One-Pot Synthesis of NiSe2 with Layered Structure for Nickel-Zinc Battery

Transition metal organic framework materials and their selenides are considered to be one of the most promising cathode materials for nickel-zinc (denoted as Ni-Zn) batteries due to their low cost, environmental friendliness, and controllable microstructure. Yet, their low capacity and poor cycling performance severely restricts their further development. Herein, we developed a simple one-pot hydrothermal process to directly synthesize NiSe2 (denotes as NiSe2-X based on the molar amount of SeO2 added) stacked layered sheets. Benefiting from the peculiar architectures, the fabricated NiSe2−1//Zn battery based on NiSe2 and the Zn plate exhibits a high specific capacity of 231.6 mAh g−1 at 1 A g−1, and excellent rate performance (162.8 mAh g−1 at 10 A g−1). In addition, the NiSe2//Zn battery also presents a satisfactory cycle life at the high current density of 8 A g−1 (almost no decay compared to the initial specific capacity after 1000 cycles). Additionally, the battery device also exhibits a satisfactory energy density of 343.2 Wh kg−1 and a peak power density of 11.7 kW kg−1. This work provides a simple attempt to design a high-performance layered cathode material for aqueous Ni-Zn batteries.


Introduction
With the continuous consumption of conventional energy resources and the increasing demand for energy, the energy crisis and the consequent environmental problems have attracted extensive attention. A large quantity of energy storage devices have been developed in the past few decades, such as supercapacitors, Li-ion batteries, and aqueous rechargeable batteries [1][2][3][4][5][6][7]. Among those aforementioned, energy storage devices such as aqueous Zn-based batteries are attracting considerable attention due to their high theoretical capacity (5851 mAh cm −3 and 820 mAh g −1 ), low redox potential (0.763 V vs. the standard hydrogen electrode (SHE)), high output voltage (such as Ni-Zn battery about 1.8 V), good compatibility with aqueous electrolytes, eco-friendliness, low cost, and abundant source of the Zn anode [8][9][10][11][12][13][14][15][16]. Compared to the high theoretical capacities of zinc anode materials, most reported cathode materials have relatively low capacities (200-300 mAh g −1 ) [1,5,[17][18][19][20][21][22]. In this scenario, the key to improving the capacity of Ni-Zn battery is to explore suitable cathode materials.
Among current cathode materials, transition metal selenides (TMSs) own higher conductivity than transition metal oxides and hydroxides. For example, the binding energy of Ni-Se is lower than that of Ni-O and Ni-S with weaker electronegativity, reducing the energy consumption of conversion reaction [23]. Additionally, due to the higher degree of reversibility of M-Se compared to M-O bonds (M = metal), TMSs always exhibit superior electrochemical performance such as high specific capacity, rate capability, and long-term cyclability [24][25][26][27][28]. Due to the porous structure and easy adjustment, MOFs are ideal precursor materials for the construction of TMSs, which have been identified as cathode candidates for zinc-based batteries. The synthesized TMSs through the above method not only own controllable nanostructures but also show greater electrochemical performance [6,7,[29][30][31][32]. For instance, CoSe 2-x -derived cobalt oxides nanoparticles synthesized by Tang et al. obtained a capacity of 7.42 mAh m −2 at 10 mA cm −2 with 0.02% cycle −1 capacity decay over 4200 cycles [19]. Porous NiCoSe 2 @NiOOH/CoOOH nanoflakes reported by Cui and his co-workers exhibited a capacity of 108.9 mAh g −1 at 1 A g −1 with 105.1% capacity retention after 4000 cycles [33]. MoSe 2 decorated Ni/Co selenide complex hollow arrays presented a capacity of 142 mAh g −1 at 2 A g −1 with~36% capacity loss after 1000 cycles [34]. However, the capacity and long cycle stability of most reported TMSs//Zn batteries have still fallen below expectations and cannot meet future practical applications. Moreover, the traditional method of constructing selenide by MOFs is relatively complicated, which requires high temperature carbonization followed by high-temperature selenization under Ar flow. Therefore, the investigation of new durable TMSs cathodes with high capacity is still highly favorable and challenging for achieving widely applicable aqueous rechargeable zinc-based batteries.
In this work, we propose a facile one-pot hydrothermal synthesis of NiSe 2 stacked layered sheets without Ni-MOFs template and high temperature heat treatment. The NiSe 2 //Zn battery was assembled by using the synthesized NiSe 2 nanosheet as the cathode and zinc as the anode of the aqueous Ni-Zn batteries, and the fabricated battery obtained a higher specific capacity of 231.6-162.8 mAh g −1 at various current densities of 1-10 A g −1 , which indicated the good rate performance. Moreover, it shows excellent cycling performance with almost no decay compared to the initial specific capacity value after 1000 cycles at 8 A g −1 . Impressively, the fabricated NiSe 2 //Zn battery shows outstanding performance with a maximum energy density of 343.2 Wh kg −1 , and a maximum power density of 11.7 kW kg −1 , substantially outstripping most of the recently reported Ni-Zn batteries.

Results and Discussions
The Ni-MOFs nanomaterials were first synthesized by a simple one-pot hydrothermal process. As shown in Figure 1a, the pristine Ni-MOFs nanomaterials own a cluster structure stacked with numerous layered sheets, and the self-assembled nanosheets can be better observed in the high magnification SEM images (Figure 1d). Based on the above synthesis process, NiSe 2 −X nanomaterials with different degrees of selenization can be obtained by adding different masses of SeO 2 into the precursor solution of Ni-MOFs synthesis. Figures 1b-c and S1 show the variation of the structure of NiSe 2 nanosheets under different SeO 2 addition conditions. It can be seen that NiSe 2 −0.5 still retains part of the original nanosheets due to the lack of SeO 2 during the synthesis process. Meanwhile, in Figure 1f, the homogeneous growth and accumulation of thick nanosheets of NiSe 2 −1 can be observed. Through the high magnification SEM images of the corresponding area (Figures 1e, f and S1b), it can be clearly observed that the thickness of nanosheets increased with the increased concentration of SeO 2 . Figure 1i-l shows the EDS images of NiSe 2 −1 nanomaterials, which indicates that Ni, Se, and C elements are uniformly distributed in the sample. In order to further explore the microstructure of Ni-MOFs and NiSe 2 -1 nanomaterials, we utilized TEM (Figures 1g and S2). Figure 1g shows that the NiSe 2 -1 nanomaterials consisted of the square-shaped nanosheets. The inset picture in Figure 1g shows the corresponding selected area electron diffraction (SAED) pattern indicating the typical polycrystalline structure of NiSe 2 −1. The lattice fringes of the NiSe 2 −1 material were visible in Figure 1h, where the characteristic lattice spacing (0.269 nm) of NiSe 2 −1 was consistent with the (210) lattice plane of NiSe 2 (JCPDS NO. 41-1495) [35]. Molecules 2023, 28, x FOR PEER REVIEW 3 of 11   (Figure 1h). Additionally, the characteristic peaks of NiSe2−0.5 sample contain part characteristic peaks of Ni-MOFs, indicating incomplete transformation, which is consistent with the SEM results. To further explore the elemental composition and valence state of the samples, XPS was used to analyze the as-synthesized samples. Figure 2b is the survey spectra of Ni-MOFs and NiSe2−1 samples. It can be clearly seen that the NiSe2−1 sample synthesized by adding SeO2 has an obvious characteristic peak of Se, indicating the successful doping of Se. The C 1s spectrum of NiSe2−1 sample can be resolved into three characteristic peaks of sp 2 -C, sp 3 -C, C=O, and O-C=O bonds, respectively ( Figure 2c) [22,23]. From the Ni 2p spectrum (Figure 2d), two peaks located at 855.9 and 873.7 eV were attributed to Ni 2+ in Ni 2p 3/2 and Ni 2p1/2 respectively. Meanwhile, the peaks located at 857.3 and 875.3 eV corresponded to the Ni 3+ in Ni 2p3/2 and Ni 2p1/2, and   Figure 1h). Additionally, the characteristic peaks of NiSe 2 −0.5 sample contain part characteristic peaks of Ni-MOFs, indicating incomplete transformation, which is consistent with the SEM results. To further explore the elemental composition and valence state of the samples, XPS was used to analyze the as-synthesized samples. Figure 2b is the survey spectra of Ni-MOFs and NiSe 2 −1 samples. It can be clearly seen that the NiSe 2 −1 sample synthesized by adding SeO 2 has an obvious characteristic peak of Se, indicating the successful doping of Se. The C 1s spectrum of NiSe 2 −1 sample can be resolved into three characteristic peaks of sp 2 -C, sp 3 -C, C=O, and O-C=O bonds, respectively (Figure 2c) [22,23]. From the Ni 2p spectrum (Figure 2d), two peaks located at 855.9 and 873.7 eV were attributed to Ni 2+ in Ni 2p 3/2 and Ni 2p 1/2 respectively. Meanwhile, Molecules 2023, 28, 1098 4 of 11 the peaks located at 857.3 and 875.3 eV corresponded to the Ni 3+ in Ni 2p 3/2 and Ni 2p 1/2 , and the others were related to the satellite peaks [22,36]. As displayed in Figure 2e, the Se 3d 5/2 , 3d 3/2 , and SeO x level appeared at 55.7, 56.9 and 60.2 eV, respectively [35,37,38]. Additionally, the spectra of O 1s for Ni-MOFs and NiSe 2 −1 samples can be deconvoluted into three peaks assigned as Se-O (530.8 eV), OH -(533.6 eV) and adsorb water (532.5 eV), respectively ( Figure S3) [35].
Molecules 2023, 28, x FOR PEER REVIEW 4 of 11 the others were related to the satellite peaks [22,36]. As displayed in Figure 2e, the Se 3d5/2, 3d3/2, and SeOx level appeared at 55.7, 56.9 and 60.2 eV, respectively [35,37,38]. Additionally, the spectra of O 1s for Ni-MOFs and NiSe2−1 samples can be deconvoluted into three peaks assigned as Se-O (530.8 eV), OH -(533.6 eV) and adsorb water (532.5 eV), respectively ( Figure S3) [35].  In order to evaluate the electrochemical performance of the prepared electrode materials, we assembled Ni-MOFs, NiSe 2 −X and Zinc plate into aqueous Ni-Zn batteries. Figure 3a shows different cyclic voltammetry (CV) curves of NiSe 2 -1 in the initial five cycles. It can be clearly observed that the current response signal increases with the number of scans, indicating the improvement of its electrochemical performance. This phenomenon can be attributed to the continuous activation of the electrode as a result of the surface chemical reaction between NiSe 2 −1 and KOH electrolyte [39,40]. Figure 3b shows the CV curves of the fabricated batteries with the working window from 1.6-2.0 V, which exhibit typical battery behavior with obvious redox peaks. As illustrated in Figure 3c, the NiSe 2 −1 displays the highest specific capacity, while the GCD curves of the as-prepared Ni-MOFs and NiSe 2 −X samples are at the same current density (3 A g −1 ). The corresponding GCD curves of NiSe 2 -1 at different current densities are depicted in Figure S4, and it can be clearly seen that the discharge specific capacity reached a remarkable 231.6 mAh g −1 at 1 A g −1 , and the discharge specific capacity was still maintained at 162.8 mAh g −1 with a high current density of 10 A g −1 , proving its excellent rate performance. Meanwhile, the obvious charge/discharge voltage plateaus appear in the GCD curves at different current densities, and the voltage range of the plateau (at 1.85-1.95 V) matches the position of the redox peaks in the cyclic voltammetry curves correspondingly. This result further confirms that the NiSe 2 -1 electrode undergoes a redox reaction, which has the typical characteristics of a battery-type material. The rate performance of the original Ni-MOFs and NiSe 2 −1 electrodes applied to Ni-Zn batteries is shown in Figure 3d. It is noteworthy that the NiSe 2 −1 exhibits an apparent increase in specific capacity during the initial cycles, which follows the same trend as that in Figure 3a. Both rate capability curves can illustrate that there is no significant decay after the rate cycling due to the good rate capability of Ni-based materials [23]. However, when the current density falls back to 1 A g −1 after 40 cycles, the average capacity retention of the NiSe 2 −1 electrode can even reach 117% compared to the initial average capacity, which is better than that of the Ni-MOFs electrode (96%) [41]. In order to evaluate the electrochemical performance of the prepared electrode materials, we assembled Ni-MOFs, NiSe2−X and Zinc plate into aqueous Ni-Zn batteries. Figure 3a shows different cyclic voltammetry (CV) curves of NiSe2-1 in the initial five cycles. It can be clearly observed that the current response signal increases with the number of scans, indicating the improvement of its electrochemical performance. This phenomenon can be attributed to the continuous activation of the electrode as a result of the surface chemical reaction between NiSe2−1 and KOH electrolyte [39,40]. Figure 3b shows the CV curves of the fabricated batteries with the working window from 1.6-2.0 V, which exhibit typical battery behavior with obvious redox peaks. As illustrated in Figure 3c, the NiSe2−1 displays the highest specific capacity, while the GCD curves of the as-prepared Ni-MOFs and NiSe2−X samples are at the same current density (3 A g −1 ). The corresponding GCD curves of NiSe2-1 at different current densities are depicted in Figure S4, and it can be clearly seen that the discharge specific capacity reached a remarkable 231.6 mAh g −1 at 1 A g −1 , and the discharge specific capacity was still maintained at 162.8 mAh g −1 with a high current density of 10 A g −1 , proving its excellent rate performance. Meanwhile, the obvious charge/discharge voltage plateaus appear in the GCD curves at different current densities, and the voltage range of the plateau (at 1.85-1.95 V) matches the position of the redox peaks in the cyclic voltammetry curves correspondingly. This result further confirms that the NiSe2-1 electrode undergoes a redox reaction, which has the typical characteristics of a battery-type material. The rate performance of the original Ni-MOFs and NiSe2−1 electrodes applied to Ni-Zn batteries is shown in Figure 3d. It is noteworthy that the NiSe2−1 exhibits an apparent increase in specific capacity during the initial cycles, which follows the same trend as that in Figure 3a. Both rate capability curves can illustrate that there is no significant decay after the rate cycling due to the good rate capability of Ni-based materials [23]. However, when the current density falls back to 1 A g −1 after 40 cycles, the average capacity retention of the NiSe2−1 electrode can even reach 117% compared to the initial average capacity, which is better than that of the Ni-MOFs electrode (96%) [41]. In order to further understand the storage mechanism of Zn 2+ , we used different electrochemical methods to explore its kinetics. Figure 4a shows the cyclic voltammetry (CV) curves of NiSe 2 −1 electrode at different scan rates. It is obvious that with the increase of the scanning rate, the peak pattern of the redox peaks does not change, but the position of the redox shifts, which is mainly attributed to polarization. Theoretically, the peak current (i) and the scan rate (v) of the CV curves follow the equation: i = av b , where a and b are variable parameters. A b value of 0.5 indicates a diffusion-controlled behavior of the Zn-ion extract/insert electrode material, and a b value of 1 indicates a control by the capacitive process. The b values can be obtained by fitting a linear relationship between log i and log v [42,43]. As exhibited in Figure 4b [44,45]. For the as-prepared NiSe 2 -1 electrode, the k 1 v value ranges from 67% to 98% at the scan range of 0.2-1 mV s −1 . Typically, the original CV curve and fitted pseudocapacitive contribution (the shaded area) at 1 mV s −1 are shown in Figure S5. In addition, the capacity contribution still dominates at the lowest scan rate. Under the influence of pseudocapacitance behavior, the active material NiSe 2 −1 tends to undergo redox reaction on the electrode surface to store charge, which can effectively protect the structure of the NiSe 2 −1 and improve the cyclic stability of the electrode material. More importantly, the near-surface pseudocapacitance effect can obtain a shorter ion transport path and a faster electron transport rate, which is conducive to enhancing the electrochemical performance of NiSe 2 −1 [23]. The electrochemical impedance spectroscopy (EIS) in Figure 4d, where the equivalent circuit for simulation was inset, shows that the Nyquist plot consists of the semicircle in the high-frequency range related to the chargetransfer resistance and the low-frequency regions related to a diffusion-limited electron transfer process of an electrode [46,47]. It can be clearly seen that with the addition of SeO 2 ( Figure S6), the structure of the material gradually becomes thicker and reacts completely. The corresponding EIS of Ni-MOFs and NiSe 2 −X (X = 0.5, 1, 2) are depicted in Figure S7. Apparently, in the high frequency region, the R ct of NiSe 2 -1 electrode is 34.31 Ω, while the R ct of Ni-MOFs electrode is 40.6 Ω; in the low frequency region, NiSe 2 -1 electrode has a higher slope of the line than Ni-MOFs electrode, which means a faster electron diffusion. Therefore, the NiSe 2 −1 electrode exhibits better impedance performance than the pristine Ni-MOFs in both low and high frequency regions.
In order to further explore the practicability of the Ni-Zn batteries prepared, we investigated the cycle performance, energy density, and power density. Figure 5a displays the comparison of the long cycle performance of Ni-MOFs and NiSe 2 −1 as the cathode electrode of Ni-Zn batteries. It can be seen that the specific capacity of Ni-MOFs is slightly higher than that of NiSe 2 −1 during the very initial cycles, corresponding to the initial rate capability curve in Figure 3e. However, it is worth noting that the specific capacity of NiSe 2 −1 has a rather high percentage increase compared to the pristine Ni-MOFs, and the increase in capacity over about 100 cycles is attributed to the surface chemistry between NiSe 2 −1 and the KOH electrolyte, which activates the electrode [39,40]. After reaching the peak specific capacity value, both materials began to decay slowly, and after 1000 cycles, NiSe 2 -1 decayed by 50% of the peak capacity (Ni-MOFs, 61%), as shown within Figure 5a. Additionally, based on the GCD curves, the energy density and power density of the NiSe 2 −1//Zn battery were calculated to obtain the Ragone plots ( Figure 5b). Noteworthily, the NiSe 2 −1//Zn battery obtained a maximum energy density of 343.2 Wh kg −1 and a peak power density of 11.7 kW kg −1 , which is superior to most of the previously reported rechargeable batteries, including Ni//Bi battery (1.4 kW kg −1 , 70.9 Wh kg −1 ) [9], Zn//OD-ZMO@PEDOT battery (273. 4  the NiSe2−1//Zn battery obtained a maximum energy density of 343.2 Wh kg −1 and a peak power density of 11.7 kW kg −1 , which is superior to most of the previously reported rechargeable batteries, including Ni//Bi battery (1.4 kW kg −1 , 70.9 Wh kg −1 ) [9], Zn//OD-ZMO@PEDOT battery (273.4 Wh kg −1 , 2.6 kW kg −1 ) [21]   the NiSe2−1//Zn battery obtained a maximum energy density of 343.2 Wh kg −1 and a peak power density of 11.7 kW kg −1 , which is superior to most of the previously reported rechargeable batteries, including Ni//Bi battery (1.4 kW kg −1 , 70.9 Wh kg −1 ) [9], Zn//OD-ZMO@PEDOT battery (273.4 Wh kg −1 , 2.6 kW kg −1 ) [21]

Synthesis of Ni-MOFs
All chemicals in this work were utilized as received without any purification (the purity of chemicals was 98%). 1 mmol Ni(NO 3 ) 2 ·6H 2 O and 1 mmol PTA were dissolved in a mixed solution consisting of 20 mL of DMF and 5 mL of 0.4 M NaOH, stirred well at room temperature and transferred into a 50 mL Teflon-lined stainless steel reactor, which was subjected to hydrothermal reaction at 160 • C for 12 h in an oven. After the reaction, the solution was cooled down to room temperature, centrifugally, and dried at 60 • C for 12 h to obtain Ni-MOFs.

Synthesis of NiSe 2 Architectures
The synthesis of NiSe 2 is based on the above synthesis method. It only needs to add different masses of SeO 2 to the precursor solution for the synthesis of Ni-MOFs. The resulting products were named as NiSe 2 −X (X = 0.5,1,2) according to the molar amount of SeO 2 added.

Fabrication of Aqueous Ni-Zn Batteries and Electrochemical Measurements
The aqueous Ni-Zn batteries were fabricated by a piece of Zn plate, the synthesized NiSe 2 −X (X = 0.5, 1, 2) electrode, Zn flakes, and a mixed aqueous consisting of 3 M KOH and 0.2 M Zn(CH 3 COO) 2 as the anode, cathode, and electrolyte. The mass loading was about 2.4 mg cm −2 . The electrochemical performances (CV, CP, EIS) of the fabricated batteries were characterized by an electrochemical workstation (CHI 660E) and Land 2001A.
All electrochemical characterization was carried out in a two-electrode system with a potential window of 1.6 to 2.0 V. Galvanostatic charge/discharge (GCD), Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were performed using an electrochemical workstation (CHI 660E). The cycle life of the Ni-Zn batteries was tested by using a Land 2001A.

Conclusions
In summary, based on the synthesis of Ni-MOF, we have developed a simple one-pot synthesis method of NiSe 2 with a layered structure. Benefiting from the unique nanosheet structure, the assembled aqueous NiSe 2 −1//Zn battery exhibits remarkable specific capacity (231.6 mAh g −1 at 1 A g −1 ), excellent rate performance (162.8 mAh g −1 at 10 A g −1 ), and prominent energy density (343.2 Wh kg −1 ) as well as excellent long-term cycling stability (almost no decay after 1000 cycles at 8 A g −1 compared to the initial specific capacity value). Moreover, NiSe 2 can be easily prepared by a one-pot hydrothermal synthesis method, which facilitates the widespread and commercial application of the energy storage devices and lays the foundation for future research and development of such materials.