Solvothermal synthesis of weakly crystalline cobalt–nickel sulfide to obtain high pseudocapacitance

In this paper, a dual solvent system of glycerol and ethylene glycerol was used to prepare nickel cobalt sulfide with different crystallinities. It was found that the crystallinity and morphology of the product had an important effect on its electrochemical performance. Therefore, the synthesis time was chosen between 1 and 18 h to control the crystallinity. The best sample (GENCS-3) was prepared within 3 hour and had flower-like structures. Its electrochemical tests showed that it had good ion transmission efficiency and excellent specific capacitance (1816 F/g at 1 A/g, 1620 F/g at 10 A/g). Especially, its retention rate could arrive at 89.4% after 1000 cycles under a current density of 5 A/g. The weak crystalline nickel–cobalt sulfide is easy to form porous structure, which can significantly increase capacitance, also maintaining good rate performance.


Introduction
As an important energy storage apparatus, supercapacitor has attracted widespread attention all over the world [1,2]. Pseudocapacitors are also called electrochemical capacitors, which store charge through a reversible redox reaction that occurs on or near the surface of their electrode material. During the reversible redox, pseudocapacitors can keep storing energy to a high energy density. [3][4][5][6][7][8] However, their electrodes are a key factor affecting their performances as supercapacitors. People are looking for good electrode materials and hope that these electrodes will not only be highly active but also environmentally friendly and affordable. In recent years, cobalt-nickel sulfide has always been a hot spot in the field of electrode materials and has been highly expected. The material has more than one morphology and its sea urchin-like, flower-like and nano-sheet morphology has also been developed and researched [9][10][11]. Our team has carried out a series of studies on the material as a capacitive electrode [10,12,13]. The higher the crystallinity of the product obtained by solvothermal synthesis with the increase of temperature and the longer the holding time, the faster the electron transfer rate of the product synthesized by solvothermal synthesis. Because the worse the crystallinity of the product, the larger the specific surface area of the product [14,15]. The research of nanoparticles has always been in a hot research direction. Because the specific surface area of nanoparticles is relatively large and the resistance to low temperature and low pressure is also excellent, it is widely used in aspects [16][17][18][19]. We have successfully prepared a serial of products of nickelcobalt compounds with different shapes, which were at last transformed into honeycomb-like cobalt nickel sulfatide through room-temperature sulfidation. And the obtained honeycomb nickel cobalt sulfatide showed a high specific capacity (1020 F/g) at large current density (30 A/g) and good cycle stability (82.35% after 3000 cycles) [20]. We also used solvothermal method in the preparation of nickel cobalt sulfide. We selected isopropanol and ethylene glycol as the mixed solvent unit and changed their ratio to observe the morphology of the corresponding products. It is interested that some of the obtained samples were also honeycomb structure, showing a capacity of 1244 F/g at 1 A/g after adding rGO [12]. Further, we also used various double-solvent systems, including EG-IPA, glycerol-IPA, EG-methanol, and EG-ethanol, to fabricate nickel cobalt sulfide. Of course, the choice of different solvents here was to understand the effect of solvent components on the morphology of the products. We did obtain various morphologies of nickel cobalt sulfide, including granular structure, hollow sphere structure, honeycomb structure, irregular sphere-like structure, and so on. As expected, the honeycomb structure samples still performed best, with capacity of up to 1464 F/g [13].
Recently, a new study has been made in recent experiments on nickel cobalt sulfide. We found that the weak crystalline samples exhibited better capacitive performance than before. In the experiments, solvothermal method was used for the preparation, and a two-solvent system of glycerol and isopropanol was adopted. In this paper, common experimental methods are used to prepare nickel-cobalt sulfides, and the crystallization of nickel-cobalt sulfides is controlled by adjusting the synthesis time to obtain an optimal combination of crystallinity and performance. The morphology and crystallinity of the products were controlled by reaction time. The results showed that all products are irregular particle with closed cavities. Of course, under short synthesis time, some cavities were opened, showing flower-like shape. The samples with the flower shape could be not crystallized well, but they could have a capacitance of up to 1800 F/g. In this experiment, the   method of one-step synthesis of nickel-cobalt sulfides is adopted, and the performance of nickel-cobalt sulfides synthesized in a short time is better. Compared with the two-step synthesis and the longer synthesis time, our method reduces the uncertainty of the synthetic products and saves the time cost. [21][22][23] 2 Experimental

Synthesis of cobalt nickel sulfide
All chemicals were purchased from high-purity commercial channels. In typical procedure, first of all

Electrochemical measurement
Electrochemical properties were tested in three-electrode electrochemical cell. The electrochemical properties of the resulting electrodes were studied by using 6 M KOH solution as electrolyte. The work electrode was fabricated from a mixture of 80 wt% cobalt-nickel sulfide material, 10 wt% acetylene black, and 10 wt% polyvinylidene difluoride binder. After blending, these mixtures were coated onto a clean nickel foam with an effective geometric area of 1 9 1 cm 2 and then dried under vacuum for 12 h at 60°C. The mass load for NiCo 2 S 4 was calculated to be around 3.0 mg/cm 2 . The platinum foil (1.5 9 1.5 cm 2 ) and a mercuric oxide electrode were used as the counter and reference electrode, respectively. The cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), and electrochemical impedance spectroscopy (EIS) were tested by CHI660D electrochemical workstation. The cycle stability was tested with LANHE test system under a current density of 5 A/g. corresponding sample. Therefore, the crystallinity of GENCS-1,3,6 samples is poor, while that of GENCS-9, 12, 15, and 18 samples is obviously better than that of the previous. Certainly, the XRD also reflects that there is a fluctuation about crystallinity around 12 h, which may be due to environmental factors.

Morphology characterization
and electrochemical characterization Figure 2a and b is the SEM pictures of GENCS-1,3 and shows that there are some opened cavities in the samples, which looks like a blooming flower. In fact, the TEM pictures, as shown in Fig. 2c, d, present that the products are some irregular nanoparticles with cavity. Our previous research suggests that when the cavity walls are thin, and at the same time, the water pressure in the cavity is large enough, the cavity walls may be broke [26,27]. Therefore, it can be inferred that the product is synthesized within 3 h, and its cavity wall is thin, and in this case, the cavity wall is easily opened by a steam breakthrough. Figure 2e, f shows the lattice stripes and the spacing of the lattice stripes, which was measured to be 0.234 nm, 0.186 nm, 0.301 nm, and 0.297 nm, respectively, corresponding to (400), (511), (220), and (311) of NiCo 2 S 4 . Some electrochemical measurements were also carried out here. It tested their charge-discharge curve (GCD) and charge-discharge curve at different current densities. Figure 2g, h shows the GCD curves and rate curves, respectively. Using the GCD discharge curve of 1 A/g current density, the discharge capacitance of the samples can be calculated to be 1874 F/g and 1816 F/g, respectively. But, at 10 A/g, the calculated value is 1260 F/g and 1620 F/g, respectively. Obviously, GENCS-1 performance declined significantly with current rising, which may be related to its poor structural stability caused by its too thin cavity wall. Certainly, these two samples must benefit from their flower-like structure to achieve high capacitance performance. The open holes will expose more active sites, thereby increasing the ion exchange rate. The SEM images of the remaining samples are shown in Fig. 3a-e. Their reaction time is 6 h, 9 h, 12 h, 15 h, and 18 h, respectively. It is observed that almost all these samples show platelet shapes, with a size of about 100 nm, but there is no any flower-like structure. This may indicate that when the reaction time is more than 6 h, not only the cavity wall becomes thicker, but the atoms on the wall also are arranged more orderly, which leads to stronger walls and all the cavities not to be broken again. The discharge test (GCD) and capacitance performance test under different current density conditions were performed on these samples, as shown in Fig. 3f and g. These data indicate that at 1 A/g current density, the capacitance of these samples is between 1000 and 1600 F/g, and when the current density rises to 10 A/g, the capacitance attenuates between 960 and 1350 F/g, showing a significant slip compared to the previous two samples, especially compared with GENCS-3. It seems that the better the crystallinity, the worse the corresponding capacitance property. On the other hand, it can be seen from Figs. 2h and 3g that the rate performance of the samples with good crystallinity will be better. Therefore, GENCS-18 has a capacitance of 1020 F/g at 1 A/g and 960 F/g at 10 A/g, showing an excellent capacitance retention rate of 94.12%.
The CV curves of the GENCS-3 sample are also tested under different scanning rates of 5 mV/s, 10 mV/s, 20 mV/s, 30 mV/s, 40 mV/s, 50 mV/s, and 100 mV/s, as shown in Fig. 4a. It can be seen in the figure that with the increase of scanning rate, the area of the curve also increases, and the position of the redox peak also moves to both ends. Due to the change of sweep speed, the polarization of the electrode occurs, and the redox peak current is also increasing. In a certain voltage range, the peak current does not reach the peak point, so the peak is moving. Figure 4b shows the GCD curves of the GENCS-3 at 1 A/g, 3 A/g, 5 A/g, 8 A/g, and 10 A/g in a potential window of 0-0.5. It can be seen that all GCD curves have symmetry under different current density, indicating that the sample has good chargedischarge reversibility. Based on Fig. 4b, the specific capacitance of GENCS-3 is calculated under different current density, showing a value of 1816 F/g, 1752 F/g, 1710 F/g, 1648 F/g, and 1620 F/g, respectively. Figure 4c provides the CV curves of all samples at 5 mV/s sweep speed, and there are obvious redox peaks of pseudocapacitance, which are mainly attributed to the Faraday redox reaction between Co 2? /Co 3? (Ni 2? /Ni 3? ) cations and OHanions. Figure 4d, e shows the impedance plots of the lowfrequency and high-frequency regions of all the b Fig. 6 All samples GENCS-1 (a), GENCS-3 (b), GENCS-6 (c), GENCS-9 (d), GENCS-12 (e), GENCS-15 (f), and GENCS-18 (g) particle size distribution samples, respectively, where the intercept of the curve with the X-axis represents the base resistance of the electrode solution. It can be seen from the figure that the slope of GENCS-3 in the low-frequency region is greater than that of other samples. This shows that its ion transmission efficiency to electrolyte is better than others. The EIS of GENCS-1, 3 has a larger slope in the low-frequency region, indicating their better diffusion behavior [28]. However, the Rs value (4.4 X) of the sample GENCS-1 is much greater than that of the sample GENCS-3 (0.5 X). Figure 4f shows the Ip*v 1=2 fitting curve of the oxidation peak of the cyclic voltammetry curve of all samples at different scan rates. Here, Ip is the peak electricity (A) of the oxidation peak or reduction peak and v is the cycle Scanning rate of ring voltammetry (mV/s). The oxidation peak current (Ip) is proportional to the square root of the scan rate, indicating that the process of OHembedding in the NiCo 2 S 4 electrode is a diffusion process. The relationship between CV peak current (Ip) and scan rate is where D OH À is the OHdiffusion coefficient. The fitting curve slope of sample GENCS-3 is 1.5, which is much larger than that of other sample fitting curves. Therefore, the diffusion coefficient (D OH À ) of GENCS-3 sample is larger than that of other samples, thus, showing excellent electrochemical performance. Figure 4g, i is retention rate chart of all samples. It can be seen that the retention rate of GENCS-3 is 89.5%, which is a little lower than that of the GENES-18. However, the difference does not exceed 5%. The degradation is due to the long-term cycling that causes certain damage to the electrode. Figure 4h shows the cycle performance of the GENCS-3 sample. It is cycled 1000 times at a current density of 5 A/g, and the performance remains 89.4%. This sample showed excellent cycle stability [12].
Here, there are two samples selected for BET testing, one with maximum capacitance and the other with minimum capacitance. Figure 5a, b shows the BET results of the samples of GENCS-3 and GENCS-18, respectively. As showing in Fig. 5a, b, both samples have many micropores, few mesopores, and few macrospores. However, the pore size of GENCS-3 is mostly in the micropore area, while the pore size distribution of GENCS-18 is not very uniform in the pore area, and the pore size fluctuates greatly. The average pore size of GENCS-3 is 7 nm, and that of GENCS-18 is 10 nm. And GENCS-3 has the surface area of 40.3953 m 2 /g, but the GENCS-18 only has a surface area of 10.2466 m 2 /g. As thought, GENCS-3 has a larger specific surface area, which is mainly due to its poor crystallinity and more defects. Of course, larger specific surface area can enlarge the effective release area between the electrode and the electrolyte ions, which is beneficial to enhance the diffusion of electrolyte ions in the electrode, thereby improving electrochemical performance [29]. GENCS-3 reason for the steep adsorption curve in the BET is that the multilayer adsorption occurs first during the adsorption process, and the agglomeration occurs when the adsorption layer on the pore wall reaches sufficient thickness. If the adsorbent has large pore size or strong molecular interaction, it may continue to adsorb to form a multilayer layer, and the adsorption isotherm will continue to rise.
All the samples were tested and analyzed. Figure 6 shows the particle size distribution of the sample, the sample GENCS-3 particle size is distributed around 1.1 lm, the rest of the sample particle size is mostly distributed between 300 to 800 nm. The particle size distribution is shown in Table 2.

Conclusions
In this paper, the double solvothermal method is used to prepare nickel cobalt sulfide nanoparticles with closed cavity in irregular shape. The synthesis time was selected to control the crystallinity. It is found that for samples with synthesis time less than 3 h, the crystallinity is poor, but there are some open cavities in the body, like flowers, which is beneficial to increase the capacitance. As the synthesis time is more than 6 h, the longer the time, the better the crystallinity of the obtained sample, and the corresponding rate performance will be better. The best sample is GENCS-3 with a short reaction time of 3hour, which has flower-like structures, and also has good ion transmission efficiency, excellent specific capacitance (1816 F/g at 1 A/g, 1620 F/g at 10 A/g), and good retention rate of 89.5%. Our study shows that nickel-cobalt sulfide with weak crystallinity is easy to form porous structure, which can significantly increase capacitance but may lose a little rate performance. Of course, the loss is controllable and does not exceed 5% Funding