Artificial Interfacial Layers with Zwitterionic Ion Structure Improves Lithium Symmetric Battery Life and Inhibits Dendrite Growth

Abstract

Lithium (Li) metal’s exceptional low electrode potential and high specific capacity for next-gen energy storage devices make it a top contender. However, the unregulated and unpredictable proliferation of Li dendrites and the instability of interfaces during repeated Li plating and stripping cycles pose significant challenges to the widespread commercialization of Li metal anodes. We introduce the creation of a hydrogen bond network solid electrolyte interphase (SEI) film that integrates zwitterionic groups, designed to facilitate the stability and longevity of lithium metal batteries (LMBs). Here, we design a PVA/P(SBMA-MBA) hydrogen bond network film (PSM) as an artificial SEI, integrating zwitterions and polyvinyl alcohol (PVA) to synergistically regulate Li⁺ flux. The distinctive zwitterionic effect in the network amplifies the SEI film’s ionic conductivity to 1.14 × 10⁻⁴ S cm⁻¹ and attains an impressive Li⁺ ion transfer number of 0.84. In situ Raman spectroscopy reveals dynamic hydrogen bond reconfiguration under strain, endowing the SEI with self-adaptive mechanical robustness. These properties facilitate a homogeneous Li flux and exceptionally suppress dendritic growth. The advanced Li metal anode may endure over 1200 h at 1 mA cm⁻² current density and 1 mAh cm⁻² area capacity in a Li|Li symmetric battery. And in full cells paired with LiFePO₄ cathodes, 93.8% capacity retention is reached after 300 cycles at 1C. Consequently, this work provides a universal strategy for designing dynamic interphases through molecular dipole engineering, paving the way for safe and durable lithium metal batteries.

1. Introduction

Li metal anodes are pivotal components of next-generation high-energy-density battery technology, frequently utilized in handheld electronics and battery-powered vehicles due to their remarkable density of energy and low redox potential [1,2]. Nonetheless, LMB encounters significant challenges during charge/discharge cycling. A primary concern is the creation and uncontrolled proliferation of Li dendrites, which can create “dead Li” [3]. The possibility for these dendrites to pierce the barrier of separation and induce short circuits within it makes them a significant safety hazard [4,5]. Moreover, solid electrolyte interfaces (SEIs) can be especially fragile and uneven because Li metal is thermodynamically unstable in organic electrolyte solvents. These SEIs are susceptible to damage during plating and stripping cycles, exacerbating the depletion of the electrolyte and active [6,7,8,9]. Therefore, significant efforts are essential to overcome these challenges in LMBs to ensure their practical applicability in the new energy sector [10].

Recent studies have explored various methods to inhibit Li dendrite growth, including modified liquid electrolytes [11,12], solid electrolytes [13,14,15], modified separators [16,17,18], and artificial protective coatings [19,20,21,22], among other methods [23,24]. Modified liquid electrolytes can create stable SEI films, but their effectiveness often decreases over time as additives are consumed [25]. Solid electrolytes generally face issues like low ionic conductivity and poor contact at the electrode interface [26]. Although modified separators are important for reducing the shuttle effect of polysulfides in Li–sulfur batteries, they still struggle with low wettability and limited Li⁺ transport [17]. Among these methods, artificial protective coatings are particularly advantageous due to their ease of fabrication, compositional flexibility, and compatibility with other strategies [19]. Ideal electrode interfacial coatings should have: (1) strong chemical stability to minimize side reactions, (2) high mechanical strength to prevent Li dendrite formation, and (3) high ionic conductivity to ensure uniform Li⁺ ion flux [27,28]. It has been found that safe, environmentally friendly, and inexpensive polyvinyl alcohol (PVA) has a high content of hydroxyl polar groups. Additionally, the PVA can interact with lithium metal ions, facilitating the formation of a stable interfacial layer. This kind of complexation reduces the possibility of direct a reaction between hydroxyl group and Li metal, thus improving the interfacial stability of PVA and Li metal [29,30]. We found several relevant reports that also used PVA or modified PVA as interface layer for the Li metal anode, and summarized the coulombic efficiency of bare Cu and PVA-Cu under different current densities and capacitance (

Table 1. Recent research progress of PVA as artificial SEI films.

Electrode Structures	Current (mA cm−2)	Capacity (mAh cm−2)	Coulombic Efficiency (Cycle)	References
Bare Cu	0.5	1	105	[25]
	1	1	50
With PVA	0.5	1	115 (+10%)
	1	1	65 (+30%)
Cu foil	1	1	150	[29]
	3	1	75
Cu/PVA	1	1	210 (+40%)
	3	1	110 (+47%)
Bare Cu	0.5	1	80	[30]
SPVA-Cu	0.5	1	100 (+25%)

). Although PVA films have good thermal and chemical stability and high mechanical strength, their ionic conductivity is low [31,32]. Therefore, they are not sufficient to meet the requirements of an ideal artificial SEI film. Conversely, zwitterionic polymers enhance Li⁺ transport via electrostatic screening yet lack structural integrity [33,34,35]. Jin et al. [36] increase the number of Li⁺ migrations by constructing a zwitterionic polymer film containing sulfonic acid and phosphate at the lithium metal negative interface using its electrostatic adsorption and dissociation effects to regulate the Li⁺ transport. In addition, within the primary chains, zwitterionic polymers contain a number of cationic and anionic groups, which have opposing electrical charges, facilitating ionic transport channels [37]. Herein, we propose a hybrid design: a hydrogen-bonded PVA network reinforced with zwitterionic methacrylate. This architecture combines PVA’s mechanical robustness with zwitterion-enabled ion regulation, addressing the trade-off between ionic conductivity and interfacial stability.

In this work, we propose a straightforward method to synthesize hydrogen bond network coatings that exhibit remarkable flexibility and ionic conductivity for stable artificial protective films on Li metal anodes. We incorporated zwitterionic polymer chains of [2-(methacryloyloxy) ethyl] dimethyl-(3-sulfopropyl) ammonium hydroxide (SBMA) and N, N′-methylenebis-(acrylamide) (MBA) into a PVA matrix through free radical polymerization, resulting in a homogeneous hydrogen bond network structure for the PSM artificial SEI film. Glycerol (Gly) enhances the density of the physical hydrogen bond network through hydrogen bonding, contributing to its high mechanical strength. Additionally, the branched chains of SBMA form intermolecular hydrogen bonds that further reinforce the network, providing flexibility and resistance to Li dendrite puncture. The presence of polar functional groups (−OH) from PVA increases the affinity for Li⁺ ions, while the sulfonic acid and quaternary ammonium groups from SBMA facilitate ionic conductivity [38,39]. Consequently, the coating demonstrated ionic conductivity and ionic transfer numbers of 1.14 × 10⁻⁴ S cm⁻¹ and 0.84. In this manuscript, we choose to coat the PSM film on the Cu collector. By modifying the Cu collector, we can create a more stable and uniform interface and we are able to isolate the electrolyte–electrode interface from the Li metal surface, giving better control over the formation and stability of the artificial SEI, which may help alleviate problems such as dendrite growth. As a result, PSM-modified Cu foil pre-deposited with Li metal demonstrates stable cycling for 1200 h in a symmetric battery at a current density of 1 mA cm⁻², exceeding the performance of bare Cu, which lasts approximately 240 h, by more than five-fold. This highlights the substantial role of the artificial SEI film in enhancing the cycle stability of LMBs.

2. Materials and Methods

2.1. Materials

Polyvinyl alcohol (PVA, polymerization degree: 2600–2800, alcoholysis degree: 99.8–100%), [2-(methacryloyloxy) ethyl]dimethyl-(3-sulfopropyl)ammonium hydroxide (SBMA, purity ≥ 98%), N,N′-methylenebis-(acrylamide) (MBA, purity ≥ 98%), glycerol(Gly, purity ≥ 99%), and ammonium persulfate (APS, purity ≥ 98%) were acquired from Aladdin Co., Ltd. (Shanghai, China). 1-Methyl-2-pyrrolidinone (NMP), and Liquid electrolyte [1 M bis(trifluoromethane sulfonyl)imide lithium salt (LiTFSI) in 1,3-dioxolane (DOL)/1,2-dimethoxyethane (DME)] were acquired from Macklin Biochemical Co., Ltd. (Shanghai, China) Lithium foil, measuring 0.55 mm in thickness, Lithium iron phosphate (LiFePO₄, LFP) cathode materials, conductive carbon black, and poly(1,1-difluoroethylene) (PVDF) were acquired from MTI Kejing Group (Heifei, China). Deionized water (DI) was self-prepared in the laboratory. All compounds were utilized without purification.

2.2. Electrode Preparations

Initially, to dissolve 2 g of PVA powder, 18 mL of DI water was heated to 120 °C for 1 h, or until the particles were entirely dissolved. The solution reached room temperature and exhibited transparency. The PVA solution was subsequently combined with 1 mL of glycerol (Gly) and agitated at room temperature for 10 min. Gly acts as a plasticizer, interacting with the hydroxyl groups of PVA through hydrogen bonding, reducing intermolecular interactions within the polymer chains and enhancing the flexibility of the resulting material. In a separate process, 1.3967 g SBMA and 0.1542 g MBA (5:1 m/m) were used as reaction monomers, and APS (2 wt% relative to the monomers) was added as a free-radical initiator to 3 mL DI water solvent. SBMA, containing a zwitterionic structure, serves as the primary functional monomer, while MBA, a bifunctional crosslinking agent, establishes covalent bridges between polymer chains. The PVA/Gly solution was then mixed with the SBMA/MBA solution to acquire a uniform PSM solution. This process integrates PVA into the copolymer matrix via hydrogen bonding and physical entanglement, stabilizing the resulting film structure. The uncoated Cu foil was affixed to glass substrate and subsequently used ethanol and DI water for rinsing more than three times. The PSM-Cu was produced by putting the PSM solution onto a bare copper substrate using a spatula, achieving a thickness of 100 μm. A protective coating was formed on the coated Cu by drying it at 60 °C for 8 h, then punched into 14 mm pieces. Additional vacuum treatment was conducted at 80 °C overnight after film formation and before battery assembly. Without SBMA/MBA solution, PVA-Cu was prepared like the control group.

2.3. Characterization

The morphologies and chemical compositions of the samples were examined using a scanning electron microscope (SEM, Hitachi S4800, Tokyo, Japan) in conjunction with an energy-dispersive X-ray spectrometer (EDS). An atomic force microscope (AFM, Agilent 5100, Santa Clara, CA, USA) instrument was used to illustrate the three-dimensional morphology of the electrode samples. The wettability tests were assessed using the DSA100 contact angle measurement device from Kris Scientific Instruments (Shanghai, China) Co., LTD. Raman spectra were acquired with a Renishaw in Via micro-Raman spectrometer from the Gloucestershire, London, UK, employing a 532 nm laser and a wavenumber range of 100–3500 cm⁻¹. The FTIR spectra were acquired using a spectrometer (INVENIO, Bruker, Germany) in the wavenumber range of 1000–4000 cm⁻¹ with a resolution of 2 cm⁻¹. The NMR spectra were obtained on an AVANCE NEO 600 MHz instrument with D₂O or DMSO-d6 as the solvent. Thermogravimetric Analysis (TGA) was conducted on a 209 F3 Netzsch analyzer. Samples weighing ~5 mg were heated from room temperature to 700 °C at a rate of 10 °C min⁻¹ under a nitrogen atmosphere. Temperature-resolved dynamic mechanical analysis (DMA) was performed employing a dynamic mechanical analyzer (TA Instruments Q800 from the Newcastle, DE, USA) at a frequency of 1 Hz, with a temperature range from −20 °C to 110 °C at a ramp rate of 5 °C min⁻¹. Tensile tests were performed using a SANS CMT6503 universal testing machine from Metz Industrial Systems (Shenzhen Branch, Shenzhen, China) Co., Ltd. The specimens were cut into dumbbell shapes with a gauge length of 5 mm and a width of 2 mm. Tests were conducted at room temperature with strain rates ranging from 10 mm min⁻¹ to 50 mm min⁻¹ to analyze the strain-rate-dependent mechanical behavior. At least three specimens were tested for each condition to ensure statistical relevance. We have clarified that adhesion was evaluated using a standard lap shear test according to ASTM D1002 [40]. Rectangular metal strips (20 mm × 5 mm) of copper, aluminum, steel, and wood were used as test substrates. PSM solution was applied as the adhesive layer and cured at 60 °C for 8 h. The shear strength was measured using a SANS CMT6503 (Metz Industrial Systems Co., Ltd., Shenzhen Branch, Shenzhen, China) universal testing machine, and three replicates were tested for each substrate to ensure repeatability and reliability. X-ray photoelectron spectroscopy (XPS) was conducted using an ESCALAB 250Xi (Shanghai Yuzhong Industrial Co., Ltd., Shanghai, China) instrument with Al K α radiation to assess the chemical composition. We used a PARSTAT P3000A (Ametek Group, San Diego, CA, USA) electrochemical machine to measure the EIS. The EIS data were analyzed and aligned with the proposed equivalent circuit. R₀ denotes the resistance of the electrolyte. The RSEI and constant phase element (C1) represent the resistance and capacitance of the interphase at the electrode surfaces, with RSEI corresponding to the semicircle observed at high frequencies. Rct and C2 denote the charge-transfer resistance and its corresponding capacitance, respectively. Ws denotes the Warburg impedance associated with the collective influence of Li⁺ ions diffusing through the electrode–electrolyte interfaces.

2.4. Crosslinking Density

The crosslinking density of the film was calculated using the Flory–Rehner equation [41].

$$V_e=-{\displaystyle \frac{\mathit{ln}\left(1-V_2\right)+V_2+x{V}_2^2}{V_1(V_2^{1/3}-V_2/2)}}$$

(1)

where $V_e$ is the crosslinking density of the hydrogel, and $x$ is the polymer–solvent interaction parameter ($x\approx 0.49$). $V_1$ is the molar volume of water. $V_2$ is the volume fraction of film in water:

$$V_2={\displaystyle \frac{m_d{\rho }_s}{\left(m_s-m_d\right){\rho }_p+m_d{\rho }_s}}$$

(2)

where $m_d$ is the mass of the sample before swelling, $m_s$ is the mass of the sample at equilibrium swelling, and ${\rho }_s$ and ${\rho }_p$ are the densities of the solvent and elastomer, respectively.

2.5. Determination of Activation Volume and Energy

The Eyring model was used to determine the activation volume and energy [42].

$${\sigma }_y\approx {\displaystyle \frac{2KT}{V_a}}\mathit{ln}{\epsilon }^{.}+{\displaystyle \frac{2E_a}{V_a}}$$

(3)

where ${\sigma }_y$ indicates the yield stress, $E_a$ represents the activation energy, ${\epsilon }^{.}$ indicates the strain rate, and $V_a$ denotes the activation volume.

2.6. Electrochemical Measurements

All electrochemical tests were conducted utilizing CR2025-type coin cells (MTI Kejing Group, Hefei, China) built within an argon-filled glove box, maintaining oxygen and water content below 0.1 ppm. The liquid electrolyte consisted of 1 M LiTFSI in a 1:1 volumetric ratio of DOL to DME, with LiNO₃ added at 2 wt%. The separator was a Celgard 2500 membrane (MTI Kejing Group, Hefei, China), and the cathode and anode were uncoated Cu foils, Cu foils coated with PVA or PSM, and Li foils to test plated/stripped Li. Charge to 0.5 V and discharge current density to 0.5 mA cm⁻²/1 mA cm⁻² were set for the coulombic efficiency test. Before assembling the Li|Li symmetric cell, our electrodes were prepared by pre-depositing Li metal electrochemically onto 8 μm-thick Cu foil substrates, which had been coated with either a PSM or PVA layer. Li metal was electrodeposited at a capacity of 5 mAh cm⁻² onto bare Cu foil, PVA-Cu, or PSM-Cu substrates. The electrolyte was a mixture of 1.0 M LiPF₆ in ethylene carbonate (EC)/dimethyl carbonate (DEC) with a volume ratio of 1:1. A slurry, the concentration of which could be adjusted using NMP, served as the electrolyte, PVDF binder made up 10% of the mixture, acetylene black was 10%, and LiFePO₄ (LFP) particles made up 80%. The LFP cathode was produced by applying the slurry to Al foils and allowing them to dry for 24 h at 60 °C. The LFP loading density was around 1.20 mg cm⁻². Li|LFP complete cells were built using LiFePO₄ cathodes and 1 mA h cm⁻² Li pre-deposited on uncoated Cu foil or Cu foil coated with PSM electrodes anodes. The resultant N/P ratio in the complete cell was controlled to be 4.9.

The PARSTAT P3000A electrochemical workstation (Ametek Group, San Diego, CA, USA) was used to record coin cell EIS throughout a frequency band (0.01 to 10⁵ Hz) with an alternating current signal (10 mV).

In order to conduct the Li⁺ ion transference number test, a pre-deposition procedure involved coating Cu foil with an artificial SEI film and then plating the electrode with Li metal. The Li @ Cu, Li@PVA-Cu, and Li@PSM-Cu electrodes were separated from one another using a separator. According to this formula, we may find the Li⁺ ion transference number [36].

$$t_{{Li}^{+}}={\displaystyle \frac{I_s(∆V-I_0{R}_0)}{I_0(∆V-I_s{R}_s)}}$$

(4)

where $I_0$ and $I_s$ denote the initial and final currents, respectively; $R_0$ and $R_s$ signify the original and final resistances; and $∆V$ represents the polarization potential (10 mV).

Electrochemical impedance spectroscopy (EIS) measured the artificial solid electrolyte interphase layer’s lithium ionic conductivity. The thin polymer sheets (PSM, PVA) saturated with liquid electrolyte were positioned at the junction of two stainless substrates. The values of the ionic conductivity $\sigma$ were calculated by applying the following formula [36]:

$$\sigma ={\displaystyle \frac{I}{SR}}$$

(5)

where $I$ denotes the thickness of the polymer film (PSM, PVA), $S$ signifies the electrode surface area, and $R$ indicates the resistance measured from EIS.

3. Results and Discussion

3.1. Electrode Preparation and Morphology

In our research, we integrated sulfobetaine zwitterion as an accessory group into a hydrogen bond network structure of PVA/P(SBMA-MBA)-Gly (PSM) coatings through initiator-induced radical polymerization. As shown in Figure 1a, the PSM is coated onto the Cu current collector rather than directly onto the Li metal surface to enhance the stability and uniformity of the SEI layer during cycling, as the Cu current collector provides a more robust substrate, facilitating the isolation of the SEI formation from potential issues directly associated with Li metal, such as dendrite growth or degradation of the Li metal surface. Figure 1b shows the construction of the coating’s network: rigid PVA clusters form the primary chain, while the monomers of SBMA and MBA undergo free-radical polymerization to establish a secondary, softer P(SBMA-MBA) chain (Figure S1). Additionally, the hydroxyl groups of Gly interact with the functional groups of both the PVA and P(SBMA-MBA) molecular chains via hydrogen bonding, thereby enhancing the physical hydrogen bond network of the coating. This unique hydrogen bond network structure is characterized by a high density of polar functional groups (-OH) and a concentration of both positive and anionic groups. Due to the porous hydrogen-bonded network and the presence of polar functional groups in the PSM film, Li⁺ ions can uniformly migrate through the film and are deposited beneath the protective layer, forming a dense, conformal lithium layer between the Cu current collector and the PSM coating (Figure 1c). Meanwhile, it preserves overall charge neutrality, which aids in preventing ion polarization at the Li anode/electrolyte interface, encouraging homogeneous Li⁺ deposition, and successfully suppresses dendrite formation. Figure S2 displays digital photographs of uncoated Cu, PVA-Cu, and PSM-Cu. Post-modification, the surface color of the Cu foil darkened and exhibited increased clarity and transparency, signifying that both PVA and PSM possess excellent film-forming capabilities. The SEM pictures of PSM-Cu and PVA-Cu (Figure 1d and Figure S3) reveal that the PSM film is more uniformly flat. Cross-sectional SEM images confirm that both PSM and PVA coatings are densely applied to the Cu foils with uniform thicknesses. The PSM film also displays a network-like morphology (Figure 1e and Figure S4). AFM measurements showed that the roughness of the PVA film is 300 nm (Figure S5), while the PSM film has a roughness of 100 nm (Figure 1f). Both films show uniform fluctuations without noticeable lumps, consistent with the SEM observations.

3.2. Structural Characterization

The EDS mapping images in Figure 2a and Figure S6 confirm the homogeneous elemental distribution within the coating layers. For the PVA-Cu, C, and O are uniformly dispersed across the surface, indicating a continuous and compact film. In contrast, the PSM-Cu exhibits an even distribution of C, O, N, and S, verifying the successful incorporation of zwitterionic SBMA units and the uniformity of the hybrid polymer network. Moreover, Figure S7 provides the EDS point analysis spectra and the corresponding quantitative data. The atomic percentages of C, O, N, and S are in good agreement with the theoretical composition of the PSM network. No signals from impurities or inhomogeneous domains were observed, which demonstrates the chemical purity and structural uniformity of the coating layer. ¹ H-NMR spectroscopy was performed (Figure S8), which clearly demonstrated the disappearance of C=C double bonds (5.5–6.5 ppm) from SBMA and MBA, alongside the emergence of new peaks (0.8–2.5 ppm) corresponding to the polymer backbone. FTIR spectroscopy -N⁺(CH₃)₂ and -SO₃ groups are shown by distinctive peaks at 1431 cm⁻¹ and 1083 cm⁻¹, respectively, confirming the successful doping of SBMA (Figure 2b) [43]. High-resolution FTIR spectra (Figure S9) additionally verified the disappearance of olefin stretching (1620–1680 cm⁻¹), collectively substantiating the successful polymerization and crosslinking under the reported conditions. Furthermore, a broad peak with hydroxyl group (O-H) at about 3300 cm⁻¹ stretching vibration was observed in all the spectra. The peaks of stretching vibration were 3305 cm⁻¹ and 3316 cm⁻¹ for PVA and SBMA, respectively. After synthesis, it was found that the O-H absorption peak of PSM film was redshifted to 3272 cm⁻¹, and the peak width became wider (Figure S10) [44]. This indicates that the O-H bonds interact with other groups (e.g., -SO₃ or C=O) through hydrogen bonding, resulting in the weakening of the bond strength and lowering of the vibrational frequency. The crosslink density of PVA and PSM was calculated using the well-known Flory–Rehner equation, which is based on equilibrium swelling experiments with deionized water [41]. Through a decrease in swelling behavior and an increase in apparent crosslink density, PSM exhibits a noncovalently crosslinked molecular structure (

Table A1. Polymer density, swelling ratio, and crosslinking density of PVA and PSM film ¹.

Samples	Polymer Density (g/cm3)	Swelling Ratio (%)	Crosslinking Density (10−4 mol/cm3)
PVA	1.25 ± 0.01	210.61 ± 0.17	2.71 ± 0.11
PSM	1.29 ± 0.03	116.70 ± 0.09	9.65 ± 0.02

¹ The error in the table is standard deviation (3 specimens for each sample).

). Raman spectroscopy detected the deformation vibration of S=O at 1039 cm⁻¹ and -N⁺(CH₃)₂ at 2978 cm⁻¹, indicating electrostatic interactions, along with O-H stretching from 3422 to 3467 cm⁻¹ (Figure 2c). These results confirm the presence of hydrogen bonding and electrostatic interactions in PSM, indicating the successful synthesis of hydrogen bond network coatings.

Additionally, for Li to be deposited into the lower layers of SEI film instead of the surface, the film must be electrically insulating [45]. Given this requirement, the excellent electronic insulating properties of PVA align well with this characteristic [46]. We investigated the top-view and cross-sectional SEM images of 5 mAh cm⁻² Li deposited on bare Cu electrodes and PSM-Cu electrodes (Figure S11), confirming that Li is uniformly deposited below the PSM film, paving the way for subsequent symmetric cell and full cell testing. Figure S11a illustrates that the Li metal on the Cu electrode displays inhomogeneous deposition, characterized by a loose structure consisting of numerous needle-like Li dendrites (Figure S11c). This morphology can lead to issues such as internal short-circuiting and reduced battery lifetime. In contrast, the Cu electrode with the PSM coating displays a flat and uniform surface after massive Li deposition (Figure S11b). Notably, no hump structures were observed after deposition of Li metal under the PSM film (Figure S11d), suggesting that the advantages of the PSM coating synergistically enhance dense and homogeneous Li deposition without significant morphological changes. XPS analyses (Figure 2d–i) demonstrated the alterations in the surface chemical states of PSM-Cu prior to Li plating and Li @ PSM-Cu subsequent to Li plating. C-F groups in the C-1s spectrogram (Figure 2d,g) and Li-F and C-F groups in the F-1s spectrogram (Figure S12) suggest anionic disintegration of LiTFSI and SEI film LiF production [47,48]. The C-O signal does not shift at 286.5 eV, so it can be inferred that the hydroxyl group of PVA forms a stable interface layer in the Li metal environment. The existence of cationic quaternary ammonium groups and the C-N bonding before and after Li deposition is evident in the N-1s energy spectrum (Figure 2e,h), indicating that the zwitterionic polymer remained relatively stable and did not decompose during battery cycling [36]. The chemical shift of the sulfonic acid groups to the high-energy region (Figure 2f,i) can be ascribed to the sulfonyl group in LiTFSI or its breakdown products, or it may result from the coulombic interactions between the underlying Li⁺ ions and the sulfonic acid groups [49].

As shown in Figure 3a, the result of TGA reveals that the PSM film possesses good thermal stability at 118.6 °C. Meanwhile, DMA measured PSM’s energy modulus of storage (G′), modulus of loss (G″), and loss factor at various temperatures (Figure 3b). The findings demonstrate that the modulus of PSM decreased with increasing temperature. This behavior is typical of physical crosslinking, which is caused by hydrogen bonding interactions [41]. According to the previous literature, the minimum G′ required to resist Li dendrite growth puncture should exceed 0.1 MPa [50,51]. Additionally, the prepared PSM films demonstrate a higher G′ compared to several other reported SEI protective films [27,52]. This further confirms that the obtained PSM films do fulfill the strength requirements. Moreover, due to the mechanical integrity of the hydrogen bond network structure, the PSM maintained a modulus not lower than the minimum G′ (0.1 MPa) even at temperature exceeding 100 °C, which is an advantageous feature for high-temperature battery operation. Consequently, the PSM film is capable of meeting the safety requirements for the battery operation at room temperature. Tensile stress–strain curves at different strain rates demonstrate the good extensibility of PSM films while maintaining high mechanical strength (Figure 3c). At 5% strain, a clear yield point indicates that the PSM film is in a glassy state. Increasing the strain rate raises tensile strength and yield stress while decreasing maximum strain. According to the Eyring equation for non-covalent bond separation, the relationship between yield stress and the square root of tension rate (Figure 3d) is nearly straight [53]. The fitted activation volume of 7.57 nm³ represents the size of the polymer fragment that produces yielding. The calculated activation energy of 13.97 KJ mol⁻¹ is the energy required for mobile segments to overcome external forces. Post-yield point, the prolonged plateau noted during plastic deformation mostly results from the sliding extension of the amorphous matrix constrained by hydrogen bonding clusters. Strain hardening transpires with stresses over 125%, resulting from the progressive unraveling of hydrogen-bonded clusters until their destruction [41,54]. The PSM-Cu also demonstrated excellent wettability with the electrolyte (Figure S13), suggesting that the zwitterionic modification improves the electrolyte’s interaction with the Cu surface. To assess the adhesion strength of PSM, lap-shear tests were conducted on metallic copper and other materials (Figure 3e and Figure S14). These results show that PSM adheres well to various metals and resists dislodgement, proposing that, with proper tuning of network elasticity and adhesion, the PSM strategy could be extended to other anodes facing volumetric expansion challenges.

The hydroxyl groups in the molecular structure of PVA have a high affinity for Li⁺ and form stable coordination complexes. This interaction not only stabilizes the Li⁺ in the electrolyte but also prevents PVA from dissolving into the solution [55]. To evaluate the chemical durability of the electrolyte with PVA and PSM, a 7-day immersion test showed that the PVA and PSM film’s quality remained stable (Figure 3f and Figure S15). Furthermore, the NMR spectra (¹H and ¹⁹F) of the electrolyte soaked for 7 days (Figure S16a,b), and no signal of dissolved PVA fragments or degradation products were observed. This indicates that PVA remains stable and does not dissolve into the electrolyte. FTIR spectroscopy further demonstrated that the PSM films before and after soaking the electrolyte were nearly identical (Figure S17). In particular, the NMR spectra (¹H and ¹⁹F) of the soaked electrolyte are presented in Figure S18a,b. From the analysis of the ¹H spectrum, the signals observed in the spectrum are attributed to the electrolyte solvents DME and DOL. Importantly, no characteristic signals of glycerol are observed in the range of δ ≈ 3.3 ppm–3.8 ppm. This absence confirms that glycerol stably remains in the PSM matrix without dissolving into the electrolyte. Furthermore, the absence of additional peaks in the δ ≈ 3.0 ppm–δ ≈ 5.0 ppm range, where potential unreacted monomers or side products might appear, provides strong evidence that the polymerization reaction is complete, and no monomer residues remain. In the ^19F spectrum, only a single sharp peak signal of LiTFSI at δ ≈ −78 ppm is observed. This signal is clear and stable, with no additional peaks corresponding to potential decomposition products. This indicates that LiTFSI in the electrolyte has not degraded or chemically reacted with the PSM, further supporting the stability of the system. The results regarding thermal stability, mechanical properties, adhesive force, and chemical stability collectively demonstrate that the PSM film has successfully achieved the desired characteristics of an artificial SEI. Therefore, it is expected that the PSM film will enhance the electrochemical performance of LMBs.

3.3. Electrochemical Performance and Coulombic Efficiency

According to the comprehensive EIS test seen in Figure 4a, the impedance of PVA and PSM is 35.1Ω and 9.2Ω, respectively. Combined with the thickness of the films, the ionic conductivity of PVA and PSM is calculated, and the results are 0.42 × 10⁻⁴ S cm⁻¹ and 1.14 × 10⁻⁴ S cm⁻¹, respectively. The ionic conductivity of PSM is significantly higher than that of PVA, surpassing that of the most advanced polymers’ SEIs such as polyurea (0.85 × 10⁻⁴ S cm⁻¹) [9] due to the large number of ionic dipole interactions that improve the ionic conductivity. A significant Li⁺ transference number of 0.84 was achieved in the Li@PSM-Cu electrode (Figure 4b), markedly surpassing previously reported values for polymer-based interphases such as 0.732 for Cu@Zn-MOF/PVA electrodes [29], as well as the 0.59 for Li@PVA-Cu and 0.45 for Li@Cu electrodes obtained in this work (Figure 4c,d). This highlights the superior ion transport regulation and selectivity enabled by the zwitterion-reinforced hydrogen-bonded PSM network. The minimizing interfacial layer thickness is crucial for preserving high energy density and rate performance in practical cells [56]. In order to investigate the effect of PSM film thickness on cell performance, we performed cycling tests on films with different thicknesses (scraper thickness: 50 μm, 100 μm, 150 μm, and 200 μm). As shown in Figure S19, the cycling stability varies significantly with film thickness, and the PSM-Cu has the best electrochemical performance when the scratch thickness is 100 μm. PSM films that are too thin may have difficulty in covering the surface of the Cu foil completely and have insufficient mechanical strength, leading to localized inhomogeneous deposition and decreasing the interfacial protection effect. When the film is too thick, it may increase the interfacial impedance, affecting the Li⁺ transport efficiency and thus reducing the electrochemical performance of the battery. To further verify the modulation of the plating/stripping process by the PSM coating, three different types of counter electrodes were used to prepare Li|Cu cells: uncoated Cu, PVA-Cu, and PSM-Cu, and subjected to constant-current charge/discharge tests. The unprotected electrode’s coulombic efficiency (CE) varied at 0.5 mA cm⁻², reaching only 70.98% after 130 cycles (Figure 4e). Reasons given for this deterioration included active Li reacting with the electrolyte, the electrolyte being consumed, which builds up “dead Li”. A 20-cycle improvement in cycle life was the only effect of coating the Cu collector using PVA. In contrast, the PSM coating effectively increased the CE to 97.7% at 300 cycles, demonstrating superior performance. Upon increasing the current density to 1 mA cm⁻², the PSM-Cu electrode still maintained similar results compared to the PVA-Cu and bare Cu electrodes (Figure S20). This suggests that PSM can effectively protect the anode, presumably thanks to the high Li affinity and excellent Li⁺ ion transport properties of the sulfonic acid groups in PSM. Additionally, we investigated the performance of PSM-Cu|Li half-cells at 0.2, 0.5, 1.0, 2.0, and 4.0 mA cm⁻² current densities. (Figure 4f). These results indicate that PSM-Cu has a lower overpotential than bare Cu at the same current density. Furthermore, with increasing current density, the overpotential gap widens between the two cells. This advantage suggests that the stable PSM-Cu electrode ensures low overpotential and uniform Li deposition [25].

3.4. SEM Analysis

After 5, 50, and 100 cycles, we disassembled the coin cell to examine the half-cell’s Li deposition condition following over time cycling at 1 mA cm⁻² current density and obtained electrode morphology maps at different cycling times (Figure 5). After five cycles, the SEM picture of the Cu electrode shows a surface that is rough and has fine dendrites surrounding the pits. On the PVA-Cu electrode, a small amount of Li deposition particles appeared on the surface, likely due to the non-uniformity of Li deposition. Surprisingly, following the implementation of the PSM film, the electrode surface was smooth and flat, showing no Li dendrite particles. The consistent deposition direction of the zwitterionic ions and the rigidity of the PSM layer are responsible for this shape. Notably, the Li@PSM-Cu electrode (high-resolution image in Figure S21) presents a remarkably smooth and dense surface without visible dendritic structures. The lithium appears to deposit in a compact and conformal manner, suggesting uniform ion flux and effective surface regulation. This improvement can be attributed to the combination of the rigid hydrogen-bonded framework and the directional alignment of zwitterionic SBMA segments, which promote homogeneous Li⁺ ion transport and suppress tip-enhanced growth that leads to dendrites. After 50 cycles, substantial pits formed on the Cu electrode, with sparse Li metal buildup observed within these pits. Aggregated Li dendrites were partially visible on the PVA-Cu electrode. Conversely, a thick and smooth exterior was maintained by the PSM-Cu, with no protruding Li dendrites, indicating excellent stability at the interface and electrode. A considerable buildup of Li metal on the exposed copper foil, known as “dead Li”, is visible in the scanning electron micrograph, which is the unprotected Cu electrode following 100 cycles; this, in turn, lowers the half-cell efficiency (CE). Meanwhile, after 100 cycles, the PVA film presents local defects and raised surfaces, which is due to repeated stripping and deposition of Li⁺ during the battery charging and discharging process, resulting in repeated expansion and contraction of the PVA film. After a long cycle, the structural integrity of the PVA layer decreases and local defects appear, which allows electrons to transfer to the PVA surface through a local conductive path and allows Li⁺ to be reduced to Li metal on the PVA surface and accumulate on the PVA surface. In sharp contrast, the PSM-Cu electrode’s outer did not exhibit any discernible dendrites. These findings indicate that PSM is highly effective in regulating Li deposition, aiding in the development of a battery that exhibits increased cycle life and great coulombic efficiency.

3.5. Symmetrical Battery

The interfacial stability provided by PSM is further demonstrated by contrasting the electrode cycle life and long-term voltage polarization in Li|Li symmetric cells during plating and stripping. The symmetric cell composed of Li@PSM-Cu has a higher stability polarization voltage of 15 mV (Figure S22a) and can be cycled for 1200 h under 1 mA cm⁻² test conditions (Figure 6a), signifying one of the premier performances among these AILs (

Table 2. A brief summary of Li metal anodes protected by artificial interphase films.

Electrode Structures	Current (mA cm−2)	Capacity (mAh cm−2)	Time (h)	Polarization (mV)	References
CPLi-4	0.5	2	1000	25	[22]
PVA/LiF	1 2	1 1	800 600	40 55	[25]
APM	1 1	0.5 1	1500 1000	100 300	[27]
Zn-MOF/PVA	1 3	1 1	580 300	51.7 42.1	[29]
PDMAPS +PMPC	1 3	1 1	1400 280	14 25	[35]
APL	2	1	200	175	[57]
NLI	1 1	1 3	1000 350	50 100	[58]
PU	1 3	1 3	1100 800	26 30.4	[59]
LiF	1	1	500	70	[60]
TPU	1	1	800	50	[61]
PSM	1 2	1 1	1200 800	15 28	This work

). Conversely, the Li@Cu|Li@Cu symmetric cell exhibits instability during cycling characterized by a gradual increase in polarization voltage and significant voltage hysteresis after only 380 h (190 cycles) at a higher polarization voltage of 30 mV (Figure S22b). This substantial increase in overpotential suggests a highly resistive interfacial coating of developing Li dendrites and “dead Li”, which further accelerates electrolyte depletion. While the PSM-modified electrode exhibits excellent electrochemical performance over extended cycling, potential degradation mechanisms during long-term operation should be considered. Repeated Li plating and stripping can cause local volumetric fluctuations at the interface, potentially leading to mechanical fatigue or the formation of microcracks within the PSM film. Such defects may eventually compromise the integrity of the protective layer and allow localized Li dendrite nucleation. After electrode cycling, scanning electron microscopy examined surface morphology. As shown in Figure S23a, after 150 cycles (300 h), loose flower-shaped Li dendrites and a significant quantity of “dead Li” were discovered on the exposed Cu electrode. In contrast, the PSM-Cu electrode surface remained smooth and dense, without any dendrites (Figure S23b). Remarkably, the Li@PSM-Cu|Li@PSM-Cu cell may be cycled for 800 h at less than 30 mV when the level of current is raised to 2 mA cm⁻² (Figure 6b). To evaluate the robustness of the PSM interfacial layer under higher current densities, we supplemented the symmetric cell cycle test data at 5 mA/cm² (Figure S24). The Li@PSM-Cu symmetric cell maintained stable voltage profiles and low hysteresis over extended cycling, indicating excellent interfacial stability and dendrite suppression capability even under high Li⁺ flux. The Li@Cu|Li@Cu symmetric cell exhibited significant voltage hysteresis after 60 h. Interfacial resistance of Li@Cu|Li@Cu symmetry cells grow substantially beyond the 50th and 100th cycles, as shown by the Nyquist plots of the symmetry cells (Figure 6c), indicating that the interfacial phase of the spontaneous SEI is unstable and forms “dead Li” easily. Conversely, the interfacial impedance of the Li@PSM-Cu|Li@PSM-Cu cell shows minimal change, confirming the presence of a stable low-resistance interfacial phase facilitated by the PSM film (Figure 6d). These results indicate that the hydrogen bond network coating effectively combines the screening function of PSM with the mechanical properties of PVA, promoting uniform deposition of Li⁺ ions, reducing volume expansion, and inhibiting Li dendrite growth, thereby contributing to the construction of a stable Li metal anode.

3.6. Asymmetric Full Battery

LFP was employed as a model cathode material to evaluate the full-cell performance of the PSM-protected Li anode. LFP was chosen due to its stable voltage plateau and commercial relevance [62]. We employed LFP as the cathode and pre-deposited Li (1 mAh cm⁻²) as the anode on both Cu and PSM-Cu collectors to construct a whole cell and validate the improvement in cycle capacity of the Li metal anode facilitated by the PSM layer. As shown in Figure 7a, the Li@Cu|LFP cell has an original strength of 135.1 mAh g⁻¹, which significantly diminished the end of 120 cycles with a 1C speed (1C = 170 mA g⁻¹), leading to gradual cell failure. This reduction can be attributable to the proliferation of Li dendrites. Conversely, the 93.8% capacity retention rate was achieved by the Li@PSM-Cu|LFP complete cell, which had a starting ability to discharge 139.3 mA h g⁻¹ and dropped to 133.5 mA h g⁻¹ after 300 constant discharges. This indicates a notable enhancement in cycling performance. The dQ/dV curves (Figure 7b,c) indicate that, as cycles increase, the peak shift in the Li@Cu|LFP whole cell reflects plateau potential variations, an increase in resistance to the embedding and disengagement of Li⁺ ions, and heightened polarization impedance. Conversely, the decay of the peak suggests a reduction in capacity per unit voltage and a shortening of the plateau region, indicating a loss of active material. Notably, as illustrated in Figure 7d, the polarization rate of the charge/discharge curve for the Li@PSM-Cu|LFP complete cell is inferior to that of the Li@Cu|LFP cell. The result indicates that the PSM film optimizes the interface during complete cell cycling, further promoting uniform deposition of Li⁺ ions. Additionally, at current densities of 0.2, 0.5, 1, 2, and 5C, the discharge capacity of the Li@PSM-Cu|LFP cell consistently exceeds that of the Li@Cu|LFP cell, regardless of the performance at different current densities (Figure 7e). This implies that the PSM film significantly enhances the utilization of LFP.

4. Conclusions

In this work, we developed a robust hydrogen-bonded network artificial SEI film (PSM) by incorporating zwitterionic P(SBMA-MBA) chains into a PVA matrix. This dual-network structure integrates physical entanglement and chemical crosslinking, resulting in excellent mechanical strength, thermal stability, and chemical durability. The presence of sulfonic acid and quaternary ammonium groups enhances Li⁺ ion conductivity (1.14 × 10⁻⁴ S cm⁻¹) and transference number (0.84), while the compact network structure promotes uniform Li⁺ deposition and effectively suppresses dendrite formation. The PSM-coated Cu substrate enables uniform Li plating beneath the protective layer, improving the interfacial stability between the anode and electrolyte. As a result, symmetric cells based on Li@PSM-Cu exhibit extended cycling life over 1200 h with low polarization at 1 mA cm⁻², significantly outperforming bare Cu. Additionally, full cells using Li@PSM-Cu paired with a LiFePO₄ cathode demonstrate excellent cycling stability, retaining 93.8% of the initial capacity (133.5 mAh g⁻¹) after 300 cycles. The PSM film also shows high electrolyte compatibility, strong adhesion, and resistance to chemical degradation, further confirming its reliability as a protective layer. Overall, this study offers a simple yet effective strategy for constructing stable interfacial films on Li metal anodes, which could play a vital role in advancing the practical deployment of high-energy-density lithium metal batteries.