Self-Assembly Synthesis of the MoS2/PtCo Alloy Counter Electrodes for High-Efficiency and Stable Low-Cost Dye-Sensitized Solar Cells

In this work, MoS2 microspheres/PtCo-alloy nanoparticles (MoS2/PtCo-alloy NPs) were composited via a novel and facile process which MoS2 is functionalized by poly (N-vinyl-2-pyrrolidone) (PVP) and self-assembled with PtCo-alloy NPs. This new composite shows excellent electrocatalytic activity and great potential for dye-sensitized solar cells (DSSCs) as a counter electrode (CE) material. Benefiting from heterostructure and synergistic effects, the MoS2/PtCo-alloy NPs exhibit high electrocatalytic activity, low charge-transfer resistance and stability in the cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) test. Meanwhile, a high power-conversion efficiency (PCE) of 8.46% is achieved in DSSCs with MoS2/PtCo-alloy NP CEs, which are comparable to traditional Pt CEs (8.45%). This novel composite provides a new high-performance, stable and cheap choice for CEs in DSSCs.


Introduction
DSSCs have been regarded as one of possible alternatives to silicon photovoltaics, owing to their sustainable energy and competent power-conversion efficiency (PCE), low cost, facile fabrication process and environmental friendliness [1,2]. Normally, a sandwich structure of DSSCs consists of a photoanode sensitized by dye sensitizers, counter electrodes and electrolytes containing triiodide/iodide (I 3 − /I − ).
There have been extensive studies of the essential components of DSSCs, such as photoanodes [3,4], dye sensitizers [5][6][7], electrolytes [8,9] and counter electrodes (CEs) [10,11]. Among them-as the catalyzers of redox couple and electrodes-CEs are expected to have imperative characters, such as high electrocatalytic activity, good electron-conduction ability and stability [12,13]. Platinum (Pt) with excellent performance on conductivity and electrocatalytic activity has been widely used as CE materials [14]. However, the traditional iodine-based electrolyte can corrode Pt into PtI 4 on the surface of electrode [15,16] leading to relatively low PCE and instability of DSSCs. Moreover, the high price and high temperature annealing of Pt CEs limits the industrialization of DSSCs. To alternate titanium(IV) tetrachloride (TiCl 4 ), acetonitrile (C 2 H 3 N) and lithium perchlorate (LiClO 4 ) were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd. (Shanghai, China). Poly (N-vinyl-2-pyrrolidone) (PVP K30) was purchased from Beijing Solarbio Biologic Technology Co., Ltd.

Synthesis of MoS 2 Microspheres
The method of synthesizing MoS 2 microspheres by using micelle template is described previously [50]. First, the PVP solution was obtained by dissolving 0.4 g of PVP (K30) into 80 mL of deionized water and stirring 30 min. Then, 1.92 g of thiourea and 0.96 g of Na 2 MoO 4 ·2H 2 O were dissolved into as-prepared PVP solution by stirring 30 min. Next, the reaction solution was transferred into a 100-mL Teflon-lined stainless-steel autoclave and heated for 24 h at 200 • C. After the hydrothermal reaction finished, the black dispersion was collected by centrifugation at 7000 rpm for 15 min, wished two times with deionized water and dried at 60 • C in vacuum.

Synthesis of the MoS 2 /PtCo-Alloy NP Composite Material
The detailed strategy of fabricating the MoS 2 /PtCo-alloy NP composite is as follows 0.5 g of as-prepared MoS 2 was dispersed into ethylene glycol (EG) by stirring 30 min. Then, 0.1 g of PVP was dissolved into the dispersion. After stirring 30 min, a certain amount of H 2 PtCl 6 and CoCl 2 (molar ratio of H 2 PtCl 6 :CoCl 2 = 1:1) was added into the mixture with stirring and ultrasonic bath for 30 min. Then, a full amount of NaBH 4 was gradually added into the prepared mixture. After stirring for 5 h, the composite was extracted by centrifuging and washing the mixture. The obtained composite was finally dried at 60 • C for 12 h in vacuum.

Preparedness of Counter Electrodes
The fabrication of MoS 2 -based CEs and Pt CEs follows different methods. The preparation of MoS 2 -based CEs was adhered to the precedent research [25]. The as-prepared MoS 2 -based materials was mixed with carbon black and PVDF at the ratio of 8:1:1 (mass ratio). The mixture was dispersed into NMP. Moreover, the dispersion was grinded until the dispersion turned to slurry with appropriate viscosity. Then, the doctor blade method was used to coat the as-obtained slurry on FTO glass. In this method, prepared solution was dropped on FTO glass, then the doctor blade was used to scrape dropped solution to form smooth and uniform film on the surface of FTO. Finally, the obtained glass with solution film heated at 500 • C for 10 min with a heat gun, in this way, CE was fabricated successfully. The material-coated FTO counter electrodes were dried at 60 • C for 12 h in vacuum and were annealed at 500 • C for 2 h in argon.
Pt CEs as a matched group in our experiments were fabricated as a traditional method. Brief, 10-mM H 2 PtCl 6 in isopropanol was coated on FTO glass by doctor blade method and heated at 500 • C for 10 min with a heat gun.

Fabrication of DSSCs
The process of photoanodes and DSSCs fabrication was described like previously studies [70,71]. To begin with, a TiO 2 blocking layer was formed on FTO glass by soaking washed FTO glass into a 40-mM TiCl 4 aqueous solution and heating at 70 • C for 30 min. Afterwards, a TiO 2 layer was coated on the pretreated FTO glass by screen printing. The thickness of TiO 2 layer was controlled by the times of repeating the screen printing. Subsequently, the TiO 2 film was roasted at 450 • C for 30 min and natural cooled in air. When it cooled down to 120 • C, the anode was immersed into 0.5-mM N719 acetonitrile Nanomaterials 2020, 10, 1725 4 of 15 and tert-butanol (1/1, v/v) solution for 24 h. Next, as-prepared anodes and counter electrodes were attached to the sandwich structure by hot-pressing with Surlyn films. Then, the iodine-based electrolyte which is the acetonitrile solution composed of 0.60-mM DMII, 0.03-M I2, 0.10-M GNCS and 0.50-M TBP was injected into the sandwich structure.

Electrochemical and Photovoltaic Measurements
All the electrochemistry measurements were implemented on an electrochemistry workstation (CHI660e, Chenhua instruments Ins., Shanghai, China). The cyclic voltammetry (CV) was conducted to estimate the performance of counter electrodes. This measurement system consists of three electrodes (a Pt auxiliary electrode, an Ag/AgCl reference electrode and a 3 cm 2 of working electrode) and iodine-based electrolyte which is the acetonitrile solution containing 10-mM LiI, 1-mM I 2 and 0.1-M LiClO 4 . The electrochemistry impedance spectroscopy (EIS) was measured with symmetric dummy cells composed of double identical counter electrodes. Moreover, they were injected the electrolyte which was same as the DSSCs electrolyte in our experiments. The active area of dummy cells was controlled at 0.64 cm 2 .The frequency range of EIS was from 1 Hz to 10 5 Hz. Moreover, tests were at 0.01 V of bias voltage and in dark. The photocurrent density-voltage (J-V) curves of the DSSCs were measured by the electrochemistry workstation. The simulative solar light (100 mW·cm −2 , AM 1.5) was produced by a Xe lamp (cel-S500/350, Beijing China Education Au-light Co., Ltd., Beijing, China).

Results
To investigate the crystal structure of samples, a series of XRD tests were conducted. The XRD patterns of MoS 2 microspheres and MoS 2 /PtCo-alloy NP composites are shown in Figure 1.
Among different types of samples, MoS 2 /PVP means the fresh material which was synthesized by the PVP micelle-assisted hydrothermal route and covered by molecular chains of PVP. Moreover, annealed MoS 2 was heated at 500 • C for 2 h in argon to remove PVP which can influence the shape and peak positions of XRD patterns. In particular, featured peaks of MoS 2 at 14.02 • (002 planes) reflect the condition of MoS 2 layer stacking (strong peak indicates order stacking of MoS 2 layer) [72]. In Figure 1 the (002) peaks of fresh samples shift towards lower angles and that peak of the annealed sample returns to 14.02 • and is diminished. The reason of peak shift is molecular chains of PVP between MoS 2 layers increase the average spacing between layers. Moreover, the peak of annealed sample is weakened, because residual carbon produced by burning PVP turned closely the stacking structure of MoS 2 to the approximately single-layer structure. Meanwhile, the MoS 2 /PtCo-alloy NP composites which contain different amount of PtCo-alloy NPs are denoted as MoS 2 /5 wt% PtCo, MoS 2 /10 wt% PtCo and MoS 2 /20 wt% PtCo. The peaks of XRD patterns roughly accord with figures of hexagonal MoS 2 peaks (JCPDS card No.37-1492) and the peak shifts occur obviously on (002), (100) and (110) MoS 2 planes. Peak position shifts apparently associate with macromolecular chain of PVP adhering, because the peak positions of MoS 2 become more accordant for the standard hexagonal MoS 2 pattern after samples were removed PVP by annealing. The (111) diffraction peak of PtCo-alloy NPs is also exhibited in Figure 1. A distinct shift of the Pt (111) diffraction peak towards a higher degree owing to the shrinking lattices of PtCo. Pt atoms substituted by smaller Co atoms normally leads to contraction of PtCo lattices which coincides with previous results [73].

Results
To investigate the crystal structure of samples, a series of XRD tests were conducted. The XRD patterns of MoS2 microspheres and MoS2/PtCo-alloy NP composites are shown in Figure 1.  This structure is beneficial to improve the catalytic activity because more edges are exposed on the surface of MoS 2 microspheres. In the TEM images of MoS 2 microspheres, many molecular chains of PVP covered microspheres. The PVP-functionalized MoS 2 microspheres create the precondition for self-assembly of MoS 2 /PtCo-alloy NP composites. Among different types of samples, MoS2/PVP means the fresh material which was synthesized by the PVP micelle-assisted hydrothermal route and covered by molecular chains of PVP. Moreover, annealed MoS2 was heated at 500 °C for 2 h in argon to remove PVP which can influence the shape and peak positions of XRD patterns. In particular, featured peaks of MoS2 at 14.02° (002 planes) reflect the condition of MoS2 layer stacking (strong peak indicates order stacking of MoS2 layer) [72]. In Figure 1 the (002) peaks of fresh samples shift towards lower angles and that peak of the annealed sample returns to 14.02° and is diminished. The reason of peak shift is molecular chains of PVP between MoS2 layers increase the average spacing between layers. Moreover, the peak of annealed sample is weakened, because residual carbon produced by burning PVP turned closely the stacking structure of MoS2 to the approximately single-layer structure. Meanwhile, the MoS2/PtCo-alloy NP composites which contain different amount of PtCo-alloy NPs are denoted as MoS2/5 wt% PtCo, MoS2/10 wt% PtCo and MoS2/20 wt% PtCo. The peaks of XRD patterns roughly accord with figures of hexagonal MoS2 peaks (JCPDS card No.37-1492) and the peak shifts occur obviously on (002), (100) and (110) MoS2 planes. Peak position shifts apparently associate with macromolecular chain of PVP adhering, because the peak positions of MoS2 become more accordant for the standard hexagonal MoS2 pattern after samples were removed PVP by annealing. The (111) diffraction peak of PtCo-alloy NPs is also exhibited in Figure 1. A distinct shift of the Pt (111) diffraction peak towards a higher degree owing to the shrinking lattices of PtCo. Pt atoms substituted by smaller Co atoms normally leads to contraction of PtCo lattices which coincides with previous results [73].
FESEM and TEM images of MoS2 microspheres and MoS2/PtCo-alloy NP composites synthesized by different methods are shown in Figures 2 and 3, respectively. MoS2 microspheres which are composed by vertical MoS2 nanoflakes expose edges of nanoflakes on the surface of microspheres. This structure is beneficial to improve the catalytic activity because more edges are exposed on the surface of MoS2 microspheres. In the TEM images of MoS2 microspheres, many molecular chains of PVP covered microspheres. The PVP-functionalized MoS2 microspheres create the precondition for self-assembly of MoS2/PtCo-alloy NP composites.  To synthesize the MoS 2 /PtCo-alloy NP composites, three different routes were conducted in our experiments, including co-reducing PtCo-alloy NPs directly, adding PVP before co-reduction in aqueous and adding PVP before co-reduction in ethylene glycol ( Figure 3). Co-reducing PtCo-alloy NPs in the MoS 2 and PtCo precursors directly (Figure 3a,b) leads to less alloys loaded and obvious aggregation. To increase the amount of loaded alloys and reduce nanoparticles aggregation, we added PVP into the mixed dispersion before co-reduction (Figure 3c,d). PVP can keep noble metal nanoparticles from aggregation [74]. Moreover, functionalized PtCo-alloy NPs can trend to self-assemble on the surface of MoS 2 , because the hydrogen-bond networks which is established by hydrophilic groups of PVP can draw PtCo-alloy NPs on the surface of MoS 2 microspheres. Moreover, ethylene glycol as the solution of reduction reaction can improve reducing effectiveness and dispersibility [75]. Thus, more uniform, highly dispersed and vast PtCo-alloy NPs loaded on MoS 2 microspheres (Figure 3e,f).  To synthesize the MoS2/PtCo-alloy NP composites, three different routes were conducted in our experiments, including co-reducing PtCo-alloy NPs directly, adding PVP before co-reduction in aqueous and adding PVP before co-reduction in ethylene glycol ( Figure 3). Co-reducing PtCo-alloy NPs in the MoS2 and PtCo precursors directly (Figure 3a,b) leads to less alloys loaded and obvious aggregation. To increase the amount of loaded alloys and reduce nanoparticles aggregation, we added PVP into the mixed dispersion before co-reduction (Figure 3c,d). PVP can keep noble metal nanoparticles from aggregation [74]. Moreover, functionalized PtCo-alloy NPs can trend to selfassemble on the surface of MoS2, because the hydrogen-bond networks which is established by hydrophilic groups of PVP can draw PtCo-alloy NPs on the surface of MoS2 microspheres. Moreover, ethylene glycol as the solution of reduction reaction can improve reducing effectiveness and dispersibility [75]. Thus, more uniform, highly dispersed and vast PtCo-alloy NPs loaded on MoS2 EDX mapping images of MoS 2 /PtCo-alloy NPs depict that Pt and Co were uniformly dispersed on the surface of MoS 2 microspheres as shown in Figure 4b,c, respectively. Additionally, the HRTEM image and XPS further demonstrate the formation of PtCo-alloy NPs. The HRTEM (Figure 4d) indicates the (111) interplanar distance of 0.221 nm was smaller than 0.226 nm which is interplanar spacing of standard Pt (111). This is the other proof that Co atoms enter into the lattices of Pt [76]. Moreover, the Pt 4f peak shifted to higher energies can be obviously observed on the XPS in Figure 5a. This phenomenon is always associated with alloy formation with Co [77]. There were different work function in Pt and Co. When the combination of them were made to form PtCo alloy, higher atomic ratio of Co to Pt lead to the rehybridization of the d-band and the sp-band, hence, as the reference level in XPS measurements, the change in Fermi level result in the shifting to higher binding energy of Pt 4f peak [78]. Nanomaterials 2020, 10, x FOR PEER REVIEW 7 of 15 spacing of standard Pt (111). This is the other proof that Co atoms enter into the lattices of Pt [76]. Moreover, the Pt 4f peak shifted to higher energies can be obviously observed on the XPS in Figure  5a. This phenomenon is always associated with alloy formation with Co [77]. There were different work function in Pt and Co. When the combination of them were made to form PtCo alloy, higher atomic ratio of Co to Pt lead to the rehybridization of the d-band and the sp-band, hence, as the reference level in XPS measurements, the change in Fermi level result in the shifting to higher binding energy of Pt 4f peak [78].  A typical synthesis route of MoS2/PtCo-alloy NP composites is shown in Figure 5b. Three steps can be concluded for the synthesis of MoS2/PtCo-alloy NP composites. First, micelle-assisted hydrothermal method was carried out to synthesize MoS2 microspheres covered by PVP. This special MoS2/PVP hybrids are easy to disperse and draw more molecular chains of PVP to form the hydrogen bonded networks, because hydrophilic amide groups of PVP on the surface of MoS2 point outward [62]. Thus, when PVP is added for the second time, the new PVP is attracted onto hydrophilic amide groups of primary PVP. Moreover, the new PVP stabilizes PtCo-alloy NPs through coordinating Pt atoms with the N and O atoms of PVP simultaneously [79]. When the co-reduction finishes, PtCoalloy NPs disperse around MoS2 microspheres. The new PVP acts as "a bridge" to self-assemble PtCoalloy NPs on the surface of MoS2. spacing of standard Pt (111). This is the other proof that Co atoms enter into the lattices of Pt [76]. Moreover, the Pt 4f peak shifted to higher energies can be obviously observed on the XPS in Figure  5a. This phenomenon is always associated with alloy formation with Co [77]. There were different work function in Pt and Co. When the combination of them were made to form PtCo alloy, higher atomic ratio of Co to Pt lead to the rehybridization of the d-band and the sp-band, hence, as the reference level in XPS measurements, the change in Fermi level result in the shifting to higher binding energy of Pt 4f peak [78].  A typical synthesis route of MoS2/PtCo-alloy NP composites is shown in Figure 5b. Three steps can be concluded for the synthesis of MoS2/PtCo-alloy NP composites. First, micelle-assisted hydrothermal method was carried out to synthesize MoS2 microspheres covered by PVP. This special MoS2/PVP hybrids are easy to disperse and draw more molecular chains of PVP to form the hydrogen bonded networks, because hydrophilic amide groups of PVP on the surface of MoS2 point outward [62]. Thus, when PVP is added for the second time, the new PVP is attracted onto hydrophilic amide groups of primary PVP. Moreover, the new PVP stabilizes PtCo-alloy NPs through coordinating Pt atoms with the N and O atoms of PVP simultaneously [79]. When the co-reduction finishes, PtCoalloy NPs disperse around MoS2 microspheres. The new PVP acts as "a bridge" to self-assemble PtCoalloy NPs on the surface of MoS2. A typical synthesis route of MoS 2 /PtCo-alloy NP composites is shown in Figure 5b. Three steps can be concluded for the synthesis of MoS 2 /PtCo-alloy NP composites. First, micelle-assisted hydrothermal method was carried out to synthesize MoS 2 microspheres covered by PVP. This special MoS 2 /PVP hybrids are easy to disperse and draw more molecular chains of PVP to form the hydrogen bonded networks, because hydrophilic amide groups of PVP on the surface of MoS 2 point outward [62]. Thus, when PVP is added for the second time, the new PVP is attracted onto hydrophilic amide groups of primary PVP. Moreover, the new PVP stabilizes PtCo-alloy NPs through coordinating Pt atoms with the N and O atoms of PVP simultaneously [79]. When the co-reduction finishes, PtCo-alloy NPs disperse around MoS 2 microspheres. The new PVP acts as "a bridge" to self-assemble PtCo-alloy NPs on the surface of MoS 2 .
During the process of preparing CEs with doctor blade method, we found that with 5 wt% PtCo and 12 wt% PtCo, surface tension of liquid is large and not easy to spread, so it cannot be wetted and converged on conductive glass. According to repeated experiments, the slurry with 20 wt% PtCo is the easiest to be coated uniformly on the electrode and its performance is also the most ideal. Therefore, CE with 20 wt% PtCo is used for the subsequent electrochemical catalytic activity and photovoltaic performance.

Cyclic Voltammetry Analysis
Due to the notable improvement of PtCo loading in cells performance, the composites with maximum PtCo loading (MoS 2 /20 wt% PtCo) in this paper was chosen as the representation of MoS 2 /PtCo-alloy NP composites to achieve CV and EIS test. At same time, for better lateral comparison, the loading content of pure Pt NPs in MoS 2 /Pt NPs is also 20 wt%. Cyclic voltammetry (CV) curves (Figure 6a) of MoS 2 /PtCo-alloy NP composites, MoS 2 /Pt NPs composites, MoS 2 microspheres, and traditional Pt electrodes in I 3 − /I − electrolyte were measured at a scan rate of 100 mV s −1 from −0.8 V to 1.4 V. Two couples of oxidation and reduction peaks are detected. The high potential peak which can be ascribed to the Reaction (1). Moreover, the low potential peak which is mainly related to performance of DSSCs is attributed to the Reaction (2) [80]. During the process of preparing CEs with doctor blade method, we found that with 5 wt% PtCo and 12 wt% PtCo, surface tension of liquid is large and not easy to spread, so it cannot be wetted and converged on conductive glass. According to repeated experiments, the slurry with 20 wt% PtCo is the easiest to be coated uniformly on the electrode and its performance is also the most ideal. Therefore, CE with 20 wt% PtCo is used for the subsequent electrochemical catalytic activity and photovoltaic performance.

Cyclic Voltammetry Analysis
Due to the notable improvement of PtCo loading in cells performance, the composites with maximum PtCo loading (MoS2/20 wt% PtCo) in this paper was chosen as the representation of MoS2/PtCo-alloy NP composites to achieve CV and EIS test. At same time, for better lateral comparison, the loading content of pure Pt NPs in MoS2/Pt NPs is also 20 wt%. Cyclic voltammetry (CV) curves (Figure 6a) of MoS2/PtCo-alloy NP composites, MoS2/Pt NPs composites, MoS2 microspheres, and traditional Pt electrodes in I3 − /I − electrolyte were measured at a scan rate of 100 mV s −1 from −0.8 V to 1.4 V. Two couples of oxidation and reduction peaks are detected. The high potential peak which can be ascribed to the Reaction (1). Moreover, the low potential peak which is mainly related to performance of DSSCs is attributed to the Reaction (2) [80].  During the process of preparing CEs with doctor blade method, we found that with 5 wt% PtCo and 12 wt% PtCo, surface tension of liquid is large and not easy to spread, so it cannot be wetted and converged on conductive glass. According to repeated experiments, the slurry with 20 wt% PtCo is the easiest to be coated uniformly on the electrode and its performance is also the most ideal. Therefore, CE with 20 wt% PtCo is used for the subsequent electrochemical catalytic activity and photovoltaic performance.

Cyclic Voltammetry Analysis
Due to the notable improvement of PtCo loading in cells performance, the composites with maximum PtCo loading (MoS2/20 wt% PtCo) in this paper was chosen as the representation of MoS2/PtCo-alloy NP composites to achieve CV and EIS test. At same time, for better lateral comparison, the loading content of pure Pt NPs in MoS2/Pt NPs is also 20 wt%. Cyclic voltammetry (CV) curves (Figure 6a) of MoS2/PtCo-alloy NP composites, MoS2/Pt NPs composites, MoS2 microspheres, and traditional Pt electrodes in I3 − /I − electrolyte were measured at a scan rate of 100 mV s −1 from −0.8 V to 1.4 V. Two couples of oxidation and reduction peaks are detected. The high potential peak which can be ascribed to the Reaction (1). Moreover, the low potential peak which is mainly related to performance of DSSCs is attributed to the Reaction (2) [80]. ); (c) relationship of redox peak current density and square root of scan rate; (d) 100-times continuous cycle scan CVs for the MoS2/PtCo-alloy NP composite CEs at a scan rate of 100 mV s −1 .
The cathodic peaks potentials of MoS 2 /PtCo-alloy NP composites, MoS 2 /Pt NPs composites, MoS 2 microspheres and pure Pt electrodes are −0.415 V, −0.495 V, −0.520 V and −0.320 V, respectively. The cathodic peaks potential of MoS 2 /PtCo-alloy NP CEs is slightly larger than that of Pt CEs, but the redox current density of MoS 2 /PtCo-alloy NP CEs is slightly larger than Pt CEs. These indicators show MoS 2 /PtCo-alloy NP CEs possess similar electrocatalytic activity and reaction velocity to Pt CEs.
The charge-transfer mechanism of the I 3 − /I − system on CE was investigated by scanning CVs at series of scan rates (Figure 6b). For MoS 2 /PtCo-alloy NP composite CEs, the cathodic peaks (corresponding to I 3 − + 2e − ↔ 3I − ) and the anodic peaks (corresponding to 3I 2 − + 2e − ↔ 2I 3 − ) regularly shifted towards negative direction and positive direction, respectively. The linear relationship between redox peak current and square root of scan rate is shown in Figure 6c. This relationship is acquired by measuring CVs with different scan rates and record the current density of peaks. It indicates only diffusion of I 3 − /I − ion pair limits redox reactions between the MoS 2 /PtCo-alloy NPs CE and electrolyte [81]. Taking repeated CV scans of CEs is a routine method to simulate DSSCs in a long time running. The extent of peak shifts in CVs curves strongly associated with electrocatalysis reduction and chemical degradation. The 100 times continuous CVs for MoS 2 /PtCo-alloy NP composites is shown in Figure 6d. The shape of the CVs curve is highly stable through repeated CVs test, attributing to the excellent electrochemical stability of MoS 2 /PtCo-alloy NP composites.
peak which can be ascribed to the Reaction (1). Moreover, the low potential peak which is mainly related to performance of DSSCs is attributed to the Reaction (2) [80].

Electrochemical Impedance Analysis
Electrochemical impedance spectra (EIS) is an important tool to analyze the dynamics of reactions in cells and the surface construct of CEs. The internal resistances of DSSCs are usually divided into series resistance (R s ), charge-transfer resistance (R ct ), diffusion resistance (Z w ) and constant phase element (CPE). DSSC with MoS 2 -based CEs had higher R s values, which could be ascribed to the different preparation technology. The diffusion of iodine-based electrolytes in the alloy composite is so rapid that, over time, the electrolyte diffuses out of the battery. In order to solve this problem, a circular CE smaller than the package area of the battery was prepared. As result of changing the location and area of CE, the electronic transmission of such electrodes was not just through CE being injecting into the electrode clamp, but passing from CE and FTO in turn, then flowing into the electrode clamp. The FTO was less conductive than CE, therefore, the existence of this process in MoS 2 -based CEs lead to higher R s compared with Pt CE. Among them, the R ct of CE/electrolyte is an index to evaluate the electrocatalytic activity, because the small R ct signifies low overpotential and high electron transferring bet ween electrolyte and CEs. Figure 7 shows the Nyquist plots of dummy cells consisted of the CE/electrolyte/CE structure. The R ct values of MoS 2 /20 wt% PtCo is 1.04 Ω, which is proximity to that of pure Pt (0.71 Ω), indicating that the electrocatalytic activity of MoS 2 /PtCo is close to that of pure Pt. Moreover, the R ct values of MoS 2 /20 wt% PtCo is smaller than that of MoS 2 /20 wt% Pt (1.76 Ω) and MoS 2 microspheres (5.18 Ω). It could be put down to the improvement of catalytic ability and electron transfer ability as the result of the metal nanoparticles filling in lamellar interstitial space. Furthermore, synergistic catalysis also plays a role in reduction of R ct . The competent charge-transfer ability of MoS 2 /20 wt% PtCo CEs displays in the EIS, owing to synergistic effects and MoS 2 /PtCo-alloy NPs heterostructure.  In which K was molar gas constant, T was temperature, e was elementary charge and J was current density of photogenerated electrons. Symbol q represented the number of electrons during the process of charge transfer with corresponding concentration cox and transfer rate ket. Nc was the density of states of TiO2 conduction band. Eref could be obtained by taking the difference between conduction band edge and the redox potential of the redox coupling electrolyte. Half-wave potentials of MoS2, MoS2/Pt, Pt and MoS2/PtCo CEs measured in CV were 121 mV, 110 mV, 105 mV and 92 mV, respectively, showing the negative shift trend. Therefore, the most negative shift of MoS2/PtCo CEs in the I3 -/Iredox energy level resulted in the increase of Voc in DSSCs, leading to the highest Voc [82]. The results of CV showed great consistency in terms of J-V curve. The performance gap between MoS2/PtCo CEs and MoS2/Pt CEs is also observed in Figure 8, because the change of electron structure improves the electrocatalytic activity and charge-transfer ability [49].     In which K was molar gas constant, T was temperature, e was elementary charge and J was current density of photogenerated electrons. Symbol q represented the number of electrons during the process of charge transfer with corresponding concentration c ox and transfer rate k et . N c was the density of states of TiO 2 conduction band. E ref could be obtained by taking the difference between conduction band edge and the redox potential of the redox coupling electrolyte. Half-wave potentials of MoS 2 , MoS 2 /Pt, Pt and MoS 2 /PtCo CEs measured in CV were 121 mV, 110 mV, 105 mV and 92 mV, respectively, showing the negative shift trend. Therefore, the most negative shift of MoS 2 /PtCo CEs in the I 3 -/Iredox energy level resulted in the increase of V oc in DSSCs, leading to the highest V oc [82]. The results of CV showed great consistency in terms of J-V curve. The performance gap between MoS 2 /PtCo CEs and MoS 2 /Pt CEs is also observed in Figure 8, because the change of electron structure improves the electrocatalytic activity and charge-transfer ability [49].

Discussion
A new and simple self-assembly method was conducted to synthesize the MoS 2 /PtCo-alloy NP composite, which was applied to DSSCs as CEs. PVP functionalizes both MoS 2 and PtCo-alloy NPs and links each other by hydrogen bonded networks. Finally, the morphology tests investigate that uniform and mass PtCo-alloy NPs were loaded on the surface of MoS 2 . The structure and composition of MoS 2 /PtCo-alloy NP composites were studied by XRD, EDS and XPS to prove that PtCo-alloy NPs successfully load on MoS 2 . Furtherly, the result of CVs and EIS shows the low charge-transfer resistance, outstanding electrocatalytic activity and excellent stability of this material, which is owing to more PtCo-alloy NPs active sites provided and good intrinsic performance of PtCo. Finally, the J-V curves were measured under 100 mW cm −2 of simulated solar illumination. Moreover, the PCE of 8.46% obtained by the DSSCs with MoS 2 /PtCo-alloy NP CEs is similar to that of DSSCs based on Pt CEs. Overall, the MoS 2 /PtCo-alloy NP composite can be regarded as an alternation for the high-cost pure Pt CEs.