1 Introduction

In the early 21st century, the latest research indicates that growing global energy demand and diminishing supply of fossil fuel sources has led to substituting traditional fuels with green and environmentally friendly alternatives [1, 2] The emerging renewable energy technology, especially when coupled with irregular energy sources like the sun and wind, might significantly reduce the environmental strain caused by energy consumption [3,4,5]. Numerous energy conversion technologies, such as converting water molecules into high-energy molecules (e.g., oxygen or hydrogen), i.e., water splitting, are essential techniques to resolve the inherent problem of intermittency associated with these energy sources [6,7,8]. Nonetheless, water splitting is considered the most optimal approach for achieving this goal, and many studies have been done on this sector in recent decades [9, 10]. Water splitting comprises two partial reactions, the cathodic hydrogen evolution reaction (HER) and the anodic oxygen evolution reaction (OER) [11,12,13,14,15,16]. The 4 electrons (e⁻) proton-associated reaction (OER) has substantial energy challenges, such as significantly lower reaction kinetics, which result in the requirement of large overpotential to proceed with the reaction [8,9,10,11]. Much research [13,14,15,16,17] has been done to develop effective electrocatalysts with lower overpotential and faster reaction rates by adjusting their morphological and electrical properties to perform OER [17, 18]. Among the numerous electrocatalysts produced so far, ruthenium (Ru) and iridium (Ir) based electrocatalysts have been widely used due to their greater efficiency for OER, but these electrocatalysts requires a higher overpotential (250–300 mV) and severely high cost limits their practical application [19, 20]. Comparable to precious metal oxides regarding OER activity, perovskites are a family of oxygen-containing materials with elevated OER performance owing to their greater flexibility and economical nature [21, 22]. ABO3 perovskite group (A is usually a rare-earth or alkaline earth metal and B is a simple or transition metal) is of tremendous interest since it exhibits several outstanding physical features including piezoelectricity, electrochemical reactivity and ferroelectricity and chemical properties [23, 24]. The stannate based perovskites (XSnO3, where X = Sr, Ca or Ba) are a type of perovskite which has garnered significant interest due to their excellent conductivity, high durability, and enhanced electrocatalytic activity in alkaline media [9]. However, certain distinctive attributes of perovskites, such as their good catalytic activity and versatility, place them at the forefront of other widely used metal oxides, i.e., transparent conductive oxides, SnO2, ZnO, and In2O3 [25]. Xiaomin Xu et al. introduced a collaborative co-doping approach to discover a range of BaCo0.9-xFexSn0.1O3-δ perovskites with adjustable electro-catalytic efficiency for OER. By adjusting the amounts of Sn and Fe dopants, BaCo0.9-xFexSn0.1O3-δ exhibits a cubical arrangement with inherent OER activity that is approximately ten times more than IrO2 with Tafel value (69 mV dec−1) [26]. The stable low valence of the Sr element and accessible redox potential of the Sn element, with its multivalent nature, makes SrSnO3 a desirable electrocatalyst for OER.

In recent research, SrSnO3 has been utilized in many applications, such as batteries and high-temperature gas sensors [27, 28]. This perovskite SrSnO3 has a cubic structure at high temperatures and boasts an elevated level of combined electronic and oxygen ionic conductivity with a low overpotential [29]. Nevertheless, SrSnO3 demonstrates varying polymorphs based on fabrication conditions and the cubic phase is not thermodynamically stable at ambient temperature [30]. Adding SrSnO3 to a carbon-based compound, including graphene oxide (GO), carbon nanotubes, graphitic carbon nitrides (g-CN), and reduced graphene oxide (rGO) makes it more durable and improves its ability to move charges at the electrode/electrolyte contact, which results in better electrical conductivity [31]. The rGO is recognized as a conducting material on account of its enhanced surface area, superior electrical conductivity, and strong chemical endurance [32]. The coupling of perovskite with carbon (C) based compounds could enhance their OER activity due to their excellent electrical conductivity, high interfacial-to-volume ratio and superior stability [33]. Although the hydrothermal procedure reduces oxygen (O2)-containing functional groups in GO, resulting in large defective regions that facilitate rapid electron transfer and the carbonyl (C=O) groups found at the corners of the GO sheets adsorb intermediate species during the reaction, act as reactive sites for the OER [34, 35] The rGO sheets are highly conductive due to their active transfer of delocalized electrons to delocalized molecular orbitals [36, 37]. Graphene-supported catalysts consisting of non-precious metal oxides have demonstrated significant potential for OER [38, 39]. Numerous compounds of perovskite oxide have already been extremely used as photocatalysts for water splitting [40, 41]. In this regard, SrSnO3 was also used as a photocatalyst in various applications [42], such as CO2 electro reduction [36], in biological samples [43], as well as in water splitting [38] due to their spatial structure, which helps in electromobility. Hence, motivated by the above-mentioned literature, we produce an rGO-based SrSnO3 composite using the hydrothermal approach, which is not yet used as an electrocatalyst for OER. The addition of rGO enhances SrSnO3 surface area which helps in reducing agglomeration and promotes the charge transfer ability by providing more active sites and also induces a synergistic effect that enhances the catalytic efficiency. Moreover, SrSnO₃ has an outstanding electrochemical characteristic, while the rGO offers extensive surface area and superior electrical conductivity. In this regard, this combination increases charge transfer and supplies more active sites for reactions. Thus, SrSnO₃/rGO composites offer a viable choice for OER electrocatalysts due to their synergistic properties, enhanced stability, improved conductivity, scalable synthesis, cost-effectiveness, and lower environmental impact.

This study was done to synthesize SrSnO3/rGO composite using the hydrothermal approach for OER. The analytical analyses revealed that the SrSnO3 has a cubic structure, which is fully dispersed on the carbon-based structure, resulting in increased surface area (130.44 m2 g−1). Moreover, the electrochemical study of electrocatalysts was done using numerous electrochemical parameters. Electrochemical studies suggested that SrSnO3/rGO composite exhibits low overpotential (199 mV) in obtaining Cd (10 mA cm−2) along with Tafel value (33 mV dec−1). This result clarifies the significance of prepared material as an electrocatalyst for OER and provides strong evidence that various methods can enhance water splitting.

2 Experimental

2.1 Reagents

Strontium carbonate (SrCO3, Sigma Aldrich, 99.9%), strontium chloride (SnCl2, Merck, 98%), citric acid anhydrous (C6H8O7, AnalaR, 99%), ammonia hydroxide (NH4OH, 31%, Sigma Aldrich), ethylene glycol ((CH2OH)2, 99%, Merck), sulphuric acid (H2SO4, 99%, AnalaR), phosphoric acid (H3PO4, 99%, Merck), graphitic powder, potassium permanganate (KMnO4, 98%, Sigma Aldrich), hydrazine solution (NH2NH2, 98%, AnalaR), ethanol (C2H5OH, 99%, Merck). Deionized water (DI) was also used in the entire fabrication method.

2.2 Preparation of SrSnO3

The pure sol-gel approach was used to synthesize pure material SrSnO3. To fabricate this material, 0.1 M of both precursor substances, SrCO3 and SnCl2, were mixed to 100 mL of ultrapure water with continuous stirring. Then, the citric acid (1.18 g), which functions as a chelating agent, was introduced into the obtained solution. To maintain the pH, 15 mL of ammonia hydroxide was poured into the obtained solution mixture drop by drop until optimum pH (9) was obtained. In the final step, ethylene glycol (5 mL) was added with constant stirring. Furthermore, the temperature was maintained at 80 °C until the water evaporated and a fine precipitate was obtained. The obtained precipitate was scraped from a beaker and collected in a crucible to annealed at 300 °C for 3 h, and finally, the annealed material was ground into fine powder.

2.3 Synthesis of rGO

Hummers’ method was used to synthesize rGO using two acids, H2SO4 and H3PO4, which were mixed in a ratio of 9:1 under constant agitating for 15 min [44]. In the prepared acidic solution, 2 g of graphitic powder was incorporated at a temperature of 22 °C under continuous stirring. Moreover, the suspension of GO was obtained by adding 6 g of KMnO4 in a solution with the aim of oxidizing the graphene powder. Afterward, 4 mL of H2O2 was inserted into a solution under steady stirring for 15 min to obtain rGO from GO and remove any residual byproduct. Furthermore, 5 mL of monohydrated hydrazine was introduced into a solution under constant agitation for 5 h. To sustain pH at 7, ultrapure water (50 mL) was mixed in a highly purified solution along with steady stirring for 15 min. The obtained mixture was cleaned multiple times with pure water and ethyl alcohol to obtain an impurities-free product. Finally, the obtained product was desiccated and preserved for further study.

2.4 Fabrication of SrSnO3/rGO

The composite was prepared via a hydrothermal technique by combining 20 mg of SrSnO3 and 20 mg of rGO in a 1:1 ratio in 0.03 L of water. Afterward, the prepared solution was persistently agitated for 30 min and moved to a hydrothermal reactor. After thermal treatment, the material was cooled and purified with ultrapure water and ethyl alcohol using centrifugation. Thus, the obtained product was dehydrated at 50 °C for 12 h. The procedure used to fabricate the SrSnO3/rGO composite was presented in Scheme 1.

Scheme 1
scheme 1

The diagrammatic depiction illustrates the hydrothermal method for producing SrSnO3/rGO composites

2.5 Physical measurement

An X-ray diffractometer (XRD, Bruker D-2 powder XRD) was applied to observe the crystallite structure of produced materials with a Cu-Kα source at 2θ from 10–80°. Moreover, the Raman study were acquired using a Raman microscope (MNSTEX PRI 100) equipped with a Helium and Neon laser beam source emitting light at a wavelength of 633.0 nm and a power delivery of 10 mW. The wavelength range covered was from 0 to 2000 cm−1. The prepared material’s morphology was analyzed with a TESCAN MIRA3 scanning electron microscope (SEM, Czech Republic). To analyze the textural properties of all prepared materials, Brunauer Emmett Teller’s using N2 as adsorbate (BET, Nova2200e Quanta chrome) was used. Moreover, the pore size distribution of all the produced material was evacuated with the Barrett–Joyner–Halenda (BJH) method (Quantchrome Autosorb iQ2).

2.6 Preparation of electrode

For the synthesis of the working electrode (WE), Ni foam was fragmented into 1 × 1 cm2 pieces and was sonicated with 5.0% HCl, acetone, ethyl alcohol, and distilled water at each stage for 15 min to eradicate the oxides’ layer. Furthermore, washed NF was dehydrated in the thermal chamber for 30 min at 60 °C. The 10 mg of prepared material was poured into 40 μL of pure water and the resultant mixture was exposed to an ultrasonic wave for 35 min. Ultimately, the resulting slurry was deposited on dry NF withg a straightforward drop caste method and desiccated for 30 min at 60 °C. The electrocatalytic efficiency of the produced material was evaluated in an alkaline solution.

2.7 Electrochemical measurement

The Metrohm AUTOLAB (PGSTAT-204) was used to investigate the electrochemical efficiency of prepared material in an alkaline media with a three-electrode workstation such as a reference electrode which is comprised of Ag/AgCl, pt wire function as auxiliary electrode, and Ni foam coated with synthesized material served as working electrode. The OER efficiency was assessed by cyclic voltammetry (CV), energy dispersive spectroscopy (EIS), chronoamperometry, and linear sweep voltammetry (LSV) using the 2.1 NOVA program. Before recording the electrochemical activity of the electrocatalyst, CV cycles at a scanning rate of 10 mV s-1 were recorded to maintain their OER efficiency. However, LSV was executed with the sweeping speed (10 mV s−1) at ambient temperature to test the electrocatalyst's OER efficiency and overpotentials were obtained at Cd (10 mA cm−2). In this research, potentials attained in the electrochemical analysis were changed to a reversible hydrogen electrode (RHE) with Eq. (1) [45].

$${E}_{RHE}={E}_{Ag/AgCl}+{E}_{Ag/AgCl}^{^\circ }+0.059pH$$
(1)

The CV graph was used to quantify electro-kinetics analysis by taking the Tafel value of a linear section of the polarization curve along with Eq. (2).

$$\eta =\alpha +\frac{2.303{RT}}{\alpha {nF}}({logj})$$
(2)

The term electrochemically active surface area (ECSA) was employed to estimate the surface area of prepared material with double-layer capacitance (Cdl). As Cdl is half of the slope, it may be found by drawing the fitting line between the scanning rate and current density difference [46]. Further investigation shows that ECSA is directly proportional to Cdl and is calculated using Eq. (3).

$${ECSA}=\frac{{Cdl}}{{Cs}}$$
(3)

Here, Cs represents specific capacitance (0.04 mF/cm2) [47].

$${Cdl}=\frac{{slope}}{2}$$
(4)

The roughness factor (Rf) was used to measure the electrocatalytic performance of prepared material with the equation (5).

$${R}_{f}=\frac{{ECSA}}{{Area}}$$
(5)

The factor “mass activity” is widely used to evaluate the performance of catalytic systems in proportion to active mass and electrode area. As mass activity increases, the performance of the catalyst also improves. Moreover, different factors like surface area, crystal structure, composition, and electrical properties are used to determine the mass activity of OER.

$${Mass\; activity}({MA})=\frac{{current\; density}(j)}{{Mass\; loading}}$$
(6)

Moreover, EIS plot was applied to evaluated the charge transfer resistance (Rct) of the material and solution resistance (Rs) of the prepared material in a frequency domain of 0.1-100000 Hz. The chronoamperometry test shows the durability of the prepared material.

3 Result and discussion

3.1 Physiochemical results

The XRD analysis was accustomed to analyzing SrSnO3 and SrSnO3/rGO crystal structures in the 2θ range of 10–80°. Figure 1a illustrates the diffractogram of SrSnO3 and SrSnO3/rGO, which give 6 sharp peaks at 21.13°, 32.51°, 44.44°, 55.72°, 66.54° and 73.29° with the plane of 100, 110, 200, 211, 220 and 013, respectively. The XRD study illustrated the cubic framework of SrSnO3 having parameters of a, b, and c, which are equal to 4.0254 Å, comparatively matching with JCPDS No # 01-074-1298. All of the SrSnO3 peaks were present in the composite’s XRD diffractogram, as shown in Fig. 1a, along with an additional peak of rGO at the 2θ range of 24.66 with plane 002 [48]. The crystallite size of pure material and composite was computed as 23.48 nm and 16.95 nm, correspondingly, using the equation named as Debye Scherer equation [49].

$$D=\frac{k\lambda }{\beta \,\cos \theta }$$
(7)

Here, D represents crystallite size, λ is the wavelength (1.54056 nm), β represents the Full-width half maxima peaks of the pattern, and θ demonstrates the diffraction angle. As a composite’s crystalline size decreases, its active sites increase, improving its electrocatalytic properties [50].

Fig. 1
figure 1

a XRD Analysis and (b) Raman analysis of all prepared materials

Figure 1b depicted Raman spectra of pure SrSnO3 and rGO-based composite SrSnO3/rGO. Raman spectra of SrSnO3 revealed the five vibration modes at 114, 147, 167, 223 and 259 cm−1 were present. From Fig. 1b, it was observed that the peaks 147 and 167 cm−1 correspond to Ag mode, which correlated to Sr–O–Sn and O–Sn–O bonds, correspondingly, while the other peaks at 114 relate to B2g and peaks at 223, 259 relate to Ag mode which closely related with cubic structure of SrSnO3 [51]. Additionally, Raman spectra of rGO revealed the existence of two bands observed at 1339 cm−1 and 1584 cm−1 that relate to the D and G band. Moreover, the D band (1339 cm−1) refers to disorders and defects, particularly at the corners of graphene that occur due to the presence of heteroatoms which disrupt the crystallization of carbon belonging to A1g phonons and the G band (1584 cm−1) related to stretching vibration of C–C bonds in-plane which show sp2 symmetry belong to Eg mode [38, 52]. Significantly, it has been observed that the D band exhibits high peak intensity compared to the G band. This can be attributed to the existence of a disorder on the rGO’s surface that supports the SrSnO3 particles [53]. Thus, the existence of all the bands of SrSnO3 and rGO in the composite spectra revealed that rGO is uniformly adsorbed on the SrSnO3 surface and the composite has been successfully synthesized.

The morphological analysis of all the prepared material was determined by SEM, as depicted in Fig. 2a, b. SEM microgram of SrSnO3 was depicted in Fig. 2a, clearly demonstrating that the obtained material had irregularly (spherical and elongated) shaped highly dense nanoparticles. Figure 2b revealed the micrograph of SrSnO3/rGO composite, demonstrating that rGO possesses a bulky sheet-like morphology. At the same time, these SrSnO3 nanoparticles were uniformly dispersed on the rGO sheets, which proved a favorable coupling of rGO and SrSnO3. The dispersion of SrSnO3 nanoparticles on the rGO surface was probably favorable in dealing with this association, which could eventually reduce the accumulation and restacking of the rGO sheets. It was observed that SrSnO3/rGO composites could prove to be advantageous for the separation of charge carriers because of their interfacial interaction, which results in increased electrocatalytic activity [54].

Fig. 2
figure 2

SEM micrograph of (a) SrSnO3 (b) SrSnO3/rGO composite

The BET technique was employed to evaluate the efficacy of newly synthesized electrocatalysts by determining the textual properties of pure SrSnO3 and SrSnO3/rGO composite. The N2 adsorption-desorption curves at 77 K were obtained for samples and the outcomes of these trends were displayed in Fig. 3a. The type IV isotherms found in adsorption/desorption isotherms with hysteresis loops (H4) [55] verify the mesoporous structure of the pure and composite with surface areas of 36 and 59 m2 g−1, respectively. Based on the measured surface area values, it can be observed that SrSnO3/rGO has a surface area more than SrSnO3, which could potentially improve the performance of the recently developed electrocatalyst. Furthermore, the BJH technique was used to evaluate the pore volume and size of all the generated material. Meanwhile, the pore volume for SrSnO3 and SrSnO3/rGO composite is found to correspondingly be 22 cm3 g−1 and 35 cm3 g−1, as shown in Fig. 3b. Hence, the pore size distribution validates the existence of a mesoporous structure in the materials. The porous and wide surface structure allows quick ions or electron interaction with the electrolyte, resulting in rapid faradaic redox reactions.

Fig. 3
figure 3

a BET curves of pure SrSnO3 material and SrSnO3/rGO composite (b) pore size distribution curves of pure SrSnO3 and SrSnO3/rGO composite

3.2 Electrochemical analysis

The catalytic efficiency of SrSnO3 and SrSnO3/rGO composite was analyzed utilizing a three-electrode workstation in an alkaline media. Figure 4a, b demonstrates the CV as well as LSV curves, which illustrates the OER activity of pure and composite material at a scan speed of 10 mV s−1 in basic media. However, the redox behavior of SrSnO3 and SrSnO3/rGO was determined using a cathodic and anodic cycle of the CV curve at sweeping speed (10 mV s−1) in a potential range of 0.0 to 2.0 as depicted in Fig. 4a. LSV measurements of pure and composite were taken to determine the material’s catalytic efficiency, as depicted in Fig. 4b. However, the SrSnO3/rGO’s increased catalytic activity showed that the composite had a lower Tafel slope and overpotential than its pure counterpart [56]. The obtained result from the CV graph showed that the pristine SrSnO3 and SrSnO3/rGO composite exhibited the overpotential of 312 mV and 199 mV, respectively, as depicted in Fig. 4c. This decline in the composite overpotential was owing to its small crystalline size, diverse morphology, and more surface area, as revealed from BET analysis in Fig. 3 [1]. Figure 4d demonstrated the Tafel plot of SrSnO3/rGO (33 mV dec−1), that was significantly less than SrSnO3 (58 mV dec−1). This reduced Tafel value revealed that the reaction proceeded fast in OER with the attachment of OH⁻ on the electrocatalyst’s surface [57].

Fig. 4
figure 4

a Polarization curve of prepared material (b) LSV curves (c) Overpotential of SrSnO3 and SrSnO3/rGO composite. d Tafel plot of SrSnO3 and SrSnO3/rGO composite

The OER procedure is thought to involve surface metal sites (Z’), which undergo a series of steps leading to an alkaline OER mechanism. First, the hydroxide anion adsorbed on the active site undergoes 1 e⁻ (electron) oxidation, resulting in Z’–OH. After a few protons and e removals, Z’–OH transforms into Z’–O. Following that, Z’–O will follow two distinct pathways to produce O2 molecules. First, Z’–O decomposes into O2 and Metal Z’, while in the upcoming step, Z’–O combines with OH⁻ to develop a Z’OOH intermediate. In the penultimate phase, Z’OOH is deprotonated to form O2, regenerating the active site (Table 1). The comparison of present work has been performed by already reported work is shown in Table 1.

$${\rm{Z}}{\rm{\mbox{'}}}+{{OH}}^{-}\to {\rm{Z}}{\rm{\mbox{'}}}{OH}+{e}^{-}$$
(4)
$${\rm{Z}}{\rm{\mbox{'}}}{OH}+{{OH}}^{-}\to {\rm{Z}}{\rm{\mbox{'}}}O+{H}_{2}O+{e}^{-}$$
(5)
$$2{\rm{Z}}{\rm{\mbox{'}}}O\to 2{\rm{Z}}{\rm{\mbox{'}}}+{O}_{2}$$
(6)
$${\rm{Z}}{\rm{\mbox{'}}}O+{{OH}}^{-}\to {\rm{Z}}{\rm{\mbox{'}}}{OOH}+{e}^{-}$$
(7)
$${\rm{Z}}{\rm{\mbox{'}}}O+{{OH}}^{-}\to {\rm{Z}}{\rm{\mbox{'}}}+{O}_{2}+{H}_{2}O+{e}^{-}$$
(8)
Table 1 Analyzing OER performance in relation to prepared materials and published literature

In analyzing the efficiency of electrocatalyst, especially metal oxide, for OER, it is necessary to evaluate its ECSA. As the structure of metal oxide is porous, which determines the specific surface area required for the adsorption of reactants, this adsorption capacity of reactants affects the catalytic behavior of the metal oxide. For example, a metal oxide with wider pores may have larger ECSA, as it can adsorb a greater number of reactants. The ECSA of pure and composite was evaluated in a non-faradic region of CV within the specific range (0.79–0.88 V) at different scan rate ranging from 10 to 50 mV s−1, as displayed in Fig. 5a, c. Moreover, the ECSA of prepared material was determined using Cdl, which has a proportional relationship between the scan rate and the Δj in the polarization curves. Figure 5b, d illustrate that the computed Cdl value of SrSnO3 and SrSnO3/rGO was 8.2 mF cm−2 and 24.5 mF cm−2, correspondingly, as estimated with Eq. 4. The higher Cdl value for the SrSnO3/rGO composite could be attributed to the dispersion of SrSnO3 over the surface of rGO, which produces highly defined channels that allow electrolytic ions to access reactive sites easily, resulting in large ECSA values [31]. The computed ECSA for SrSnO3 and SrSnO3/rGO composite were 205 cm2 and 612.5 cm2, correspondingly. The high OER efficacy of SrSnO3/rGO may be owing to numerous variables such as its high active sites, well-defined channels and large ECSA. The greater surface area results in the exposure of additional catalytic sites, leading to increased mass activity [58]. Catalysts with greater mass activity can generate a larger amount of oxygen per unit mass. This implies that a reduced amount of catalyst is necessary to attain an equivalent degree of performance, which is especially significant for enhancing resource effectiveness. The calculated Rf values of pure and composite were 410 and 1225, correspondingly. This shows that the contact area of SrSnO3/rGO composite enables simple access for electrolyte ions, hence boosting the electrochemical efficiency.

Fig. 5
figure 5

a ECSA of pristine material at various scanning speeds (10–50 mV s−1). b Cdl plot of SrSnO3 (c) ECSA of composite material at various scanning speeds (10–50 mV s−1) and (d) Cdl plot of SrSnO3/rGO composite

The resistance of all the prepared materials can be determined by EIS in a frequency series of 0.1–100 kHz at 5 mV. The EIS plot of pristine SrSnO3 and SrSnO3/rGO composite gave Rs (solution resistance) and Rct across electrolyte-electrode interaction, as demonstrated in Fig. 6a. The Rs value of SrSnO3 and SrSnO3/rGO was roughly 1.33 Ω and 1.23 Ω, and the Rct of SrSnO3 and SrSnO3/rGO composite was around 3.69 Ω and 0.16 Ω, correspondingly. This decline in Rct and Rs values for composite compared to the pristine results in the increased conductivity of the material and the fast dynamics of OER. Moreover, the chronoamperometry stability tests was applied to assess the durability of the composite, which was used as an electrocatalyst during water splitting, as displayed in Fig. 6b. The chronoamperometry of the composite demonstrated that at the first 10 h, the current density of SrSnO3/rGO indicated a little drop and then remained enduring for 35 h at Cd of 67 mA cm-2. The LSV stability experiment was conducted to assess the durability of the composite at 10 mV s−1 for 3000th cycles during OER, as depicted in Fig. 6c. Furthermore, the LSV curve of the first SrSnO3/rGO composite cycle was analogous to the 3000th one. After 3000th cycles of stability tests, the LSV curve demonstrated that the composite exhibited excellent stability during the water oxidation reaction with a slight reduction in current density. Thus, no obvious shifts were detected in the LSV peak of SrSnO3/rGO composite after the 3000th cycle.

Fig. 6
figure 6

a EIS of SrSnO3 and SrSnO3/rGO composite. b Chronoamperometry (CA) of SrSnO3/rGO composite. c LSV stability of SrSnO3/rGO composite

4 Conclusion

This investigation presents a new hydrothermal method to fabricate the SrSnO3/rGO composite, which functioned as an effective electrocatalyst for the OER. The Physical characterizations were used to evaluate the produced materials’ morphology, textual properties, and structure. The analysis found that the rGO-based SrSnO3 composite had an overpotential of 199 mV, resulting in a Cd of around ~10 mA cm−2 and a Tafel gradient (33 mV dec−1). The electrocatalytic activity of SrSnO3/rGO composite was validated by EIS and chronoamperometry study, demonstrating high conductivity assigned to the expanded surface area confirmed by cubic structure and stability for 35 h. This present research shows a novel strategy for increasing the performance of stannate-based catalysts by employing rGO as an appropriate material for producing several active sites. The prepared composite serves as a competent electrode material for OER performance and offers new possibilities for various applications.