Li1.3Al0.3Ti1.7P3O12 activated PVDF solid electrolyte for advanced lithium–oxygen batteries

Abstract

Lithium‐ion composite solid electrolyte membranes embedded with Li_1.3Al_0.3Ti_1.7P₃O₁₂ and poly(vinylidene fluoride) are prepared using a facile casting method. Furthermore, we added LiI as an active agent for decomposing the anode product. The synergy resulted in a high conductivity of 7.4 mS·cm⁻¹ and lithium‐ion mobility of 0.59 and a reduction of the overpotential to 0.86 V for lithium–oxygen batteries (LOBs). The membrane has enhanced Young's modulus of 6.6 GPa that effectively blocked the lithium dendrite growth during the battery operation and puncturing to the membrane led to a significant LOB cycle life of 542 cycles. Meanwhile, Li|Li symmetrical battery overpotential maintained at 42 mV after 470 h of operation.

INTRODUCTION

Aprotic lithium–oxygen batteries (LOBs) have attracted increasing attention as promising energy storage solutions due to their impressive theoretical energy density (approximately 3600 Wh kg⁻¹). However, several technical hurdles still impede their further development, such as the sluggish decomposition kinetics of lithium peroxide (Li₂O₂), electrode surface passivation resulting from the formation of thin-film products, carbon cathode decomposition, electrolyte degradation, and various complications associated with the lithium anode. Lithium metal (Li) exhibits an exceptionally high theoretical specific energy (3860 mAh g⁻¹) and a low oxygen reduction potential (−3.04 V vs. the standard hydrogen electrode), making it a good candidate for battery anodes. In Li–O₂ batteries, Li₂O₂ forms and decomposes at the cathode. Difficulty in decomposing the generated discharge product Li₂O₂ leads to poor round-trip efficiency and short cycle life of the battery. The discharge product causes the battery charging overpotential to increase dramatically, resulting in a high operating voltage, which reacts with the electrolyte and then causes the electrolyte to dry up therefore affecting the LOB's cycle life. Catalysts are commonly used to decompose Li₂O₂ by acting as electron and hole acceptors to reduce the charging overpotential of LOBs. In the electrochemical step, the soluble redox mediator (RM) is oxidized (reaction ()), then whereupon it undergoes a chemical reaction with Li₂O₂ (reaction ()). 1 $\mathrm{RMs}\rightleftharpoons \mathrm{R}{\mathrm{Ms}}_{\left(\text{soluble}\right)}^{+}+{\mathrm{e}}^{-}.$ 2 ${\mathrm{Li}}_2{{\mathrm{O}}_2}_{\left(\text{solid}\right)}+{2\mathrm{RMs}}_{\left(\text{soluble}\right)}^{+}\rightleftharpoons 2\mathrm{RMs}+{\mathrm{O}}_2+2\mathrm{L}{\mathrm{i}}^{+}.$

Generally, they can be divided into solid-phase catalysts and liquid-phase catalysts. Liquid-phase catalysts have been widely used in LOBs due to their good diffusion rate in solution as RMs, RMs include lithium halide, N-methyphenothiazine (MPT), tetrathiafulvalene (TTF), 2,2,6,6-tetramethylpiperidinyloxyl (TEMPO). They are oxidized to RM⁺, which provides electron holes for Li₂O₂ and further oxidizes it to O₂ during charging.

A key component affecting the performance of LOBs is the lithium metal anode. Uneven mass and charge transfer within the nonuniform solid electrolyte interphase (SEI) layer leads to uncontrolled growth of lithium dendrites. These dendrites can penetrate the SEI layer and continually deplete the electrolyte and lithium, facilitating reactions between the lithium metal and contaminant by-products, ultimately leading to sudden short circuits and safety issues. Composite solid-state electrolyte membranes have been previously applied to LOBs, but mainly focused on enhancing ionic conductivity, neglecting the inhibition of lithium dendrite growth, whereas Young's modulus is a key factor in evaluating the inhibition of lithium dendrites. Ideal composite solid-state electrolyte membranes need to have the following conditions, the membrane's base polymer is supposed to have flexibility and mechanical strength, good contact with lithium metal, and produce uniform Li⁺ transport channels. The polymer also has good compatibility with the electrolyte. The mixed solid electrolyte should also be uniformly distributed to avoid localized destruction of Li⁺, while the solid electrolyte should have good stability in air and water. Sodium superionic conductor (NASICON)-type ceramic polymers, such as Li_1.5A1_0.5Ti_1.5(PO₄)₃ (LATP), stand out as the most suitable choice for constructing composite solid electrolyte membranes due to their excellent air stability. However, LATP cannot be in direct contact with Li metal due to the reduction of Ti⁴⁺ by Li. Therefore, a polymer substrate is needed to form a stable interfacial layer with Li metal. Nonetheless, inherent characteristics of Li, including dendrite growth and volume changes during battery operation, pose significant challenges. These phenomena accelerate lithium consumption and elevate the battery's internal resistance, ultimately leading to premature failure. As Li₂O₂ is difficult to decompose, the battery generates excess charging overpotential, which impairs the performance of the battery, and the addition of a catalyst to the anode side of the battery can effectively decompose the discharge product.

LiI has been studied as a liquid RM can effectively reduce the voltage during the charging process. I₃⁻ acts as an electron acceptor and a RM. It showed the effect of the decomposition of discharge products on the performance of LOBs. However, during battery cycling, the production of I₃⁻ can occur, and this makes it susceptible to reacting with the lithium anode. This reaction generates LiI₃, which not only damages the lithium anode but also renders the I⁻ ineffective. Conventional commercial glass fiber (GF) membrane has a porous structure with high conductivity, which can satisfy the environment where lithium ions are free to shuttle, but due to the porous structure of GF membranes, it causes I₃⁻ to undergo the shuttling effect of RMs. Wu et al. prepared a protective separator with zeolite molecular sieves (4 Å) by adding zeolite to a poly(vinylidene fluoride-hexafluoropropylene)-based membrane, to suppress the shuttle effect of TEMPO, the LOBs with this membrane had an extended cycle life up to 170 cycles at a current density of 250 mA g⁻¹. Chen et al. used Nafion, PEO into a two-dimensional backbone structure to form a negative barrier for the purpose of blocking I₃⁻ shuttling, and a of cycle life of 472 cycles which is 3.8 times improvement compare with GF was obtained at a current density of 500 mA g⁻¹. Li et al. utilized a scalable and flexible membrane consisting of polytetrafluoro thylene@polystyrene (PTFE@PS), which possesses outstanding oxygen permeability and resistance to moisture. This membrane serves as an effective barrier, safeguarding both the lithium anode and the entire battery system against corrosion caused by moisture (H₂O). Remarkably, when operated at a capacity of 500 mAh g⁻¹, the LOBs equipped with the PTFE@PS membrane exhibited a five times improvement in cycling performance compared with the LOBs lacking this membrane. Sun et al. prepared polyetherketone nanofibrous membranes using chemically induced crystallization, and LOBs using this prepared membrane had high cycling stability (194 cycles at 500 mAh g⁻¹), and these polyetherketone nanofibrous membranes had a protective effect on the anode, as well as being able to regulate the Li⁺ flux. Yuan et al. coated lithium metal using an organic–inorganic interlayer-reinforced PVDF-HFP with a high Li⁺ transfer number (0.62) and Young's modulus (6.17 GPa). The Li/Li symmetric battery could be operated stably for at least 2190 h and LOBs could be operated for 214 cycles, protecting the lithium metal anode and improving the ionic conductivity. Protect the Li-metal anode from parasitic reactions caused by RMs and impurities diffusing from the cathode. LATP has gained our attention as a solid electrolyte membrane that exhibits better stability in air and water, but conventional solid electrolyte membranes are difficult to be applied in practice due to the brittleness of their own ceramics, interfacial issues of electrode contact, and other reasons.

In this work, we reduced the charging overpotential of LOBs by adding LiI, blocked the shuttle effect of RMs using composite lithium-ion membranes, and retained I⁻ effectively to play a role in the positive electrode while effectively protecting lithium metal. We prepared lithium-ion composite membranes by using polyvinylidene fluoride (PVDF) as a base polymer and LATP solid electrolyte as an inorganic filler, which utilizes the good compatibility of PVDF polymer with lithium surface, the formation of LiF bonds to homogenize Li⁺, protects lithium metal from side reactions, and provides highly stable lithium anode for LOBs. Compared with ordinary PVDF polymer membranes, lithium-ion composite membranes have higher lithium-ion mobility (t_Li⁺ = 0.59), higher conductivity (7.4 mS cm⁻¹), and LOBs obtained by using lithium-ion composite membranes can be operated for 542 cycles, which is much larger than that of the traditional commercial GF membranes (80 cycles).

EXPERIMENTAL

Materials and chemicals

Li_1.3Al_0.3Ti_1.7P₃O₁₂ purchased from Shenzhen Kejing Zhida Technology Co. PVDF (Kynar Flex® 2801-00) was purchased from Arkema, France. Multi-walled carbon nanotubes (MWCNTs, ≥98%), N, N-dimethylformamide (DMF, ≥99.5%) were purchased from Sinopharm Chemical Reagent Co., Ltd., China. Dimethyl sulfoxide (DMSO, 99.9%) and propylene carbonate (PC, 99.7%) were bought from Sigma-Aldrich. GF separator (d = 18 mm, GF/D, Whatman) and carbon paper (TGP-H-060, Toray) were used directly as purchased. Lithium perchlorate (LiClO₄, ≥99.99%, Sigma-Aldrich) and lithium iodide (LiI, 99.9%, Sigma-Aldrich) was dried in a vacuum oven at 120°C for 12 h before adding into 1.0 mol L⁻¹ LiClO₄/DMSO and LiI/LiClO₄/DMSO (containing 1.0 mol L⁻¹ LiClO₄ and 0.05 mol L⁻¹ LiI) electrolytes. Molecular sieves (4 Å, Sigma-Aldrich) after activation were added to the electrolytes for the purpose of removing any moisture for 1 week prior to use. Lithium sheets (d = 14 mm, Tianjin Zhongneng Lithium Industry Co., Ltd.) are immersed in 0.1 mol L⁻¹ LiClO₄/PC solution for at least 3 days.

Synthesis of LCM composite membranes

The LATP solid electrolyte was first mixed with DMF solvent using an ultrasonic sound bath at 25°C for 30 min; PVDF pellets were added followed by further mixing for 4 h. Vacuum filtration was carried out for 4 h to remove any air bubbles. The obtained slurry was poured onto a glass plate and membranes with different thicknesses were cast using a doctor blade followed by drying in an oven at 60°C for 4 h. The membranes were then sliced into d = 16 mm for LOBs using a slicer and soaked in 1 M LiClO₄/DMSO for at least 24 h before use. The ratio of PVDF pellets to DMF solvent was approximately 16%, and the LATP:PVDF mass ratios of 1:3, 1:2, and 1:1 were labeled lithium-ion conducting membrane (LCM)-1, LCM-2, and LCM-3, respectively.

Battery assembly and testing

Briefly, 10 mg of MWCNTs were uniformly dispersed in 30 mL of ethanol solution under ultrasound to form an ink-like slurry, which was subsequently sprayed onto carbon paper using a spray gun. The carbon paper was then cut to d = 1 cm² (with loading of 0.1 mg MWCNTs) and dried in a vacuum drying oven at 80°C for 12 h and placed in a vacuum chamber for use as a positive electrode. The batteries were assembled in the following order: negative electrode—PVDF/LCM—GF (injected electrolyte)—positive electrode, in a glove box filled with argon gas for the assembly of CR2032 type batteries with holes (Nanjing Jiumen Automation Technology Company), where the electrolyte added is 130 μL and 0.05 M LiI is 0.0065 mmol in the electrolyte. In which LCM-2 membrane possesses stable interfacial contact and excellent battery conductivity as the main research object. Battery testing was carried out in a 99.9% pure oxygen atmosphere, the current density of the conventional battery cycle was 1 A g⁻¹, corresponding to a battery capacity of 1000 mAh g⁻¹ and the battery capacity is calculated based on the content of MWCNTs. A fixed capacity of 1000 mAh g⁻¹ was used to test the multiplier performance at 3, 5 A g⁻¹ and full discharge performance was tested using 0.1 mA, 2 V as the abort test condition.

SSN|SSN symmetric battery was assembled, GF, PVDF, and LCM was used as a membrane, where a 1 M LiClO₄ electrolyte was injected and assembled using a CR2032 nonporous battery case. SSN|SSN symmetric batteries were used to study the improvement of ionic conductivity by LCM membranes, and the Li|Li symmetrical battery testing was utilized to explore the influence of the LCM-2 membrane on the Li anode.

Characterizations

A field emission scanning electron microscope (Nova NanoSEM450) was used to study the morphology of the material surfaces and an x-ray diffractometer (Bruker D8A A25 X) to characterize the structural properties of the materials. Raman spectrometer (532 nm, DXR2, Thermo Scientific) is used for the discharge product analysis. An electrochemical workstation (CHI 760E, Shanghai Chenhua Instrument Co., Ltd.) was used for electrochemical performance testing, an x-ray spectrometer (Thermo Scientific K-Alpha+, x-ray source: Al Kα micro-focused monochromatic source) was used to characterize the elemental properties of the membranes, a Fourier transform infrared (FTIR) spectrometer (NICOLET IS 10) was used to characterize the functional groups of the membranes, and a contact angle measuring instrument (JC2000D1, Shanghai Zhongchen Digital Technology Equipment Co., Ltd.) was used to measure the hydrophilic nature of the membrane surface. Battery performance testing with Battery Tester (CT-4008T-5V10mA Neware Technology Limited). In order to demonstrate the puncture resistance of the membrane against lithium dendrites, the Young's modulus of the membrane was measured by AFM (SPM-9700HT).

Electrochemical testing

Ionic conductivity

The electrochemical impedance spectroscopy (EIS) was carried out using a frequency range of 1 MHZ to 0.1 Hz, an AC voltage of 5 mV and a test temperature of 25°C. The ionic conductivity can be calculated using the below equation. $\sigma =\frac{d}{R_{\mathrm{b}}A}.$ where σ is the ionic conductivity, d is the thickness of the membrane, R_b is the solution resistance, and A is the effective contact area of the SSN electrode.

Lithium-ion transfer number

The lithium-ion transfer number (t_Li⁺) was obtained by using constant potential polarization and electrochemical impedance mapping, and t_Li⁺ can then be calculated by measuring the change in current before and after polarization, combined with information from the impedance mapping EIS before and after polarization, using equation as follows: $t_{\mathrm{L}{\mathrm{i}}^{+}}=\frac{I_{\mathrm{s}}\left(\mathrm{\Delta }V-I_{\mathrm{i}}{R}_{\mathrm{i}}\right)}{I_{\mathrm{i}}\left(\mathrm{\Delta }V-I_{\mathrm{s}}{R}_{\mathrm{s}}\right)}.$ where t_Li⁺ is the lithium-ion transfer number, I_s and I_e are the initial and steady state currents for constant potential polarization. ΔV is the constant potential polarization voltage: mV, R_s, and R_e are the resistance of the diaphragm before and after constant potential polarization.

I⁻ permeation testing

An H-shaped apparatus was used to evaluate the permeability of I⁻ by adding 7 mL of 0.05 M LiI containing 1 M LiClO₄ electrolyte to the left side of the H-shaped electrolytic battery and 7 mL of 1 M LiClO₄ electrolyte to the right side, with the septum in the middle. A solution of H₂O₂ with a mass fraction of 0.5% starch was prepared as the I⁻ detector, and 5 μL of the solution on the right side was aspirated and added to the detector during the test; if it turned blue, I⁻ was present.

Electrolyte adsorption test

The electrolyte adsorption rate is used to measure the adsorption capacity of the membrane to the electrolyte. The liquid electrolyte absorption (δ) is assessed according to the below equation. $\delta \left(\%\right)=\frac{W_{\mathrm{e}}-W_0}{W_0}.$ where δ is the electrolyte adsorption rate and W₀ and W_e represent the mass before and after adsorption of the electrolyte, respectively.

RESULTS AND DISCUSSION

The SEM image of the original PVDF membrane is shown in Figure , where tiny cracks appeared on the surface, that is too big to block I₃⁻ (0.514 nm) at the microscopic scale, Figure  shows the thickness of the PVDF membrane of about 60 μm, and the particles of the LATP are around 600 nm in Figure . The LATP particles can be added with different contents for varied ionic conductivity. Figure  shows the crystal structure of the LATP solid-state electrolyte, where Li⁺ can undergo rapid transport through the lattice. We fabricated LCM-1, LCM-2, and LCM-3 by incorporating LATP particles with various contents. It has been revealed that within a specific range of LATP particle addition, the conductivity and performance of the composite membrane were enhanced. However, it is important to concurrently assess the compatibility of LATP particles with the lithium anode. When LATP particles were added up to 50%, the membrane's surface became excessively convex with LATP, negatively impacting the stability of the lithium anode. Consequently, we determined that a 33% LATP content in LCM-2 represented the optimal ratio, and SEM images showed that the surface filler particles of LCM-2 had good homogeneity and interfacial contact. Figure  presents the SEM image of the LCM-2 membrane, predominantly composed of a blend of LATP particles and PVDF polymer. The membrane of LCM-2 exhibits a flat and dense structure, devoid of significant pores, which could hinder the infiltration of detrimental molecules/ions, preventing them from reacting with the lithium anode. Furthermore, the flat membrane structure promotes the even diffusion of Li⁺. The EDS surface (Figure ) and cross-sectional (Figure ) images show that LATP particles are homogeneously mixed with PVDF, where C and F elements represent within the PVDF; P, Al and Ti represent on the LATP.

[IMAGE OMITTED. SEE PDF]

The peak of the PVDF membrane in the XRD in Figure  appears at 20.5°. The XRD of Figure  shows LATP particles corresponding to the PDF cards of LiTi₂(PO₄)₃, and the XRD of the LCM-2 membrane shows the coexistence of the peaks of PVDF and LATP, which suggests that the crystalline phase of LATP was not disrupted during the preparation process and that the LATP particles were well-mixed with the PVDF polymer, while we found in Figure  infrared spectroscopy (FTIR) that the normal PVDF-based membrane exhibits CH bending (1402 cm⁻¹) and CF bending (1166 cm⁻¹). As the mass ratio of LATP increases, the CF peak at 1166 cm⁻¹ of LCM-2 shifts to 982 cm⁻¹, because of the tightened interaction between LATP and PVDF. At the same time, the shift of the peaks illustrates the disruption of the crystallinity of PVDF membrane and the increase in the proportion of amorphous regions of the polymer due to the addition of LATP. Because of the addition of LATP, the t_Li⁺ of the LCM-2 membrane in Figure  reached 0.59. Figure  shows the improved ionic conductivity of the LCM membranes, which has been increased from 3.3 mS cm⁻¹ of ordinary PVDF-based membranes to 4.6 mS cm⁻¹ for the one with LCM-1 membrane, 7.4 mS cm⁻¹ for the one with LCM-2 membrane, and 8 mS cm⁻¹ for the one with LCM-3 membrane. The addition of a small amount of solid electrolyte filler can effectively improve the lithium-ion mobility of the membrane, while the ionic conductivity shows a tendency to increase and then level off with the increase of the proportion of LATP. This may be because more vacant defects are provided with a small amount of active filler, making it easy for ions to jump continuously, and the solid electrolyte itself provides a large amount of Li⁺, increasing the concentration of Li⁺ at the interface with the polymer. When the active filler content is higher than 40%, the active filler appears to pile up, at which point most of the ion transfer is concentrated between the filler-filler transfer, thus reducing the ionic conductivity, and we believe that LCM-2 membrane is the optimum ratio, shrinking the gap with the normal GF membrane of 17 mS cm⁻¹ difference.

Characterization of LCM membranes

We were able to prepare LCM membranes on a large scale by the solution casting method, and Figure  shows the continuous and uniform membranes obtained under this method, and Figure  shows the contact angle of LCM-2 membrane with the electrolyte (DMSO) of 13.5, indicating that this membrane has good permeability toward the electrolyte which can provide a good internal ion channel environment (Figure ). The contact angles for the pristine PVDF membranes, (DMSO and water) were 11.2° and 87.2°, respectively. Figure  indicates that the GF membrane has an electrolyte adsorption rate of 811.2%, which can withhold plenty of electrolyte to be consumed during battery cycling, therefore we chose GF membrane matrix material for the LCM membranes preparation. LCM membranes have an electrolyte adsorption rate of only 13.2%; the small amount of electrolyte helps to fill some pores of the membrane and facilitates the ions transfer. Compared with GF and PVDF membranes, LCM-2 membranes have lower porosity and enhanced I⁻ blocking. The LiI catalyst was tested for I⁻ transmission to show that the LiI catalyst was able to sustain the effect on the positive side, but not losing it to the negative side. An H-type device with 0.05 M LiI containing 1 M LiClO₄ on the left side and 1 M LiClO₄ on the right side was used to analyze the permeability of the membranes. As it can be seen in Figure , a minor amount of the I⁻ ions permeated through the GF membranes at hour 6, the color of the left size solution (H₂O₂ with 0.5% starch blue) changed to light blue, at the 12-h time point, a substantial amount of I⁻ ions transferred, leading to a pronounced dark blue color. Conversely, the LCM-2 membrane exhibited exceptional resistance to I⁻ ion permeation throughout the 12-h permeability test, showcasing its robust I⁻ blocking properties.

[IMAGE OMITTED. SEE PDF]

Performance of LOB with LCM membranes

Figure  shows an extension of the cycle life of LOBs with LiI catalyst, achieving 542 cycles with an overpotential of 0.86 V when employing the LCM-2 membrane. Figure  shows the cycle life testing result of LiI LOBs with LCM-1 and LCM-3 membranes is 417 and 392 cycles, respectively, indicating that the LCM-1 membrane is inadequate to form a “local blocking” effect, whereas the LCM-3 membrane, with an excess of solid electrolyte particles, may have contact with the lithium metal and potentially affect the battery's performance. In comparison, Figure  indicates the LOBs without LiI show only 113 cycles with the overpotential raised to 1.57 V with the same LCM-2 membrane is used. This outcome suggests that LiI successfully reduced the battery's charging potential and enhanced the efficient decomposition of cathode products. This difference is primarily attributed to the charging potential, as it can be that in Figure , the charging potential of LOBs after 20 cycles is only 3.78 V, whereas in Figure , it elevated to 4.3 V after the same number of cycles, leading to continued charging of the electrolyte under high-voltage conditions. The enhanced ionic conductivity of the LCM-2 membrane is attributed to the presence of the LATP solid electrolyte, while the Lewis acid–base interaction impedes anion mobility and results in the uniform deposition of lithium ions. For the GF membranes with LiI, the cycle life of the conventional GF membrane in Figure  ran only 80 cycles, and the LOB with the PVDF membrane ran only 131 cycles shown in Figure . This indicates that LiI does not play a role in the continuous decomposition of the anode product in the battery assembled with GF membrane and PVDF membrane. Figures  show that the discharge curves of GF and PVDF membranes are lower than the discharge voltage of 2.0 V, which indicates that the lithium anode has lost the ability to detach lithium ions, and both GF and PVDF membranes are unable to prevent harmful substances from attacking the lithium anode surface. The discharge potential of LOBs in Figure  is kept at around 2.7 V within 113 cycles, but the charge potential reached 4.5 V, and the charging time was short. Figure  shows that the addition of LiI into the LiClO₄/DMSO electrolyte resulted in increased peak currents for the formation and decomposition of discharge products, suggesting that iodide accelerated the ORR and OER processes and that the two sets of peaks centered on 3.3 and 3.8 V versus lithium voltages in the tests of LCM-2 membranes were attributed to the I₃⁻/I⁻ and I₂/I₃⁻ redox couples, which experienced the equilibrium reactions of I₃⁻ + 2e⁻ ⇌ 3I⁻ and I₂ + I⁻ + e⁻ ⇌ I₃⁻, respectively. It is shown that the LCM-2 membrane can retain LiI, at the positive electrode. We also compared the battery cycle performance of the LOBs with GF and LCM-2 membranes at 1A, 3A, and 5A multipliers, Figure  shows the cycle performance curves of LOB with LCM-2 membranes of 542, 201, and 144 cycles at 1A, 3A, and 5A, respectively, and the performance of the one with GF membranes of 80, 53, and 35 cycles at 1A, 3A, and 5A, respectively. The LCM-2 membranes were able to maintain their excellent battery performance at high charge/discharge rates.

[IMAGE OMITTED. SEE PDF]

Figure  shows the full discharge performance of the LOB using GF membranes at a full discharge capacity of 3782 mAh. The full discharge performance of LOB with LCM-2 membranes showed a capacity of 43 755 mAh (Figure ), while the insertion of Figure  indicates that the full discharge capacity in argon is only 68 mAh, indicating that the ORR contributes dramatically to such a high full discharge capacity. The LCM-2 membranes can effectively improve the stability of the battery during cycling, resulting in a significant improvement in cycling, multiplier, and full discharge capacity.

Mechanistic characterization of membranes in the circulation

The elemental composition of the LCM-2 membrane after 80 LOB cycles was analyzed using XPS. The XPS was calibrated using C at a binding energy of 284.8 eV. Figure  shows the CC, CO, and CF bonds binding energies of 284.8, 286.8, and 290 eV, respectively, and the corresponding substances are (CH₂CF₂)_n and Li₂CO₃, where Li₂CO₃ may be produced due to the decomposition of the electrolyte. Figure  shows that the binding energies of the LiF and CF bonds of element F are 684.8 and 687.4 eV, respectively. Whereas the element F of the fresh LCM-2 membrane in Figure  shows only the CF peak at 286.6 eV, indicating that PVDF membrane and the lithium surface had more intimate contact with tight bonds after cycling and generated LiF which can homogenize the intercalation of lithium and prohibits the lithium dendrites growth.

[IMAGE OMITTED. SEE PDF]

Figure  shows the fine spectrum of O1s orbit with peaks at 528.5, 531.6, and 531.8 eV corresponding to TiO, AlO, and LiO, respectively, which also confirms the presence of Li₂CO₃. Figure  also shows the presence of LiF, and the binding energy of Li at 55.7 eV. Figure  shows that the binding energies of Ti 2p_3/2 and Ti 2p_1/2 are at 458.8, 464.3 eV, and Figure  shows that the binding energy of the AlO bond of the element Al is located at 74.5 eV, and the presence of the elements Ti and Al shows that the LATP solid electrolyte is stable in the battery cycle, but their signals are weaker than the other elements and can only be observed in the single element spectrum, which is not shown in the full spectrum of Figure .

Characterization of LOBs cathode

Figure  shows the SEM image the morphology of multi-walled carbon nanotubes (MWCNTs) on the pristine cathode, showcasing their strong adhesion to the carbon paper. Figure  shows the SEM of the GF membrane during discharge, where the cathode carbon paper is covered with discharge products, followed by their decomposition upon charging (as seen in Figure ). After 80 cycles, the battery assembled with the GF membrane fails, as demonstrated in Figure , which illustrates a significant amount of deposit covering the cathode surface. Raman spectra in Figure  indicates that the discharge products primarily consist of LiOH (460 cm⁻¹) and Li₂O₂ (798 cm⁻¹), with the intensity of LiOH peaks exceeding that of Li₂O₂, suggesting LiOH as the predominant product. For the LCM-2 membrane LOBs, Figure  reveals that carbon nanotubes remain clearly visible after just one cycle, indicating that LiI accelerates the decomposition of discharge products and facilitates rapid product dissolution. In Figure , discharge products cover the carbon cathode surface after 542 cycles, appearing as small granular deposits. Raman spectra in Figure  show that these granular products are primarily Li₂O₂, with a higher intensity of Li₂O₂ peaks compared with LiOH, signifying Li₂O₂ as the main product. Remarkably, the LCM-2 membrane LOB experiences failure not during discharge but during the charging process. Wide-ranging SEM mapping in Figure  reveals that extensive granular products entirely cover the positive electrode. We posit that I⁻ ions are continually consumed during the cycling process, as reflected in the battery's charging voltage reaching 4.3 V at 330 cycles. The ultimate failure of the LCM-2 membrane battery is attributed to insufficient charging time, resulting in an excess of challenging-to-decompose products on the positive electrode, confirming complete I⁻ consumption.

[IMAGE OMITTED. SEE PDF]

Characterization of the lithium anodes

In Figure  displays a smooth lithium surface, while Figure  reveals that the thickness of the lithium sheet is 383 μm. As shown in Figure , after cycling the LCM-2 membrane LOB for 80 cycles, a relatively flat lithium surface under ordinary camera, while the SEM image provides a detailed microstructure view, revealing the appearance of small particles on the lithium surface. Figure 's EDS image indicates that these particles primarily stem from trace amounts of LiOH, even though LiOH peaks were not detected in the XRD spectra in Figure . The presence of elements C and F in Figure  suggests that a minor amount of residual polymer remains on the lithium surface, confirming the close contact between the LCM membrane and the lithium sheet. In Figure , the LCM-2 membrane LOB results in a lithium sheet thickness of 353 μm after 80 cycles. However, Figure  revealed a completely pulverized lithium surface under an ordinary camera after cycled 80 cycles for the LOB with GF membrane. The XRD image in Figure  also demonstrates that the lithium metal surface has undergone complete pulverization, leading to the generation of LiOH. Additionally, Figure  shows the SEM image that highlights the extent of lithium sheet pulverization, leaving only a residual thickness of 161 μm. Unfortunately, this lithium sheet is nonfunctional due to the surface being covered by LiOH, which is the reason cause of the GF membrane's failure during battery cycling. Figure  shows the lithium surface of the LCM-2 membrane LOB after 542 cycles. Images in Figure  indicate that the lithium surface turned to a yellowish color, likely originating from by-products deposited on the lithium surface resulting from electrolyte decomposition. The SEM image shows the development of a thin membrane on the lithium metal surface. In the XRD image of Figure , lithium metal remains the dominant phase, with a small quantity of LiOH present. In Figure , the thickness of the lithium sheet is still 245 μm. The changes observed in the lithium flake's surface and thickness suggest that the LCM-2 membrane effectively protects the lithium anode.

[IMAGE OMITTED. SEE PDF]

Performance evaluation of Li⁺ stripping and plating

Solid polymer electrolytes with high mechanical modulus have been applied to control the generation of lithium dendrites during battery operation, the Young's modulus of traditional polymers is below 1 GPa, that is not able to meet the requirements of lithium dendrite blocking, and the purpose of lithium dendrite blocking can be achieved by adding the optimal proportion of inorganic fillers to increase the Young's modulus. Figure  shows that the Young's modulus of t PVDF membrane is 640 MPa, Figure  shows that due to the addition of LATP particles, the Young's modulus of LCM-1 membrane increases to 1.4 GPa, but it is not enough to meet the requirement of lithium dendrimer blocking, and by increasing the proportion of inorganic filler, Figure  shows that the Young's modulus of LCM-2 membrane reaches 6.6 GPa, which is more than 10 times of that of the PVDF membrane, and it achieves the goal of resisting lithium dendrimers. The AFM images of Figure  and Figure  show that the surface roughness of LCM-1, LCM-2, and LCM-3 membranes are 197.6 nm, 313.4 nm and 1.1 μm, respectively, for different LATP contents, but the highest content of LCM-3 membrane is not conducive. The Li|Li symmetric battery cycle of Figure  also showed that the LCM-2 operated at a low overpotential of about 42 mV for more than 470 h, whereas the PVDF membrane failed at 270 h. Figure  shows that the PVDF membrane had a similar voltage distribution to that of the LCM-2 membrane at the first 6 h with an overpotential of about 21 mV, whereas the PVDF membrane at 276 h exhibited a larger voltage shock. The potential of PVDF membrane showed a large voltage oscillation, and the overpotential dropped below 2 mV, and a short circuit of Li|Li symmetric battery occurred. The voltage distribution of LCM-2 membrane was stable at 273–277 h with an overpotential of 36 mV, and even at 470 h, it still preserved a stable voltage distribution with an overpotential of 42 mV, and the voltage distribution of LCM-2 membrane was stable. Figure  shows that the Li|Li symmetric cycle life of GF, LCM-1, and LCM-3 membrane is 210, 420, and 470 h. Among them, the LCM-1 membrane was not able to resist lithium dendrite and led to short circuits, while the LCM-3 membrane has poor contact with lithium sheet due to the large surface roughness, which leads to a gradual increase of the overpotential in Li|Li symmetric battery cycling, and the overpotential of Li|Li symmetric battery reached 160 mV at 470 h. Figure  shows that the LCM-2 membrane has a uniform fast voltage distribution, and uniform fast lithium-ion channels, while PVDF membrane has low lithium-ion mobility. Figure  shows a comparison of our work with others, and the radar plot shows that we outperform the comparative literature in LOBs in terms of both full discharge and cycling multiplier performance.

[IMAGE OMITTED. SEE PDF]

CONCLUSION

Composite LCM membranes have been successfully synthesized by incorporating LATP solid electrolyte as an active filler into PVDF polymer. The charging overpotential of LOBs is reduced, the shuttle effect of RMs is blocked with the addition of LiI as an active catalyst, and the LCM-2 membrane's conductivity is significantly improved at 7.4 mS cm⁻¹ compared with PVDF membranes with a conductivity of 3.3 mS cm⁻¹. Simultaneously, the addition of LATP led to a reduction in the porosity of the PVDF membrane and an enhancement in t_Li⁺ to 0.59. This improvement also contributed to a more effective blockage of I⁻. When applied in LOBs, the LCM-2 membrane demonstrated an impressive long cycle life of 542 cycles, along with a high specific capacity of 43 755 mAh. This underscores the efficacy of the solution casting method as a novel approach for the scalable production of durable LOBs membranes.

AUTHOR CONTRIBUTIONS

Caizheng Ou contributed to the investigation and writing—review and editing of the manuscript. Xiaoteng Liu was responsible for the conceptualization, supervision, and funding acquisition. Kun Luo and Xiaoqin Zou provided supervision and resources. Hao Zhang handled data curation. Dan Ma contributed to supervision. Hailiang Mu managed project administration. Xiangqun Zhuge, Yurong Ren, Maryam Bayati, and Ben Bin Xu conducted validation.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the financial support from Jiangsu Specially-Appointed Professor Fund by Jiangsu Education Department, Science and Technology Plan Project of Changzhou (No. CQ20D2EHPA034), the UK Engineering Physics and Science Research Council (Grant No. EP/S032886/1), RSC research grant (R21-220003), the National Natural Science Foundation of China (Nos. 51874051, 52111530139, and 22375031), the Jilin Natural Science Fund for Excellent Young Scholars (20230508116RC), the Fundamental Research Funds for the Central Universities (JGPY202103 and 2412023YQ001).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

References

Wu X, Yu W, Wen K, et al. Strategies to suppress the shuttle effect of redox mediators in lithium‐oxygen batteries. J Energy Chem. 2021;60:135‐149. doi:10.1016/j.jechem.2020.12.034

Dou Y, Xie Z, Wei Y, Peng Z, Zhou Z. Redox mediators for high‐performance lithium–oxygen batteries. Natl Sci Rev. 2022;9(4): [eLocator: nwac040]. doi:10.1093/nsr/nwac040

Mu H, Zhuge X, Ren G, et al. Dual functional mesoporous silica colloidal electrolyte for lithium‐oxygen batteries. Chem Eng J. 2023;455: [eLocator: 140761]. doi:10.1016/j.cej.2022.140761

Mu H, Luo K, Pang Y, et al. Mesoporous SiO₂ anode armour for lithium oxygen battery. Chem Engl J. 2023;475: [eLocator: 146489]. doi:10.1016/j.cej.2023.146489

Tułodziecki M, Leverick GM, Amanchukwu CV, et al. The role of iodide in the formation of lithium hydroxide in lithium–oxygen batteries. Energy Environ Sci. 2017;10(8):1828‐1842. doi:10.1039/C7EE00954B

Amanchukwu CV, Harding JR, Shao‐Horn Y, Hammond PT. Understanding the chemical stability of polymers for lithium–air batteries. Chem Mater. 2015;27(2):550‐561. doi:10.1021/cm5040003

Li X, Liu J, Zhang J, Wu S, Zhang D, Shui J. Interface engineering of NiCo₂O₄/BCN nanotube for performance enhancement of lithium‐oxygen battery. Chem Eng J. 2021;411: [eLocator: 128403]. doi:10.1016/j.cej.2021.128403

Zhu X, Wu Y, Wang Z, et al. Hierarchical architecture: a novel, facile and cost‐efficient strategy to boost electrochemical performance of Li‐O₂ battery cathodes. Chem Eng J. 2022;450: [eLocator: 138462]. doi:10.1016/j.cej.2022.138462

Zhang W, Tang S, Chen Z, et al. The controllable construction of nanochannel in two‐dimensional lamellar film for efficient oxygen reduction reaction and lithium‐oxygen batteries. Chem Eng J. 2022;430: [eLocator: 132489]. doi:10.1016/j.cej.2021.132489

Zhao C, Yan Z, Zhou B, et al. Identifying the role of Lewis‐base sites for the chemistry in lithium‐oxygen batteries. Angew Chem Int Ed. 2023;62(32): [eLocator: e202302746]. doi:10.1002/anie.202302746

Ko Y, Kim K, Yoo J, et al. Redox mediators for oxygen reduction reactions in lithium–oxygen batteries: governing kinetics and its implications. Energy Environ Sci. 2023;16(11):5525‐5533. doi:10.1039/D3EE02880A

Wang X‐X, Guan D‐H, Miao C‐L, Kong DC, Zheng LJ, Xu JJ. Metal–organic framework‐based mixed conductors achieve highly stable photo‐assisted solid‐state lithium–oxygen batteries. J Am Chem Soc. 2023;145(10):5718‐5729. doi:10.1021/jacs.2c11839

Cao X, Zhang Y, Lu C, et al. Synergistic dual atomic sites with localized electronic modulation enable high‐performance lithium–oxygen batteries. Chem Eng J. 2023;466: [eLocator: 143351]. doi:10.1016/j.cej.2023.143351

Zhang W, He G. Solid‐electrolyte interphase chemistries towards high‐performance aqueous zinc metal batteries. Angew Chem Int Ed. 2023;62(13): [eLocator: e202218466]. doi:10.1002/anie.202218466

Guo X, He G. Opportunities and challenges of zinc anodes in rechargeable aqueous batteries. J Mater Chem A. 2023;11(23):11987‐12001. doi:10.1039/D3TA01904G

Chen J, Chen G, Zhao S, et al. Robust biomass‐derived carbon frameworks as high‐performance anodes in potassium‐ion batteries. Small. 2023;19(7): [eLocator: 2206588]. doi:10.1002/smll.202206588

Wang Q, Lu T, Liu Y, et al. Li⁺ migration and transformation at the interface: a review for stable Li metal anode. Energy Storage Mater. 2022;55:782‐807. doi:10.1016/j.ensm.2022.12.043

Zhang Y, Zuo T‐T, Popovic J, et al. Towards better Li metal anodes: challenges and strategies. Mater Today. 2020;33:56‐74. doi:10.1016/j.mattod.2019.09.018

Zhang J, Zhao Y, Sun B, et al. A long‐life lithium‐oxygen battery via a molecular quenching/mediating mechanism. Sci Adv. 2022;8(3): [eLocator: eabm1899]. doi:10.1126/sciadv.abm1899

Sun Z, Tan Y, Zhan Y, et al. An organic redox mediator with a defense‐donor for lithium anode in lithium‐oxygen batteries. Small Methods. 2023;7(2): [eLocator: 2201289]. doi:10.1002/smtd.202201289

Sun Z, Lin X, Dou W, et al. Redox mediator with the function of intramolecularly disproportionating superoxide intermediate enabled high‐performance Li‐O₂ batteries. Adv Energy Mater. 2022;12(12): [eLocator: 2102764]. doi:10.1002/aenm.202102764

He J, Hong H, Feng Q, et al. Conjugated cobalt polyphthalocyanine with defective π‐π extended structure for enhanced rechargeable li‐oxygen batteries. Chem Eng J. 2022;444: [eLocator: 136544]. doi:10.1016/j.cej.2022.136544

Kim H, Lim YS, Kim JH. Highly porous structured Sr‐doped La₂NiO₄/NiO composite cathode with interconnected pores for lithium‐oxygen batteries. Chem Eng J. 2022;431: [eLocator: 134278].

Shen J, Wu H, Sun W, et al. In‐situ nitrogen‐doped hierarchical porous hollow carbon spheres anchored with iridium nanoparticles as efficient cathode catalysts for reversible lithium‐oxygen batteries. Chem Eng J. 2019;358:340‐350. doi:10.1016/j.cej.2018.10.038

Oh YJ, Kim JH, Park S‐K, Park JS, Lee JK, Kang YC. Highly efficient hierarchical multiroom‐structured molybdenum carbide/carbon composite microspheres grafted with nickel‐nanoparticle‐embedded nitrogen‐doped carbon nanotubes as air electrode for lithium‐oxygen batteries. Chem Eng J. 2018;351:886‐896. doi:10.1016/j.cej.2018.06.166

Oh YJ, Kim JH, Kang YC. Yolk–shell‐structured manganese oxide/nitride composite powders comprising cobalt‐nanoparticle‐embedded nitrogen‐doped carbon nanotubes as cathode catalysts for long‐life‐cycle lithium–oxygen batteries. Chem Eng J. 2019;373:86‐94.

Landa‐Medrano I, Olivares‐Marín M, Pinedo R, Ruiz de Larramendi I, Rojo T, Tonti D. Operando UV‐visible spectroscopy evidence of the reactions of iodide as redox mediator in Li‐O₂ batteries. Electrochem Commun. 2015;59:24‐27. doi:10.1016/j.elecom.2015.06.019

Kwak W‐J, Hirshberg D, Sharon D, et al. Understanding the behavior of Li‐oxygen cells containing LiI. J Mater Chem A. 2015;3(16):8855‐8864. doi:10.1039/C5TA01399B

Feng N, Mu X, Zhang X, He P, Zhou H. Intensive study on the catalytical behavior of N‐methylphenothiazine as a soluble mediator to oxidize the Li₂O₂ cathode of the Li‐O₂ battery. ACS Appl Mater Interfaces. 2017;9(4):3733‐3739. doi:10.1021/acsami.6b14889

Chen Y, Freunberger SA, Peng Z, Fontaine O, Bruce PG. Charging a Li‐O₂ battery using a redox mediator. Nat Chem. 2013;5(6):489‐494. doi:10.1038/nchem.1646

Bergner BJ, Schürmann A, Peppler K, Garsuch A, Janek J. TEMPO: a mobile catalyst for rechargeable Li‐O₂ batteries. J Am Chem Soc. 2014;136(42):15054‐15064. doi:10.1021/ja508400m

Pan Z, Brett DJ, He G, et al. Progress and perspectives of organosulfur for lithium–sulfur batteries. Adv Energy Mater. 2022;12(8): [eLocator: 2103483]. doi:10.1002/aenm.202103483

Chen R, Tang H, Dai Y, et al. Robust bioinspired MXene–hemicellulose composite films with excellent electrical conductivity for multifunctional electrode applications. ACS Nano. 2022;16(11):19124‐19132. doi:10.1021/acsnano.2c08163

Wang J, Huang G, Yan J‐M, et al. Hybrid solid electrolyte enabled dendrite‐free Li anodes for high‐performance quasi‐solid‐state lithium‐oxygen batteries. Natl Sci Rev. 2021;8(2): [eLocator: nwaa150]. doi:10.1093/nsr/nwaa150

Li Y, Zhang H, Zhang R, et al. Lithiophilic seeds and rigid arrays synergistic induced dendrite‐free and stable Li anode towards long‐life lithium‐oxygen batteries. J Energy Chem. 2022;73:268‐276. doi:10.1016/j.jechem.2022.06.021

Wu X, Wu H, Guan S, et al. A molecular sieve‐containing protective separator to suppress the shuttle effect of redox mediators in lithium‐oxygen batteries. Nano Res. 2023;16(7):9453‐9460. doi:10.1007/s12274‐023‐5663‐5

Chen Z‐F, Lin X, Xia H, et al. A functionalized membrane for lithium–oxygen batteries to suppress the shuttle effect of redox mediators. J Mater Chem A. 2019;7(23):14260‐14270. doi:10.1039/C9TA03133B

Li Z, Liu Y‐E, Weng S, et al. Oxygen‐permeable and moisture‐proof membrane for stable Li‐O₂/air batteries in humid working environment. Energy Storage Mater. 2023;58:94‐100. doi:10.1016/j.ensm.2023.03.018

Sun Y, Chen K, Zhang C, et al. A novel material for high‐performance Li‐O₂ battery separator: polyetherketone nanofiber membrane. Small. 2022;18(21): [eLocator: 2201470]. doi:10.1002/smll.202201470

Yuan D, Ji C, Zhuge X, et al. Organic‐inorganic interlayer enabling the stability of PVDF‐HFP modified Li metal for lithium‐oxygen batteries. Appl Surf Sci. 2023;613: [eLocator: 155863]. doi:10.1016/j.apsusc.2022.155863

Pogosova MA, Krasnikova IV, Sanin AO, et al. Complex investigation of water impact on Li‐ion conductivity of Li_1.3Al_0.3Ti_1.7(PO₄)₃—electrochemical, chemical, structural, and morphological aspects. Chem Mater. 2020;32(9):3723‐3732. doi:10.1021/acs.chemmater.9b04419

Chen X. ALD Processes and Applications to Nanostructured Electrochemical Energy Storage Deveces. College Park: University of Maryland; 2013.

Li Z, Wu X, Yu X, et al. Long‐life aqueous Zn–I₂ battery enabled by a low‐cost multifunctional zeolite membrane separator. Nano Lett. 2022;22(6):2538‐2546. doi:10.1021/acs.nanolett.2c00460

Johnson L, Li C, Liu Z, et al. The role of LiO₂ solubility in O₂ reduction in aprotic solvents and its consequences for Li‐O₂ batteries. Nat Chem. 2014;6(12):1091‐1099. doi:10.1038/nchem.2101

Gao X, Chen Y, Johnson L, Bruce PG. Promoting solution phase discharge in Li‐O₂ batteries containing weakly solvating electrolyte solutions. Nat Mater. 2016;15(8):882‐888. doi:10.1038/nmat4629

Monroe C, Newman J. The impact of elastic deformation on deposition kinetics at lithium/polymer interfaces. J Electrochem Soc. 2005;152(2): [eLocator: A396].

Liu S, Rao Y, Jang H, Tan P, Lu N. Strategies for body‐conformable electronics. Matter. 2022;5(4):1104‐1136. doi:10.1016/j.matt.2022.02.006

Ren Y, Zhou Y, Cao Y. Inhibit of lithium dendrite growth in solid composite electrolyte by phase‐field modeling. J Phys Chem C. 2020;124(23):12195‐12204.

Liu Q, Yu J, Guo W, et al. Boosting the Li|LAGP interfacial compatibility with trace nonflammable all‐fluorinated electrolyte: the role of solid electrolyte interphase. EcoMat. 2023;5(4): [eLocator: e12322]. doi:10.1002/eom2.12322

Safanama D, Chien SW, Jenssen K, Tan MY, Ding N, Fam DWH. Fibre‐reinforced multifunctional composite solid electrolytes for structural batteries. Compos Sci Technol. 2023;241: [eLocator: 110134]. doi:10.1016/j.compscitech.2023.110134

Chen W, Luo Z, Zhuge X, et al. Protecting lithium anode with ionic liquid modified poly (vinylidene fluoride) single ion conducting separators for iodide‐assisted lithium oxygen batteries. J Energy Storage. 2022;50: [eLocator: 104580]. doi:10.1016/j.est.2022.104580

Luo K, Zhu G, Zhao Y, et al. Enhanced cycling stability of Li–O₂ batteries by using a polyurethane/SiO₂/glass fiber nanocomposite separator. J Mater Chem A. 2018;6(17):7770‐7776. doi:10.1039/C8TA00879E