electrode prepared by commercial Ir black was used as a comparison, and the test temperature was 80 °C. The polarization curve of the electrolytic cell is shown in the Fig. 6. At the current density of 1.0 A cm−2, the electrolysis voltage is 1.720 V, which is better than the performance of commercial Ir black film electrode (1.0 A cm−2@1.775 V). Ir mass normalization was performed to improve the electrode performance. The mass-specific activity of Irx@Au0.25Ir0.75−x electrode reached 9.44 A mgIr−1@2.0 V, which was better than the mass-specific activity of commercial Ir black (5.6 A mgIr−1), indicating that Irx@Au0.25Ir0.75−x catalyst showed great potential in improving the mass-specific activity of precious metals. Irx@Au0.25Ir0.75−x and commercial Ir black showed similar redox current curves, indicating that both showed good OER catalytic activity and electrochemical activation polarization. AC impedance test in 1.45, 1.5, and 1.6 V electrolytic cell running state is shown in the figures. The high-frequency impedance is usually related to the sum of the electronic conductivity of the catalytic layer, the proton conductivity of the membrane, and the contact resistance among the electrolytic cell components. The ohmic impedances of the two groups of electrodes are 0.166 Ω cm2 and 0.192 Ω cm2, respectively, indicating that the Irx@ Au0.25Ir0.75−x catalytic layer exhibits faster electronic conductivity. On the one hand, it is due to the good conductivity of the Au nanoparticles in the catalytic layer. On the other hand, defects in the alloy catalyst can also lead to enhanced conductivity. The charge transfer resistance (Rct) reflects the electrochemical reaction activity of the electrode. At 1.5 V electrolysis voltage, the Rct of the two groups of electrodes is 0.195 Ω cm2 and 0.325 Ω cm2. The Irx@Au0.25Ir0.75−x electrode has a small and low electrochemical transfer resistance, indicating that its membrane electrode has the highest OER catalytic activity. The membrane electrode was used for investigating the stability of the catalyst. The results are shown in the figure. At the current density of 0.5 A cm−2, the electrolysis voltage of the electrolytic cell only increased by 40 mV after 100 h, indicating that the membrane electrode exhibited good stability. Further, combined with other electrochemical characterization methods, the performance changes in the electrolytic cell after stable operation were analyzed, that is, the CV and AC impedance spectroscopies of the electrolytic cell were tested. After 100 h of stable operation, the oxidation peak currents that correspond to the electrode potential of 1.0 V increased from 4.5 to 5.2 mA cm−2. The AC impedance test showed that the charge transfer resistance in the Nyquist diagram was 0.206 Ω cm2 (increased by 5.6%), showing good stability.

In the lifetime operation process, factors, such as gas and liquid fluid erosion, electrochemical oxidation, and other factors cause changes in performance. Ir catalysts in the alloy structure of Irx@Au0.25Ir0.75−x electrode will undergo loss, agglomeration, intra-membrane migration, and structural reconstruction. On the one hand, loss and agglomeration reduce the number of active sites of the catalyst. On the other hand, the reconstruction of the catalyst often brings the synergistic effects of the electronic structure, thereby improving its intrinsic activity. The heterogeneous structure of the alloy catalyst remains stable during the stable operation. The dissolution and reconstruction of the Ir catalyst only have a small effect on the performance degradation of the electrolytic cell. The agglomeration of cathode Pt catalyst and the decrease in proton conductivity caused by the accumulation of metal ions in Nafion membrane are also the dominant factors that degrade the performance of electrolytic cell.

4 Conclusion

In summary, we constructed Ir@AuIr catalyst with Ir-rich core and AuIr alloy shell H2O-ethylene glycol liquid phase reduction progress, which can be used as a highly efficient and stable anodic OER catalyst for PEMWE. In the Irx@Au0.25Ir0.75−x catalyst, the Ir core can expose more active sites, and the strong electronic interplay tailored the electronic configuration of Ir active center, which exhibits excellent electrocatalytic properties, such as specific activity, mass activity, and stability. The Ir@AuIr electrocatalyst demanded merely 235 mV overpotential at 10 mA cm−2, which was 73 mV lower than commercial Ir black and 2.61-fold higher mass activity for Irx@Au0.25Ir0.75−x than commercial Ir black. Additionally, the electrolytic voltage has almost no increase after 50 h multi-potential test for Ir@AuIr electrocatalyst, indicating excellent stability. From the electronic regulation effect, the Irx@Au0.25Ir0.75−x electrocatalyst exhibited higher effectiveness and stability within PEMWE than commercial Ir black. The catalyst prepared in this work cannot only be employed for OER but also provided feasibility for electrocatalytic reactions.