In recent decades, global warming and energy crises have become critical, and the need for inexhaustible clean energy technologies has become urgent [1, 2]. In 1991, Grätzel [3] firstly reported dye-sensitized solar cells (DSSCs). DSSCs have aroused interest from researchers owing to their low price, excellent theoretical power conversion efficiency (PCE), environmental friendliness, and multi-color transparency [4, 5]. In general, DSSCs consist mainly of a dye-sensitized photoanode, a counter electrode (CE), an electrolyte-based on iodine, and an external circuit [6]. The CE, a critical component of DSSCs, affects the catalytic reduction of redox couples and plays an important role in the photovoltaic conversion efficiency. A superior electrocatalytic activity for the regeneration of redox couples, excellent electrical conductivity, and good mechanical stability are characteristics of a promising CE in the electrolytic cell system [7, 8]. Usually, Pt is utilized as a standard CE in DSSCs on the basis of its outstanding electrocatalytic activity and ability [4, 9]. However, due to the expensive and rare properties of Pt, large-scale commercialization is unrealistic. Therefore, it is important to exploit low-cost, efficient, large-scale commercial, and stable alternative Pt-free CE materials.

In the past years, researchers have made great efforts to find alternatives to Pt, such as alloys, metal oxides, and sulfur compounds, and to achieve excellent catalytic activity and conductivity of DSSCs via modulating their crystal structures, morphologies, or chemical compositions [10-13]. Metallic sulfur compounds with ultrathin nanostructures have received considerable attention due to their short transport paths and high surface area in chemical/physical processes [14]. In recent years, antimony sulfide (Sb₂S₃) has been applied in DSSCs due to its excellent light absorption coefficient (up to 10⁵ cm^‒1) [15]. For example, Yong et al. [16] reported that Sb₂S₃-based solar cells constructed on the sensitized device structure achieved a high PCE of 7.5%. Sun and co-workers [17] synthesized a hybrid catalyst composed of mesoporous carbon nanoparticle films and Sb₂S₃ , and its PCE reached 6.69%. Sb₂S₃ is considered one of the most promising CE materials due to its 1.5-2.2 eV gap width, which covers a large proportion of the visible spectrum [18, 19]. Besides, Sb₂S₃ is the main component of nontoxic antimony minerals, which are abundant in the earth [20]; they are also applied in other fields as solar energy conversion materials [21], catalysts [22], and photoconductive detectors [23]. Generally, graphene possesses a large specific surface area, remarkable carrier mobility, and excellent thermal/chemical stability, which makes it a suitable candidate for improving the electron transfer efficiency and electrocatalytic activity [24-28]. In addition, the introduction of graphene into the CEs of DSSCs effectively raises the electron transfer of the TiO₂ conductive band to the external circuit, decreases the charge recombination, and enhances the conversion efficiency [29, 30]. Based on the aforementioned characteristics, graphene is considered a perfect backing material for growing the nanomaterials applied in the CEs of DSSCs [31].

In this study, we fabricated uniform flower-like Sb₂S₃ via an uncomplicated procedure after adjusting the experimental conditions (Sb source and pH value). Afterward, with the addition of graphene in the synthesis, the morphology of the products changed from a flower-like structure to nanosheets, and the Sb₂S₃ nanosheets on the reduced graphene oxide (Sb₂S₃@RGO) were synthesized. We assessed the photovoltaic property and electrocatalytic activity of the as-synthesized samples used as CEs of DSSCs. It was confirmed experimentally that Sb₂S₃@RGO exhibited an excellent PCE of 8.17%, which was better than those of the flower-like Sb₂S₃ (7.38%), RGO (2.66%), and standard Pt CEs (7.75%). We provide a convenient approach to synthesizing an alternative material for Pt in the field of DSSCs. To the best of our knowledge, this is the first time that Sb₂S₃ nanosheets on RGO have been applied as a CE in DSSCs.

GO nanosheets were prepared according to the modified Hummers method [32]. Briefly, 2 g of graphite powder was added to 100 mL of concentrated H₂SO₄ at 0 ℃. Gradually, 6 g of KMnO₄ was added to the above mixture. Subsequently, the obtained suspension was continuously stirred for 15 h at 35 ℃. Thereafter, 200 mL of distilled water was gradually added, and the temperature was maintained at 98 ℃ for 2 h. When the temperature dropped to 60 ℃, 10 mL of H₂O₂ (30%) was added into the suspension in a dropwise manner to thoroughly eliminate the additional KMnO₄. After the above process, a bright yellow mixture was obtained, which was centrifuged and washed with conductivity water until the supernatant was neutral. The final precipitate was GO. The GO powder was obtained by ultrasonication of the GO precipitates dispersed in water, followed by drying for 10 h at 50 ℃.

The GO powder obtained above was uniformly dispersed in ethylene glycol under ultrasound, and the final concentration was 0.1 mg/mL; thereafter, 0.228 g of antimony trichloride (SbCl₃) was added to 20 mL of ethylene glycol under stirring to form a clear solution. Afterward, 0.485 g of L-cysteine and 20 mL of ethylene glycol containing graphene were successively added into the solution. After 10 min of stirring, 0.4 g of SDS (sodium dodecyl sulfate) was added as the template. The pH of the reaction system was adjusted to 10 with a 10 mol/L NaOH aqueous solution. Subsequently, the well-mixed liquid was transferred into a Teflon-lined autoclave and kept at 150 ℃ for 24 h. When the temperature of the autoclave was lowered to room temperature, the precipitate obtained by centrifugation was washed several times with deionized water and ethanol. The obtained samples were desiccated under vacuum at 60 C all-night and then calcined at 200 ℃ for 2 h under argon protection.

The crystalline structures of the obtained samples were analyzed by X-ray diffraction (XRD, Rigaka D/max2500) with Cu K_α radiation (λ = 1.540 56 Å). Scanning electron microscopy (SEM, Nanosem 430, FEI) was utilized to measure the morphology of the samples, and transmission electron microscopy (TEM, Tecnai G2 F20, operating voltage of 200 kV, FEI) and energy-dispersion spectroscopy (EDS) were utilized to characterize the detailed microstructure information and element composition of the samples. The Brunauer-Emmett-Teller (BET) specific surface areas of the samples were studied by the BET equation in a Tristar 3000 nitrogen adsorption apparatus.

The mixture composed of 0.1 g of the sample and 0.025 g of PEG20000 (polyethylene glycol 20, 000) was continuously ground in ethanol to obtain a smooth mud. The obtained mud was coated on a clean fluorine-doped tin oxide (FTO) glass substrate with a doctor blade to form a film. The film was stabilized by drying at room temperature for one day in a vacuum oven. Subsequently, Sb₂S₃@RGO CEs were obtained from the conductive glass covered with film heated under argon protection at 400 ℃ for 2 h. For comparison, Sb₂S₃@RGO CE and RGO CE were prepared by the same process as above.

The TiO₂ film was fabricated with a commercial TiO₂ sol by the same procedure. The dye-sensitized TiO₂ electrodes were obtained by soaking the FTO glass plates with TiO₂ films in a solution of N719 dye (di-tetrabutylammonium cis-bis (isothiocyanato), bis(2, 2'-bipyridyl-4, 4'-dicarboxylato) ruthenium (Ⅱ)), and ethanol in the dark for one day. DSSCs were assembled in a cell composed of a TiO₂ photoanode, CE, and liquid electrolyte. Surlyn 1702 (a thermoplastic frame) was used as the interval between the two electrodes. To avert leakage of the electrolyte solution, paraffin wax acted as a sealant to clamp the two electrodes together. Other DSSCs were prepared with Sb₂S₃ or RGO, or Pt while other processes remained unchanged.

The Zahner IM6 electrochemical workstation was applied to measure all the electrochemical performance. The photocurrent-voltage curves were obtained with a Keithley digital source meter (Keithley 2410, USA). Cyclic voltammetry (CV) curves were recorded using a three-electrode system on the electrochemical station, in the range of -0.4-1.2 V, at a scan rate of 25 mVs^‒1.

Electrochemical impedance spectroscopy (EIS) testing was performed under the following conditions: zero bias potential, a frequency range of 0.1 Hz-1 MHz, and a 5 mV sinusoidal AC voltage signal. The obtained impedance spectra were analyzed with the A-view 2.0 software. The Tafel-polarization curves were measured in a symmetric virtual cell employed for the EIS experiments, using the same electrolyte as in the DSSCs. The voltage range was −1.0-1.0 V, and the scan rate was 20 mVs^-1.

The morphology of the Sb₂S₃ sample changes with the pH value in the experimental process when the Sb source is invariable. In Fig. S1, antimony triacetate is employed as the Sb source. Fig. S1(a) shows that the Sb₂S₃ morphology is identical to an integrated nanorod structure (pH = 3). When the pH is 6, the morphology of the sample is an irregular sphere (Fig. S1b). Nanosheet structure begins to appear when the pH is 8 (Fig. S1(c)), although there is no uniform structure with nanosheets in Fig. S1(d) (pH = 10). According to the XRD analysis (Fig. S1(e)), the diffraction peaks of the sample are consistent with that of the standard card (JCPDS card No.42-1393) only when the pH is 3. Antimony chloride is used as the Sb source in Fig. S2. Similarly, the morphology of the sample changed from the irregular rod-like shape (pH = 3, Fig. S2(a)) to nanospheres (pH = 6, Fig. S2(b)), followed by nanosheets structure (pH = 8, Fig. S2(c)). Differently, the nanosheets form uniform flower-like structures when the pH is 10 (Fig. S2d). The results of XRD analysis show that the diffraction peaks of samples match the values well in the standard card (JCPDS card No. 42-1393), except for the sample obtained at a pH of 3 (Fig. S2(e)). The seriousness of the results indicates that the Sb source and alkaline environmental conditions of the synthesis are indispensable in forming Sb₂S₃ with a uniform flower-like structure composed of nanosheets.

Subsequently, graphene was introduced in the synthesis, and Sb₂S₃@RGO composites with a nanosheet structure (Fig. 1(a)) were successfully synthesized using antimony chloride as the source of Sb at a pH of 10; other conditions remain unchanged. Fig. 1(b) shows that the graphene nanosheet is considerably visible in the Sb₂S₃@RGO composites. In the inset of Fig. 1(b), the lattice spacing of 0.358 nm is consistent with the d-spacing between the adjacent (130) crystallographic planes of the Sb₂S₃@RGO composite. The elemental mapping (EM) of the sample was used to examine the compositional uniformity of Sb₂S₃@RGO. Fig. 1(c) demonstrates that the C, S, and Sb elements of the Sb₂S₃@RGO composites were uniformly distributed. The diffraction peaks of Sb₂S₃@RGO are noticeable, clear, and in agreement with the values from the standard card (JCPDS card No. 42-1393) (Fig. 2(a)), and there are no characteristic peaks of other impurities. The elemental composition of the as-fabricated Sb₂S₃@RGO composite is further analyzed by EDS. Fig. 2(b) shows that there are C, S, and Sb elements in the Sb₂S₃@RGO composite, and C originates from graphene.

Figs. 3(a) and 3(b), respectively, display the aperture structure and pore diameter distribution of Sb₂S₃ nanospheres and Sb₂S₃@RGO nanosheets, analyzed by the nitrogen sorption method. The two isotherms have similar hysteresis loops and both belong to the type Ⅳ sorption isotherm. Using BET calculation, the specific surface of Sb₂S₃ is ~41.72 m²g^‒1, while that of the Sb₂S₃@RGO composites is ~44.53 m²g^‒1. The above results can be ascribed to the addition of RGO, which provided more active surface sites and increased the adsorption of the reactants owing to its large surface area [33]. According to the results from the Barrett-Joyner-Halenda (BJH) pore-size distribution test, the average pore diameters of both samples are 31.08 and 22.65 nm, respectively. The results indicate that both Sb₂S₃ and the Sb₂S₃@RGO composites possess a mesoporous structure, which provides vast access to the internal electrocatalyst, thereby improving the catalytic capacity of CE and promoting the electrochemical reaction [34].

EIS is utilized to assess the electrochemical properties of the electrode material, which can reflect information about the impedance from the chemical reaction kinetics between the electrode interface and the electrolyte [35]. The EIS curves of various CEs are displayed in Fig. 4, and the matching data calculated from the results are shown in Table 1. The ohmic series resistance (Rs) is directly connected with the electrolyte property and CE sheet resistance, acquired from the X-intercept in the high-frequency region of EIS. The R_s values of the Sb₂S₃@RGO composites, Sb₂S₃, RGO, and Pt are 10.75, 11.12, 12.28, and 11.5 Ω, respectively. The R_s of the Sb₂S₃@RGO composites is lower than that of Pt, suggesting the relatively high conductivity of the composites. The charge transfer resistance (R_ct) is crucial to the electron exchange between the CE and the electrolytes [36]. The lesser R_ct value of the Sb₂S₃@RGO composites (1.099 Ω) compared to those of Sb₂S₃ (1.40 Ω), RGO (7.88 Ω), and Pt (1.28 Ω) indicates the better catalytic activity of the CEs and a high PCE of the DSSCs combined with these CEs [37]. The above results indicate that the Sb₂S₃@RGO composites possess an outstanding electrocatalytic ability in the reaction of I₃^- to I^-.

CV was carried out with a three-electrode system to evaluate the electrocatalytic performance of Sb₂S₃@RGO. In addition, the electrocatalytic properties of Sb₂S₃, RGO electrode, and Pt electrode were detected under the same test conditions. All the CV curves (-0.3-1.65 V) have two pairs of redox peaks; the left pair (Ox₁/Red₁) is attributed to the reaction of I₃^-/I^-, and the right one (Ox₁/Red₁) belongs to the reaction of I₂/I₃^- (Fig. 5). The peak-to-peak separation (E_pp) and peak current density (J_A) are two vital data for assessing the electrochemical property of the catalytic electrodes [38]. From Table 1, the relatively high J_A and low E_pp values belong to the Sb₂S₃@RGO CE, indicating the higher electrocatalytic activity than that of the standard Pt CE [39]. In addition, the results show that the Sb₂S₃@RGO CE presents a higher electrocatalytic performance than that of the pure Sb₂S₃ CE, which suggests that adding RGO into the Sb₂S₃ sample improved the electrocatalytic activity. To investigate the electrochemical invertibility and stability of the Sb₂S₃@RGO CE, successive CV experiments were carried out at a scan rate of 25 mVs^‒1. There is a negligible change in E_pp and J_A, illustrating that the Sb₂S₃@RGO CE possesses excellent cyclic reversibility and stability (Fig. 6).

Tafel polarization testing was carried out to further analyze the electrocatalytic properties of the samples. Fig. 7 displays four Tafel curves of the cells based on Sb₂S₃@RGO, Sb₂S₃, RGO, and Pt. The limiting diffusion current density (J_lim) refers to the junction of the Y-axis with the cathodic branch, which is closely related to the diffusion rate of the I₃^-/I^- redox couples [40]. The exchange current density (J₀) is acquired from the slope of the anodic or cathodic branches by linear extrapolation [41]. The value of J₀ is positively correlated with the catalytic activity of the cathode for I₃^- [42]. As shown in Table 1, the largest values of both log J_lim (2.12 mAcm^‒2) and log J₀ (1.16 mAcm^‒2) are generated from the Sb₂S₃@RGO composite CEs, which indicates that the Sb₂S₃@RGO nanosheets possess a superior electrocatalytic property to that of the Pt CEs. Additionally, the results of the Tafel polarization testing are in line with the outcome of the CV.

The photocurrent density-voltage (J-V) curves of the DSSCs based on the Sb₂S₃@RGO composites, Sb₂S₃, RGO, and Pt are displayed in Fig. 8, and the related test parameters are summarized in Table 2. From Fig. 8, we can see that the DSSCs assembled with the Sb₂S₃@RGO composites possess a PCE of 8.17%, which is higher than those of the DSSC models with Pt CEs (7.75%), RGO CEs (2.66%), and Sb₂S₃ CEs (7.38%). The detailed corresponding test data in Table 2 shows that the Sb₂S₃@RGO CE exhibited a higher fill factor (FF = 0.68) and short-circuit current (J_sc = 16.23 mAcm^‒2) than those of the Pt CE (FF = 0.66, J_sc = 15.87 mAcm^‒2). The high catalytic activity for I₃^- reduction led to the enhanced J_sc value [35] and the higher FF value benefitted from the relatively small series resistance (R_s) in the EIS experiment [31], which eventually led to the superior PCE value of the DSSC based on Sb₂S₃@RGO. The result of the Sb₂S₃@RGO CE from photoelectric conversion efficiency testing matches the results of CV, polarization, and EIS. All the above results have verified its high-efficiency electrocatalytic performance.

In summary, by changing the Sb source and pH value as well as introducing graphene, we successfully synthesized Sb₂S₃@RGO composites with a nanosheet structure by a facile and economical hydrothermal method. Based on the electrochemical experimental results, we summarize that the composites, as CE materials, displayed excellent electrochemical and photovoltaic properties applied in DSSCs. Notably, a higher PCE (8.17%) of the Sb₂S₃@RGO composites was achieved, compared with that of the standard Pt (7.75%). The obtained results also indicated that the addition of RGO not only changed the morphology but also improved the electrocatalytic performance of the Sb₂S₃@RGO composites when compared with that of the rare Sb₂S₃. We believe the current work could serve as a reference and provide guidance for designing high-performance DSSCs.

Acknowledgments

This work was funded by the Tianjin science and technology support key projects (18YFZCSF00500), and the National Science Fund for Distinguished Young Scholars (21425729) from the National Natural Science Foundation of China.