Mesoporous Co x Sn(1-x)O2 as an efficient oxygen evolution catalyst support for SPE water electrolyzer.

SPE water electrolysis is a promising method of hydrogen production owing to its multiple strengths, including its high efficiency, high product purity and excellent adaptability. However, the overpotential of the oxygen evolution reaction process and consumption of Ir during charging in SPE water electrolysis will inevitably result in large energy loss and then high cost. Under these circumstances, we propose a novel 40IrO2/Co x Sn(1-x)O2 (x = 0.1, 0.2, 0.3) anode catalyst, where the Co x Sn(1-x)O2 support is synthesized by a hydrothermal method and IrO2 is synthesized by a modified Adams fusion method. After modifying the component of Co x Sn(1-x)O2, the 40IrO2/Co x Sn(1-x)O2 exhibits an increased specific surface area, electrical conductivity and surface active sites. Moreover, a single cell is fabricated by Pt/C as cathode catalyst, 40IrO2/Co x Sn(1-x)O2 as anode catalyst and Nafion 117 membrane as electrolyte. The 40IrO2/Co0.2Sn0.8O2 exhibits the lowest overpotential (1.748 V at 1000 mA cm-2), and only 0.18 mV h-1 of voltage increased for 100 h durability test at 1000 mA cm-2. Consequently, Co x Sn(1-x)O2 is a promising anode electrocatalyst support for an SPE water electrolyzer.

Introduction

Hydrogen is regarded as one of the promising solutions for developing clean energy and solving the thorny environmental problems present on the Earth [1]. Water splitting for hydrogen generation is a major component of modern clean energy technologies [2], such as water-alkali electrolyzers [3], solid polymer electrolyte (SPE) water electrolyzers [4] and photocatalytic/photo-electrochemical water splitting [5-8]. SPE water electrolyzers offer us an effective and simple method to produce hydrogen through reusing the surplus electric power generated by renewable energy (such as wind and photovoltaic power) [9]. As a result, SPE water electrolyzers have gained a lot of research attention [10-12]. Nevertheless, the conspicuous weakness of SPE water electrolysis should not be ignored, including the majority increases of the activation overpotential in the oxygen evolution reaction (OER) process [13]. Although some highly active and stable noble metal-based catalysts have been developed, such as RuO2 and IrO2 for OER, these materials are still far from large-scale application because of their high cost and scarcity. Therefore, multiple studies have been devoted to develop novel, highly efficient and low-cost catalysts (e.g. La2NiMnO6 [14], Ni-Fe-layered double hydroxide [15] and Ternary Ni–Co–Fe blue analogue [16]). However, these new catalysts too have drawbacks, such as a complicated preparation process and poor durability. Additionally, many research efforts have been made to reduce the amount of these noble metals, such as designing their bulk structural parameters (i.e. grain size, morphology, and dimensions) [17-19], tailoring composition (such as introducing foreign elements or oxides into the structure of noble metal catalysts) [20], and adopting supports (i.e. carbon-based [21], SnO2 [22], TiO2 [10], TiC [23]).

It is well known that an appropriate support will favour a noble metal-based material to achieve a better dispersion and greater surface area, which not only reduces the usage of noble metal but also maintains high activity for the catalysts [23]. As supports, carbon-based materials have recently received a lot of attention due to their high specific surface area and excellent electric conductivity. Unfortunately, carbon-based materials are easily electrochemically oxidized at potential above 0.206 V versus SHE [24], leading ultimately to the unsustainability of carbon-based supports in SPE water electrolyzer [25]. Therefore, the development of a highly stable and active support for the OER is still a research focus. Tin oxide (SnO2), as a corrosion-resistant support, has been reported to promote the dispersion of noble metal-based materials and provide more surface active sites of catalysts [26]. Wang et al. reported that SnO2 in the IrO2/SnO2 catalyst could act as a Brønsted base and accept protons from the IrO2 sites, which is favourable to the enhancement of catalytic performance of IrO2/SnO2 catalyst [22]. Unfortunately, the current reported OER activity of SnO2-based supports is still not significant because of the poor electronic conductivity of SnO2. Heteroatom doping is a method of enhancing the catalytic activities of catalysts due to the charge delocalization mechanism [27]. Co is a transition element with incomplete electron shell and possesses interesting catalytic properties [28-30]. As far as we know, the effect on the electrocatalytic activity of SnO2 with varying Co content doped as support for IrO2 has been rarely studied.

In this study, IrO2 supported by mesoporous CoSn(1−O2 (x = 0.1, 0.2, 0.3) as an anode catalyst for SPE water electrolysis has been exploited. CoSn(1−O2 (x = 0.1, 0.2, 0.3) is successfully synthesized by a hydrothermal method and IrO2 is synthesized by using a modified Adams fusion method [4,31]. The structure, morphology, specific surface area, electrical conductivity and surface active sites of the prepared samples have been thoroughly determined by various characterization techniques. The electrocatalytic activity of the prepared samples as anode catalysts in single cells is also tested. The prepared samples with Co doping show low overpotential (1.748 V at 1000 mA cm−2) and excellent stability.

Experimental section

Preparation of Co-doped SnO2 support

The supports of CoSn(1−O2 (x = 0, 0.1, 0.2, 0.3) were synthesized by the hydrothermal method and then processed by heat treatment. One gram of hexadecyl trimethylammonium bromide (Sinopharm, Shanghai, China) was dissolved in a mixture of 20 ml of ethanol (Sinopharm) and 20 ml distilled water. Then, significant amounts of tin chloride (Sinopharm) and cobalt acetate (Sinopharm) (molar ratio (Sn2+ + Co2+): CTAB = 1 : 1.05) were added to the above solution, and continuously stirred to get a homogeneous solution. A certain amount of ammonia water (Sinopharm) was added dropwise to the solution under vigorous stirring at room temperature. The Ph value of the solution was maintained at 9 and under stirring for 2 h. After it was stirred, the mixture was transferred to a stainless Teflon-lined 100 ml autoclave and kept at 180°C for 24 h in an oven. The resulting yellow precipitate was collected by centrifugation, washed with distilled water and ethanol several times to remove the impurities, and then dried in a vacuum oven at 60°C. The dried samples were processed under 350°C for 5 h in a muffle furnace. After cooling to room temperature, CoSn(1−O2 nanoparticles were finally obtained.

Preparation of supported catalysts

In this work, the IrO2 loading was applied as 40 wt %, 0.196 g chloroiridic acid (Hesen, Shanghai, China), 5 g of ultrafine sodium nitrate (Sinopharm) and 0.12 g as-prepared support powder were added into 10 ml of isopropyl alcohol (Sinopharm), which was stirred to obtain uniform suspension. Then the suspension was ground for 6 h by planetary ball milling. The obtained slurry was dried in a vacuum oven at 60°C, and treated at 400°C for 1 h in a muffle furnace with a heating rate of 5°C min−1. After heat treatment, the powder was washed in turn with 0.1 M HCl, distilled water, and ethanol to eliminate residual impurities, and dried in a vacuum oven at 70°C overnight. The resulting material was denoted as 40IrO2/CoSn1−O2 (x = 0.1, 0.2, 0.3). The preparatory method of unsupported IrO2 was similar to that of the 40IrO2/CoSn1−O2 samples.

Physical characterization

X-ray diffraction (XRD) was performed to obtain the crystal structure and phase purity of all the prepared samples using a Bruker D8 Advance X-ray diffractometer (Bruker, Karlsruhe, Germany) with a Cu-Ka radiation source (λ = 0.154056 nm). Transmission electron microscopy (TEM) and high-resolution TEM (HR-TEM) images were carried on a JEOL 2010F microscope (JEOL, Tokyo, Japan). The specific surface area and pore size distribution of the as-prepared samples were recorded with the measurement of nitrogen adsorption isotherm at 77 K using a Micromeritics ASAP 2020 analyzer (Micromeritics, Norcross, Georgia, USA). Electrical conductivity measurements were carried out on cylindrical pellets compressed from the powder samples at 30 MPa between two copper electrodes. The substrate area was restricted to 1 cm2 while the thickness of the pellet was measured by a Vernier caliper. The value of resistivity was immediately measured by a JG-ST2258A resistivity tester (Jingge Electronic, Suzhou, Jiangsu, China) by inputting the thickness-area ratio as a parameter, followed by conversion to conductivity.

Electrochemical characterization

The half-cell electrochemical evaluation of different samples was investigated by a three-electrode measurement in the N2-saturated 0.5 M H2SO4 electrolyte. A reversible hydrogen electrode (RHE) and a platinum wire acted as the reference and counter electrode, respectively. The working electrode was prepared by a catalyst layer coating on the glassy carbon disc (GCE, 5.6 mm in diameter). Briefly, the catalyst layer was fabricated as follows: 12.46 mg of 40IrO2/CoSn1−O2 powder was dispersed in 2 ml methanol/Nafion (50 : 1, wt.%) mixed solution and uniform ultrasound solution was obtained by ultrasound. The loadings of all samples on glassy carbon were controlled at 0.1 mg cm−2. Cyclic voltammetric (CV) measurements were performed on a CHI 760E Electrochemical Workstation at a scanning rate of 50 mV s−1 between 0.05 and 1.35 VRHE. The potential range of the linear sweeping voltammetry (LSV) curve was from 1.4 to 1.65 V versus RHE at a scan rate of 50 mV s−1, and the rotation rate of the working electrode was 1600 rpm.

The membrane electrode assemblies (MEAs) were prepared by the spray method and assembled using a Nafion117 membrane (DuPont, Wilmington, Delaware, USA) adopted as an SPE, the prepared samples were used as anodic electrocatalysts and a commercial 60 wt.% Pt/C (Johnson Matthey, London, UK) catalyst acted as cathodic electrocatalyst. Prior to the assembly, the Nafion117 membrane was cleaned by H2O2 solution, distilled water, and H2SO4 solution at 80°C for 1 h for each step. The sprayed catalyst inks were fabricated by the mixture of the obtained samples, Nafion solution, isopropyl alcohol and deionized water, and sonicated for 2 h to get a homogeneous suspension. The Nafion loading on each side of the membrane was maintained at 25 wt.%. The noble metal Ir and Pt loading on the membrane was 2.5 mg cm−2 for the anode and 0.5 mg cm−2 for the cathode, respectively. Finally, the MEAs (with an effective area of 3.645 cm2) were assembled into a home-made single cell water electrolyzer, as shown in the schematic diagram in scheme 1. Ti mesh and plates were made as current collector and bipolar plate for the anode side, respectively. Carbon paper and Ti plates were used as current collector and bipolar plate for the cathode, respectively. Deionized water with a preheated 80°C temperature was pumped by a peristaltic pump to the anode compartment. At atmospheric pressure, the polarization curves of the single cells were measured by a Motech LPS305 programmatic DC power supply. The electrochemical impedance spectrums (EIS) for the single cells were conducted at 0.3 A cm2 in the frequency range of 0.1 Hz to 10 kHz (amplitude = 80 mV) and recorded with a Solartron Analytical 1260 impedance analyzer coupled to a Solartron Analytical 1287 potentiostat.

Scheme 1.

Schematic of the SPE water electrolyzer structure.

Results and discussion

The XRD patterns of the prepared samples are shown in figure 1. Figure 1a shows that all the diffraction peaks of SnO2 are matched well with the characteristic peaks of tetragonal rutile structure (JCPDS 41–1445) [32]. The diffraction peaks located at approximately 26.6°, 33.9°, 37.9°, 51.8°, 54.7°, 57.8°,62.6°, 65.9°, 71.3° and 78.7° represent the (110), (101), (200), (211), (220), (002), (221), (301), (202) and (321) planes, respectively. It is noted that the diffraction peaks of SnO2 doped with varying Co content show similar shapes to the pure SnO2, and no second phase about Co is detected, which confirms that Co ions were successfully doped into SnO2 [33]. Figure 1b displays the XRD patterns of the unsupported IrO2 and 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3). The typical peaks of the unsupported IrO2 could be matched well with the tetragonal rutile structure [31]. After IrO2 supported on CoSn1−O2, the 40IrO2/CoSn1−O2 exhibit similar shapes with CoSn1−O2, implying that the loaded IrO2 will not affect the crystal structure of CoSn1−O2 by our presented modified Adams fusion treatment. The lattice parameters (a = b, c) of SnO2 and CoSn1−O2 (as listed in table 1) decrease with the increasing Co-doping concentration. This is because the radius of Co2+ (0.072 nm) is smaller than that of Sn4+ (0.083 nm) at a coordination number of 6 [34]. Furthermore, the grain sizes of all the obtained samples were calculated using the by Debye­–Scherrer equation, L = Kλ/(βcosθ), where K is the constant (0.89), λ is the wavelength of the X-ray radiation (Cu Kα = 0.15406 nm), β is the line width at half maximum height and θ is the diffracting angle. The calculated grain sizes of all the prepared samples are listed in table 1. The grain sizes of CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) gradually decrease from 8.28 nm to 5.52 nm with the Co content increasing, which may be attributed to the fact that the highly Co doped induced a segregation at the grain boundaries, which leads to a decrease in the grain size [35].

Figure 1.

XRD patterns of the (a) CoSn1−O2; (b) 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) and unsupported IrO2.

Table 1.

The lattice constant, grain sizes, BET surface area and BJH adsorption average pore diameter results of CoSn1−O2 (x = 0, 0.1, 0.2, 0.3).

	lattice constant
samples	a = b (Å)	c (Å)	grain size (D) nm	BET surface area (m2 g−1)	BJH adsorption average pore diameter (nm)
SnO2	4.7367	3.1908	8.28	72.19	7.88
Co0.1Sn0.9O2	4.7194	3.1849	7.08	104.01	8.48
Co0.2Sn0.8O2	4.7084	3.14430	6.54	131.92	10.59
Co0.3Sn0.7O2	4.6918	3.1273	5.52	132.22	7.47

The morphologies and particle sizes of the 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) and unsupported IrO2 were characterized by the TEM images. As shown in figure 2a, the prepared SnO2 is in an irregular shape and the average particle size is approximately 10.2 nm. Figure 2b shows that the unsupported IrO2 exhibits a quasi-spherical shape and serious aggregation of nanoparticles with a broad particle size distribution, which would result in a poor catalytic activity. In contrast to the unsupported IrO2, the IrO2 nanoparticles supported on CoSn1−O2 present a quasi-spherical shape with sub 3 nm sizes as the darker dots in figure 2c–f. Meanwhile, IrO2 nanoparticles are observed to be well-dispersed on the CoSn1−O2 supports. It is noted that the particle sizes of CoSn1−O2 supports decrease gradually with the increase of the Co-doped content, which is in agreement with the XRD results and previous report [34]. The reduced CoSn1−O2 particle sizes would expose more surface area for the dispersion of IrO2 and avoid the serious aggregation of IrO2 nanoparticles. Therefore, the catalytic activities of prepared samples could be anticipated to enhance with the Co-doped content. Figure 3 exhibits the HR-TEM images of SnO2 and 40IrO2/Co0.2Sn0.8O2. The HR-TEM image in (figure 3a) confirms the SnO2 particles present an irregular shape with a mean particle size of about 10.2 nm. The lattice fringe is about 0.335 nm, corresponding to the (110) planes of SnO2. The HR-TEM image of figure 3b reveals the IrO2 in darker dots with the mean particle size of about 2.25 nm, and it disperses evenly on the surface of Co0.2Sn0.8O2. The lattice fringes are about 0.258 and 0.335 nm corresponding to the (101) plane of IrO2 and (110) plane of Co0.2Sn0.8O2, respectively. Additionally, the presence, relative amount and homogeneous distribution of the Ir and Co elements in these samples were verified by EDX mapping (electronic supplementary material, figure S1–S5). According to the obtained results, the atomic ratio of Co/Sn is around 8.1 for Co0.1Sn0.9O2, 4.0 for Co0.2Sn0.8O2 and 2.7 for Co0.3Sn0.7O2, which is consistent with the designed cobalt content.

Figure 2.

TEM images of (a) unsupported IrO2, (b) SnO2, (c) 40IrO2/SnO2 (d) 40IrO2/Co0.1Sn0.9O2, (e) 40IrO2/Co0.2Sn0.8O2 and (f) 40IrO2/Co0.3Sn0.7O2.

Figure 3.

HR-TEM images of (a) SnO2 and (b) 40IrO2/Co0.2Sn0.8O2.

It is well known that the specific surface areas play an important role in the catalytic activity of a catalyst. Thus, to identify the impact of Co doping and IrO2 loading on the specific surface areas and pore size of SnO2, nitrogen adsorption–desorption measurements were carried out. Figure 4 and electronic supplementary material, figure S7 represent the N2 adsorption–desorption isotherms of the CoSn1−O2 and 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3), the inset are the corresponding Barrett–Joyner–Halenda (BJH) pore size distribution curves. Typical Langmuir type IV with an inherent hysteresis loop at relative high pressure is detected for all samples, which suggests that the pores between the nanoparticles are mainly constructed by mesoporous structures. The specific surface areas of the samples are calculated by Brunauer–Emmett–Teller (BET). It can be seen that the specific surface areas increase significantly from 72.19 m2 g−1 for SnO2 to 104.01, 131.92 and 139.22 m2 g−1 for Co0.1Sn0.9O2, Co0.2Sn0.8O2 and Co0.3Sn0.7O2, respectively (table 1). The enhancement of specific surface areas of CoSn1−O2 might be attributed to the decreased particles sizes resulting from Co-doped SnO2, as demonstrated by the TEM results. As listed in electronic supplementary material, table S1, the specific surface areas for 40IrO2/SnO2, 40IrO2/Co0.1Sn0.9O2, 40IrO2/Co0.2Sn0.8O2 and 40IrO2/Co0.3Sn0.7O2 are 53.21, 87.54, 91.25, 93.06 m2 g−1, respectively. The specific surface areas of CoSn1−O2 are higher than those of 40 IrO2/CoSn1−O2 (x = 0.1, 0.2, 0.3), which might be due to the pore blocking of CoSn1−O2 samples with IrO2 loading. From pore size distributions curves, the pore sizes of CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) samples are mainly about 9–10 nm, whereas the mainly pore sizes of 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) samples are mainly about 3–5 nm and 6–7 nm, which could be a result of the IrO2 nanoparticles that occupied a certain amount of pore volume, resulting in a decrease in the pore size [36]. The Co-doped SnO2 samples exhibit a high specific surface area and evident porosity, which might be advantageous for the efficient catalytic performance of the prepared samples.

Figure 4.

N2 adsorption isotherms and pore size distributions (inset) of the CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) supports.

The XPS spectra were evaluated to reveal the elemental states of each element in the 40IrO2/Co0.2Sn0.8O2 sample, as shown in figure 5. The high-resolution Ir4f spectrum (figure 5a) shows spin–orbit doublets at approximately 61.9 eV and 64.9 eV which can be classified as Ir4f7/2 and Ir4f5/2, respectively [37]. The binding energy of Ir4f7/2 at 61.9 and Ir4f5/2 at 64.9 eV is in the form of the Ir4+ and Ir4f7/2 at 62.9 and Ir4f5/2 at 65.9 eV is in accordance with Ir3+. Clearly, the atomic ratio of Ir4+ is higher than that of Ir3+, the ratio of Ir4+/Ir3+ is 1.42. This indicates that the majority of Ir element in the crystal lattice is Ir4+ cations. The Co2p spectrum (figure 5b) is deconvoluted into two main components at 781.2 eV for Co2p3/2, 797.8 eV for Co2p1/2, and two satellite peaks (noted as ‘Sat.’) are also detected. The presented satellite peaks and the difference of 15.1 eV of the Co2p3/2 and Co2p1/2 imply that the majority of cobalt is in the states of Co2+.[38] figure 5c exhibits the high resolution of Sn3d spectrum. The peaks located at about 487.3 eV and 495.8 eV are the representatives of Sn3d5/2 and Sn3d3/2, and no other peak is detected, confirming the chemical state of Sn is only in tetravalence [39]. The O 1 s spectrum (figure 4d) could be divided into three major peaks: O1 (approx. 529.4 eV), O2 (approx. 531.5 eV) and O3 (approx. 532.9 eV), which are according to the metal-oxygen bonding, oxygen vacancies and hydroxyl species of water molecules adsorbed on the surface, respectively [40].

Figure 5.

The high-resolution XPS spectrum of (a) Ir4f, (b) Co2p, (c) Sn3d and (d) O1s in 40IrO2/Co0.2Sn0.8O2.

High electrical conductivity of support is favourable to the supported catalysts for the enhancement of catalytic performance. Prior to the catalytic performance test, the electrical conductivities of all the prepared samples were measured and listed in electronic supplementary material, table S2. The electrical conductivities of SnO2, Co0.1Sn0.9O2, Co0.2Sn0.8O2 Co0.3Sn0.9O2, 40IrO2/SnO2, 40IrO2/Co0.1Sn0.9O2, 40IrO2/Co0.2Sn0.8O2, 40IrO2/Co0.3Sn0.9O2 and unsupported IrO2 are 1.95 × 10−6, 2.02 × 10−5, 9.51 × 10−5, 6.94 × 10−5, 6.08 ×10−2, 3.15 × 10−1, 1.02 × 100, 8.13 × 10−1 and 1.02 × 101 S cm–1, respectively. The electrical conductivities of Co-doped SnO2 are much higher than those of the pure SnO2, suggesting a favourable effect of the Co dopant on the improvement of the SnO2 conductivity. After the IrO2 loading, the conductivity of 40IrO2/CoSn1−O2 (x = 0.1, 0.2, 0.3) is enhanced because of the excellent electrical conductivity of IrO2. However, the electrical conductivities of Co0.3Sn0.9O2 and 40IrO2/Co0.3Sn0.9O2 are lower than those of Co0.2Sn0.8O2 and 40IrO2/Co0.2Sn0.8O2, which might be due to the increasing impurity scattering centres with the enhancement of Co contents that impede the electron transport, decrease the carrier mobility and reduce electrical conductivity [41].

Electrochemical properties

To evaluate the effect on the surface active sites of catalysts with Co doping, the cyclic voltammograms (CVs) of 40IrO2/CoSn1−O2 were measured in N2-saturated 0.5 M H2SO4 at a scan rate of 50 mV s−1, as shown in figure 6a. For comparison, the CVs of the pristine SnO2 and unsupported IrO2 were also tested at the same condition, and the current densities of all the samples were normalized to the IrO2 loading. The pristine SnO2 shows very low current densities because of the poor electrocatalytic activity. The CVs of CoSn1−O2 samples were also tested (electronic supplementary material, figure S6 (a)) and exhibited higher current densities with Co doping than that of SnO2, but still unsatisfactory. The shapes of the voltammogram of 40IrO2/CoSn1−O2 are similar to those of the unsupported IrO2 but with a higher current density. The CVs of all samples present a broad redox peak of the reversible oxidation and reduction on the IrO2 surface, which suggests a typical pseudo-capacitive behaviour. The voltammetric charge of 40IrO2/CoSn1−O2 and unsupported IrO2, as a function of scan rates, was calculated by the following [42]:where I is the current density obtained in CV curves, v = 50 mV s−1 is the scan rate, m is the loading of noble metal Ir on the glassy carbon electrode, E is the scan potential between –0.148 and 1.15 versus Ag/AgCl. The sequence is 40IrO2/Co0.3Sn0.7O2 (333.2 C g(Ir)−1) > 40IrO2/Co0.2Sn0.8O2 (314.3 C g(Ir)−1) > 40IrO2/Co0.1Sn0.9O2 (252.6 C g(Ir)−1) > 40IrO2/SnO2(238.5 C g(Ir)−1) > IrO2 (203.5 C g(Ir)−1). This confirms the positive effect on the increment of surface active sites with Co doping. The enhancement of surface active sites could be ascribed to the decreased particle sizes of SnO2 supports that provide more sites for IrO2 dispersion. Consequently, according to the calculated voltammetric charge, it is expected that the OER catalytic activity of the prepared samples would be improved as the Co-doped samples. The overall OER performance of the samples will be discussed in the following single cell test results. Figure 6b shows the LSV polarization curves of the pristine SnO2, unsupported IrO2 and 40IrO2/CoSn1−O2 samples in N2-saturated 0.5 M H2SO4 at a scan rate of 50 mV s−1. The potentials at the current density of 10 mA cm−2 are listed in table 2. The measured potentials are 1.577, 1.570, 1.557, 1.541 and 1.554 V versus RHE for unsupported IrO2, 40IrO2/SnO2, 40IrO2/Co0.1Sn0.9O2, 40IrO2/Co0.2Sn0.8O2 and 40IrO2/Co0.3Sn0.7O2, respectively. It is clearly observed that 40IrO2/Co0.2Sn0.8O2 reveals the lowest overpotential on mass activity, indicating Co-doped support could favour the increment of the active substance and enhance the OER performance. Electronic supplementary material, figure S6 (b) displays the LSV polarization curves of the CoSn1−O2 samples. It is clearly observed that the overpotential of CoSn1−O2 samples is higher than that of 40IrO2/CoSn1−O2 samples, indicating the low catalytic activity of supports.

Figure 6.

(a) Cyclic voltammetry curves of 40IrO2/CoSn1−O2, pristine SnO2 and unsupported IrO2 in N2-saturated 0.5 M H2SO4 solution at a scan rate is 50 mV s−1 and (b) LSV curves of 40IrO2/CoSn1−O2, pristine SnO2 and unsupported IrO2.

Table 2.

The obtained values of voltammetric charge (C/g), the potentials at 10 mA cm−2, cell potential at 1 A cm−2, RΩ and Rct of the prepared samples.

		LSV		EIS (mΩ cm2)
samples	voltammetric charge (C g−1)	The potentials at 10 mA cm−2	cell potential at 1 A cm−2	RΩ	Rct
40IrO2/SnO2	238.5	1.570	1.847	152.0	40.3
40IrO2/Co0.1Sn0.9O2	252.6	1.557	1.812	114.2	37.0
40IrO2/Co0.2Sn0.8O2	314.3	1.541	1.748	74.7	32.1
40IrO2/Co0.3Sn0.7O2	333.2	1.554	1.77	85.3	34.3
unsupported IrO2	203.5	1.577	1.713	63.1	52.6

Electrolysis cell performance

After the MEAs assembled in single cells, the OER performance of the 40IrO2/CoSn1−O2 catalysts was further characterized by I-V polarization measurement from 0.01 to 1 A cm−2 at 80°C, as displayed in figure 7. Again, the OER performance of unsupported IrO2 was detected at the same condition for comparison. For the low current density (less than 0.1 A cm−2), the cell voltage of the 40IrO2/CoSn1−O2 increases and unsupported IrO2 increases nonlinearly along the current density, which is mainly affected by activation polarization. Once the polarization current density increases gradually, the disparity in the OER performance is mainly due to ohmic resistance and the polarization resistance, as presented by the linear increase of potential for all samples. Compared to the 40IrO2/SnO2, the cell voltage of 40IrO2/CoSn1−O2 shows a superior performance. The 40IrO2/Co0.2Sn0.8O2 exhibits optimal activity at 1A cm−2, followed by the order of 40IrO2/Co0.3Sn0.7O2, 40IrO2/Co0.1Sn0.9O2, 40IrO2/SnO2, in that order. However, the OER performance of the 40IrO2/CoSn1−O2 catalysts is still lower than that of unsupported IrO2 due to the lower amount of Ir loading. The sequence in OER performance at 1 A cm−2 is IrO2 (1.713 V) < 40IrO2/Co0.2Sn0.8O2 (1.748 V) < 40IrO2/Co0.3Sn0.7O2 (1.770 V) < 40IrO2/Co0.1Sn0.9O2 (1.812 V) < 40IrO2/SnO2 (1.847 V). Notably, the OER performance decreases when the Co-doping level reaches to the point x = 0.3. This could be attributed to the increasing impurity scattering centres with the enhancement of Co contents that impede the electron transport and decrease the carrier mobility, leading to the degradation of OER performance [41].

Figure 7.

Polarization curves of single cells equipped with 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) and unsupported IrO2 at 80°C.

The charge transfer resistance (Rct) is a critically important parameter in reflection of reaction kinetics in electrocatalytic performance for a catalyst, and a lower Rct implies a faster reaction rate [43]. To further reveal the intrinsic charge transfer properties of IrO2 supported by varying Co-doped SnO2 as supports, the EIS measurements were conducted at 0.3 A cm−2 in the single cells. Figure 8 exhibits the Nyquist plots of 40IrO2/CoSn1−O2 and unsupported IrO2, and the appropriate equivalent circuit model is shown in the inset. Rct is the charge transfer resistance of a faradic process occurring at the interface of catalyst and electrolyte, which is evaluated from the large semicircles by the difference between low- and high-frequency intercepts on the real axis [44]. The ohmic resistance (RΩ) is the series resistance of all the geometry compositions in a cell. RΩ can be calculated from the intercept extending from the high-frequency side of the curve of the real axis [45]. The calculated values of the Rct and RΩ are listed in table 2. It can be seen that the unsupported IrO2 shows a lower RΩ (63.1 mΩ cm2) because of the excellent electrical conductivity than 40IrO2/CoSn1−O2. The Rct value of 40IrO2/Co0.2Sn0.8O2 is 32 mΩ cm2 and exhibits the lowest charge transfer resistance, followed by 40IrO2/Co0.3Sn0.7O2 (34 mΩ cm2), 40IrO2/Co0.1Sn0.9O2 (37 mΩ cm2), 40IrO2/SnO2 (40 mΩ cm2) and unsupported IrO2 (52 mΩ cm2) in sequence. It is anticipated that the significant increase of the exposed surface area of IrO2 could enhance surface active sites and charge transport property, leading to the improvement of OER performance of 40IrO2/CoSn1−O2.

Figure 8.

Nyquist diagrams of 40IrO2/CoSn1−O2 (x = 0, 0.1, 0.2, 0.3) and unsupported IrO2 at 0.3 A cm−2 and 80°C.

Long-term durability is a critical parameter for the commercialization of catalysts. The durability test of the unsupported IrO2 and 40IrO2/Co0.2Sn0.8O2 catalysts in a single cell was measured at current densities of 1 A cm−2 at 80°C for 100 h. As displayed in figure 9, both the cell potentials of the samples reveal different degrees of increase. The cell potential of the unsupported IrO2 shows a relatively uniform upward trend, rising from 1.713 to 1.755 V after 100 h, and the average decay is 0.42 mV h−1. The cell voltage of 40IrO2/Co0.2Sn0.8O2 increases slightly during the initial 40 h, but later remains almost constant at 1.766 V, and the average degradation rate is 0.18 mV h−1. According to Kötz [46], the alternation between Ir(III) and Ir(IV) during the catalysis of OER plays a crucial role in the oxidation of the hydroxyl, while Ir(VI) is easily prone to corrosion according to the following [47]:This implies that the supports are beneficial in the formation of stable Ir oxide during the modified Adams fusion, and that Co-doped content might further enhance the durability of 40IrO2/CoSn(1−O2 catalyst. Additionally, recent studies on OER performance and stability of iridium-based catalysts have been summarized and listed in table 3. It can be seen that the cell voltage and degradation rate of 40IrO2/Co0.2Sn0.8O2 are comparable to or superior to those of the previously reported iridium-based catalysts (such as IrO2/V-doped TiO2 [31] and IrO2-ATO [48]). This further confirms that the prepared Co-doped SnO2 as a support for IrO2 catalysts is a potential candidate for the practical application of SPE water electrolyzer.

Figure 9.

Durability test of the unsupported IrO2 and 40IrO2/Co0.2Sn0.8O2 sample in a single cell at 1 A cm−2 at 80°C.

Table 3.

OER performance and stability reported in the literature for iridium-based catalysts.

references	anode catalyst	Ir loading (mg cm−2)	cathode catalyst	Pt loading (mg cm−2)	operating current (A cm−2)	operating temperature (°C)	cell voltage	electrode fabrication process	active area (cm2)	test time (h)	degradation rate (μV h−1)
Hao et al. [31]	IrO2/V -doped TiO2	2.5	Pt/C	0.5	1	80	2.015 V (1 A cm−2)	spraying	3.65	50	1980
Puthiyapura et al. [48]	IrO2-ATO	2	Pt/C	0.5	1.0	80	1.80 V (1 A cm−2)	spraying	1
Rakousky et al. [49]	IrO2 and TiO2	2.25	Pt/C	0.8	2.0	80	1.84 V (2 A cm−2)	commercial	17.64	1150	194
Zeng et al. [50]	Ir black	0.873	Pt/C	1.0	0.25	80	1.728 V (2 A cm−2)	spraying	4	300	52
Faustini et al. [51]	Ir0.7Ru0.3Ox	1.8	Pt/C	0.5	1	80	1.680 V (1 A cm−2)	spraying	6.25
Jorge et al. [52]	gCNH-IrO2	1.2	Pt/C	4	1	80	1.93 V (1 A cm−2)	spraying	7.07
this study	IrO2/Co0.2Sn0.8O2	2.5	Pt/C	0.5	1	80	1.776 V (1 A cm−2)	spraying	3.65	100	180

Conclusion

A varying amount of Co-doped SnO2 as anode support for IrO2 has been successfully prepared and then characterized by series methods. The particle sizes of the prepared CoSn(1−O2 (x = 0.1, 0.2, 0.3) samples were decreased with the increase of Co-doping content, which provided more sites for the well-dispersed IrO2. Also, the prepared samples exhibited increased high specific surface areas. The 40IrO2/Co0.2Sn0.8O2 exhibited the lowest overpotential with a cell potential of 1.748 V at 1000 mA cm−2 and showed a good stability during the 100 h operating at the current density 1 A cm−2 at 80°C. The decreased overpotential of 40IrO2/Co0.2Sn0.8O2 could be ascribed to the increment of surface active sites, the enhancement of electrical conductivity and the higher charge transfer, as verified by CVs and EIS measurement results. Consequently, the CoSn(1−O2 shows a promising alternative support for anode catalysts in the SPE water electrolyzer.