Synthesis and characterization of the FCOF film.
The imine-linked FCOF thin films were prepared through a solvothermal procedure (Fig.2a). In a typical process, two monomers (2,3,5,6-tetrafluoroterephthaldehyde (TFTA) and 1,3,5-tris (4-aminophenyl) benzene (TAPB)), are dissolved in a dioxane/mesitylene (D/M) mixture and condensed in a solvothermal tube, using acetic acid as the catalyst. To obtain a highly crystalline and continuous FCOF film, reaction conditions i.e. the proportion of solvent mixture, the concentration of catalyst and the reaction time needs to be controlled (Supplementary Fig. 1-4). When the ratio of D: M solvent in the mixtures is optimized to 1:9 (v/v) with 1.5M acetic acid catalyst, continuous bright orange FCOF films without any insoluble COF particulates are uniformly attached to the tube inner wall, indicating the FCOF films is successfully achieved (Supplementary Fig. 1). After soaking in pure water overnight, the free-standing FCOF films detach from the tube wall due to surface tension (Supplementary Fig. 5). The growth mechanism of FCOF film during solvothermal processes is mainly attributed to the fusion of numerous nanospheres that are formed during a co-condensation reaction (Supplementary Fig. 2), as revealed by the time-dependent morphology evolution (Supplementary Fig. 4). In addition, such uniform FCOF films can grow on various substrates such as copper (Cu), silicon, stainless steel foil/grids, nickel and titanium foils by placing the targeted substrates in the reaction solutions, as shown in Supplementary Fig. 6. This is beneficial for structural and property characterization after being transferred to various substrates during subsequent post-processing.
The crystal structure of the FCOF film is identified by two-dimensional wide-angle X-ray scattering (2D WAXS) measurement (Fig. 2b). From the integrated WAXS curves, the peaks at q=0.20, 0.35, 0.40, 0.53 A-1 correspond to the planes (100), (110), (200) and (210), respectively, consistent with a previous report34, confirming the high degree of crystallinity of the as-prepared film. High-resolution transmission electron microscopy (HRTEM) images clearly show the lattice fringes of the FCOF film at a spacing of ~0.34nm (Fig. 2c), which represents the π-π stacking distance. Field emission scanning electron microscopy (FESEM) images showed a smooth and defect-free film suspended on a copper grid (Fig. 2d), and the folds at the edges also reflect the flexibility of the film to a certain extent (Fig. 2e). Brunauer-Emmett-Teller (BET) measurement indicates the surface area of the film is as high as 723 m2g-1, and the main pore size distribution is 2-3 nm in diameter (Fig. 2f), which is well consistent with the WAXS result. The chemical structure of the FCOF film was further confirmed. The emerging peak at 1614 cm-1 in FTIR spectra (Fig. 2h) is assigned to the newly formed C=N imine stretch vibrations. The peak intensity at 1705 cm-1 assigned to C=O stretching weakens in FCOF film, indicating the consumption of the aldehyde groups of TFTA monomers34. High-resolution X-ray photoelectron spectroscopy (XPS) of the N1s spectra (Fig. 2g) shows that the weak peak at 399.89 eV arises from the N-H bonds35, revealing the small amount of residues of the amino groupss34, consistent with the FTIR result (N-H peak at ~3400 cm-1). The high intensity of the peak at 398.96eV in the N1s spectra (Fig. 2g) further confirm the formation of C=N (imine) bonds35 in the FCOF film.
The F atoms are the crucial elements within the FCOF films for achieving high-performance Zn anodes. The element mapping obtained from energy dispersive X-ray spectroscopy (EDX) indicates that F is evenly distributed in the film (Supplementary Fig.7). The F content is estimated to be 8.25 atomic %, in according to the XPS result (Supplementary Fig.8). In addition, the thickness of the film is adjustable by means of controlling the concentration of monomers. As determined by AFM analysis, the FCOF films had thicknesses of ~100, ~300 and ~500nm (Supplementary Fig. 9). The FCOF@Zn anode is fabricated via a pulling method in acetone solvent using Zn foil as substrate. After drying, the FCOF film tightly adhered the surface of the Zn foil and did not detach even under rolling, bending or unfolding of the Zn (Fig. 2j, 2i). In addition, as determined by nano-indentation measurements (Fig. 2k), the high quality two-dimensional FCOF crystalline films showed a remarkable elastic modulus exceeding 30GPa and an average hardness of over 1.2GPa, which is an order of magnitude higher than a recently reported TiO2 and polyvinylidene difluoride (PVDF) hybrid matrix36 (2.67GPa). The good mechanical strength is greatly beneficial to buffer volume expansion and retard dendrite propagation during the dissolution/deposition of Zn anodes.
Good Zn2+ conductivity through a protective layer is highly desired for Zn anodes. To experimentally determine the ion transport behavior of the FCOF films, the ionic conduction was calculated based on electrochemical impedance spectroscopy (EIS) results (Supplementary Fig.10 a and b). The ion conductivity (24.19 mS/cm) for the 100nm FCOF film coated glass fiber separator is 1.7 times higher than the bare glass fiber separator (14.12mS/cm). The result indicates that Zn2+ transport is enhanced by the FCOF film37,38. According to the equivalent circuit fitting results (Supplementary Fig.10 c and d), Zn anodes coated by the FCOF films with different thicknesses revealed lower charge transfer resistance (Rct) than bare Zn. In particular, the Rct of Zn anodes coated by the 100 nm thick FCOF film is lowest (90 Ω), about half that of bare Zn (180 Ω). Apparently, the Zn2+ transport is increased by the fluorinated 1D nanochannels. This is mainly due to the F atoms surrounded within the nanopores that endow the film a strong hydrophobic effect. Consequently, the fluorinated nanochannels inside the FCOF film appear to assist the de-solvation of hydrated Zn2+, thereby increasing the Zn2+ transport39. This is further supported by the Zn2+ transference number (ZTN) tests (Supplementary Fig.11), wherein the FCOF films show a higher ZTN (0.7) compared with conventional glass fiber separators (0.4).
In addition, the FCOF film is beneficial for aqueous electrolyte (2M ZnSO4) corrosion resistance of the Zn surface. The impedance of the FCOF@Zn symmetric cells increases from 180 to 600Ω within 8 hours (h) after cell assembly (Supplementary Fig. 12a). For bare Zn symmetric cells, on the contrary, the impedance increases dramatically from 200 to over 10000 Ω after 8 h (Supplementary Fig. 12b). The increase in impedance implies that the continuous corrosion of Zn by the electrolyte results in large amounts of by-products deposited on the surface. Time-dependent XRD patterns (Supplementary Fig.12c) show a peak at around 8° appears, corresponding to the by-product species Zn4SO4(OH)6·5H2O (JCPDS# 39-0688)11. When plain Zn anodes are immersed in aqueous electrolyte for 48 h, the peak intensity of by-products increases sharply. Much less irreversible by-product is accumulated on the surface of FCOF@Zn anodes during the same time duration.
Horizontal Parallel Zn platelet deposition enabled by the FCOF film.
In addition to the fast ion conduction and suppression of side reactions features, the morphology and texture of Zn deposits has been proven to have large impact on the cycling life of Zn batteries. Attaining an even planar deposition can ensure the batteries running for a prolonged time without short circuiting. To investigate the deposition morphology of Zn underneath the FCOF film, the Ti/Zn or FCOF@Ti cells are employed. As shown in Fig. 3a, b, the Zn deposits underneath the FCOF film exhibits platelet morphology and the platelets are stacked horizontally in response to a controlled capacity of 1mAh/cm2. Meanwhile, for the bare Ti without FCOF film protection (Fig. 3c, d), disordered, distributed and irregularly-shaped Zn dendrites are observed on the surface. When further increasing the used capacity to 2mAh/cm2, similar consistent morphological characteristics of the two samples are still maintained. The XRD results reveal the intensity of (002) plane located at 2θ=36.3° is highest for the Zn deposits underneath FCOF films (Fig. 3e, f), while the bare Zn deposits show (101) planes dominating the peak intensities (Fig. 3g, h). This change in the dominant peaks implies that the FCOF films on Zn anodes influence the preferred orientation of the Zn deposits. The orientation of the Zn deposits can also be quantified by calculating the texture coefficient40,41 (Tc, Supplementary Fig. 13). The Tc(002) of Zn deposits underneath the FCOF film is 19.2, much higher than that of the deposits on bare Zn (11.5), verifying the preferential growth on the (002) plane of Zn modulated by an FCOF film. X-ray diffraction pole figures were used to further identify the texture information of Zn deposits. The (002) pole figure (Fig. 3k) of Zn underneath FCOF films shows a sharp intensity concentration around ψ = 0-20°, indicating that the Zn platelets have a preferred textured based on (002) planes, and are nearly paralleled to the electrode substrate24,42 (Fig. 3i). In contrast, the random distributed of bare Zn deposits leads to a broad distribution of grain orientations, and the corresponding (002) pole figure (Fig. 3j) shows almost uniform distribution of diffraction intensity along the radial direction, indicating its random (non-preferential) texture.
In addition, the 2D WAXS patterns of deposited Zn underneath FCOF film show some strong, discrete diffraction spots in the ring plane (Fig. 3l, m), while for bare Zn deposits, the WAXS results are continuous diffraction rings (Fig. 3n, o). This indicates that the bare Zn deposits are polycrystalline and randomly oriented, whereas the Zn grain size influenced by FCOF films is larger and more oriented5. The structure of the Zn platelets was characterized by HRTEM and selected area electron diffraction (SAED). As shown in Fig. 3p,q and Supplementary Fig. 14, the diffraction patterns of the SAED results can be indexed into diffraction spots of the [001] zone. The HRTEM image of Fig. 3r further shows two d-spacings of 0.230 nm and 0.133 nm with an interfacial angle of 90°, corresponding to the (100) and (1-20) planes, respectively. The HRTEM result is in accord with the indexed SAED diffraction spots. To further verify the indexing results, an atomic arrangement model of Zn along the [001] direction was simulated, as shown in Fig. 3s. Obviously, the indexed result matches well with the theoretical crystal model. According to the above results, we conclude that the exposed hexagonal planes of the Zn platelet are predominately (002) planes.
Performance evaluation of the high-rate and long-life zinc anode.
The planar Zn deposition morphology, fast Zn2+ transport, and corrosion resistance properties enabled by the FCOF film are expected to greatly improve the electrochemical performance of Zn anodes. The reversibility of Zn anodes can be measured by a procedure that wherein a specific amount of Zn is plated on the substrate and then stripped away. Coulombic efficiency (CE) is an important index to evaluate such reversibility. The CE using the half cells in FCOF@Ti/Zn and Ti/Zn configurations is measured. At a moderate current density (1mAhcm-2, 5mAcm-2, Supplementary Fig.15a-c), the FCOF@Ti/Zn cells produced CE values of ~98.4% on average, with stability over 480 cycles. By contrast, the Ti/Zn with no FCOF cells ran for only 30 cycles, and their CE was around ~95.1%. When the current density is increased to an ultrahigh current density of 80 mAcm-2(Fig. 4a), the FCOF@Ti/Zn cells still exhibited a high CE of approaching 97.2% on average within for 320 cycles, whereas the CE of the Ti/Zn cells decreases rapidly after 95 cycles. When further increasing the capacity to 2mAhcm-2 at a current density of 40mAcm-2, the FCOF@Ti/Zn cells showed CE of 97.3% for over 250 cycles, much higher than that of the Ti/Zn cells (~35 cycles, 84.1%). Remarkably, as evidenced from the AFM height and phase imaging (Fig. 3e, f, Supplementary Fig.16a and c), the horizontally arranged platelet morphology of the Zn deposits underneath FCOF well remained well after 100 cycles during Zn plating/stripping processes (1mAhcm-2, 5mAcm-2). The average height difference (along X and Y axis) is only 170 nm (Supplementary Fig.16e), indicating the surface of the Zn deposits underneath FCOF is very flat and homogeneous. However, the Zn deposits on bare Ti after 100 cycles showed fluctuating and rough patterns with a much higher height difference of 710nm (Supplementary Fig.16b, d and f). During the Zn plating/stripping process, the H+ from the decomposition of water will receive electrons and then evolve H2, which could induce an increase of OH-. The generated OH- will react with Zn2+, SO42-, and H2O to form by-products such as Zn(OH)2 or Zn4SO4(OH)6·nH2O on Zn surface16. Raman spectroscopy was carried out to reveal the components on Zn deposits surfaces after cycling (100 cycles at 1mAh/cm2, 5mA/cm2). Sharp peaks at 1152, 1110, 1011 and 967 cm-1 are observed on the Zn deposits on bare Ti (Supplementary Fig.17a), which implies that the by-product should be the Zn4SO4(OH)6·5H2O43. In contrast, the peaks of the Zn deposits underneath FCOF are not obvious and their intensity is much lower (Supplementary Fig.17b), indicating less by-product accumulation on its surface. Raman mapping (8×10μm area, Fig. 4g) of the dominated peak at 967 cm-1 reveals that the counts variation for the Zn deposits underneath FCOF is within 13-726, which is one to two orders of magnitude smaller than for Zn deposits on bare Ti (Fig. 4h, counts: 1200-8400). It has been reported that the Zn deposits with high percentage of (002) planes parallel to the substrate could provide higher corrosion resistance than other planes44. Combined with the F endowed hydrophobic properties, the water-related side reactions could be largely suppressed in the Ti@FCOF@Zn cells. In addition, the bare Ti is not adequate for regulating the zinc deposition behavior or suppressing by-products accumulation, causing elevated voltage hysteresis or short circuiting of the batteries, as evidenced by the voltage fluctuation in Ti/Zn cells during cycling (Supplementary Fig.15c-e). Whereas the voltage files of FCOF@Ti/Zn remained stable at various levels of current density (Supplementary Fig.15a, Fig. 4c and d). Meanwhile, FCOF@Ti/Zn anodes display long-term stability of the Zn plating/stripping process even at ultrahigh current density up to 80 mAcm-2, larger than that of most previous studies (Supplementary Table 1).
To evaluate the stability of the Zn anodes, the FCOF@Zn symmetric cells showed prolonged cycle life for over 1700h at 1 mAhcm-2 and 5 mAcm-2, which was nearly 13 times the performance figures of the bare Zn anodes (Fig. 4i). The FCOF@Zn symmetric cells showed lower voltage hysteresis (FCOF@Zn: 60 mV vs. bare Zn: 80 mV), which we mainly attribute to the enhanced Zn2+ transport within the 1D fluorinated nanochannels. Under elevated current densities of 8 and ,40mAcm-2 (Supplementary Fig.15e and Fig. 4j), the FCOF@Zn symmetric cells could sustain repeated deposition/dissolution processes without obvious significant fluctuations in the voltage-time curves. However, the bare Zn symmetric cells suffer short-circuits after a few limited cycles. The excellent performance of FCOF@Zn symmetric cells at ultrahigh current density (40mAcm-2) is also far superior to most previous reported values (below 10mAcm-2, Supplementary Table 2). The highly stable electrochemical performance of the FCOF@Zn anodes indicates that dendrite formation is largely suppressed. To identify the suppression of dendrite growth in FCOF@Zn anode, transparent home-made Zn/Zn symmetric cells with or without FCOF were assembled to realize in situ monitoring of the Zn deposition process using an optical microscope. Zn deposition was performed under a current density of 20mAcm-2 for 35 mins. As shown in Fig. 4k, after an initial 5 mins of deposition, nonuniform Zn morphology with some protuberances appears on the bare Zn surfaces. These protuberances remained and grew into needle-like dendrites in the following deposition process. In contrast, the deposition on FCOF@Zn is smooth as evidenced in Fig. 4l. No obvious Zn dendrites were observed, even after 35 mins deposition. The microscopic morphologies of the Zn anodes after cycling at 1 and 5 mAhcm-2, for 500h were also investigated. As shown in Supplementary Fig. 18, the FCOF@Zn anodes show that dendrite-free morphology and parallel platelet-morphology is consistently maintained. In contrast, protuberant Zn dendrites are found randomly distributed on bare Zn surfaces. Moreover, the HRTEM and FTIR results (Supplementary Fig. 19) show the FCOF films maintain good crystallinity and chemical structure stability after cycling. Consequently, it can be concluded that the multifunctional F nanochannels greatly improve the Zn2+ kinetics and deposition morphology, which results in high-rate and long life FCOF@Zn anodes.
Full cell performance and flexible device demonstration.
We next evaluated the electrochemical performance of full cells in which FCOF@Zn or Zn anodes were paired with high mass-loading (~8mgcm-2) manganese dioxide (MnO2) cathodes. For the FCOF@Zn/MnO2cells, cyclic voltammetry curve (CV) curves demonstrated a larger current density at 0.1mVs-1 and a smaller voltage gap between typical redox peaks than in Zn/MnO2 cells (Supplementary Fig. 20a). This implies that the FCOF@Zn/MnO2 cells possessed higher specific capacity and better charge transfer capability6. EIS results further confirmed that, the impedance of the FCOF@Zn/MnO2 cells (~100 Ω) is lower than that of Zn/MnO2cells (Supplementary Fig. 20b). Therefore, at a current density of 3C, the FCOF@Zn/MnO2 cells revealed a high initial reversible specific capacity of 130 mAhg-1, while the Zn/MnO2 cells attained only 120 mAhg-1 (Supplementary Fig. 20c). The FCOF films clearly endowed stable cycling of the Zn anodes, and retained a capacity of ~92 % and stable charge/discharge curves after 1000 cycles (Fig. 5a and Supplementary Fig. 20d, e). This is nearly four times higher than the Zn/MnO2 cells (capacity retention: 20%). Reducing the capacity ratio of the negative electrode to the positive electrode (N/P) during full cell operation is a key parameter to achieve high energy density14,30,45. In previous studies, many systems chose to use thick zinc foil (≥100um) paired with low mass loading cathodes to assemble full cells, the N/P reported in these studies is typically higher than 50, which is not beneficial for achieving high energy density. In our case, the excellent performance of the FCOF@Zn anodes allowed us to further evaluate the cycle performance of full cells under harsh conditions. Using thin FCOF film-protected Zn plates as anodes (the thin Zn plates is rolled to desired thickness to satisfy the required N/P condition), FCOF@Zn/MnO2 cells with N/P=10:1 and N/P=5:1 showed stable specific capacity at current density of 4 mAcm-2 for over 300 and 200 cycles, respectively (Fig. 5b). The Zn platelet morphology after cycling and stable charge-discharge curves indicate the FCOF films enabled great performance improvements in Zn anodes (Supplementary Fig. 21). To evaluate the electrochemical performance of the aqueous Zn batteries for commercial applications under practical conditions, lean electrolyte addition and high areal capacity cathode is needed (inset of Fig. 5c). Here, the electrolyte-to-capacity ratio (E/C) used was controlled at 12 μL mA h-1 and the MnO2 mass loading was increased to 16 mgcm-2. The assembled full cell still showed considerable capacity of 0.5 mAhcm-2 after 250 cycles at a current density of 3mAcm-2. The gravimetric energy density of the cell is 130Wh/kg (based on the total mass of the Zn anode and the MnO2 cathode), which is significantly increased (by approx. 6.5 times) compared with many reported Zn/MnO2 cells using low mass loading cathodes and thick Zn foils30,39 (Fig. 5d). It should be noted that the cell still delivered an energy density of 55Wh kg-1 when including the electrolyte weight. Further optimization of other key components such as separator membrane and electrolyte may improve the energy density of the cell.
To further demonstrate the application prospects of the FCOF@Zn anodes for constructing realistic, smart, high-performance aqueous Zn batteries, we assembled a flexible transparent battery for device demonstration. Fig. 5e and Supplementary Fig. 22 shows the structural schematic diagram of the transparent battery. The MnO2 cathode and FCOF@Zn anode are fixed to the flexible PVC substrate, and glass fiber is used as the separator. All layers are sandwiched and the battery is then assembled by thermal sealing. The cycling performance of the flexible battery under different bending conditions is shown in Fig. 5h-j. The EIS results (Supplementary Fig. 23) and the charge and discharge curves (insets of Fig. 5i, j) remain nearly unchanged at bending angles of 0°, 45° and 60°, respectively, indicating its good mechanical stability and flexibility. To create a more realistic scenario, a flexible FCOF@ Zn/MnO2 battery was used to power a wearable bracelet for lighting a light emitting diode (LED) indicator (Fig. 5f, g), showing its promising application in portable wearable electronic devices.