Morphology–Property Effect in Electrocatalysis Based on Ni 1- χ Se@Graphene Series with Speciﬁc Stoichiometric Ratio

Due to their special role in surface atoms, charge transfer channels and catalytic active sites, morphologies of electrocatalysts produce morphology–property effects in electrocatalysis. To study this effect, a series of speciﬁc stoichiometric ratios of nickel and selenium, Ni 1- χ Se, with different surface atom concentrations and morphologies, were synthesized on graphene via several thermal-reduction methods. In addition, we systematically investigated the morphology–property effect of a Ni 1- χ Se series based on electrochemical impedance spectra (EIS), cyclic voltammetry (CV), and Tafel polarization experiments. The Ni 1- χ Se nanoparticles demonstrated superior performance in electrocatalysis than that of Ni 1- χ Se nanoplates. After nanoplates were assembled to form nanospheres, Ni 1- χ Se nanospheres exhibited higher catalytic activity in terms of reducing I 3 − and multiple times faster charge-transfer velocities than those of Ni 1- χ Se nanoplates. Synthetically electrocatalytic properties of Ni 1- χ Se series were also measured as counter cells (CEs) of dye-sensitized solar cells (DSSCs). Ni 1- χ Se nanoparticles showed a higher power conversion efﬁciency ( PCE ) (7.33%) in a DSSC cell than when using a Pt CE (7.02%). The performance of Ni 1- χ Se nanoplates (6.57%) was worse than that of Ni 1- χ Se nanoparticles and Pt CEs. Simultaneously the self-assembled Ni 1- χ Se nanospheres (7.37%) exhibited PCE similar to that found with Ni 1- χ Se nanoparticles. © The Author(s) 2015.

Due to the great effect of morphology on electrocatalytic properties, controlling the morphologies of nanomaterials to develop highly effective electrocatalysts has been a subject of intense research. [1][2][3] Morphology-property effects mainly focus on the different concentrations and types of surface atoms resulting from different morphologies. Surface atoms greatly affect the adsorption energy and reaction activation energy of electrocatalysts. 4 In addition, they cause differences in the shapes and amounts of catalytic active sites. The collision among reactants and catalytic active sites is the key segment in the catalytic reaction process. 5 Thus, morphology exerts an influence on the activity and selectivity of electrocatalysts. Recent research has revealed that electrocatalysts with different morphologies exhibited different charge transfer abilities due to the different charge transfer channels. [6][7][8] Our group has reported that microsphere NiSe 2 /RGO showed a better electrocatalytic performance than that of octahedron NiSe 2 /RGO, and mesoporous Ni 0.85 Se owned a high catalytic activity and fast electron-transport ability. 9,10 More research is required to further explore the mechanism of morphology-property effect in electrocatalysis reaction.
Many researchers have focused on transition metal chalcogenides (TMCs) with a wide range of potential applications in the field of material science. [11][12][13] Specific properties of TMCs rely on their distinctive electronegativity, valence bond, and structure, as well as outer electrons of metal and chalcogens, etc. Furthermore, the great covalency of the metal-chalcogens interactions reduces the relative charge on the metal ion, thus enhancing metal orbital diffuseness and favoring M-M bonding. 14 Because of the aforementioned structure characteristics, TMCs have been shown to be excellent electrocatalysts due to their outstanding charge transfer and catalytic properties. [15][16][17] In previous reports, the TMCs with stoichiometries similar to those of Ni 0.85 Se, Co 0.85 Se, and Co 9 S 8 are always good catalysts with excellent activity for the oxygen reduction reaction. [18][19][20] The specific stoichiometries result in abundant unsaturated atoms and unique electronic configuration, which give rise to outstanding active catalytic sites and fast electron-transfer channels. These properties are beneficial to obtain excellent electrocatalytic properties in a heterogeneous electrocatalytic reaction. 21,22 However, due to the mutability of the valence of chalcogens, the synthesis of this type of TMCs is difficult to control. 23,24 Here we report the systematic investigation of morphologyproperty effect in electrocatalysis based on Ni 1-χ Se (χ = 0-0.15) with z E-mail: gaoguandao@nankai.edu.cn; liul@nankai.edu.cn different morphologies. Ni 1-χ Se nanoparticles, Ni 1-χ Se nanospheres, and Ni 1-χ Se nanoplates were combined with graphene to obtain excellent electrocatalytic property via a convenient solvothermal-reduction method. In the synthesis process, adjusting selenium sources (selenium powder and Na 2 SeO 3 ), solvents, and proportion of Ni/Se were performed to control the morphology and valence of Ni 1-χ Se. Then, Ni 1-χ Se@graphene series were treated as thin films to carry out electrochemical measurements. From the analysis of EIS, CV, and Tafel polarization, the electrocatalytic property of Ni 1-χ Se nanoparticles was better than Ni 1-χ Se nanoplates, and Ni 1-χ Se nanospheres presented similar and/or better electrocatalytic properties in comparison to Ni 1-χ Se nanoparticles. Furthermore, the PCE of Ni 1-χ Se@graphene series were also measured as counter cells (CEs) of DSSCs. We found that the PCE values of Ni 1-χ Se series presented an order of Ni 1-χ Se nanospheres (7.37%) > Ni 1-χ Se nanoparticles (7.33%) > Pt (7.02%) > Ni 1-χ Se nanoplates (6.57%).

Experimental
Synthesis of Ni 1-χ Se@graphene series.-In a typical synthesis of nanoparticulate Ni 1-χ Se, the original molar ratio of NiCl 2 · 6H 2 O and selenium powder was selected to be 1:1. First, 0.952 g (4 mmol) NiCl 2 · 6H 2 O, 1 g EDTA, 1 g NaOH, and 5 ml of GO water suspension (10 mg ml −1 ) were dissolved in 30 mL ethylene glycol in Teflonlined autoclave of 50 mL capacity. After stirring for 1 h, 0.316 g (4 mmol) selenium powder was added to the above mixture. Then, 5 ml hydrazine hydrate was added dropwise. The resulting solution was sealed and heated at 180 • C for 15 h. Thereafter, the autoclave was allowed to cool to room temperature. The product was washed with water and absolute ethanol to remove impurities, and then dried at 60 • C. The synthetic procedure of nanoplates Ni 1-χ Se was similar to that of Ni 1-χ Se nanoparticles. The differences were that the amount of NiCl 2 · 6H 2 O was changed to 0.4754 g (2 mmol), and solvent was changed to distilled water. During synthesizing Ni 1-χ Se nanospheres, the nickel source and selenium sources were NiSO 4 (0.2628 g) and Na 2 SeO 3 (0.789 g), respectively. The other experimental processes were the same as those for the Ni 1-χ Se nanoparticles.
Characterization of obtained samples.-For 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 morphologies of samples were studied by field-emission scanning electron microscopy (FE-SEM, H775 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) equipped with energy dispersive spectrometer (EDS) to measure the elements in the samples. The purity of the samples was performed by the X-ray photoelectron spectroscopy (XPS) analysis (PHI5000 VersaProbe).
Fabrication and characterization of Ni 1-χ Se@graphene series thin films.-A Ni 1-χ Se@graphene series materials slurry was made in ethanol by mixing 0.1 g Ni 1-χ Se@graphene powder with 0.025 g PEG20000 which was used as a dispersant as well as a binder and stirred continuously. Then, a thin film was made using a doctor-blade to wipe slurry on FTO conductive glass (LOF, TEC-15, 15 W per square). After the film was steady, the conductive glass with the thin film was heated at 400 • C for 1 h under the protection of argon, and the Ni 1-χ Se@graphene thin film was achieved.
All of the electrochemical measurements were measured with the Zahner IM6 exelectrochemical workstation. Photocurrent−voltage curves were conducted in simulated AM 1.5 illumination (100 mW cm −2 , Trusttech CHF-XM-500W) with a Keithley digital source meter (Keithley 2410, USA). Cyclic voltammetry (CV) was recorded with a three electrode system on the ex-electrochemical workstation. Pt was used as the counter electrode, and Ag/AgCl was used as the reference electrode. The solution of 10.0 mM LiI, 1.0 mM I 2 and 0.1 M LiClO 4 in acetonitrile served 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 to 1 MHz. The amplitude of the sinusoidal AC voltage signal was 5 mV. The analyses of the resulting impedance spectra were conducted using Zview 2.0. Tafel polarization measurement was employed in a symmetrical dummy cell which was used in the EIS experiments. The electrolyte was the same as the electrolyte of DSSC (0.05 M I 2 , 0.1 M LiI, 0.6 M 1, 2-dimethyl-3-propylimidazolium iodide (DMPII), and 0.5 M 4-tert-butyl pyridine with acetonitrile as the solvent). The scan rate was 20 mV s −1 , and the voltage range was −1.0 to 1.0 V.

Results and Discussion
Characterization of Ni 1-χ Se@graphene series .-X-ray diffraction (XRD) was performed to verify the phase structures of Ni 1-χ Se products. Figure 1 shows the XRD patterns of Ni 1-χ Se nanoparticles, Ni 1-χ Se nanospheres and Ni 1-χ Se nanoplates. All of the diffraction peaks in three patterns can be readily indexed as pure-phase compounds of hexagonal Ni 1-χ Se (JCPDS no. 18-0888, vertical line in Fig. 1). The diffraction peak at about 16.84 • also belongs to hexagonal Ni 1-χ Se; however, the data derived from XRD analysis software (Jade) does not provide the relevant information. No impurities, such  as selenium, nickel oxide or other phases of nickel selenides, were observed, indicating that the products were pure phase.
The morphologies and structures of the as-prepared Ni 1-χ Se samples were studied by SEM and TEM. The TEM images of Ni 1-χ Se nanoparticles in Figs. 2a and 2b revealed that the nanoparticles with diameters of ∼20 nm dispersed in situ on graphene uniformly. Careful observation also finds that the shadows in Figs. 2a and 2b were graphene. The well-dispersed nanoparticles on the graphene increased the number of catalytic active sites. 25 As the SEM image (Fig. 2d) shows, each Ni 1-χ Se nanosphere, with the size of 400∼500 nm, is composed of numerous primary nanoplates. In the TEM image of Fig. 2e, Ni 1-χ Se nanospheres can be observed to be porous among nanoplates. The porous structure is beneficial to accommodating I 3 − molecules between nanoplates through the intercalation process, as well as the collision among I 3 − and catalytic active sites. The morphology and structure of Ni 1-χ Se nanoplates are shown in Figs. 2g and 2h, and Ni 1-χ Se nanoplates show irregular plate-like morphology with a side length of 30-60 nm. Furthermore, it can be clearly observed that nanoplates grow on graphene nanosheets. SEM and TEM images demonstrate that Ni 1-χ Se nanoparticles, Ni 1-χ Se nanospheres, and Ni 1-χ Se nanoplates have been successfully synthesized in situ on graphene. More detailed information about morphologies of three Ni 1-χ Se samples are shown in supporting information, Se Figure S1 through S3.
The chemical composition of these samples was further investigated by energy dispersive spectrometer (EDS) (Figs. 3a-3c). EDS profiles indicate the existence of Cu, Ni, Se (Fig. 3f). The Cu peak is from the copper wire mesh, while the Ni and Se peaks are from Ni 1-χ Se products. The peaks of the elements Ni and Se were detected in the EDS pattern and the molar ratio was calculated as 0.90:1, 0.89:1, and 0.86:1 for Ni 1-χ Se nanoparticles, Ni 1-χ Se nanosheres, and Ni 1-χ Se nanoplates, respectively.
Electrocatalytic property of the Ni 1-χ Se@graphene series.-EIS experiments were carried out on two symmetric thin films of Ni 1-χ Se@graphene samples to confirm the charge transfer and electrolyte molecules diffusion between CEs materials and electrolytes. As is shown in Fig. 4, two features are observable in the recorded spectra: a low-frequency feature due to the electrolyte diffusion and a high-frequency arc due to the charge transfer in the interface. 26 The Nyquist plots obtained with EIS measurements can be simulated by the equivalent circuit (inset in Fig. 4) and the relevant values are summarized in Table I. In an equivalent circuit, series resistance (R s ) is deduced by high-frequency cut on the real axis (Z axis), and chargetransfer resistance (R ct ) is derived from the radius of high-frequency semicircle on the real axis. 27 The impedance spectra indicates that the R ct value of Ni 1-χ Se nanospheres (0.50 ) is smaller than Ni 1-χ Se nanoparticles (0.73 ), Pt (1.28 ) and Ni 1-χ Se nanoplates (1.60 ) CEs. In addition, the R ct value is in a contradictory relationship with the charge-transfer ability. From the analysis of R ct values, Ni 1-χ Se nanoplates had lower charge-transfer ability than Ni 1-χ Se nanoparticles. Nevertheless, the connected nanoplates increased charge transfer channels and decreased the electrons hop among 2D nanoplates structure, thus Ni 1-χ Se nanospheres exhibited multiple times faster charge-  transfer velocity than Ni 1-χ Se nanoplates, even better than Ni 1-χ Se nanoparticles and Pt. In order to understand the catalytic activity of the resultant film, the electrochemical behavior of Ni 1-χ Se@graphene series has been studied by CV. The electrochemical measurements indicate that all of the samples exhibit a perfect reversible response, a characteristic of Ni 1-χ Se electrodes with rapid electron-transfer kinetics. 28 CV profiles of three Ni 1-χ Se electrodes at a scan rate of 25 mV s −1 are presented in Fig. 5. Two pairs of oxidation and reduction peaks are obviously observed for Ni 1-χ Se@graphene series and Pt electrodes. From left to right, the two couples of the redox current peaks in the CV curves correspond to the I 3 − /I − and I 3 − /I 2 redox reactions. 29 The catalytic activities of different CEs can be analyzed by the peak-to-peak  separation (E pp ) and the peak current density (J A ) in lower potential. 30 E pp is negatively correlated with the standard electrochemical rate constant and has positive correlation with overpotential losses. In addition, the E pp values presented the trend of Ni 1-χ Se nanoplates (399 mV) > Ni 1-χ Se nanoparticles (351 mV) > Ni 1-χ Se nanospheres (255 mV), which reveal the reverse order of catalytic activity. The better catalytic activity of Ni 1-χ Se nanoparticles than Ni 1-χ Se nanoplates results from the more catalytic active sites. After nanoplates assembled to form nanospheres, the interior reaction space of nanospheres increased the collision among I 3 − and catalytic active sites. Simultaneously, the assembled nanoplates decreased the diffusion distance of I 3 − /I − redox and enhanced the activity of catalytic active sites. Therefore Ni 1-χ Se nanospheres showed enhanced catalytic activity compared to Ni 1-χ Se nanoplates. Furthermore, Ni 1-χ Se nanospheres also showed a higher J A value than Ni 1-χ Se nanoparticles and Ni 1-χ Se nanoplates, which further reveals the best catalytic activity of Ni 1-χ Se nanospheres among Ni 1-χ Se@graphene series. 31 Results of CV experiments agree with the EIS experiments, and relevant parameters are shown in Table I.
The CV experiments of three electrodes at different scan rates from 10 to 100 mV s −1 were also performed and CV profiles are shown in Figs. 6a-6c. Depending on the scan rate, outward extension of all the peaks can be observed. The CV measurements at different scan rates indicate that all the Ni 1-χ Se samples exhibit a perfect reversible response, a characteristic of Ni 1-χ Se electrodes with rapid electron-transfer kinetics. 28 The anodic and cathodic current density as a function of the square root of scan rate for the three samples is shown in Fig. 6d, with Ni 1-χ Se nanospheres electrode displaying the highest current density. All of the lines of three Ni 1-χ Se electrodes in anodic and cathodic zone are almost linear, as shown in Fig. 6d, except the mismatch of Ni 1-χ Se nanospheres within experimental error. The linear relationship indicates that the redox reaction of I − /I 3 − cou-ples at the three Ni 1-χ Se electrodes belongs to a diffusion-controlled transport and obeys the Randles−Sevcik equation. 32 Tafel polarization experiments were carried out with two identical as-prepared thin films to measure the information of electrochemical polarization and concentration polarization arising from different charge exchange and electrolyte diffusion rates at the CE/electrolyte interface. The results are shown in Fig. 7, and the corresponding parameters are listed in Table I. The polarization zone of the Tafel polarization curve corresponds to the high-frequency region of Nyquist  plots, and the diffusion zone relates to the low-frequency region. The large slope for the anodic or cathodic branches indicates a high exchange current density (J 0 ), which indicates good catalytic activity toward triiodide reduction of CEs. 33 As shown in Table I, the calculated J 0 values of Ni 1-χ Se nanospheres, Ni 1-χ Se nanoparticles, and Ni 1-χ Se nanoplates were 0.65 log (mA cm −2 ), 0.55 log (mA cm −2 ), and 0.49 log (mA cm −2 ), respectively. The intersection of the cathodic branch with the Y-axis can be considered as the limiting diffusion current density (J lim ), which is determined by the diffusion properties of the redox couple and the CE materials. 34 As shown in Table I, the values of J lim also follows an order of Ni 1-χ Se nanospheres > Ni 1-χ Se nanoparticles > Ni 1-χ Se nanoplates. Both J 0 and J lim match well with the results of CV and EIS experiments, and Tafel polarization experiments further demonstrate the catalytic activity for reducing I 3 − and diffusion coefficient of I 3 − in Ni 1-χ Se series. Current density-voltage (J-V) measurements under the AM 1.5 G condition were carried out to evaluate the synthetically electrocatalytic properties of Ni 1-χ Se series, which are shown in Fig. 8. The corresponding values of open-circuit voltage (V oc ), short-circuit current (J sc ), fill factor (FF), and PCE are summarized in Table II. Note that the DSSC with Ni 1-χ Se series also presents a significant trend of J sc in Ni 1-χ Se nanospheres > Ni 1-χ Se nanoparticles > Pt > Ni 1-χ Se nanoplates, which is in good agreement with the tendency of PCE. Due to the high charge-transfer ability and many catalytic active sites of nanoparticles, 18,35 Ni 1-χ Se nanoparticles exhibited excellent electrocatalytic performance with high PCE (7.33%) and J sc (14.80 mA cm −2 ) values. EIS, CV, and Tafel polarization experiments demonstrated that Ni 1-χ Se nanoplates had a lower charge transfer ability and catalytic activity than Ni 1-χ Se nanoparticles. The poor electroctalytic property of Ni 1-χ Se nanoplates resulted in slow and incomplete reduction of I 3 − , and impeded photogenerated electron transfer at the Ni 1-χ Se nanoplates/electrolyte interface. Thus, Ni 1-χ Se nanoplates showed lower J sc (12.80 mA cm −2 ) and PCE (6.57%) than Ni 1-χ Se nanoparticles. As mentioned above, the assembled structure of nanospheres possessed large internal surface area and can accommodate I 3 − molecules among nanoplates through intercalation process. Together with the enhanced collision among I 3 − and catalytic active sites as well as the improved charge transfer among nanoplates, Ni 1-χ Se nanospheres exhibited higher J sc (15.20 mA cm −2 ) and PCE (7.37%) than those of Ni 1-χ Se nanoplates, even were similar with those of Ni 1-χ Se nanoparticles. Moreover, the J sc and PCE values of Ni 1-χ Se nanoparticles and Ni 1-χ Se nanospheres were higher than those of Pt CE (J sc = 14.00 mA cm −2 , PCE = 7.02%).

Conclusions
Ni 1-χ Se@graphene series with specific stoichiometry ratio and different morphologies were successfully synthesized in situ on graphene with different selenium sources, solvents, and proportions of Ni/Se. The results demonstrate that Ni 1-χ Se nanoparticles exhibit better electrocatalytic property than that of Ni 1-χ Se nanoplates. However, the self-assembled structure of nanospheres endowed Ni 1-χ Se nanospheres with enhanced charge transfer ability and catalytic activity. Simultaneously, Ni 1-χ Se nanospheres showed similar or better electrocatalytic property in comparison with Ni 1-χ Se nanoparticles. In summary, this research not only facilitates the synthesis of TMCs with specific stoichiometry ratio, but also proposes electrocatalytic mechanism of morphology-property effect in electrocatalytic property.