Ion Selective and Water Resistant Cellulose Nanofiber/MXene Membrane Enabled Cycling Zn Anode at High Currents

Aqueous rechargeable zinc ion batteries (ZIBs) are regarded as a promising candidates for next‐generation energy storage devices but strongly hindered by the limited utilization of the zinc metal anode (below 5%) due to the active water/anion corrosion. Herein, an ion selective and water‐resistant cellulose nanofiber (CNF)/MXene composite membrane has been developed through molecular sieving to restrict active water and anions from the electrode/electrolyte interface through dehydration of zinc ions, avoiding the water/anion‐induced corrosion/decomposition. In this way, the CNF/MXene@Zn anode exhibits significantly enhanced coulombic efficiency (99.5 % at 10 mA cm‐2) and low voltage hysteresis. Moreover, coated with CNF/MXene composite membrane, zinc symmetric batteries can be operated at the extremely high current of 100 mA cm‐2 and ultra‐high Zn utilization of 88.2% to achieve record‐high cumulative plating capacity of 12 Ah cm‐2. Furthermore, the full vanadium dioxide (VO2) |CNF/MXene@Zn batteries exhibit a high capacity of 357 mAh g‐1 at 2 A g‐1 and retain 93.3% of the capacity after 500 cycles. Moreover, at negative/ positive capacity (N/P) ratio of 2.8, the CNF/MXene membrane coated zinc is able to stably cycle for 100 cycles, demonstrating the potential for high energy zinc battery. This designed CNF/MXene membrane enables ZIBs as viable energy storage devices for practical applications.

direction, leading to a loosen and porous surface structure. [14] The continuous growth of zinc dentrites may eventually pierce through separators, resulting an internal shorting failure. During the discharging process, the Zn dendrites are prone to break from the substrate and become "dead Zn." [15] This could result in severe capacity fading and poor coulombic efficiency (CE). Currently, numerous strategies have been developed to suppress the corrosion reactions and the growth of zinc dentrites, such as hybrid electrolyte preparation, [16] nanostructured Zn anode construction, [17] and stable protective layer introduction. [18] Among these strategies, the use of an interface layer is a promising method. The interface layer can be fabricated on the surface of zinc metal to guide the stripping and plating processes of zinc. For example, Archer et al. reported that graphene nanosheets can drive the deposition of Zn with a locked in-plane growth of Zn (0002). [19] Moreover, Chen et al. synthesized the nitrogen-doped graphene oxide as interface modification layer on zinc metal. [15] The parallel graphene layer directs the deposition of Zn crystals in (0002) planes. In addition, in situ Raman spectroscopy demonstrated that the directional plating of Zn metal can effectively hinder the side reactions as well as hydrogen evolution. With a similar layered structure, MXene nanosheets were also studied as protective layers for zinc anodes. Niu et al. coated the zinc foil with a uniform MXene layer on the surface via a reducing/ assembling strategy. [18] The MXene layer effectively reduces the Zn nucleation energy barrier through the favorable charge redistribution effect, leading to a uniform electric field. Therefore, MXene-integrated Zn shows higher capacity retention and lower polarization potential in ZIBs. These studies provide novel insights to develop advanced anodes for ZIBs by designing thin protective interface layers. However, water/ anion-induced corrosion is an another typical issue to affect the long-cycling stability of ZIBs, and the problem remains largely unexplored. The repeated Zn plating/stripping upon cycling is accompanied with the decomposition of water since the standard potential of Zn/Zn 2+ is greatly lower than that of hydrogen evolution. The decompositon of active water occupies the active sites for zinc plating/stripping, leading to even higher local current density and further deteriorating the dendrite formation. Moreover, the decomposition of anions results in the decrease in the concentration of the zinc ions in the electrolyte, which is hazardous to the ionic conductivity in the electrolyte, as well as the increase in the local pH value in the electrode/electrolyte interface, hindering the zinc ion transport at high depth-of-discharge (high utilization of zinc anode). [20] Therefore, it is highly desirable to develop a water restricted and ion selective ultrathin protective membrane for aqueous zinc metal batteries. [18,21] Herein, we developed an ion selective and water resistant cellulose nanofiber (CNF)/MXene composite membrane to modulate the stripping and plating processes of zinc ions. Benefiting from the rich functional group on the CNF surface, the CNF/MXene coating layer enables the molecular sieving with cation/anion selectivity and restricting the water transport, avoiding corrosion upon zinc stripping/plating, and facilitates the zinc ions to pass through the film, leading to corrosionfree zinc plating/stripping. In this way, the CNF/MXene@Zn anode exhibits attractive coulombic efficiency (99.7%), low voltage hysteresis, and long cycle life (≈3000 h) in Zn/Zn symmetric batteries. Noteworthy, at the extremely challenging current density of 100 mA cm −2 , the symmetric batteries are able to operate at a record-high capacity of 100 mAh cm −2 , which corresponds to a ultra-high Zn utilization rate of 88.2%. When being coupled with a VO 2 cathode, the full battery exhibits a high capacity (403 mAh g −1 at 0.5 A g −1 ), outstanding rate performance, and long cycling stability (80 mAh g −1 after 2000 cycles at 30.0 A g −1 ). Moreover, at the severe condition of N/P ratio of 2.8, the CNF/MXene membrane coated zinc is able to stably cycle 100 times, demonstrating the potential for high energy zinc battery. The CNF/MXene membrane with ion sieving performance to screen out anions and restrict active water in the electrode/electrolyte interface demonstrates a significant advance in modulating the stripping and plating processes, having large potential to be applied in other metal batteries.

Preparation of Exfoliated Ti 3 C 2 T x MXene Nanosheets
Exfoliated Ti 3 C 2 T x MXene nanosheets were synthesized based on the previous published literature. [22] First, 2 g lithium fluoride (LiF) were slowly added to 20 mL of 12 m HCl solution under continuous stirring. After the LiF powder was dissolved completely, 1 g Ti 3 AlC 2 powder was slowly added into the solution. The mixture was then continuously stirred at 45 °C for 24 h. Afterward, the solid residue was collected by centrifuging and washed with 1 m HCl solution for several times to remove the residual HF. Next, the solid residue was washed with deionized water for several times until the pH value became to 6.0. After the chemical etching in LiF/HCl hybrid solution, MXene formed an accordion-like morphology. Subsequently, the collected solid residue was dispersed into 200 mL deionized water, and ultrasonicated for 2 h under argon atmosphere in an ice bath. The ultrasonic exfoliation procedure was carried out to further peel off the MXene into monolayers. Finally, the exfoliated MXene nanosheet suspension was collected by centrifuging at 3500 r min −1 for 1 h. The concentration of MXene was about 2 mg mL −1 in the suspension.

Fabrication of Cellulose Nanofiber/MXene Membrane Coated Zinc Foil
CNFs (0.2 g, 15 wt% water suspension, University of Maine, Orono, ME, USA) were first added to 10 mL deionized water. After being stirred for 2 h and sonication for half an hour, a uniform CNF suspension was obtained. Then, 5 mL of prepared MXene suspension were added to the prepared CNF suspension. After being stirred for 0.5 h, a uniform CNF/MXene mixed suspension was obtained. CNF/MXene suspension was casted onto the surface of a bare zinc foil (diameter = 10 cm). The thickness of the coating layer was controlled by varying the concentration of the CNF/MXene suspension. www.advenergymat.de www.advancedsciencenews.com

Preparation of VO 2 as a Cathode Material for Zinc Ion Batteries
VO 2 nanofibers were synthesized based on an established method. [23] First, oxalic acid (5 g) was dissolved in 60 mL deionized water. After being stirred for 10 min, 2.4 g divanadium pentoxide were added to the solution. A transparent blue solution was obtained after the solution was stirred at 25 °C for 1 h. Then the solution was transferred to a Teflon-lined autoclave for heating at 180 °C for 12 h. After the autoclave was cooled, the precipitate was collected with centrifugation and washed with water and ethanol alcohol for three times, respectively. The final product was obtained after being dried at 70 °C for overnight.

Material Characterization
X-ray diffraction (XRD) spectra of all the samples were acquired from a Rigaku MiniFlex XRD instrument (RIGAKU, Austin, TX, USA) using Cu Kα radiation (λ = 1.5405 Å) at current of 40 mA and voltage of 40 KV with the 2θ range from 5° to 90° and a scan rate of 1° min −1 . To investigate the chemical state of MXene before and after exfoliation, spectra of X-ray photoelectron spectroscopy (XPS) were collected on an AXIS165 spectrometer (MRFN, Manchester, UK). The microstructure of CNF/MXene membrane was observed using an FEI Quanta 3D FEG field emission scanning electron microscope (SEM) (FEI, Boston, MA, USA) at an acceleration voltage of 20 kV. Energy dispersive spectroscopy (EDS) mapping was also done with the SEM instrument. The morphology and crystallographic structures of all the samples were further investigated with transmission electron microscopy (TEM), high resolution transmission electron microscopy (HRTEM) images. The corresponding selected area diffraction patterns (SAEDs) were collected using a JEM-1400 (JEOL USA Inc., Peabody, MA, USA). Thermogravimetric analysis was done using a Q50 analyzer (TA Instruments Inc., New Castle, DE, USA) with the heating rate of 1 °C min −1 from 30 to 800 °C in a nitrogen atmosphere. Atomic force microscopy (AFM) was conducted using a Bruker Dimension Icon XR (Bruker Nano Inc., Tucson, AZ, USA).

Electrochemical Measurements
The electrochemical data were collected using 2032-coin cells with the prepared cathode, zinc metal with 100 µm in thickness (otherwise specified) as the anode, 3 m Zn(CF 3 SO 3 ) 2 aqueous solution as the electrolyte, and glass fiber membrane as the separator. The cathode electrode is composed of VO 2 nanofibers, carbon black, and polytetrafluoroethylene binder at a weight ratio of 6:3:1. Galvanostatic charge/discharge electrochemical tests were performed using an eight-channel LAND battery analyzer (CT3001A, LAND Electronics Corporation, Wuhan, China) with the voltage range of 0.2-1.6 V. Cyclic voltammetry (CV) was scanned with an electrochemical workstation (CHI 760e) at 0.1 mV s −1 in the range of 0.2-1.7 V. Electrochemical impedance spectroscopy (EIS) data were collected by applying an AC potential of 5 mV amplitude with the frequency range of 0.01-100 kHz.

Theoretical Calculation
The spin-polarized calculations were performed using the Dmol3 package. [20,24] The exchange-correlation function was set to M06-L. [25] All electrons are included in all calculations for the treatments of core electrons in DMol. The Double numerical plus polarization (DNP-4.4) basis set was used for the description of the atomic orbital. [26] The convergence tolerance was set to 1) 1.0 × 10 −6 Ha for energy, 2) 4.0 × 10 −3 Ha Å −1 for the maximum force, and 3) 5.0 × 10 −3 Å for the maximum displacement. The thermal searing of 0.02 Ha was applied to the orbital occupation to speed up convergence.

Results and Discussion
As illustrated in Figure 1a,b, the CNF/MXene coating layer enables the molecular sieving to prevent/restrict the active water/ anion from the electrode/electrolyte interface, avoiding corrosion upon zinc stripping/plating, and facilitates the zinc ions to pass through the film, leading to corrosion-free zinc plating/ stripping. Initially, to measure ion selectivity, current-voltage (I-V) curves were obtained with a concentration gradients of Zn(CF 3 SO 3 ) 2 electrolyte across the membrane (Figure 1c). The stainless-steel plates were used as electrodes for both sides. By assuming that each ion species contributes a current given by the Nernst-Planck equation, which is parametrized by an effective diffusion constant D i (different for each ion species i), the GHK model presents a quantitative measure of selectivity that is useful for comparing selectivity among different membranes. [27] The reversal potential is related to the selectivity of the pore via the GHK voltage equation where S GHK is the selectivity ratio, c high and c low are the solution concentrations in the fluid reservoirs, e is the electron charge, k B is the Boltzmann constant, and T is the solution temperature. The mobility of ions through the membrane remains similar (2.63 vs 2.24 mS for CNF-MXene-based membranes and glass fiber membrane, respectively), while the Zn 2+ / CH 3 SO 3 − selectivity ratio S GHK was calculated to be 5 compared with that the glass fiber with selectivity of only 1, suggesting the excellent cation/anion selectivity and preventing the anions from direct contacting the electrode/electrolyte interface. It is believed that the cellulose with rich surface functional groups (COH and CCOOH) ( Figure S1, Supporting Information), which would be deprotonized in the electrolyte, and prefer to adsorb the cations and conduct cations rather than anions. In such a way, the anion decomposition/corrosion reaction can be prevented. Additionally, the desolvation energy barriers of Zn 2+ through the membrane are studied by density functional theory calculations (Figure 1d,e). In Zn(CF 3 SO 3 ) 2 electrolyte, each Zn 2+ is coordinated with 6 H 2 O molecules, which delivers a ultrahigh desolvation energy of Zn 2+ . In the CNF-MXene-based membranes, Zn 2+ first can be coordinated with two COH terminals, which are linked on the backbone structure of CNFs, and then can be coordinated with H 2 O molecules, which leads to desolvation process. The lower desolvation energies suggests that the membrane would fix the active water and prevent the active water from the electrolyte/electrode interface and facilitate to improve the kinetics of Zn plating/stripping. [21a,28] Thus, it can be anticipated that the membrane with merits of desolvation and ion selective would be beneficial to the performance of zinc metal battery.
MXene monolayers were synthesized by a chemical etching method followed with a sonication process. The exfoliated MXene displays a smooth and thin nanosheet morphology with the thickness of only several nanometers and a lateral size of hundreds of micrometers. (Figure S2, Supporting Information) Pure CNFs present a weak peak at 16.2°, and a dominant peak at 23.0°, corresponding to (110) and (200) planes, respectively, as shown in Figure 2. In the XRD pattern of CNF/MXene composite membranes, the characteristic peaks from CNFs and the dominant diffraction peak (6.46°) from exfoliated MXene are both present. XPS results are illustrated in Figure 2b and Figure S2, Supporting Information. As shown in Figure S3, Supporting Information, the Ti 2p components were fitted as Ti(I), Ti(II), Ti(III) that could be attributed to Ti bound to carbon at binding energy of 454.6 (460.7) eV, Ti atoms bonded with O at 455.2 (461.6) eV and OH at 456.3 (462.5) eV, respectively. Ti(I), Ti(II), Ti(III) species constitute the majority fraction (≈89.7 mol%) of the XPS spectrum related to Ti 2p. The Ti 2p 3/2 component centered at 459.1 eV attributes to Ti atoms with a formal valence of IV (TiO 2 ). This came from surface oxidation and comprised just about 10.3 mol%. The peaks of Ti 2p and C 1s spectra confirm the existence of TiC and TiO bonds, suggesting the formation of Ti 3 C 2 (OH) 2 after exfoliation. The C1s XPS spectra can be divided into four peaks: 281.5, 284.6, 286.6, and 288.5 eV, corresponding to CTi, CC, CO, and OC = O bonding, respectively. The strong peak at 284.6 eV is attributed to the dominant carbon content from CNFs. The existence of 286.6 and 288.5 eV are due to numerous functional groups (COH) from MXene nanosheets and CNFs. The abundance of functional groups enables firm bonding connection between MXene nanosheets and CNFs, as well as, intimate contact between CNF/MXene membrane and the zinc foil. Figure 2c illustrates the SEM image and EDS mapping of the prepared CNF/MXene membrane. As shown in Figure 2a, the entangled CNFs are firmly combined with MXene nanosheets, forming a dense film. The EDS mappings demonstrate that the elements C, Ti, O, and F are homogeneously distributed in the entire area in Figure S4, Supporting Information, suggesting a uniform composition of the prepared composite film. Contact angles (10 µL of 3 m Zn(CF 3 SO 3 ) 2 electrolyte solution) were obtained on the bare zinc foil and the CNF/MXene@Zn foil. As displayed in Figure 2d, compared with the contact angle on bare Zn foil (92.13°), the CNF/MXene membrane coating led to a much lower contact angle (31.00°), reflecting the improved wettability and electrolyte accessibility. The spacing between layers is only 1.3 nm, which is consistent with the reported literatures. [29] These results intuitively demonstrate the successful exfoliation of MXene nanosheets as prepared. The SAED pattern in Figure 2f reveals the combination of characteristic halo ring pattern from amorphous CNFs and the spot pattern from crystal exfoliated MXene nanosheets. The spot pattern demonstrates that the prepared MXene has a hexagonal symmetry. The combination of amorphous CNFs and crystalline MXene can be further distinctly observed from the inset FFT images in the HRTEM image in Figure 2g. The haloring pattern in FFT image can be ascribed to the amorphous CNFs. The bright spot array can be well indexed as the [0001] zone axis of the hexagonal MXene crystals. The close interface layer further confirms the tight connection between CNFs and MXene. Furthermore, zeta potential was tested to investigate the surface charge of the prepared CNF/MXene suspension. The average zeta potential value is −15.9 mV, exhibiting an obvious negative surface charge.
To study the effect of the CNF/MXene membrane on the Zn stripping and plating behavior, the morphology of Zn foils with/without CNF/MXene membrane were observed after Zn stripping and deposition. After being stripped for 20 min, bare zinc anode presents serious corrosion/dendrite formation with numerous pores on the surface (Figure 3a,b). After plating for 20 min, numerous irregular dendrites appeared on the bare Zn anode (Figure 3c,d). This is attributed to strong local gradients in the electric field on the surface of the bare zinc metal anode during the Zn stripping and plating processes. Zn stripping or nucleation depends on the electric field distributions at the interface between anode and electrolyte. During the discharging process, zinc ions are preferentially stripped away at areas with higher electric fields, which gradually result in serious corrosion on the surface. During the charging process, zinc ions preferentially nucleate at locations with enhanced electric fields. [30] The enhanced local electric field forms a higher charge region, promoting more deposition of zinc ions, gradually evolving plot (Figure 3c,d).
In comparison, with CNF/MXene protective layer, the Zn anode shows a flat surface with some laminate-like scratches on the surface after stripping for 20 min, effectively avoiding serious corrosion. (Figure 3e,f). After plating for 20 min, the Zn anode exhibited a plate-like plating morphology without dendrite, (Figure 3g,h), and the CNF/MXene protective layer These results reveal that the CNF/MXene coating layer can regulate the Zn stripping and plating behavior. After coating with CNF/MXene protective layer, the homogeneous stripping and plating of zinc ions are achieved, leading to a flat and smooth surface. AFM was applied to observe the altitude intercept between bare Zn anode (about 2 µm) and CNF/MXene protected zinc anode (about 400 nm) after stripping/plating for 20 min. (Figure 3i-l) The deposited zinc is further detected by the focused ion beam/SEM to reveal the place of zinc deposition in Figure S6, Supporting Information. The CNF/MXene layer and zinc layer are obvious and are further confirmed by the EDX mapping. In the EDX mapping, the S element is only observed on the surface of the coating layer, while no S is captured below the MXene layer, which further supports the cation/anion selectivity of the coating layer.
Moreover, ex situ XPS work was carried out to detect the surface residence in the zinc plate as shown in Figure S6, Supporting Information. The S 2p spectrum in the bare zinc plate shows dominate peak at the binding energy of 170.5 eV, corresponding to SO/SZn bonds, suggesting the intensive decomposition of the anions (CF 3 SO 3 − ) in the zinc plate surface. On the contrast, no obvious peak is observed in the S 2p spectrum of the zinc plate after 100 cycles with the protection of the CNF/MXene coating layer, revealing the minimum decomposition of anions and confirming that the CNF/MXene coating layer suppressed the anion-induced decomposition. In addition, linear sweep voltammetry curves are further used to investigate the corrosion reaction (hydrogen evolution) in the surface of zinc plate in Figure S7, Supporting Information. The CNF/MXene@Zn electrode shows an increased corrosion potential of -0.95 V, which is much higher than that of bare zinc anode (−0.97 V). The increased corrosion potential reflects that the corrosion reaction of water is greatly suppressed by the CNF/MXene coating layer. To further backup our claim, we carried out the GC to detect the gas evolution in the electrode/electrolyte interface. In a typical analysis, a constant current of 10 mA cm −2 is applied and the zinc anode with CNF/MXene coating shows greatly reduced hydrogen evolution rate (≈0.8 mmol cm −2 h −1 ), which also strengthens the reduced corrosion of zinc plate ( Figure S9, Supporting Information).
Coulombic efficiency (CE) is an important parameter to evaluate the reversibility and stability of the Zn anode. It is calculated based on the ratio between the stripping zinc ions to the plated zinc ions at each cycle. In Figure 4a and Figure  S9, Supporting Information, the Cu//CNF/MXene@Zn asymmetric batteries show a stable and high average CE (>99.0%) for over 300 cycles at a fixed plating capacity of 1 mAh cm −2 . In a sharp contrast, without protective layers, CE value of the bare Zn/Cu symmetric cells failed after only 28 h, revealing the unstable zinc stripping and plating behaviors on copper substrate. The greatly improved CE is attributed to the uniform stripping and plating of zinc and inhibition of side reactions owing to the CNF/MXene protective layers. In addition, the electrochemical Zn plating/stripping performance was evaluated using symmetric Zn//Zn batteries with and without CNF/MXene coating layers. As shown in Figure 4b,c, the bare Zn//Zn batteries show an overpotential of about 80 mV at a low current density of 1 mA cm −2 . The batteries suffer obvious fluctuation after cycling for 20 h and rapid short circuit after only 70 h (Figure 4d). In comparison, the symmetric cells with CNF/MXene coating on zinc anode surface show a low overpotential of only 60 mV and an extremely stable and long cycling for over 1200 h at a current density of 1 mA cm −2 . Figure 4c shows a comparison of the rate performance of the Zn symmetric batteries with and without CNF/ MXene protective layers at a fixed capacity of 1 mAh cm −2 . It is clear that the batteries with CNF/MXene-protected Zn anodes exhibit a flat and stable voltage plateau at each current density. When the current density increases from 1 to 5 mA cm −2 , the overpotentials of the symmetric batteries increase from 58 to 110 mV. When the current density is reduced back to www.advenergymat.de www.advancedsciencenews.com 1 mA cm −2 , the overpotential returns to about 58 mV, showing excellent rate capability and reversibility. In order to verify the fast kinetics of the CNF/MXene membrane, the EIS spectrum of Zn/Zn symmetrical battery without and with CNF/Mxene coating was obtained at different temperatures as shown in Figure S11, Supporting Information. The log(1/R ct ) against 1000/T data were then plotted for zinc anode with and without CNF/Mxene coating. The data show that the plot obeys the Arrhenius equation A, frequency factor; E a , activation energy; R, gas constant; T, absolute temperature ( Figure S12, Supporting Information). This result suggests that R CT can be described as the thermal activation process with the activation energy E a for interfacial charge transfer. The CNF/MXene membrane shows greatly reduced activation energy of 0.14 eV (vs 0.26 eV), suggesting the facilitated interfacial resistance for zinc ion transport for CNF/MXene coating layer. The positive role of CNF/MXene membrane is further demonstrated in Cu/Zn asymmetric cells and Zn/Zn symmetric cells tested at ultra-high current. As shown in Figure 5a, at the high current density of 5 mA cm −2 for the high areal specific capacity of 5 mAh cm −2 , the Cu-CNF/MXene@Zn asymmetric cells can deliver an ultra-stable cycling performance with a high average CE (about 99.43%) for about 200 h. The voltage profiles of the asymmetric cells at selected cycles are shown in Figure S11, Supporting Information. Apparently, the Cu-bare Zn cells witness a sudden short circuit at the initial cycle with CE value of only 10%. When the current density increased to 10 mA cm −2 for 10 mAh cm −2 , the Cu-CNF/MXene@Zn asymmetric cells still maintain a high cyclic stability and high CE value (about 100%) for more 100 h. (Figure 5b) Notably, the enhanced stability is achieved in the asymmetric cells equipped with CNF/MXene membranes. The performance of CNF/MXene@Zn symmetric cells at high current was further investigated. As shown in Figure 5c, the CNF/MXene@Zn symmetric cells present high cyclic stability at the current density of 10, mA cm −2 for areal specific capacity of 10 mAh cm −2 . To verify the positive role of CNF/MXene membrane on the utilization of Zn, we further tested the CNF/MXene@Zn symmetric cells at the extremely high current of 100 mA cm −2 . The theoretical areal specific capacity of the CNF/MXene@Zn electrode is 113.34 mAh cm −2 . As the cycling performance and voltage profile shown in Figure 5d and Figure S12, Supporting Information, the CNF/MXene@Zn symmetric cells present low overpotential of 164 mV and steady cyclic performance for 120 h at the areal specific capacity of 100 mAh cm −2 , corresponding to 88.2% of the total theoretical areal capacity. The EIS spectrum ( Figure S16, Supporting Information) and the non-rectangular shape confirm that the cell is not soft short circuited. [31] The cumulative plated capacity (CPC) is proposed as an indicator to evaluate the stability of symmetric cells. Encouragingly, with CNF/MXene membrane, the symmetric cells achieve the CPC value as high as 12 Ah cm −2 , significantly higher than those of the previously reported value. [32] (Figure 5e) SEM images of the CNF/MXene@Zn electrode after cycling were obtained to further study the usage of zinc anode. In the SEM images shown in Figure S17, Supporting Information, and the EDS mapping shown in Figures S18, Supporting Information, the integrity of the electrode is well maintained. The zinc anode was still well coated with CNF/MXene membrane after stripping/plating for 120 h ( Figure S19, Supporting Information). Compared with the electrode before cycling ( Figure S4, Supporting Information), the thickness of CNF/MXene@Zn electrode only increased from 150 to 182 µm, suggesting that minimum porous zinc dendrite is formed (Figure 5f).
To evaluate the performance of CNF/MXene membrane in practical applications, full cells are assembled using CNF/MXene@Zn as anode and the synthesized VO 2 nanofibers as cathode. The VO 2 //CNF/MXene@Zn batteries can operate with negligible self-discharge ( Figure S20, Supporting Information). Figure 6a illustrates the initial CV curves of VO 2 //Zn batteries and VO 2 //CNF/MXene@Zn batteries at a scan rate of 0.1 mV s −1 . Compared with VO 2 //Zn batteries, the CV curves of VO 2 //CNF/MXene@Zn batteries exhibit improved current density and anodic/cathodic peaks shifting to more negative/ positive potentials, revealing that the CNF/MXene membrane enhances the conductivity of batteries. Therefore, the VO 2 // CNF/MXene@Zn batteries present an improved rate capability with the current density ranging from 0.5 to 30 A g −1 . As shown in Figure 6b,c, the batteries achieve a high specific capacity of 403 mAh g −1 at 0.5 A g −1 . With the current density increased to 30 A g −1 , it still retains the specific capacity of 90 mAh g −1 , about threefold higher than the capacity of VO 2 //Zn batteries (27.7 mAh g −1 ). With the current density returns 0.5 A g −1 , the specific capacity returns to 393 mAh g −1 , showing an excellent rate capability. The remarkable rate performance is ascribed to the low charge transfer resistance due to the CNF/MXene membrane coating layer. Figure 6d,e illustrate the cycling performance of the batteries at the current density of 2 A g −1 and 30 A g −1 , respectively. At 2 A g −1 , the VO 2 //CNF/MXene@Zn batteries deliver a high initial specific capacity of 357 mAh g −1 , and still stabilize at 333.2 mAh g −1 after 500 cycles, retenting 93.2% of the initial capacity. The VO 2 /Zn batteries drop to Figure 6. Electrochemical performance of full cells using CNF/MXene coated zinc metal as anode and the synthesized VO 2 as cathode. a) CV curves for VO 2 //Zn batteries and VO 2 //CNF/MXene@Zn batteries at a scan rate of 0.1 mV s −1 . b) Typical voltage profile and c) rate capabilities of VO 2 // CNF/MXene@Zn batteries at the current density ranging from 0.5 to 30 A g −1 . Long cycling performance at the current density of d) 2 A g −1 and e) 30 A g −1 . f) VO 2 //CNF/MXene@Zn batteries at low N/P ratio of 2.8. Surface morphology of g) bare Zn electrodes and h) the zinc anode with CNF/ MXene layer and after 500 cycles at 2 A g −1 .
www.advenergymat.de www.advancedsciencenews.com 270 mAh g −1 after 500 cycles, with the capacity retention of 82.6% (Figure 6f). At the extremely high current density of 30 A g −1 , the specific capacity increases as the cycling continues due to the activation process. After reaching the maximum activated capacity of 110 mAh g −1 at around 200th cycle, the batteries exhibit a capacity degradation, maintaining the specific capacity of 80 mAh g −1 after 2000 cycles, indicating an impressive long-term durability with 100% coulombic efficiency. In a sharp contrast, without CNF/MXene membrane coating, the VO 2 /Zn batteries can only deliver the specific capacity of about 35 mAh g −1 . This probably results from the low Zn 2+ plating/ stripping kinetics and the associated side reactions. Moreover, the VO 2 /Zn batteries can also operate stably at 1 C ( Figure S21, Supporting Information). The improved kinetics of Zn stripping/plating is verified by the EIS analysis before and after 500 cycles. As shown in Figure S22, Supporting Information, both the initial and cycles R ct of VO 2 //CNF/MXene@Zn batteries are smaller than those of the VO 2 //Zn batteries. (60.2 Ω vs 218.5 Ω, and 84.3 Ω vs 320.5 Ω, respectively.) This demonstrates the faster charge transfer and enhanced plating/stripping kinetics with CNF/MXene coating layer. In addition, after 500 cycles, the resistance for VO 2 //Zn batteries shows a large change (from 218.5 to 320.5 Ω). Such obvious fluctuations in the R ct suggest uneven Zn plating and stripping upon cycling. With CNF/MXene membrane coating on the surface of zinc anode, the R ct remains relatively stable after 500 cycles (from 60.2 to 84.3 Ω), revealing that the CNF/MXene membrane can support stable and corrosion-free Zn plating/stripping behaviors. Figure 6g,h shows the surface morphology of bare Zn electrodes and the zinc anode with CNF/MXene coating layer after 500 cycles at 2 A g −1 . It is obvious that the bare zinc anode experienced a serious corrosion and forms a large amount of "dead Zn" due to the uneven Zn plating/stripping during cycling. While, with CNF/MXene membrane coating, the zinc anode is able to cycle free from the active water and anion corrosion and presents a corrosion-free and plate-like plating morphology, indicating that the CNF/MXene membrane enables the uniform zinc deposition and inhibiting the growth of zinc dendrite. To further prove zinc anode cycling at severe conditions, the full cell with the N/P ratio of 2.8 was tested for both VO 2 //CNF/MXene@Zn batteries and VO 2 @Zn batteries. The mass loading of VO 2 cathode is about 6 mg cm −2 with the capacity of 2.1 mAh cm −2 , while zinc foil with a thickness of 10 mm (5.8 mAh cm −2 ) was used as anode. As shown in Figure 6f, at the current density of 1 A g −1 , VO 2 //CNF/MXene@ Zn batteries and VO 2 @Zn batteries show a high closed initial capacity of about 365 mAh g −1 . However, the capacity of VO 2 @ Zn batteries dropped to only 54% of its initial capacity after 100 cycles. On the contrast, the VO 2 //CNF/MXene@Zn batteries show a relatively stable cycling performance at low N/P ratio with slight capacity decay, retaining 86.2% of its initial capacity.

Conclusion
In this work, an ion selective and water resistant CNF/MXene composite membrane is developed to modulate the stripping and plating processes of zinc ions through ion sieving. The CNF/MXene membrane on the surface of zinc metal enables the molecular sieving effect to restrict the active water and screen-out anions from the electrode/electrolyte interface, avoiding corrosion upon zinc stripping/plating, while facilitating the zinc ions to pass through the film leading to corrosion free zinc plating. Therefore, the CNF/MXene@Zn anode exhibits obviously enhanced average coulombic efficiency (99.0%), low voltage hysteresis, and long cycling stability (≈3000 h). At the extremely high current density of 100 mA cm −2 , the symmetric batteries can still deliver a recordhigh capacity of 100 mAh cm −2 , corresponding to a ultra-high Zn utilization of 88.2%. It is worth noticing that the porous dendrite formation is the minimum with zinc plate thickness increasing from 150 to 182 µm. When being coupled with a VO 2 cathode, the full battery exhibits a high capacity (403 mAh g −1 at 0.5 A g −1 ), outstanding rate performance, and long cycling stability (80 mAh g −1 after 2000 cycles at 30.0 A g −1 ). This CNF/ MXene membrane with ion sieving performance to reject anions and restrict active water in the electrode/electrolyte interface demonstrate a significant effect in modulating the stripping and plating processes, showing a great potential to be applied in other metal batteries.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.