The Morphology–Property Effect and Synergetic Catalytic Effect of CdS as Electrocatalysts for Dye-Sensitized Solar Cells

1Agro-Environmental Protection Institute, Ministry of Agriculture, Tianjin 300191, People’s Republic of China 2College of Resources and Environment, Northeast Agricultural University, Haerbin 150036, People’s Republic of China 3Center forAircraft Fire and Emergency, Economics and Management College, Civil Aviation University of China, Tianjin 300300, People’s Republic of China 4Tianjin Key Laboratory of Environmental Remediation and Pollution Control, College of Environmental Science and Engineering, Nankai University, Tianjin 300071, People’s Republic of China

catalytic effect, need further study for the optimal electrocatalytic property.
In this paper, we investigated the morphology-property effect and synergetic catalytic effect of CdS as CE materials. Firstly, we synthesized CdS nanoparticle, CdS nanorod, CdS nanorod grown on graphene (CdS nanorod/RGO) and CdS nanoparticle grown on graphene (CdS nanoparticle/RGO) through facile solvothermal method. For investigating the property effect, the electrocatalytic performance of four kinds of CdS samples was measured as CEs in DSSCs under the same conditions. As to the effect of morphology, CdS nanorod exhibited higher electrocatalytic performance than CdS nanoparticle, similarly, the power conversion efficiency of CdS nanorod/RGO (7.19%) was higher than CdS nanoparticle/RGO (6.55%). In addition, the synergetic catalytic effect between CdS and graphene is obvious, two CdS/RGO composites owned much better catalytic activity and electrical conductivity than pure CdS nanoparticle and CdS nanorod.

Experimental
Synthesis of GO nanosheets.-Graphene oxide (GO) nanosheets were made by a modified Hummers method. 36 In detail, graphite powder (2 g) was put into 100 mL of cooled (0 • C) concentrated H 2 SO 4 , followed by the slow addition of KMnO 4 (6 g), a slight exotherm may be produced in this process. The suspension was then stirred at 35 • C for 12-15 hours. Afterwards, 200 mL of distilled water was added and the temperature was kept at 96 • C for 2 h. The temperature was reduced to 60 • C, and H 2 O 2 (30%, 10 mL) was injected into the suspension to completely react with the excess KMnO4, which yielded a bright yellow mixture. The solid product was separated by centrifugation, and then washed with HCl (5%) several times and with water until the pH value of the supernatant was nearly 6, and graphene oxide was obtained. The collected precipitate was dispersed in water, then sonicated and subsequently concentrated to obtain a GO suspension, and kept at 50 • C for 10 h and GO powder was obtained. The X-ray diffraction and transmission electron microscopy figures of graphene oxide and RGO are shown in supplementary materials, Figure S1  nanosheets through a solvothermal treatment: 0.4 g CdCl 2 · (5/2H 2 O) and 5 ml of GO water suspension (the concentration is 10 mg/ml) were added into 15 ml deionized water, with stirring for a magnetic stirrer about 1 h. Then 0.4 g Na 2 S · 9H 2 O was dissolved in 15 ml deionized water, and the Na 2 S · 9H 2 O aqueous solution was dropwise transferred in the above solution. After stirring about 30 min, the mixed solution were transferred into a 50 mL Teflon-lined stainless steel autoclave, and heated at 150 • C for 3 h. After being washed by deionized water and ethanol, the obtained CdS nanoparticle/RGO samples were dried in air.
Synthesis of CdS nanoparticle.-The synthetic procedures of CdS nanoparticle samples were almost the same with the procedures of CdS nanoparticle/RGO samples. The only difference is that the 5 ml of GO water suspension is changed to 5 ml of distilled water.
Synthesis of CdS nanorod/RGO.-Firstly, 1.0 g CdCl 2 · (5/2H 2 O) and 5 ml of GO water suspension were added into 35 ml tetraethylenepentamine. Then, after stirring about 1 h, 0.8 g thioacetamide was added and stirred for about 30 min. Finally, the autoclave was sealed and heated for 3 h at 180 • C.
Synthesis of CdS nanorod.-For synthesizing CdS nanorod, just like the synthesis of CdS nanorod/RGO, but the 5 ml of GO water suspension is changed to 5 ml of distilled water.
Characterization of obtained samples.-For characterizing the obtained samples, the crystallinity and composition of the samples were characterized by X-ray diffraction (XRD, D/max-2500, JAPAN SCIENCE) with Cu Kα radiation (λ = 1.54056 Å). The morphology of samples was studied by field-emission scanning electron microscopy (FE-SEM, Nanosem 430, FEI). More detailed insight into the microstructure of the sample was given by high-resolution transmission electron microscopy (TEM, Tecnai G2 F20, operating at 200 kV, FEI).
Fabrication of DSSCs.-CE materials slurry was made in ethanol by mixing 0.1 g CE materials powder with 0.025 g PEG20000 which was used as dispersant as well as binder and stirred continuously. Then a film was made through using a doctor-blade to wipe CE materials slurry on FTO (Fluorinedoped Tin Oxide) conductive glass (LOF, TEC-15, 15 W per square). After the film was steady, the conductive glass with film was heated at 400 • C for 1 h under the protection of argon, and the counter electrode was gotten.
A commercial TiO 2 sol (Solaronix, Ti-Nanoxide T/SP) was used to prepare the TiO 2 film on FTO also through the doctorblade method, and the film was soaked in an N719 dye solution (RuL2(NCS)2:2TBA(L = 2,2'-bipyridyl-4,4'-dicarboxylic acid)) (in ethanol) for 24 h to obtain dye-sensitized TiO 2 electrodes. DSSCs were assembled by injecting the electrolyte into the aperture between the dye-sensitized TiO 2 electrode and the counter electrode. The liquid electrolyte composed of 0.05 M I 2 , 0.1 M LiI, 0.6 M 1, 2-dimethyl-3propylimidazolium iodide (DMPII), and 0.5 M 4-tert-butyl pyridine with acetonitrile as the solvent. Surlyn 1702 was used as the spacer between the two electrodes. The two electrodes were clipped together and solid paraffin was used as the sealant to prevent the electrolyte solution from leaking. The effective cell area was 0.25 cm 2 . The standard Pt CE used in this work was prepared by sputtering Pt on FTO, and was purchased from Dalian Heptachroma Solar Tech Co., Ltd.  as the electrolyte. Electrochemical impedance spectra (EIS) analysis was conducted at zero bias potential and the impedance data covered a frequency range of 0.1 Hz-1 MHz. The amplitude of the sinusoidal AC voltage signal was 5 mV. The analyses of the resulting impedance spectra were conducted using the software Zview 2.0. Tafel-polarization measurements were employed in the same symmetrical dummy electrode and electrolyte with EIS experiments, with a scan rate of 20 mV s −1 and a voltage range of −1.0 to1.0 V.

Results and Discussion
X-ray powder diffraction analysis.-For characterizing the phase purity and structure of the obtained samples, X-ray powder diffraction (XRD) was performed. As shown in Figure 1 (112) planes of the greenockite CdS, respectively. No characteristic peak was observed for other impurities in the all samples, demonstrating that pure crystalline of four CdS samples were formed via the solvothermal process. No diffraction peaks for RGO can be observed in the CdS/RGO composites is due to that the main characteristic peak of RGO at about 25.0 • may be overlapped by the (100) peak of hexagonal CdS. 37,38 Scanning electron microscopy and transmission electron microscopy analysis.-The morphology and structure of the synthesized products have been investigated using SEM and TEM techniques. Figure 2a and 2e present overall SEM images of CdS nanoparticle/RGO and CdS nanorod/RGO, respectively, revealing the uniform sprinkle of nanoparticles and nanorods on graphene nanosheets. The uniform distribution of CdS samples on graphene sheets is more evident from the TEM images shown in Figure 2b and 2f. The apparent contrast can be observed between the gray edge portion and dark particles or rods, and reveal that the nanoparticles or nanorods were grown on graphene nanosheets. As displayed in Figure 2c, the size of CdS nanoparticles are about 10 to 20 nm. In can be observed in Figure  2g 39 Furthermore, it is noted in Figure 3B that the G band of as-synthesized GO shifts to lower frequency in CdS/RGO composites, and 9 cm −1 and 6 cm −1 red shifting were respectively observed in CdS nanorod/RGO and CdS nanoparticle/RGO. Meanwhile, the ID/IG ratio of CdS nanorod/RGO and CdS nanoparticle/RGO increased to 1.17 and 1.11, in comparison with GO (0.84). It can be indicated that GO was reduced and its native sheet structure was restored in CdS nanorod/RGO and CdS nanoparticle/RGO from the red-shifted G band and the increased ID/IG ratio. 37

DSSCs performance analysis of different CEs.-To investigate
the interface characterization of the DSSCs fabricated with different CEs, electrochemical impedance spectra (EIS) experiments were employed using symmetric electrodes with two identical electrodes (CE/electrolyte/CE). The Nyquist plot of CdS nanoparticle/RGO, CdS nanorod/RGO, CdS nanoparticle and CdS nanorod are presented in Figure 4, and corresponding parameters are summarized in Table I. The impedance properties, namely, series resistance (R S ), charge transfer resistance (R ct ), and Nernst diffusion impedance (Z N ) can be obtained from the Nyquist plot in Figure 3. The first semicircle at high frequency was assigned to the charge transfer at the interface between the electrode and the electrolyte, whereas the second semicircle at low frequency was primarily associated with the Nernst diffusion of I 3 − within the electrolyte. 40,41 CdS nanorod/RGO exhibited lower R ct  (5. 16 cm −2 ) than CdS nanoparticle/RGO (R ct = 6.59 cm −2 ), the similar comparison happens to the comparison between pure CdS nanorod (R ct = 36.60 cm −2 ) and CdS nanoparticle (R ct = 102.90 cm −2 ). On the other hand, the introduction of graphene induces an obvious improvement in the charge transfer and electrocatalytic activity, which can be directly reflected by the comparison of first arc between CdS nanoparticle/RGO, CdS nanoparticle and CdS nanoparticle, CdS nanorod. 42 The resistance results revealed that the CdS nanorod had better electrical conductivity compared with CdS nanoparticle, and the introduction of graphene obviously enhanced the charge transfer ability of CdS. The EIS in light illumination of four CdS samples were performed, corresponding images and parameters are shown in Figure S7 and Table S1. Different with previous reported EIS measurements in light illumination of photoanode performed with photoanode/electrolyte/CE system, EIS measurements in light illumination of CE was performed with CE/electrolyte/CE system. [43][44][45] EIS results shows that R s and R ct values of four CdS samples in light illumination were smaller that those without light illumination. In our opinion, light illumination promote the electron jump from the valence band to conduction band, and then increased the amount of electron in conduction band for exchange current between CdS samples and electrolyte molecules in space charge region. Thus the charge-transfer across the CdS CE/electrolyte interface were improved.
To further characterize the charge-transfer properties of four CdS electrodes, Tafel polarization was carried out using the same symmetrical electrodes for EIS measurements. The information regarding the exchange current density (J 0 ) and the limiting diffusion current density (J lim ) can be obtained from the curves, as shown in Figure 5. 46 The slopes of Tafel curves based on different CdS samples in Figure 4 shows the trend of CdS nanorod/RGO > CdS nanoparticle/RGO > CdS nanorod > CdS nanoparticle. Higher J 0 means that the CdS with the morphology of nanorod own better catalytic activity for I 3 − reduction than those of nanoparticle, same condition happens to the comparison between CdS RGO and CdS 4.7 The variation of J 0 obtained from Tafel curves is in agreement with the change tendency of R ct derived from the EIS spectra. Furthermore, the J lim obtained from the current density at low slope depends on the diffusion coefficient of I − /I 3 − redox couples at the CE/electrolyte interface. Obviously, the J lim ( Figure S5) of four CdS CEs also follows an order of CdS nanorod/RGO (2.15 log (mA cm −2 )) > CdS nanoparticle/RGO (2.05 log (mA cm −2 )) > CdS nanorod (1.82 log (mA cm −2 )) > CdS nanoparticle (1.76 log (mA cm −2 )), which is in good accordance with that of Z N derived from EIS characterization.
In Eqs. 1 and 2, R ct is the charge-transfer resistance obtained from EIS, R is the gas constant and F is Faraday's constant, D is the diffusion coefficient of I 3 − , l is the spacer thickness, C is the I 3 − concentration, N A is Avogadro's constant, and e and n have their usual meanings. 18,48 Table I. Corresponding EIS, Tafel polarization and photovoltaic parameters of the DSSCs assembled with four CdS electrodes and Pt.

CEs
V oc (mV)  Cyclic voltammetry (CV) was carried out to evaluate the electrochemical catalytic activities of the CdS and Pt CEs. Figure 6 shows the CV curves of four CdS CEs and Pt CE for the iodide species (I − /I 3 − ). Catalytic activities of different CEs can be analyzed by the peak-to-peak separation (E pp ) and the peak current density in lower potential. 49 E pp in lower potential is negatively correlated with the standard electrochemical rate constant and has positive correlation with overpotential losses. 50 The peak current density in lower potential is positively correlated with the electrocatalytic reaction rate. 26 As shown in Figure 6, the lower E pp and higher peak current density reveals the better electrocatalytic activity for reducing I 3 − of CdS nanorod/RGO CE than CdS nanoparticle/RGO CE. Furthermore, the similar trend can be observed between CdS nanorod CE and CdS nanoparticle CE. Results of CV experiments also matched well with the EIS experiments, and manifested the morphology-property effect and synergetic catalytic effect of CdS CEs.
Current density-voltage (J-V) measurements under the AM 1.5 G condition were carried out to compare the comprehensive electrocatalytic properties of four CdS CEs and Pt CE. The J-V curves of DSSCs with the above CEs are displayed in Figure 7 and the corresponding photovoltaic performance parameters are summarized in Table I. As shown in Table I, the short-circuit current (J SC ) and fill factor (FF) values of the four CdS electrodes are found in the same sequence  with power conversion efficiency (PCE) values, which can represent the charge transfer ability, catalytic reaction velocity and utilization efficiency of photogenerated electrons at the CEs. 25 The PCE value of the DSSCs based on CdS nanorod/RGO is 7.19%, which is higher than that of CdS nanoparticle/RGO (6.55%). As to the pure CdS, CdS nanorod (5.13%) also exhibited higher PCE value than that of CdS nanoparticle (4.74%). The photovoltaic performance comparison between nanoparticle and nanorod further demonstrated that CdS with the morphology of nanorod own better electrocatalytic property than nanoparticle, which is due to that high crystallinity and the onedimensional nature of nanorod assisted the transfer of charge from one side to another side. 37 This favorable feature of nanorod accelerated the transfer velocity and utilization efficiency of photogenerated electrons, accompying with fast electrocatalytic reaction, which is demonstrated by the values of R ct , J SC and FF. On the other hand, the photovoltaic performance of RGO is shown in Figure S6, with the J SC and PCE values of 12.81 mA cm −2 and 4.04%. Thus, CdS/RGO CEs exhibited both higher J SC , FF and PCE values than CdS and RGO CEs, revealing the good synergetic catalytic effect between graphene and CdS. The PCE and J SC values of CdS nanorod/RGO are highest among CdS electrodes, even slightly inferior to the commercial Pt electrode (PCE = 7.44%, J SC = 748 mV).
In addition to high electrocatalytic activity of as CdS/RGO electrocatalysts of DSSC, electrochemical stability is also an important characteristic from the practical viewpoint. Herein, 50 consecutive cycles of CV measurements were used to evaluate electrochemical stability of CdS nanorod/RGO and CdS nanoparticle/RGO. As shown in Figure 8, after 50 consecutive cycles of CV testing, the CdS nanorod/RGO CE did not exhibit any significant decline in peak current densities and showed relatively high stability, indicating that CdS nanorod/RGO can coexist with the I − /I 3 − redox couples under working conditions. 51 In P316 ECS Journal of Solid State Science and Technology, 7 (6) P311-P316 (2018) comparison, CdS nanoparticle/RGO CE exhibited worse electrochemical stability than CdS nanorod/RGO CE, and its consecutive cycles of CV shows obvious change in peak current density and E pp . The high electrocatalytic activity and excellent electrochemical stability reveals that CdS nanorod/RGO owns comprehensively higher electrocatalytic performance than CdS nanoparticle/RGO. The enhanced electrocatalytic activity and long-term stability of CdS/RGO CEs demonstrated that it is a good candidate electrocatalyst to substitute expensive Pt material for DSSCs.

Conclusions
In summary, CdS nanoparticle, CdS nanorod, CdS nanorod/RGO and CdS nanoparticle/RGO have been successfully synthesized by facile solvothermal methods. It is noteworthy that the morphologyproperty effect and synergetic catalytic effect of CdS as CE materials are noticeable. Furthermore, the EIS, CV, Tafel polarization and photocurrent density−voltage experiments demonstrated the two property-effects of CdS as CE materials from different perspectives. This work provides new insight to design efficient CE materials with optimal electrocatalytic property and fabricate low cost and good electrochemical-stable DSSC.