Li-Ni Metal Oxides Processed with Rapid Atmospheric-Pressure-Plasma Jet for Flexible Gel-Electrolyte Li-Ion Hybrid Supercapacitors

We use screen printing to deposit LiCl + Ni(NO3)2·6H2O pastes on a flexible carbon cloth substrate and then calcine it using a nitrogen atmospheric-pressure-plasma jet (APPJ). Through the high-temperature treatment by APPJ, pastes can be rapidly converted into Li-Ni oxides (LNOs). The LNOs on carbon cloth are then used as the electrodes of flexible gel-electrolyte Li-ion hybrid supercapacitors (Li-ion HSCs). The best areal capacity of 21.076 mC cm−2, as measured by cyclic voltammetry, is achieved by APPJ treatment at 620 °C for 480 s. To demonstrate the flexibility of the Li-ion HSCs, the Li-ion HSCs were bent at different curvatures to measure its performance. After bending test, the capacity remains >93% under bending with a curvature of up to 2 cm−1.

Atmospheric-pressure plasma (APP) is a chemically active medium that can be used for not only surface treatment and modification but also rapid processing of materials. [1][2][3][4] Compared with low-pressure plasma (LPP), APP does not require costly vacuum chambers and pumps that require frequent maintenance. 5,6 Frequently encountered APP technologies includes corona discharge, dielectric barrier discharge (DBD), and APP jet (APPJ). 6 A corona discharge is heterogeneous and has low current density. DBD is frequently used for large-area low-temperature materials processing. 7 An APPJ flow has been used in different fields such as biomedical applications, 8 etching and deposition, 9,10 nanoparticle generation, 11 and supercapacitors (SCs). 12 An SC is a type of energy storage device that has a higher power output and longer cycle lifetime and is more ecofriendly than a battery. Further, it has higher energy density than a conventional capacitor. [13][14][15] Its energy storage mechanisms can be catalogized as electric double-layer capacitance (EDLC), pseudocapacitance (PC), and hybrid. [16][17][18][19] EDLC is the most common mechanism, wherein charge storage relies on surface energy ion adsorption and desorption. 20 Most of EDLC commonly use porous sponge-like materials as electrodes, with carbon being the most frequently utilized material, such as activated carbon (AC), graphene, and carbon nanotubes (CNTs). However, because EDLC stores charge on the surface of electrodes, its energy density is affected by the specific surface area of the electrode. The current challenges for EDLC include low energy density and the presence of dielectric absorption effects. [21][22][23] PC can store more energy than EDLC because it involves charge transfer between a guest ion and a host active material. 24 Metal oxides, such as manganese dioxide and ruthenium dioxide, are commonly used as electrode materials for PC. Due to the use of the Faraday effect for charge storage in PC, chemical and physical changes occur during the charging and discharging processes, leading to the limited cycle life.
A hybrid SC (HSC) is an energy storage device with characteristics that lie between those of a battery and an SC.
The characteristics of HSC are similar to those of PC, expect that the electrode materials exhibit the Faraday effect. 25- 30 An HSC's energy density could be higher than that of a battery-type electrode SC, and its power density could be higher than that of a capacitive electrode SC. 31 To achieve better performance, HSC typically utilize electrode materials that consist of a combination of different components. These components include both Faradaic materials, which exhibit battery-like behavior, and non-Faradaic materials. In addition to using composite materials, both symmetrical and asymmetrical electrode assembly are notable features of HSC. [32][33][34][35][36] Various lithium transition metal oxides such as LiCoO 2 , LiNiO 2 , and LiMnO 2 have been investigated for use in HSCs. 37 Among them, LiNiO 2 has a higher discharge capacity and lower cost and is more ecofriendly than LiCoO 2 . 38-40 LiNiO 2 can be synthesized by various methods including solid-state synthesis, 41 sol-gel method, 42 hydrothermal method, 43 and emulsion method. 44 However, most of these synthesis methods are time-consuming and involve high-temperature reactions.
In our previous study, we used an APPJ to rapidly synthesize rGO-SnO 2 and rGO-MnO x SCs. 45,46 We also investigated how the performance of the LiMn 2 O 4 HSCs changed after treatment with APPJ for different processing times. 47 The above research results all indicate that proper plasma treatment time can effectively assist in the synthesis of materials and enhance SC performance. With the advancement of technology, wearable devices and portable electronic products have gained significant attention in recent years, and the flexibility of batteries has become a crucial challenge. [48][49][50] Therefore, in this study, we used the APPJ to rapidly synthesize Li-Ni oxide (LNO), replacing the time-consuming processes used previously, and assembled it into HSCs. In addition to conducting electrochemical tests to evaluate its performance, we also performed bending tests with different curvatures to assess its mechanical properties.  45,46,51 After the mixture was stirred at 850 rpm for 24 h, it was condensed by using a rotatory evaporator at 55°C for 5 min to obtain the pastes.
Fabrication of Li-ion HSCs. -Figures 1a-1c show the fabrication process of the LNO electrode. First, the pastes were screen-printed on a  carbon cloth three times with a printed area of 1.5 cm × 2 cm. Second, the pastes on carbon cloth were dried at 100°C for 10 min in an oven. Finally, the APPJ was used to convert the pastes into LNOs on carbon cloth. The nitrogen flow rate and APPJ temperature were 46 slm and 620°C, respectively, with processing times of 0, 30, 60, 180, and 480 s.
Characterizations of LNO and Li-ion HSCs.-The water contact angle was measured using an optical goniometer (Model 100SB, Sindetake). The surface morphology of the LNO electrode was examined by scanning electron microscopy (SEM; JSM-7800 F Prime, JEOL) with energy-dispersive X-ray spectroscopy (EDS). The LNO crystallinity was inspected by X-ray diffraction (XRD; D2 Phaser, Bruker) with a Cu-Kα source. The surface chemical components were analyzed by X-ray photoelectron spectroscopy (XPS; Sigma Probe, Thermo VG Scientific) with an Al-Kα X-ray source. The LNO electrochemical performance was estimated by cyclic voltammetry (CV) and galvanostatic charging/discharging (GCD) under a two-electrode configuration. Both were measured using an electrochemical workstation (PGSTAT204, Metrohm Autolab). CV measurement were performed with a potential window of 0-0.8 V and potential scan rates of 200, 20, and 2 mVs −1 . CV measurement of the LNO Li-ion HSCs with potential scan rates of 200, 20, and 2 mV s −1 . To evaluate the areal capacity (Q c ) of the active material on the electrode. The areal capacity (Q c ) is calculated as 53 where ΔV is the potential window and C A , the areal capacitance that is calculated as 54 where A is the screen-printed area of the carbon cloth electrode; v, the potential scan rate; ΔV, the potential window; and I(V), the response current. GCD measurements were performed with a potential window of 0-0.8 V and charging/discharging currents of 4, 2, 1, 0.5, and 0.25 mA. To evaluate the areal capacity Q c (μA h/cm 2 ) of LNO Liion HSCs can be calculated as 55 where ΔE is the width of the potential window and C, the areal capacitance that is calculated as 56 where I is the charging/discharging current (A); A, the electrode area; and T, the discharging time.

Results and Discussion
Water contact angle.- Figure 2 shows the water contact angle of carbon cloth and LNO on carbon cloth. The pristine carbon cloth is hydrophobic with a water contact angle of 128.99°, as shown in Fig. 2a; this value agrees with a previous research result. 57 After printing LiCl-Ni(NO 3 ) 2 pastes, the sample becomes hydrophilic, and the water droplet penetrates the sample in 15 s. This is because the ethyl cellulose in the pastes acts as amphiphilic substance that modifies the carbon cloth surface, as shown in Fig. 2b. 58 After APPJ treatment on LiCl-Ni(NO 3 ) 2 , the surface becomes even more hydrophilic, and the water droplet penetrates the samples immediately, as shown in Figs. 2c-2f. The hydrophilicity can promote contact at the electrolyte/electrode interface, thereby enhancing the areal capacity of the Li-ion HSC. 59 SEM.- Figure 3 shows SEM images. Ethyl cellulose decomposes at 312°C and is burned out after APPJ treatment. We can find that with an increase in plasma treatment time, it can be observed that the crystallization changes from white dot-like particles with poor crystalline structure. As the plasma treatment time increases, there is a gradual reduction in the number of white particles, as shown in Figs. 3a-3f. Figures 3g-3l show SEM images with a higher magnification rate (10,000×). A longer APPJ treatment produces a relatively more uniform particle size distribution on the carbon cloth. 60 When the APPJ treatment time is lower than 180 s, white crystal with many unreacted compounds can be seen on the carbon cloth. These are probably lithium carbonate and nickel oxide, because a short APPJ treatment results in the incomplete combustion of lithium chloride and nickel nitrate. 61 At 180 s, a relatively complete monoclinic layered rock-salt structure is produced; nonetheless, but some white crystals remain. 62 Almost no white crystals are present, and the monoclinic layered rock-salt structure indicative of (1-x)LiNiO 2 -xLi 2 NiO 3 solid solution is more complete. 63 XRD.- Figure 4 shows the XRD patterns of LNO with and without APPJ treatment. The XRD diffraction peak of carbon cloth is found at 2θ = 24.5°. After APPJ treatment, the diffraction peak corresponding to (104) appears; it indicates the presence of Li 1−x Ni 1+x O crystalline phase. The synthesis of LiNiO 2 usually requires a long processing time and high-temperature calcination, which results in the synthesis of LiNiO 2 via a solid-state reaction between Li 2 CO 3 and NiO. However, non-stoichiometric LiNiO 2 is frequently synthesized in reaction between Li 2 O and NiO. [64][65][66] Most LiNiO 2 is compared with the intensity ratio of I(003)/I(104); 42,67-69 however, in our case, the (003) peak has very low intensity. The low intensity of (003) indicates the cubic crystal structure. 66 With APPJ treatment for 480 s, the full-width at half-maximum of the (104) diffraction peak decreases, and the material is closer to a rock-salt structure.
XPS.- Figure S1 shows the survey scan XPS spectra of LNO Liion HSCs on carbon cloth; Ni, O, C, N, Cl and Li can be identified clearly. It indicates that LNO is deposited on the carbon cloth. Figure S2 shows the C1s spectra; here, the C-C, C-O, C=O, and O-C=O binding energies are 284.5 eV, 286.2 eV, 288 eV, and 292.5 eV, respectively. 70,71 Ni-carbide is found at 282.5 eV. 71 Figure S3 shows the Li1s spectra; Ni3p (∼67 eV) is seen because NiO is present instead of layered LiNiO 2 . 60 Table I shows the corresponding bonding content in the O1s spectra in Figs. 5a-5e. As the APPJ treatment time increases, the oxygen lattice content increases. This suggests that the APPJ can enhance the metal oxygen as the APPJ treatment time increases. The presence of C-O and C=O indicates ethyl cellulose that interferes with the electrochemical reaction, and each bonding content decreases as the APPJ treatment increases to 480 s. In order to analyze the surface differences of the electrodes before and after electrochemical measurements, we compared the O1s spectra with 480-s APPJ treatment before and after measurements conducted at different potential windows (0.8 V and 2 V), as shown in Figs. 5f-5g. Table II shows the corresponding bonding contents in the O1s spectra before and after electrochemical measurements under different potential windows. We can observe that, after electrochemical measurements, there is a peak at 531 eV attributed to active oxygen, which is originated from residual lithium as surface impurities. This is due to the reduction of Ni 3+ to Ni 2+ in lithium nickel oxide, leading to the oxidation of lattice oxygen (O 2− ) into active oxygen (O − ). 75 In  previous studies, it has indicated that during charging-discharging process, lithium nickel oxides generates CO 2 , which then combines with active oxygen to form Li 2 CO 3 . 76 However, we did not observe the CO 3 2− peak in the electrochemical measurement under the potential window of 0.8 V, as shown in Fig. 5f. When the potential window is increased to 2 V, we can observe the peak of CO 3 2− and a decrease in the percentage of active oxygen, as shown in Fig. 5g and Table II. Furthermore, the reduction in the proportion of active oxygen is consistent with the increase in Li 2 CO 3 . From this, it can be inferred that the LNO electrode is unlikely to generate gas at the potential of 0.8 V, and it is required to increase potential to 2 V in order to induce gas generation. Figure 6 shows the Ni 2p 3/2 spectra; it consists of a main peak (∼855 eV) in which the intensity is contributed by both Ni 2+ and Ni 3+ and the intensity of the satellite (∼858-868 eV) is positively correlated with Ni 2+ . The Ni ion is in the 3+ state in LiNiO 2 ; however, Ni 3+ is unstable, and the valence 2+ and 3+ valence states mix. 77 To analyze the Ni 2+ /Ni 3+ ratio, we calculate the Ni 3+ where I(Ni2p 3/2 ) is the area of the Ni2p 3/2 main peak and I(satellite), the area of the satellite. The concentration comparison of Ni 3+ is shown in With APPJ processing for 480 s, the LNO Li-ion HSCs show the highest areal capacity of 21.076 mC cm −2 under a potential scan rate of 2 mV s −1 . Compared with the areal capacity of untreated LNO Liion HSC, that of LNO Li-ion HSCs treated by APPJ for 480 s increases by nearly 30 times. A shorter APPJ processing time results in smaller areal capacity. We limit our APPJ processing time to 480 s because the advantage of APPJ its rapid processing capability.
To clarify the charging behavior, quantitative kinetics analysis is used based on CV curves at various sweep rates 79 where i is the current according to the power law relationship with the sweep rate v, and both a and b are variable parameters where the b-values are determined from the slope of the plot of log v-log i. A b-value of ≈1 indicates a capacitive process (either EDLC or pseudocapacitive), and one between 0.5 and 1 indicates battery-type behavior. 80 Quantitative kinetics analysis is used based on CV curves at various sweep rates, as shown in Figs. 7g-7k. As the APPJ treatment increased, the behavior became closer to that of a capacitor, as shown in Fig. 7l. Furthermore, battery-type behavior is clearly seen at low potential range.
In the GCD results shown in Fig. 8, a GCD plateau is seen with an APPJ treatment time less than 180 s, indicating more battery-like behavior.    Table VII. From Figs. 8a-8e, it can be observed that there is a significant areal capacity loss during the electrochemical measurement with three electrode configuration. Table II shows an increase in the proportion of active oxygen after electrochemical measurements, indicating the reduction of Ni 3+ to Ni 2+ . Therefore, the areal capacity loss can be attributed to the formation of Ni 2+ byproducts. 75 According to the b-values, the GCD curves show a plateau; this is the battery-type material characteristic at low potential with APPJ  processing time shorter than 180 s. As the APPJ treatment time increases, the GCD curves become more symmetrical and triangular, suggesting that the performance of LNO Li-ion HSCs is more capacitance-like. Figure 8k shows Ragone plots analyzed based on the GCD results. The energy density (E A ) and power density (P A ) are calculated as where C A is the areal capacitance calculated by GCD results; ΔV, the potential window; and T, the discharging time of the GCD curve. APPJ treatment for 480 s under a charging/discharging current of 0.25 mA results in the highest energy density of 1.86 μWh cm −2 .

EIS of Li-ion
HSCs.- Figure 9 shows the EIS results and equivalent circuit model. [81][82][83] The equivalent circuit consists of the solution resistance R s ; charge transfer resistance R ct , and Warburg resistance W o used for fitting the curve in the low-frequency part and the double-layer capacitance CPE1. The Nyquist diagram is seen to consist of two parts. First, Li-ion HSCs are controlled by the electrochemical reaction; therefore, the Nyquist diagram appears semicircular at high frequency. Then, ion diffusion takes over, and the Nyquist diagram shows a straight line having a slope related to the ion diffusion rate at low frequency. Table VIII lists the fitting parameters for the EIS results. Because ethyl cellulose cannot be well removed when the APPJ treatment time is shorter than 60 s, the Nyquist diagrams show a straight line at both high and low frequencies. As the APPJ treatment time was increased to 480 s, an obvious semicircle appeared at high frequency and R ct (1.282 Ω) reached its minimum. R ct decreases as the APPJ treatment time increases.
Stability and bending test. -Figures 10a and 10b show the results of a 1000-cycle CV stability test and capacity retention under a potential scan rate of 20 mV s −1 for HSCs with 480-s APPJ treatment. The capacity retention was 42.8% after 1000 cycles and >50% after 700 cycles with APPJ treatment time of 480 s, as shown in Table IX. Figure 10c shows the mechanical bending stability test for HSCs with 180-s and 480-s APPJ treatments. In this experiment, we employed different curvatures as the bending test instead of different bending angles. [84][85][86] For 480-s APPJ treatment, the capacity retention rate was >99.99% under the maximum test curvature (2.0 cm −1 ), and the areal capacity did not vary by more than 5% for all bending curvatures, as shown in Table X. For 180-s APPJ treatment, the capacity retention rate was >93% under the maximum test curvature (2.0 cm −1 ). The APPJ-processed LNO Liion HSCs functioned well even under bending. The comparison of the performance of flexible SCs made on carbon cloth are listed in Table XI.

Conclusions
We investigate nitrogen-APPJ-processed flexible gel-electrolyte LNO Li-ion HSCs. The capacity of 480-s APPJ-treated flexible gelelectrolyte LNO Li-ion HSCs under a potential scan rate of 2 mV s −1 increases to 21.076 mC cm −2 . This is 30 times higher than that in the unprocessed case. The capacity retention rate is >50% after a 700-cycle CV test under a potential scanning rate of 20 mV s −1 . The capacity remains >93% under bending with a curvature of up to 2 cm −1 . As the APPJ treatment time increases, the GCD curves becomes more symmetrical and triangular, suggesting that the LNO Li-ion HSCs are more capacitor-like. With 480-s APPJ treatment, the highest energy density of 1.86 μWh cm −2 is achieved under a charging/discharging current of 0.25 mA.

Acknowledgments
This work is financially supported by the Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan (112L9006). JZC gratefully acknowledges funding support from the National Science and Technology Council in Taiwan (NSTC 111-2221-E-002-088-MY3). This work is partly supported by the National Science and Technology Council in Taiwan under grant no. NSTC 111-3116-F-002-005. The SEM experiments were conducted by Yuan-Tzu Lee at the Instrument Center National Taiwan University. The XRD experiments were conducted by You-Zeng Lin at the Precious Instrumentation Center, National Taiwan University of Science and Technology.