NiFeRu Layered Double Hydroxide and Its Derivatives Supported on Graphite Foam as Binder-Free Cathode for Nonaqueous Li-O2 Batteries

Li-O2 batteries (LOBs) generally referred as Li-air batteries (LABs) potentially have a high specific capacity among all kinds of metal-ion batteries. However, it suffers severely from sluggish discharge/charge kinetics. A binder-free integrated cathode is fabricated which contains NiFeRu layered double hydroxide (NiFeRu LDH) nanosheets grown on a 3D conductive graphite foam (NiFeRu LDH@GF) in this work. This integrated 3D cathode exhibits a specific capacity of 3,085 mAh g−1 and a cycle life of 46. The electrochemical performances of NiFeRu LDH derived metal oxides (DMO@GF) by calcining NiFeRu LDH@GF at 350°C, 450°C, and 550°C are also been investigated. The 450°C obtained DMO@GF (450-DMO@GF) exhibits the highest catalytic activity for oxygen reduction/evolution reaction. The unique floccule-like and cross-connected structure of the 450-DMO@GF and its low charge transfer resistance are responsible for the outstanding electrochemical performance.


INTRODUCTION
The modern e-society has put increasing demand for new energy sources that can enable long-lasting power and fast charging and/or release of the energy. Among various available batteries, Li-O 2 batteries (LOBs), which are also called Li-air batteries (LABs), have the highest theoretical specific energy of 3,505 Wh kg −1 . This specific energy is ∼10 times that of Li-ion batteries (Bruce et al., 2012) and higher than other Li metal batteries (Xue et al., 2019;Sun et al., 2020). Therefore, LOBs have attracted immense interest since 1996 as the "holy grail" of batteries (Abraham and Jiang, 1996). However, the development of LOBs are slow because of their low energy efficiency caused by decomposition of nonaqueous electrolyte and carbon-based cathode (Freunberger et al., 2011;Ottakam Thotiyl et al., 2013), short cycle life caused by nonrecovery of the reaction surface/interface (Lu et al., 2016), and poor power performance led by low kinetics of electron/Li + /O 2 transport during the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR) (Liu et al., 2015). These issues are all related to the properties of cathode materials, and achieving cathode with high electrochemical performances is still a big challenge. During the discharging process, O 2 should be first reduced at the gas/liquid/solid (O 2 /Li + /e − ) three-phase reaction interface of cathode. Then, the reduced O 2 − will combine with Li + to form Li 2 O 2 . In the charging process, Li 2 O 2 should be decomposed to O 2 and Li + , which guarantees reuse of the reaction interface of cathode during the following cycles. In addition to the Li anode and electrolyte, these two discharging/charging processes are dependent on the catalytic activity of cathode for ORR and OER. The slow kinetics of ORR and OER taking place in the cathode will undoubtedly give rise to poor battery performances. Thus, there is consensus that a stable cathode material with high ORR and OER activity is the key player of LOBs.
To realize a stable cathode and a superior Li-O 2 battery, numerous research efforts concerning catalyst design and catalytic mechanism have been devoted to the electro-catalytic materials field (Chang et al., 2017). Various cathode materials based on carbon materials (Qin et al., 2019), heteroatom doping carbon material (Jiang et al., 2019), noble metal , and metal oxides (Yuan et al., 2019) have been used to the LOBs. The aim is to enhance the performances of LOBs by promoting the reaction kinetics associated with Li 2 O 2 generation/decomposition.
In particular, the larger overpotential during the charging process than discharging process reveals more sluggish OER kinetics . Layered double hydroxide (LDH) has become the protagonist in different areas including photochemistry , adsorption (Gu et al., 2015), drug delivery (Ribeiro et al., 2014), and water electrolysis due to its advantages such as more exposed catalytic active sites than 0D or 1D materials, large surface-to-bulk ratios, and easily controllable layered structure adjustment (Cai et al., 2019). Especially, LDHs have attracted increased attention because of their OER catalytic activity since 2013 (Gong et al., 2013). This implies that LDHs are promising in accelerating the kinetics during the charging process for LOBs. However, there are only limited research so far on LDH cathode of LOBs. Recently, cathodes of nonaqueous LOBs based on CoNiFe LDH and conductive RuO 2.1 super-lattice composites have been fabricated by casting these composites with polyvinylidene fluoride (PVDF) binder on stainless steel mesh (Lu et al., 2020). It is worth mentioning that LDHs usually have poor conductivity and insufficient active edge sites. Therefore, binder-free cathodes with LDHs supported on conductive substrate will be more superior.
Herein, in this work, we design and realize an integrated binder-free cathode for LOBs by directly growing LDH on a graphite foam (LDH@GF). NiFe LDH doped by Ru (NiFeRu LDH) is selected since catalysts based on trimetal usually offer higher catalytic activity for OER in comparison with bimetalcentered LDH, and Ru has high adsorption capabilities toward oxygen-containing intermediates (Seh et al., 2017). In addition, their derived metal oxides (DMOs) supported on the GF (DMO@ GF), by calcination under different temperatures, are also investigated as cathodes of LOBs since transitional metal oxides (e.g., NiO (Tong et al., 2015)) and spinel oxides (e.g., NiCo 2 O 4 (Sun et al., 2014)) have exhibited catalytic activity in LOBs.

Preparation of Cathode Materials
The 3D graphite foam (GF) substrate was prepared by using the chemical vapor deposition (CVD) method using the nickel foam as a template according to our previous work (Chao et al., 2014;Song et al., 2020). After etching the nickel template using a solution of 1 M FeCl 3 and 0.5 M HCl, a free-standing GF (∼0.5 mg cm −2 ) was obtained. Then, a piece of GF was immersed into a mixed solution of 98% HNO 3 and 70% H 2 SO 4 with a mole ratio of 1:3. After washing with deionized water, this hydrophilic-treated GF was immersed into ethylene glycol for 1 h and then transferred into a teflon autoclave containing a mixed solution of Ni(NO3)2·6H2O/Fe(NO3) 2·9H2O/RuCl3/NH 4 F/urea (mole ratio is 5:4:1:35:35). The hydrothermal reaction was carried out at 120°C for 12 h. Then, the NiFeRu LDH supported on GF (NiFeRu LDH@GF, ∼1.3 mg cm −2 ) cathode was obtained after cooling naturally and washing with deionized water. The fabrication process for NiFeRu LDH@GF cathode is shown in Figure 1.
The DMOs supported on the GF (DMO@GF, ∼1.0 mg cm −2 ) cathode was achieved through heating NiFeRu LDH@GF in a quartz tube furnace under Ar gas at 350°C, 450°C, and 550°C for 4 h.

Characterization
X-ray diffraction (XRD) was performed on an X-ray diffractometer (Bruker-AXS D8 Advance, CuKα, λ 1.54186). The morphology and microstructure of the cathode were observed through an transmission electron microscope (JEOL 2100 F HRTEM) and scanning electron microscope (JEOL 7600F FESEM). Raman spectra were measured using a Raman system with a laser wavelength of 532 nm (WITec-CRM200).

LOBs Assembly and Electrochemical Tests
The homemade Swagelok-type cells which were assembled under pure argon gas (<0.5 ppm H 2 O and O 2 ) in a glovebox were used for the electrochemical measurements. A free-standing NiFeRu LDH@GF cathode or a DMO@GF cathode and a Li anode were separated by a Whatman glass fiber separator. The glass fiber separator was first impregnated with 1 M bis(trifluoromethane) sulfonamide (LiTFSI). Tetraethylene glycol dimethyl ether (TEGDME) was used as the solution. The galvanostatic discharge/charge tests were conducted using a Neware battery tester. Electrochemical impedance spectroscopy (EIS) was carried out using CHI660E. The frequency range is from 1 kHz to 1 Hz, and an AC amplitude of 5 mv was applied.
The Raman technique was carried out to further identify composition of the as-prepared cathode. The G band (1,575 cm −1 ) in the Raman spectrum ( Figure 2B) is identified for the graphite (Tuinstra and Koenig, 1970). D band (1,348 cm −1 ) of graphite is not observed, which indicates that Frontiers in Energy Research | www.frontiersin.org November 2020 | Volume 8 | Article 612093 the GF is highly ordered (Ferrari et al., 2004). The broad band located at around 547 cm −1 and the faint broad feature at around 458 cm −1 are in accordance with the observation of NiFe-LDH (Lu et al., 2014). The doping of Ru is responsible for the small shift of Raman peaks for NiFeRu-LDH in comparison with those of NiFe-LDH. Field emission scanning electron microscopy (FESEM) was utilized to reveal the morphology of NiFeRu-LDH. NiFeRu-LDH grows uniformly on the surface of GF ( Figure 2C). Numerous vertically aligned nanosheets with size of 300-600 nm and thickness of 20-40 nm are observed ( Figure 2D). This interconnected structure of nanosheet may be helpful in improving the conductivity and providing more active edge sites of NiFeRu-LDH. The transmission electron microscopy (TEM) photograph further uncovers the nanosheet morphology of NiFeRu-LDH ( Figure 2E), and the corresponding energy-dispersive X-ray (EDX) elemental mapping discloses the existence of Ru in the nanosheets ( Figure 2F-H). This 3D-structured porous binder-free cathode with a large amount of nanosheets can provide a large reaction interface and abundant highly efficient catalytic active sites for discharging products accommodation.
The electrochemical activity of NiFeRu-LDH@GF is revealed by using the galvanostatic discharge/charge method. Home-made Swagelok-type cells are used to characterize the electrochemical performance. The Li metal is used as anode, and the free-standing NiFeRu-LDH@GF without additional current collectors is the cathode. The cathode and anode are separated by a glass fiber containing electrolyte (details are in the Experimental Section). The first galvanostatic discharging and charging curves of LOBs at various current densities are shown in Figure 3A. The specific discharge capacities of LOBs at 200 mA g −1 , 300 mA g −1 , and 400 mA g −1 are 3,085 mAh g −1 , 2,155 mAh g −1 , and 1,085 mAh g −1 , respectively. This result indicates that the specific capacity is sensitive to the current density, which is because of the limited conductivity of discharging products and O 2 transport limitation (Viswanathan et al., 2011;Zhu et al., 2013). The mid-capacity potential (defined as the potential at half of the specific capacity) for the discharging process at 200 mA g −1 is 2.49 V, which is close to 2.46 V at 300 mA g −1 and 2.43 V at 400 mA g −1 . This means the catalytic activity of NiFeRu-LDH@GF for ORR is not the controlling step for the discharging process. However, the mid-capacity potential for the charging process increases from 4.20 V at 200 mA g −1 to 4.32 V at 400 mA g −1 . The obviously increased overpotential implies that the catalytic activity of NiFeRu-LDH@GF is not sufficient to OER.
To reveal the cycling performance, the NiFeRu-LDH@GF cathode is tested at a current density of 200 mA g −1 , and a fixed specific capacity of 500 mAh g −1 is a commonly used parameter in LOBs Zhang et al., 2018). The overpotential gradually increases for the charging process, and the terminal voltage is higher than 4.4 V after 46 cycles ( Figure  3B). Meanwhile, the terminal voltage for the discharging process is about 2.15 V. The performance up to 46 cycles is similar to that of CoNiFe-LDH cathode reported recently (Lu et al., 2020). These   (311) and (400) facets of the NiFe 2 O 4 , respectively, and the (111) and (200) facets of NiO at the peak position of 37.2°and 43.3°are also observed in Figure 4A ( Xue et al., 2020). These results prove that NiFeRu-LDH has changed to metal oxides including NiFe 2 O 4 and NiO. In addition, the diffraction peak intensity of DMOs increases. In addition, the full width at half maximum (FWHM) decreases when the calcination temperature increases. This means more metal oxides are formed, and their crystal size is larger at higher calcination temperature. The diffraction peaks of Ru oxides are not detected because of the low Ru atom ratio in NiFeRu-LDH.
Raman spectra of DMO@GF further identify the DMOs (Figure 4B). The G band (1,575 cm −1 ) of graphite in Raman spectrum still exists, which is consistent with the result of XRD ( Figure 4A). The Raman peaks at 188 cm −1 , 305 cm −1 , 468 cm −1 , and 549 cm −1 are related to the symmetric and antisymmetric bending of oxygen atom in iron-oxygen bond at FeO 6 octahedral voids (Kumar et al., 2020). The band at 673 cm −1 is responsible for the symmetric stretching vibrations of oxygen atoms against Ni ion in NiO 4 tetrahedra. The NiO-related Raman bands at 508 cm −1 and 1,014 cm −1 are also observed in Figure 4B (Qiu et al., 2019). Furthermore, the characteristic peaks of RuO 2 at 528 cm −1 , 646 cm −1 , and 716 cm −1 are included in the broad bands between 450 cm −1 and 750 cm −1 (Huang and Pollak, 1982). These results predict a conclusion that NiFeRu-LDH has transformed to metal oxides on GF.
Galvanostatic discharge/charge method was further utilized to investigate the electrochemical performances of DMO@GF cathode by calcining NiFeRu-LDH@GF at 350°C, 450°C and 550°C (denoted as 350-DMO@GF, 450-DMO@GF and 550-DMO@GF). The specific discharge capacities of 350-DMO@ GF, 450-DMO@GF and 550-DMO@GF electrodes are 2,718 mAh g −1 , 3,700 mAh g −1 , and 2,555 mAh g −1 , respectively ( Figure 5A). Obviously, only the 450-DMO@GF exhibits higher specific discharge capacity than that of NiFeRu-LDH@GF ( Figure 3A). Meanwhile, the mid-capacity potential for the discharging process of 450-DMO@GF is 2.55 V which is close to 2.56 V of 350-DMO@GF and 2.50 V of 550-DMO@GF. This means the discharging over-potential of DMO@GF, especially 450-DMO@GF, is smaller than that of NiFeRu-LDH@GF (2.49 V). More importantly, the mid-capacity potential of  Frontiers in Energy Research | www.frontiersin.org November 2020 | Volume 8 | Article 612093 5 DMO@GF for the charging process (4.02 V for 350-DMO@GF, 3.97 V for 450-DMO@GF and 4.01 V for 550-DMO@GF) is much lower than that of NiFeRu-LDH@GF (4.20 V). The cycling properties of 450-DMO@GF cathode are further researched since it has highest specific discharging capacity and smallest charging over-potential ( Figure 5B). The cycle life of 58 is realized for 450-DMO@GF cathode at a current density of 200 mA g −1 and a fixed specific capacity of 500 mAh g −1 . The terminal voltage for charging and discharging process is 4.08 V and 2.25 V. These cycling performances of 450-DMO@GF are better than those of NiFeRu-LDH@GF ( Figure 3B). These results imply that the NiFeRu-LDH DMOs have higher catalytic activity for ORR and OER process in comparison with NiFeRu-LDH. It has been reported that metal oxides such as NiCo 2 O 4 , NiO and RuO 2 have better electrochemical performances than those of layered-double-hydroxides (Chang et al., 2017), which is consistent with our results.
To reveal the differences of electrochemical performance for NiFeRu-LDH DMOs, the morphology and structure of DMO@ GF are analyzed by SEM. At a low temperature of 350°C, the vertically aligned nanosheets of NiFeRu-LDH partially changed to floccule-like products because of the phase transformation from NiFeRu-LDH to DMOs ( Figure 6A). Then, all the nanosheets of NiFeRu-LDH change to floccule-like DMOs that are cross-connected after calcining 450°C ( Figure 6B). The floccule-like DMOs further grow to thick nanosheets since the crystal grains grow up accompanying the temperature rising, and no floccule-like products are observed at 550°C ( Figure 6C). Obviously, the high specific surface combined with the crossconnected structure of the floccule-like DMOs is beneficial to improve the electrocatalytic activity and meanwhile to lower the overpotential. Therefore, the 450-DMO@GF electrode has the best charge-discharge performance among three DMO@GF electrodes.
EIS is further introduced to research the mechanism of differences in kinetic characteristic among DMO@GF and LDH@GF (Figure 7). The impedance spectra of LOBs are interpreted and proposed by Laoire. The equivalent circuit is Rs (C (R p W), where Rs is the electronic resistance from the cathode, the current collector, and electrolyte resistance, C is the capacitive contribution of the cathode and anode, R p is the charge transfer resistance of the two electrodes, and W is the linear Warburg element that is attributed to the diffusion of the O 2 and Li + species to the electrode (Laoire et al., 2011). Obviously, the charge transfer resistance R p of three DMO@GF electrodes is smaller than that of LDH@GF, which implies that DMO@GF electrodes have higher catalytic activity in LOBs. In addition, the charge transfer resistance of the 450-DMO@GF electrode is smallest among the three DMO@GF electrodes, which is responsible for its outstanding electrochemical performance. These results are consistent with the discharge/charge characteristic of DMO@GF and LDH@GF (Figures 3, 5).

CONCLUSION
In summary, we present an integrated binder-free cathode for LOBs based on Ru-doped NiFe layered double hydroxide grown on a GF (NiFeRu LDH@GF) through hydrothermal reaction. The vertically aligned NiFeRu LDH nanosheets combined with the ultralight and conductive GF substrate provide a specific capacity of 3,085 mAh g −1 at 200 mA g −1 and a cycle life of 46. Furthermore, we also investigate the NiFeRu LDH DMOs supported on a GF (DMO@GF) by calcining NiFeRu LDH@ GF at 350°C, 450°C, and 550°C. The 450°C-obtained DMO@GF (450-DMO@GF) shows the best electrochemical performances among the three DMO@GFs. The specific capacity of 450-DMO@GF is 3,700 mAh g −1 , and its cycle life is 58. SEM reveals that the floccule-like 450-DMO@GF has a high specific surface combined with a cross-connected structure. EIS further reveals that the charge transfer resistance of the 450-DMO@GF electrode is smallest. Our results provide a new kind of integrated binder-free cathode to achieving LOBs with high performance.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary materials, further inquiries can be directed to the corresponding author/s.