Synthesis of MoIn2S4@CNTs Composite Counter Electrode for Dye-Sensitized Solar Cells

A ternary and composite MoIn2S4@CNTs counter electrode (CE) with a hedgehog ball structure was synthesized by using a facile one-step hydrothermal method. The composite MoIn2S4@CNTs film possesses large specific surface area through N2 adsorption-desorption isotherms test, which is advantageous to adsorb more electrolyte and provide larger active contact area for the electrode. In addition, the composite MoIn2S4@CNTs CE exhibits low charge transfer resistance and fine electrocatalytic ability made from a series of electrochemical tests including cyclic voltammetry, electrochemical impedance, and Tafel curves. Under optimal conditions, the DSSC based on the MoIn2S4@CNTs-2 composite CE achieves an impressive power conversion efficiency as high as 8.38%, which remarkably exceeds that of the DSSCs with the MoIn2S4 CE (7.44%) and the Pt electrode (8.01%). The current work provides a simplified preparation process for the DSSCs.


Background
In recent decades, it is urgent to exploit and utilize renewable energy substituting the conventional fossil fuels with the severe energy shortage and environmental degradation increasing [1,2]. Dye-sensitized solar cell (DSSC) has attracted wide-spread research by virtue of its environmental friendliness, facile preparation process, brilliant photovoltaic performance, and so on [3,4]. The counter electrode (CE), as one of the significant components of a DSSC, plays the role of gathering electrons from external circuit and catalyzing the reduction reaction of I 3 − to I − in the liquid electrolyte [5,6]. Generally, ideal CE materials contain the merits of high electrical conductivity and remarkable catalytic activity. However, platinum (Pt) as a prevalent and efficient CE material is confined to large-scale commercialized application because of the major weaknesses of scarcity, expensiveness, and poorly long-term stability [7,8]. Therefore, a great many of efforts have been made to develop easily accessible, cost-effective, and Pt-like catalytic activity applied in DSSC for years [9].
Up to now, various kinds of outstanding alternative materials have been proposed, such as carbonaceous materials [10,11], transition metal chalcogenides [12], conductive polymers [13], metal alloys [14], and their compounds [15,16]. Among them, the binary transition metal chalcogenides attracted a great deal of attention due to their unique structure and chemical properties. For instance, the synthesized MoS 2 on FTO substrate exhibits a sandwich-layered structure, larger surface area, and more active edge sites leading to extremely photoelectric performance as a CE for DSSC [17]. Meanwhile, extensive research work focused on the catalytic activity for I 3 − reduction also has been made for WS 2 [18], FeS 2 [19], CoS [20], and NiS 2 [21], which were comparable to or even better than that of the Pt electrode. Nevertheless, the inherent characteristic of these materials, such as low electrical conductivity and only two-fixed chemical compositions, hampered further improvement in their catalytic activity [22]. Hence, numerous methods aimed to overcome the abovementioned shortcoming were adopted to synthesize multinary transition metal chalcogenides by component elements adjusting, structure designing, and morphology tailoring. Fortunately, considerable multinary transition metal chalcogenides have achieved significant enhancements in catalytic ability for DSSCs, such as NiCo 2 S 4 [23], MIn 2 S 4 (M = Fe, Co, Ni) [22], CuInS 2 [24], CoCuWS x [25], and Ag 8 GeS 6 [26], which catalytic ability are obviously much better than that of their binary counterparts.
Moreover, it is widely acknowledged that carbon nanotubes (CNTs) exhibit considerably novel characteristics of large specific surface area, superb electrical conductivity, high mechanical strength, and photochemical stability, which are widely used in the synthesis and modification of other materials [27]. Regrettably, CNTs show poor electrocatalytic activity for I 3 − reduction, which greatly limits their application independently in a DSSC device. Fortunately, a large number of studies have shown that the composite CE modified with CNTs all obtained greatly improved photoelectric performance for DSSCs [18,28,29]. Liu et al. have reported a flowerlike hierarchical structure of Cu 2 MnSnS 4 /CNT (CMTS/ CNT) CE via solvothermal method in DSSC gained a photoelectric conversion efficiency of 8.97%, much higher than that of the DSSCs with CMTS (6.21%) and Pt (8.37%) CEs [29].
Based on the above considerations, in this study, a MoIn 2 S 4 @CNT composite CE of DSSC with hedgehog ball structure was synthesized by using a facile one-step hydrothermal method and expected to improve higher device performance. Scanning electron microscope results show that different CNT contents result in visible changes on morphology. According to a series of electrochemical characterizations including cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), and Tafel curves tests, the MoIn 2 S 4 @CNTs CE indicates a remarkable catalytic activity and fine charge transfer resistance.
The DSSC assembled with the MoIn 2 S 4 @CNTs CE with suitable content achieves a superior power conversion efficiency of 8.38%, which is better than that of the DSSC based on the Pt CE (8.01%).

Preparation of porous TiO 2 photoanodes
The colloid of TiO 2 was prepared as our previous work [30]. The dye-sensitized TiO 2 photoanodes were fabricated as follows: Firstly, 3-M tape (50 μm thick) with exposed area of 0.283 cm 2 was attached on FTO. Subsequently, the as-prepared TiO 2 colloid was coated on by using a blade-coating method. Secondly, the drying TiO 2 electrode was sintered at 450°C for 30 min in muffle furnace. Afterwards, the TiO 2 electrode was immersed in a 40-mM titanium tetrachloride (TiCl 4 ) aqueous solution at 70°C for 30 min, and then annealed in air at 450°C for 30 min. After cooling down to room temperature, the TiO 2 electrode was immersed in dye Z907 (0.3 mM) absolute ethanol solution for 24 h to adsorb sufficient dyes and obtained the resultant dyesensitized TiO 2 photoanode.

Fabrication of ternary MoIn 2 S 4 @CNTs CE
The MoIn 2 S 4 thin films were grown directly onto FTO substrates by a simple approach referring to our previous report [31]. In a typical preparation, 0.0696 g Na 2 MoO 4 ·2H 2 O, 0.169 g InCl 3 ·4H 2 O, and 0.1394 g TAA were diffused in 30 ml deionized water under ultrasonicating for 2 h until all the reactants were dissolved. The pre-cleaned FTO substrates were put into a 100-ml Teflon-lined stainless steel autoclave with the conductive side facing up before transferring the above precursor mixture into it. After being sealed, the autoclave was placed in an oven and heated under 200°C for a reaction time of 15 h. The FTO glass substrates covered with MoIn 2 S 4 materials were taken out from the autoclave, washed with ethanol, deionized water, and then dried in air under 60°C for 12 h.
In order to study the impact of the CNT contents on the fabricated composite CE and the performance of the DSSC, different contents of CNTs adding to the precursor were conducted, including three samples in which the amount of CNTs were 10, 20, and 30 mg, respectively, keeping the other reagents and fabrication processes unchanged. The above samples were marked MoIn 2 S 4 (0 mg), MoIn 2 S 4 @CNTs-1 (10 mg), MoIn 2 S 4 @CNTs-2 (20 mg), and MoIn 2 S 4 @CNTs-3 (30 mg).
For comparison, a pyrolysised Pt CE was employed as the reference CE. The H 2 PtCl 6 in isopropanol solution (0.50 wt%) was dropped onto the surface of the FTO glass, and then sintered at 450°C in muffle furnace for 30 min to fabricate the Pt CE.

Fabrication of the DSSCs
The DSSCs with sandwich structure were constructed by clipping the sample CEs (including MoIn 2 S 4 , various MoIn 2 S 4 @CNTs, and Pt CEs) together with as-prepared dye-sensitized TiO 2 photoanode. Surlyn was used as spacer between the electrodes and followed by filling the interspace with liquid redox electrolyte that consisted of 0.60 M tetrabutylammonium iodide, 0.10 M lithium iodide, 0.05 M iodine, and 0.50 M 4-tert-butyl-pyridine acetonitrile solution.

Characterizations
Chemical element composition of the samples was characterized by using X-ray photoelectron spectroscopy (XPS) analysis (Kratos Axis Ultra). The morphological features of samples were observed by field emission scanning electron microscopy (FESEM, JSM-7001F). BET-specific surface area method was employed using a JW-K analyser by nitrogen absorption to test the surface area and pore size distribution. The other relevant electrochemical properties were researched by a CHI660E electrochemical workstation. The electrochemical impedance spectroscopy (EIS) was performed in a frequency range of 0.1-10 5 Hz with a disturbed amplitude of 5 mV. The photovoltaic performances of the DSSCs were carried out by measuring current density-voltage (J-V) characteristic curves under irradiation of 100 mW cm −2 from the solar simulator (CEL-S500, Beijing China Education Au-light Co., Ltd).

Composition and Morphology
The XPS is employed to examine the surface compositions and chemical states of each element in MoIn 2 S 4 and MoIn 2 S 4 @CNTs films. The spectrum of survey data in MoIn 2 S 4 and MoIn 2 S 4 @CNTs-2 samples is demonstrated in Fig. 1a to verify the presence of Mo, In, S, and C (as a reference) elements. In addition, all spectra of the four samples, calibrated by the C 1s peak at the binding energy of 284.6 eV [17], are analyzed via the Gaussian fitting method showing in Fig. 1b and c. The C 1s peak appeared in MoIn 2 S 4 sample is well known to have originated from the adventitious carbon caused by exposure to the air. In the Mo 3d region, two major peaks at 228.8 and 232.1 eV are assigned to Mo 3d5/2 and Mo 3d3/2 of the MoS 2 [32], respectively. This result confirms that Mo element is in its IV oxidation state, which is reduced to Mo 4+ (MoIn 2 S 4 ) from Mo 6+ (Na 2 MoO 4 ) [33]. The doublet peaks at binding energies of 445.2 and 452.2 eV correspond to In 3d5/2 and In 3d3/2 [24,34]. As for the XPS spectra of S 2p, the peaks located at 161.8 and 163.1 eV belong to S 2p3/2 and . The above results are well in agreement with our previous studies [31]. Moreover, no other element or extra peak is found in the survey, which reconfirms that the samples synthesized have the similar chemical compositions and structure.
The surface morphologies of as-prepared MoIn 2 S 4 and MoIn 2 S 4 @CNTs nanofilms are observed by SEM images in Fig. 2. In Fig. 2a, MoIn 2 S 4 sample displays a petal-like nanosheet structure with a uniform, smooth, and dense surface. Unlike MoIn 2 S 4 nanofilms, the hedgehog ball structure is found in MoIn 2 S 4 @CNT samples in Fig. 2b-d, and the average diameter of the MoIn 2 S 4 @CNTs nanospheres is around 890 nm. It is easy to observe that so many network nanosheets are fully grown on the FTO substrate. Compared to the MoIn 2 S 4 @CNT samples with low (MoIn 2 S 4 @CNTs-1) and high (MoIn 2 S 4 @CNTs-3) CNT contents, the sample with moderate content (MoIn 2 S 4 @CNTs-2) exhibits more hedgehog balls and nanosheets array on the network of the MoIn 2 S 4 . A good contact performance between the network nanosheet array structure with large number hedgehog balls on FTO substrate facilitates the reduction of I 3 − attributed to its good conductivity and vast catalytic active sites, which can be predicted that the MoIn 2 S 4 @CNTs-2 CE will achieve better performance than that of the MoIn 2 S 4 , MoIn 2 S 4 @CNTs-1, and MoIn 2 S 4 @CNTs-3 CEs. Moreover, dissimilar morphologies between MoIn 2 S 4 and MoIn 2 S 4 @CNTs indicate that CNTs play a pivotal role in controlling morphology of the samples. Figure 3 shows the XRD patterns of various samples. Among them, the diffraction peaks at 25 [33]. From the MoIn 2 S 4 @CNTs-2 and MoIn 2 S 4 samples, the peaks of the above discussed appear well in the both samples. Compared to the MoIn 2 S 4 sample, the peak for the CNTs at 25.74°can be seen clearly in MoIn 2 S 4 @CNTs-2 XRD patterns. Thus, it can be inferred that the MoIn 2 S 4 @CNTs-2 materials are synthesized successfully and there are no impurities introduction. N 2 adsorption-desorption isotherms are measured and showed in Fig. 4 to explore the specific surface areas and pore features. Generally, the larger specific surface area facilitates more convenient charge transmission on the CE/electrolyte interface [35]. It demonstrates from Fig. 4 that the samples possess an evident hysteresis loop of type IV adsorption-desorption behavior, and their corresponding data calculated from the Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) method are tabulated in Table 1. By comparison, it is easy to find that the specific surface area and average pore diameter of the MoIn 2 S 4 @CNT samples are much better than that of the MoIn 2 S 4 . Among the three MoIn 2 S 4 @CNT samples, MoIn 2 S 4 @CNTs-2 shows the largest specific surface area of 66.80 m 2 g −1 and the smallest average pore diameter of 17.8 nm, which can be ascribed to the excellent hedgehog ball structure after doping of moderate CNTs. It is reasonable to believe that the MoIn 2 S 4 @CNTs-2 CE will obtain the fine catalytic activity and hence achieve highly efficient device performance.

Electrochemical properties
CV measurements are carried out to study the electrocatalytic behavior of the as-obtained samples in the potential range of − 0.6 to 1.0 V at a scan rate of 60 mV s −1 for the MoIn 2 S 4 and MoIn 2 S 4 @CNTs CEs, and the calculated values are summed up in Table 1. The left peaks of the two pairs of oxidation and reduction peaks in each CV curve in Fig. 5a are ascribed to the equation , which determines the performance of electrocatalytic activity of the CE materials, especially in DSSCs [23,36]. The values of the negative reduction peak current density (J pc ), a key parameter in CV test, follows the orders of Pt (3.80 mA cm −2 ) < MoIn 2 S 4 (4.31 mA cm −2 ) < MoIn 2 S 4 @CNTs-1 (4.68 mA cm −2 ) < MoIn 2 S 4 @CNTs-3 (5.09 mA cm −2 ) < MoIn 2 S 4 @CNTs-2 (7.47 mA cm −2 ). Evidently, the MoIn 2 S 4 CE itself has good catalytic activity, and the MoIn 2 S 4 @CNTs-2 CE exhibits much higher J pc than that of the Pt, MoIn 2 S 4 CEs, and the other two kinds of MoIn 2 S 4 @CNTs CEs attributed to its distinctive surface morphology, the synergistic effect of CNTs doped, and the larger surface area. The results indicate that the MoIn 2 S 4 @CNTs-2 CE has fine electrocatalytic activity for the I − /I 3 − redox couple in DSSC CEs. Figure 5b shows the CV curves of the MoIn 2 S 4 @CNTs-2 CE at the scan rate of 60 mV s −1 Fig. 5 a CVs of the various CEs at a scan rate of 60 mV s −1 . b 50 cycles CV of the MoIn 2 S 4 @CNTs-2 CE Fig. 6 CVs of the MoIn 2 S 4 @CNTs CEs at different scan rates a MoIn 2 S 4 @CNTs-1, b MoIn 2 S 4@ CNTs-2, c MoIn 2 S 4 @CNTs-3, and d the influences of the scan rate on the redox peaks current density and they are almost no changing with 50 cycles, which indicates that the MoIn 2 S 4 @CNTs-2 CE possesses excellent electrochemical stability. Figure 6a-c show the CVs of MoIn 2 S 4 @CNTs CEs at different scan rates. With the scan rates increasing from 20 to 120 mV s −1 , the oxidation and reduction peaks shift towards positive and negative direction owing to the fast diffusions of I − /I 3 − redox couple on the surfaces of CEs and the large electrochemical polarization [28]. Furthermore, Fig. 6d shows the relationship between the anodic and cathodic peaks current densities of the left peak pairs against the square root of sweep rates. The well-fitted linear relations indicate that the redox reaction of I − /I 3 − is dominated by diffusion-controlled ion transport [22,25]. Figure 7a shows the Nyquist plots of the different samples to further insight into the kinetics of the interfacial charge transfer process. The EIS data made from the fitted curves with a model equivalent circuit of the insert is listed in Table 2. Typically, the Nyquist plots contain two semicircles, the first semicircle on the left represents the charge transfer resistance R ct at the CE and electrolyte interface, and the second semicircle corresponds to the Nernst diffusion impedance in the electrolyte, while the intercept of the curve in the high-frequency region on the real axis is known as the series resistance R s . In general, R s and R ct are two vital parameters for evaluating the catalytic activity of CE in DSSC. A small R s indicates good contact between the catalyst and the substrate, and hence the resistance of the entire device is also small [29,37,38]. Meanwhile, little R ct stands for a high charge transfer rate. As presented in Table 2, the R s values of the MoIn 2 S 4 , MoIn 2 S 4 @CNTs-1, MoIn 2 S 4 @CNTs-2, and MoIn 2 S 4 @CNTs-3 CEs are 24.77, 23.16, 18.96, and 19.58 Ω cm 2 , respectively. Obviously, all the composite MoIn 2 S 4 @CNTs CEs have the smaller R s than that of the MoIn 2 S 4 CE, indicating that the conductivity of the MoIn 2 S 4 @CNTs CEs is enhanced after doping CNTs. Furthermore, among the four CEs, the trends of R ct is MoIn 2 S 4 > MoIn 2 S 4 @CNTs-1 > MoIn 2 S 4 @CNTs-3 > MoIn 2 S 4 @CNTs-2, which suggests an inverse order of electrochemical impedance and catalytic ability of the CEs. MoIn 2 S 4 @CNTs-2 CE possesses the lowest R ct value can be attributed to the synergy of the CNTs with fine conductivity and MoIn 2 S 4 with excellent catalytic ability, resulting in more effective reduction of triiodide on the CE/electrolyte interface. Apparently, the conductivity and catalytic ability of the MoIn 2 S 4 @CNTs composite CEs has been improved greatly than that of the MoIn 2 S 4 CE, and the result is completely consistent with the BET and CV tests.
The Tafel polarization curves of the various CEs are measured as shown in Fig. 7b, and the corresponding parameters values are summarized in Table 2. Normally, a standard Tafel curve includes two significant parameters named exchange current density (J 0 ) and limiting diffusion current density (J lim ). J 0 is related to the catalytic reduction reaction. The larger the J 0 is, the better the catalytic effect. J lim is also positively related to the diffusion efficiency of electrolyte. The larger J lim indicates the faster diffusion of I 3 − ions [29,37]. As presented in Fig. 7b and Table 2, the J lim and J 0 are all in  orders of Pt < MoIn 2 S 4 < MoIn 2 S 4 @CNTs-1 < MoIn 2 S 4 @CNTs-3 < MoIn 2 S 4 @CNTs-2, suggesting that the catalytic activity of the MoIn 2 S 4 @CNTs-2 CE has enhanced greatly after doping CNTs. Among above CEs, the MoIn 2 S 4 @CNTs-2 CE gains the best catalytic activity compared to the others. The largest J lim and J 0 of the MoIn 2 S 4 @CNTs-2 CE can be ascribed to its large specific surface area made from hedgehog ball structure and enhanced conductivity by CNTs doping.

Photovoltaic performance of the DSSCs
For contrast, the DSSCs with the MoIn 2 S 4 , MoIn 2 S 4 @CNTs, and Pt CEs are prepared with the uniform photoanodes and electrolyte. The J-V characteristic curves are measured under 1 sun (AM 1.5 G, 100 mW cm −2 ) and the corresponding photovoltaic parameters values are listed in Table 3. Four key parameters including short-circuit current density (J sc ), open-circuit voltage (V oc ), fill factor (FF), and power conversion efficiency (η) are usually adopted to assess the photovoltaic performance of the DSSCs. The FF and η of the DSSCs are calculated according to the Eqs. (1) and (2): where P in is the incident light power and J max (mA cm -2 ) and V max (V) are the current density and voltage at the point of maximum power output in the J-V curves, respectively.
As can be seen from Fig. 8, the DSSCs with MoIn 2 S 4 and CNTs CEs have power conversion efficiency of 7.44% and 3.62%. Compared to the DSSCs with the MoIn 2 S 4 and CNTs CEs, the DSSCs assembled with the three MoIn 2 S 4 @CNTs CEs exhibit improvement J sc and η values. Moreover, the J sc and η values of the DSSCs based on the MoIn 2 S 4 @CNTs CEs increase with the CNT contents increasing from 10 to 20 mg. While further increases the CNT content to 30 mg, it results in a slight descent for the J sc and η. Compared with the Ptbased DSSC, all the three MoIn 2 S 4 @CNTs-based DSSCs exhibit improved η of 8.16%, 8.31%, and 8.38%, which are higher than that of the Pt-based DSSC (η of 8.01%) under the same condition. Especially, the DSSC assembled with the MoIn 2 S 4 @CNTs-2 CE shows the best photovoltaic performance and achieves a η of 8.38%, and its corresponding J sc of 17.17 mA cm −2 , V oc of 0.745 V, and FF of 0.655. The improved photoelectric property of the DSSC with the MoIn 2 S 4 @CNTs-2 CE was due to the synergetic effect of the CNTs and the MoIn 2 S 4 .

Conclusions
Ternary MoIn 2 S 4 and MoIn 2 S 4 @CNTs counter electrodes are fabricated on FTO substrate by using a facile one-step hydrothermal method and served in DSSCs. Under optimal conditions, the DSSCs based on the MoIn 2 S 4 @CNTs CEs all achieve good power conversion efficiency. Especially, the DSSC with the MoIn 2 S 4 @CNTs-2 composite CE exhibits a good power conversion efficiency of 8.38%, which is much higher than that of the DSSCs with the MoIn 2 S 4 CE (7.44%) and the Pt electrode (8.01%). The enhanced photoelectric property of the DSSC with the MoIn 2 S 4 @CNTs-2 CE was due to the synergetic effect of the CNTs and the MoIn 2 S 4 . Meanwhile, the synergetic effect of the MoIn 2 S 4 @CNTs CE in electrochemical performance has confirmed from a series of electrochemical test including cyclic voltammetry, electrochemical impedance, and Tafel curves. The composite MoIn 2 S 4 @CNTs film possesses large specific surface area through N 2 adsorptiondesorption isotherms test, which is advantageous to adsorb more electrolyte and provide larger active contact area for the electrode. The facts of the MoIn 2 S 4 @CNTs CE served in DSSC broadens the potential applications of transition metal complex semiconductors in the field of optoelectronic chemistry.