Suppressing concentration polarization in lithium battery composite polymer electrolytes via piezo-assisted electromechanical coupling effect

Abstract

The inherent limitation of Li⁺ transport and resulting severe concentration polarization in solid-state electrolytes have hindered the practical application of lithium metal batteries. Addressing this challenge, here we create a piezoelectric polymeric composite electrolyte based on poly(vinylidene fluoride) blended with 0.5Ba(Zr_0.2Ti_0.8)O₃-0.5(Ba_0.7Ca_0.3)TiO₃, which exploits volume fluctuations of lithium metal negative electrodes during cycling to activate a piezo-assisted electromechanical coupling effect. The resulting gradient piezo-field within the prepared electrolyte selectively accelerates Li⁺ while impedes anions movement, thereby effectively suppressing concentration polarization fundamentally. Consequently, the prepared electrolytes exhibit relatively low concentration polarization, enabling a high critical current density of 3.7 mA cm^–2, stable Li plating/stripping even at high current density of 2 mA cm⁻², and prolonged cycling stability of Li | |Ni_0.8Co_0.1Mn_0.1O₂ full cell over 2600 times at the specific current of 900 mA g⁻¹ within a potential window of 2.8 to 4.5 V. This work proposes a mechanical-electrochemical conversion strategy by constructing a piezoelectric electrolyte that actively utilizes the unavoidable volume fluctuation of lithium metal to minimize Li⁺ concentration gradient, offering a promising pathway towards high-performance lithium metal batteries.

Concentration polarization limits cycling stability of cells at high currents. Here, authors propose a mechanical-electrochemical conversion strategy by constructing a piezo electrolyte that actively utilizes the unavoidable volume fluctuation of lithium metal to minimize concentration polarization.

Introduction

Solid-state electrolytes with high safety are popular emerging in the applications of lithium metal battery for achieving high-energy-density¹. Particularly, polymer composite electrolytes (PCEs), strategically engineered through the rational integration of polymer matrices and functional fillers, have emerged as transformative candidates for solid-state lithium metal batteries (SSLMBs)^2,3. Critical challenges, including subthreshold ionic conductivity, insufficient mechanical strength, continuous interfacial side reaction, etc., have been extensively investigated for the commercialization of SSLMB^4–6. Although some achievements have been obtained, the state-of-art SSLMBs still suffer from poor cycling stability, especially at high current densities. This is highly related to the growth of lithium dendrites that lead to short-circuit failures. As proposed in the Chazalviel model^7,8, the cation concentration in the vicinity of negative electrode depletes continuously until to zero after a certain period (defined as Sand’s time) when the current density is above a critical value, which creates a large space charge and electric field at the electrode, leading to unavoidable growth of dendrites. Besides, with the existence of non-equilibrium ion distributions and substantial concentration polarization, other poor performances are prone to be companied with cells, such as low conductivity, narrow operating voltage range, etc. Therefore, strategic suppression of concentration polarization is an important cornerstone for long cycle stability of SSLMBs at high current densities.

The concentration gradient (ΔC) is defined below⁹,

$$\mathrm{\Delta }C\approx \frac{{LD}_{-}}{{FD}_{+}}$$ 1

where L is the thickness of the electrolyte, F is the Faraday constant, D₊ and D₋ are the diffusion coefficients of Li⁺ and anion. Detailed derivation procedures are provided in the supplementary information (SI). According to the equation of ΔC, rational elevated D₊ coupled with restricted D₋ constitutes essential design criteria for mitigating electrolyte concentration gradients, thereby fundamentally suppressing concentration polarization. Strategies to restrain concentration polarization have mainly focused on decreasing D₋, including inhibition of the anion migration through the introduction of anion-trapping acceptors, Lewis acidic fillers, and steric-sieving metal-organic-frames, as well as covalently bonding anions through the use of polymeric lithium salt, ion-containing polymers, etc.^9,10. These strategies can reduce concentration polarization to some extent. However, electrolytes with restricted anionic mobility are accompanied with low ionic conductivity and inferior interfacial composition¹⁰. In addition, enhancing D₊ is also effective to restrain concentration polarization, which can be achieved by introducing an additional driving force to accelerate Li⁺ movement based on the electrochemical potential gradient. Consequently, the critical key to mitigate concentration polarization of PCEs, without scarifying their ionic conductivity and interfacial stability, is developing a strategy to optimize diffusions of Li⁺ and anion. However, such optimization to simultaneously enhance D₊ while reducing D₋ is quite challenging and rarely reported.

As mentioned above, lithium dendrites are prone to generate at high current densities due to the occurrence of concentration gradient¹¹. Besides, the volume of the lithium metal negative electrode varies infinitely during cycling because of lithium plating/stripping and by-product generation¹². These coupled phenomena lead to significant volumetric fluctuation of the lithium negative electrode, companying with deleterious consequences on cells, such as break of solid electrolyte interface (SEI), high stress on electrolytes, etc. Reducing volumetric fluctuation of lithium negative electrode has been widely attended, but this phenomenon cannot be completely averted^13,14. Therefore, it is worthwhile and interesting to strategically exploit the intrinsic volumetric fluctuation of the lithium negative electrode and associated dynamic stress on electrolytes to optimize the electrochemical performances of SSLMBs. Piezoelectricity enables bidirectional energy interconversion between mechanical stress and electrical potential¹⁵, which is applied in a wide range of applications such as energy harvesting, microelectromechanical systems, and medical devices^16,17. Recently, employing locally distributed piezo-field to promote Li⁺ diffusion or homogenize Li⁺ deposition have emerged in electrochemical energy storage^18–20. However, an extrapolation about the application of a piezo-field to suppress concentration polarization in cells by regulating the migration of both Li⁺ and anions has received limited attention.

In this work, a piezoelectric PCE was prepared by incorporating 0.5Ba(Zr_0.2Ti_0.8)O₃–0.5(Ba_0.7Ca_0.3)TiO₃ (BCZT) ceramic with prominent piezoelectric properties into a polyvinylidene fluoride (PVDF) matrix. Furthermore, a piezo-assisted electromechanical coupling effect is proposed in the piezoelectric PCE (Fig. 1), where a piezo-field gradual decreasing from negative electrode to positive electrode is generated on the piezoelectric electrolyte arising from the volume fluctuation of negative electrode. Then, the gradient piezo-field functions as an additional driving force to accelerate Li⁺ transport and hamper anions migration, thus mitigating the concentration polarization and achieving energy conversion. Attributing to the piezo-assisted electromechanical coupling effect, the PVDF-BCTZ PCEs with mitigated concentration polarization demonstrate high critical current density (CCD) of 3.7 mA cm⁻², stable Li plating/stripping at high current density of 1 and 2 mA cm^–2, as well as prolonged cycling stability of Li||Ni_0.8Co_0.1Mn_0.1O₂ (NCM811) full cell at the specific current of 900 mA g^–1 and elevated cut-off voltage (4.5 V).

Fig. 1. Schematic illustration of piezo-assisted electromechanical coupling effect.

a In the non-piezo PCE, the transport of ions is driven by electrochemical potential, leading to a high concentration gradient. b In the piezo PCE, the expansion of lithium metal negative electrode during charging leads to spatially heterogeneous piezo-potential gradient in electrolyte, which assists ions transport and results in a low concentration gradient.

Results

Preparation and characterization of PCEs

BCTZ is chosen as the piezoelectric filler since it is lead-free and possesses relatively high piezoelectric coefficient^21–23. To isolate the role of piezoelectricity, cubic BaTiO₃ (BTO) with non-piezoelectricity is chosen as control. BCTZ was synthesized by electrospinning followed by annealing treatment. From scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HRTEM) observations, the annealed BCTZ nanofibers show a statistically average diameter of 236 ± 30.7 nm (Fig. 2a), which turn into nanostick morphology after grinding (inset of Fig. 2b). The synthesized BCTZ exhibits tetragonal symmetry, as evidenced by the lattice spacing in HRTEM and selected area electron diffraction (Fig. 2b), which is a non-centrosymmetric structure with spontaneous polarization. To validate the dielectric properties of fillers, the BCTZ and the cubic BTO were processed into dense pellets via spark plasma sintering (Fig. S1). From broadband dielectric spectroscopy (BDS) measurement, the BCTZ pellet has near 10 times higher dielectric constant at 10 Hz (5.0 × 10³ vs. 5.9 × 10²) than cubic BTO pellet (Fig. S2a). As shown in Fig. S2b, the BCTZ pellet shows narrower and steeper polarization-electric field (P-E) loops compared to cubic BTO pellet, which corresponds to more obvious polarization change under applied electric fields and higher spontaneous polarization strength²³. With high dielectric permittivity and large polarization, high piezoelectric properties would be anticipated for BCTZ since piezoelectric coefficient is positively correlated with remanent polarization (P_r) and dielectric constant²⁴.

Fig. 2. Structural and piezoelectric characterization of electrolytes.

a SEM image and b HRTEM image of synthesized BCTZ. c XRD patterns of prepared electrolytes and PDF cards of fillers. d dielectric profiles of pure PVDF film, composite PVDF-BTO film, and composite PVDF-BCTZ film without adding Li salt. e P-E curves of PVDF-BCTZ film without adding Li salt. f Phase-voltage profiles and amplitude-voltage profiles of PVDF-BTO PCE and PVDF-BCTZ PCE containing Li salt. g Schematic diagram of piezoelectric testing and corresponding atomic displacement under external force. h Open-circuit voltage versus time of pure PVDF film, PVDF-BTO film, and PVDF-BCTZ film without adding Li salt sandwiched between two pieces of stainless steel.

Flexible PCEs were prepared by incorporating fillers in PVDF matrix using a solution-casting method. In order to optimize filler content, PCEs containing 2.5, 5, 10, and 15% BCTZ in weight were prepared for ionic conductivity test and the 10 wt% BCTZ PCE exhibits the highest ionic conductivity (Fig. S3). Thus, 10 wt% was chosen as the most appropriate filler fraction in PCE for the following investigations. As shown in Fig. S4, the PVDF-BCTZ PCE presents a dense structure that provides a continuous path for ion transportation. X-ray diffraction (XRD) patterns in Fig. 2c demonstrate the successful incorporation of functional fillers in the PVDF matrix, with BCTZ and BTO showing pure tetragonal and pure cubic phases, respectively. Tensile testing and nano‑indentation (Fig. S5) reveal that PVDF‑BCTZ PCE and PVDF-BTO PCE possess comparable mechanical properties, which are higher than those of PVDF electrolytes²⁵.

Since LiFSI can interfere with dielectric tests (Fig. S6), salt-free films were chosen for dielectric measurement. The PVDF-BCTZ film exhibits a high dielectric constant of 18.5 at 10 Hz (Fig. 2d) from BDS results, which is higher than that of PVDF film (8.4) and PVDF-BTO film (12.4). From P-E loops, PVDF-BCTZ film shows a P_r of 3.1 μC cm⁻² at 300 kV mm⁻¹, which is higher than 2.7 μC cm⁻² of PVDF film (Fig. 2e vs Fig. S7). The increased dielectric constant and P_r would probably contribute to a high piezoelectricity.

In piezoelectric measurements, low-frequency stress causes mobile ions to drift and obscure the signal, while high-frequency excitation prevents ionic movement, isolating the pure piezoelectric response²⁶. Piezoresponse force microscopy (PFM) measurement with high-frequency excitation was performed on electrolyte containing lithium salt (Fig. 2f). The PVDF-BCTZ PCE exhibits a phase-voltage hysteresis loop with 180° phase switching and pronounced butterfly-shaped hysteresis loop, in clear contrast to the weak response of PVDF-BTO PCE²⁷. In addition to the microscopic piezoelectric response provided by PFM, a macroscopic piezoelectric response with low-frequency excitation was also performed on film without lithium salt using a custom-built electromechanical coupling setup (Fig. S8). As schematically illustrated in Fig. 2g, applied mechanical stress induces the displacement of positive-charge and negative-charge centers within the perovskite lattice, which dynamically modulates the local dipole moments of fillers through piezoelectricity, and the corresponding piezo-generated charges were detected via high-precision electrometer. From Fig. 2h, the PVDF-BCTZ film shows a high peak-to-peak piezoelectric response of ~400–600 mV under varied dynamic impact forces, which is comparable to commercial PolyK Piezo film (Fig. S9), indicating a piezoelectric response of the PVDF-BCTZ film. On the contrary, weaker responses of PVDF-BTO and PVDF films could arise from the triboelectricity. Consequently, both microscopic and macroscopic piezoelectric measurements confirm the piezoelectric properties of PVDF-BCTZ PCE, indicating the generation of a built-in piezo-field in response to external stresses.

Mechanism of the piezo-assisted electromechanical coupling effect

Thermogravimetric analysis (TGA) was carried out to detect the residues DMF in electrolytes based on the second weight loss step shown in Fig. S10. The DMF residues in PVDF, PVDF-BTO, and PVDF-BCTZ electrolytes are calculated to be about 12.2, 15.7, and 13.3 wt.%, respectively. The DMF residue amount in PVDF-BCTZ PCE is between those of PVDF and PVDF-BTO electrolytes with small deviations, which indicates that the difference in the effect of DMF on the properties of these electrolytes is negligible. The coordination states of FSI^– are determined via Raman analysis in Fig. 3a, b, and the content of free FSI^–, contact pairs, and aggregate cluster in PVDF-BCTZ PCE are 45.5, 38.2, and 16.3%, respectively. The PVDF-BCTZ PCE exhibits the highest free FSI⁻ content among PVDF (17.1%) and PVDF-BTO (38.6%) electrolytes, which is attributed to its large polarization facilitating the dissociation of Li salt. The PVDF-BCTZ PCE exhibits the highest ionic conductivity (7.13 × 10^–4 S cm⁻¹) at 25 ± 0.5 °C and the lowest E_a (0.198 eV) from the Arrhenius plot compared to PVDF electrolyte (3.01 × 10^–4 S cm^–1, 0.320 eV) and PVDF-BTO PCE (6.36 × 10^–4 S cm⁻¹, 0.268 eV) (Fig. 3c). The higher ionic conductivity of PVDF-BTO PCE than PVDF-BCTZ PCE at elevated temperature is probably due to its higher P_r at high temperature (Fig. S11).

Fig. 3. Electrolyte properties and the investigation of piezo-field-assisted ion transport.

a Raman spectra of three electrolytes and b corresponding quantification results of the FSI^– states in electrolytes. c Arrhenius plots of three electrolytes, the inner figure presenting their ionic conductivity at 25 ± 0.5 °C. d TUNA current of PVDF-BCTZ electrolyte under 5 V, the inner figure illustrating the schematic of PFTUNA technique. e Ionic conductance under external torsion and their fitting lines, the inner figure presenting slope values of fitting lines. f Concentration polarization (η_con) value (column chart) and its proportion to total polarization (line chart) of assembled symmetric Li||Li cells under varied current density. g, h Generated piezo-field, i, j Li⁺ concentration (Conc.), k, l Li⁺ concentration variation (ΔConc.) along the edge of (g, i, k) PVDF-BTO PCE and (h, j, l) PVDF-BCTZ PCE from simulation.

Besides, the electronic conductivity of PVDF-BCTZ PCE is 1.0 × 10^–10 S cm^–1 (Fig. S12), which is 10⁶ lower than its ionic conductivity, meaning that the PVDF-BCTZ PCE conducts only ions instead of electrons. Moreover, the PVDF‑BCTZ PCE presents relatively higher Li⁺ transference number (t_Li+) (Fig. S13) as well as ⁷Li and ¹⁹F diffusion coefficients than PVDF electrolyte (Fig. S14), indicating the embedded BCTZ fillers can facilitate Li ions mitigation.

The cross-sections of lithium metal in a Li|| NCM811 full cell at the fully charged and discharged states (Fig. S15) indicate the obvious thickness change of lithium metal. Furthermore, stress variation of a Li||NCM811 pouch cell (size = 50 × 50 mm², positive electrode loading = 1.8 mA h cm^–2) was recorded according to an attached in situ sensor (Fig. S16), which shows that the stress increases/decreases with lithium plating/stripping. The thickness-change induced interfacial stress could be accumulated up to GPa²⁸, which we suppose is sufficient to induce a pronounced piezoelectric response. Peakforce-TUNA operation mode was performed to investigate the influence of local polarization on Li⁺ flux, using an PFTUNA atomic force microscope probe as the working electrode and lithium metal as the counter electrode. As schematically shown in inset of Fig. 3d, an Li⁺ flux in the electrolyte is converted to a current (namely, TUNA current) at the AFM probe during the reduction of Li⁺ to Li²⁹. The height morphology of detected PCE area is seen in Fig. S17, where PVDF and BCTZ are distinguished based on their modulus. From Fig. 3d, the PVDF matrix exhibits a significantly higher responded TUNA current compared to the BCTZ filler because BCTZ is non-ionic conductive. Furthermore, the PVDF boundary adjacent to the BCTZ filler displays a markedly elevated TUNA current relative to the PVDF bulk, probably because the Li⁺ is preferentially attracted by BCTZ with pronounced polarity. On the contrary, the TUNA current is uniform throughout the PVDF electrolyte (Fig. S18), suggesting that results in Fig. 3d is not from the boundary-induced effect. Besides, the TUNA current here should be independent of interfacial polarization, since which should be aligned to the direction of the electric field^30,31. Similar phenomenon has been demonstrated in piezo-catalysis, where anions and cations are accelerated by the positive and negative charges on the dipoles, which in turn improves the catalytic efficiency^32–34. Combined with the in situ stress evolution and the TUNA current distribution, it is demonstrated that the volume change of the lithium metal during cycling leads to internal piezo-field (changed dipoles) in the electrolyte due to its suffered stress, which attract Li⁺.

To directly demonstrate the effect of piezo-field on ion transport, assembled Cu|PVDF-BCTZ|Cu and Cu|PVDF-BTO|Cu pouch cells were placed between two stainless steel plates and subjected to various external stresses for EIS test at around 25 °C. As seen from Fig. 3e, the overall ionic conductance of the PVDF-BCTZ PCE is higher than that of the PVDF-BTO PCE, which is consistent with their ionic conductivities measured in coin cells. Furthermore, the greater slope of the conductance-force fit line for PVDF-BCTZ PCE compared to PVDF-BTO PCE implies that PVDF-BCTZ PCE has a stronger conductivity response to external dynamic stresses. This phenomenon arises from the generated piezo-field, which drives cations toward the negative end of dipoles and anions toward the positive end of dipoles, thereby facilitating ion transport. Galvanostatic charge/discharge-relaxation processes at current densities from 0.1 to 0.5 mA cm^–2 were performed, as shown in Fig. S19a, to identify the inhibition effect of piezo-field on concentration polarization. As illustrated in Fig. S19b, instantaneous voltage steps upon current on/off originate from ohmic polarization (η_ohm) and electrochemical polarization (η_elec), whereas the gradual evolutions of voltage during relaxation corresponds to concentration polarization (η_con)³⁵. The η_con value and its proportion to total polarization (η_con/(η_con+η_ohm+η_elec), η_con%) at various current densities (column and line chart in Fig. 3f, respectively) demonstrate that the η_con and the η_con% values of PVDF-BCTZ PCE are lower than the corresponding values of PVDF and PVDF-BTO electrolytes. Besides, the PVDF-BCTZ PCE presents both lower increments of η_con and η_con% from 0.1 to 0.5 mA cm^–2 compared to PVDF and PVDF-BTO electrolytes (Fig. S19c). Furthermore, cells using piezoelectric PVDF‑BCTZ PCE displays a capacity increase under dynamic stress (Fig. S20), in clear contrast to cells based on PVDF-BTO PCE. All signs indicate that piezoelectric PVDF-BCTZ electrolyte exhibits improved mitigation of concentration polarization.

Theoretical simulations further verify the mechanical-electrochemical conversion between the internal pressure generated by the piezoelectric electrolyte and the reduced concentration polarization. The PVDF-BCTZ and PVDF-BTO PCE models (Fig. S21) were built, and the stress suffered from the volumetric fluctuation of lithium metal (Fig. S22), the piezo-field distribution (Figs. S23 and 24), as well as the Li⁺ concentration distribution (Figs. S25 and 26) during the charging process were investigated via Finite Element Analysis (more simulation details are provided in SI). As shown in Fig. S27, since lithium is deposited on the negative electrode, the stress on the PCE increases during charging and is higher near the negative electrode than that near the positive electrode. Due to the presence of non-piezoelectric BTO and piezoelectric BCTZ fillers, no piezo-field generates on PVDF-BTO PCE throughout the process (Fig. 3g), whereas the piezo-field along the thickness of PVDF-BCTZ PCE (Fig. 3h) varies with the change of stress. Different piezoelectric responses lead to completely divergent Li⁺ concentration distribution on PCEs. For the PVDF-BTO PCE, the Li⁺ concentration (Conc.) near negative electrode and positive electrode decreases and increases, respectively, over charging, with concentration gradient gradually broadening (Fig. 3i). However, for PVDF-BCTZ PCE, the Li⁺ concentration near two electrodes present similar regulations but with smaller concentration gradient (Fig. 3h), which suggests the weak concentration polarization. Moreover, the change of Li⁺ concentration within one minute (denote as the concentration variance, ΔConc.) as function of the charging time was determined to reflect the Li⁺ transport rate. As shown in Fig. 3k, l, the bluer the color means the smaller ΔConc., indicating the slower Li⁺ transport. These results show that the Li⁺ transport in PVDF-BCTZ PCE (Fig. 3l) is faster than that in PVDF-BTO PCE (Fig. 3k). Similarly, the piezoelectric field near negative electrode in PVDF-BCTZ PCE also attracts anions and inhibits their migration towards the positive electrode, synergistically alleviating the concentration polarization.

Integrating the experimental and simulation results, the piezo-assisted electromechanical coupling effect that dynamically inhibits concentration polarization is elucidated in Fig. 1. The gradient in the electrochemical potential of ions serves as primary driving force for ion transport in electrolytes³⁶, as the case of non-piezoelectric PVDF-BTO PCE (Fig. 1a). In contrast, the Li⁺ plating induced gradient stress partially aligns the dipoles and generates a gradient piezo-field (Fig. 1b), where the negative electrode feature more aligned or polarized dipoles than the positive electrode side. As a result, Li⁺ is more effectively drawn toward the negative electrode, while anion migration toward the positive electrode is suppressed to reduce concentration polarization.

Electrochemical performance of assembled cells

The galvanostatic voltage profiles of Li||Cu half cells (Fig. S28) show that the Cu|PVDF-BCTZ|Li has not only lower nucleation overpotential (μ_nuc) but also smaller mass-transfer controlled plateau potential (μ_pla) than Cu|PVDF-BTO|Li, indicating lower nucleation energy barrier and faster charge transfer^37,38. Tafel curves and the corresponding exchange current density (j) indicate that the Li|PVDF-BCTZ|Li symmetric cell exhibits the highest charge-transfer kinetic and the fastest Li⁺ transport across the formed SEI (Fig. 4a). As shown in Fig. 4b, the CCD of the PVDF-BCTZ PCE is as high as 3.7 mA cm⁻², higher than that of the PVDF-BTO PCE (1.98 mA cm⁻²) and the PVDF electrolyte (1.09 mA cm⁻²). Considering the comparable mechanical properties (Fig. S5), the superior CCD value of PVDF-BCTZ PCE may be attributed to the effect of piezo-field on suppressing concentration polarization. The CCD value of BCTZ embedded PVDF electrolyte present in this work is competitive with most of the reported state-of-the-art PVDF-based composite electrolytes (Fig. 4c)^39–48. The Li|PVDF-BCTZ|Li symmetric cells ran stably for 2200 h at a current density of 0.1 mA cm⁻² with a small overpotential (Fig. 4d), which outperforms greatly than the controlled Li|PVDF-BTO|Li (680 h) and the Li|PVDF|Li symmetric cells (480 h). From the enlarged view (inset of Fig. 4d), once the current is switched on, the voltage of Li|PVDF-BTO|Li rises for a while, whereas the voltage of Li|PVDF-BCTZ|Li remains constant, which indicates the concentration polarization generated in PVDF-BCTZ electrolyte is smaller than that generated in PVDF-BTO electrolyte³⁵. To verify the ability of electrolytes for migrating concentration polarization, Li plating/stripping at high current density were conducted. The Li|PVDF-BCTZ|Li cell exhibited stably operation for 320 h at 1 mA cm⁻², while the lifetimes of Li|PVDF-BTO|Li and the Li|PVDF|Li cell were 180 and 68 h, respectively (Fig. 4e). Similarly, the inset of Fig. 4e also shows faster voltage stabilization for PVDF-BCTZ electrolyte compared to PVDF-BTO electrolyte, suggesting a smaller concentration polarization again. Even at a higher current density of 2 mA cm^–2, Li|PVDF-BCTZ|Li cell shows a comparable lifetime of 150 h, while the Li|PVDF-BTO|Li and the Li|PVDF|Li cells fail rapidly (Fig. S29). All these results prove that the PVDF-BCTZ PCE with the inherent piezo-field is characterized by low concentration polarization, allowing stably Li plating/stripping in Li||Li cells at high current densities.

Fig. 4. Electrochemical performance and stability of electrolytes.

a Tafel curves with corresponded j values and b CCD profiles and values of symmetrical cells with PVDF, PVDF-BTO, and PVDF-BCTZ electrolytes. c A summary of CCD values for PVDF-based composite electrolytes reported in recent years. Galvanostatic cycling profiles of symmetrical cells with PVDF, PVDF-BTO, and PVDF-BCTZ electrolytes at d 0.1 mA cm⁻², 0.1 mA h cm^–2 and e 1 mA cm^–2, 0.5 mA h cm^–2. f CV curve of NCM811|PVDF-BCTZ|Li at a scan rate of 0.05 mV s⁻¹. g Rate performance, cycling stabilities at h the specific current of 360 mA g^–1 within 2.8 to 4.3 V, and i the specific current of 900 mA g⁻¹ within 2.8 to 4.5 V of the Li||NCM811 solid-state full cells.

Linear sweep voltammetry (LSV) curve (Fig. S30) suggests that the embedding of BCTZ in PVDF electrolytes broadens electrochemical window from 4.53 V of pure PVDF electrolyte to 4.71 V. FTIR spectra (Fig. S31) show the shift of C = O and O = C-N peaks for DMF after introducing BCTZ, indicating an altered local chemical environment for DMF. Density functional theory (DFT) calculations (the details provided in Supplementary Data 1 and the results present in Fig. S32) show a higher adsorption energy of DMF on BTO/BCTZ (−2.527 eV) than that on PVDF chains (−0.314 eV). This helps to restrain DMF decomposition and thus extends the electrochemical stability window⁴⁹.

To evaluate the performance of the piezoelectric electrolyte in practical batteries, full cells were assembled with NCM811 as the positive electrode and Li metal as the negative electrode. Cyclic voltammogram curves of NCM811|PVDF-BCTZ|Li full cell displays a better redox reaction reversibility and smaller polarization (the voltage difference between the redox peaks) compares to that of NCM811|PVDF|Li cell and NCM811|PVDF-BTO|Li cell (Figs. 4f and S33). The effect of piezo-fields on the rate performance of assembled batteries is demonstrated in Fig. 4g. The NCM811|PVDF-BCTZ|Li delivers discharge capacities of 183.7, 174.0, 162.8, 138.6, 100.3 and 61.5 mA h g^–1 at 0.1, 0.2, 0.5, 1, 5, and 10 C within 2.8 to 4.3 V (1 C = 180 mA g^–1), respectively, and the discharge capacity recovers to initial value at different rates, revealing the superior rate recoverable performance. Although NCM811|PVDF|Li and NCM811|PVDF-BTO|Li cells present good rate recoverable performance, their absolute discharge capacities exhibit significant rate-dependent attenuation, particularly at elevated rates.

To test the operational stability of piezoelectric electrolytes in full batteries, long cycle tests under various conditions were conducted. The NCM811|PVDF-BCTZ|Li full cell delivers 1350 cycles with a capacity retention of 77% at the specific current of 360 mA g^–1 and 25 °C within 2.8 to 4.3 V, while NCM811 | PVDF-BTO|Li and NCM811|PVDF|Li full cells end up in short circuit after 750 and 300 cycles, respectively. Remarkably, under fast charging/discharging at the specific current of 900 mA g^–1 with a high cutoff voltage of 4.5 V, the NCM811|PVDF-BCTZ|Li full cell exhibits an initial discharge capacity of 134.7 mA h g⁻¹ and a 70% retention capacity from 100 to 2500 cycles. While the NCM811||Li full cells using PVDF-BTO and PVDF electrolytes cycled with short lifespan and low discharge capacity. A drop in the initial capacity in Fig. 4i is probably due to the generation of bending layers with dislocations on the outer surface of NCM811 particles, which arises from the Li⁺ concentration gradient at the positive electrode|electrolyte interface during high current charging/discharging⁵⁰. Matching with a higher loading of positive electrode (3 mg cm^–2), the NCM811|PVDF-BCTZ|Li stably runs for 90 cycles, while the NCM811|PVDF-BTO|Li and NCM811|PVDF|Li cells present short circuit after 30 and 20 cycles (Fig. S34). These results confirm full cells assembled with PVDF-BCTZ PCEs exhibit high discharge capacity and long cycle stability, particularly at high rate.

Interphases and structures of cycled electrodes

Li metals in Li||Li systematic cells after 50 times Li plating/striping at 0.1 mA cm⁻², 0.1 mA h cm⁻² were analyzed to characteristic the Li deposition behavior and SEI composition. The cycled Li metal matched with PVDF electrolyte displays a fluctuating and loose morphology with Li dendrites (Fig. 5a), while the one matched with PVDF-BCTZ electrolyte presents a relatively flat and dense morphology (Fig. 5b). Besides, cryogenic transmission electron microscopy (cryo-TEM) images (Fig. S35) reveal a thinner SEI layer formed using PVDF-BCTZ PCE than using PVDF electrolyte. The interfacial chemistries on the cycled Li metal were detected by X-ray photoelectron spectroscopy (XPS) depth profile analysis. In Fig. 5c, d, the higher F%/C% ratio and S%/C% ratio indicate that the SEI formed with PVDF-BCTZ PCE consists of an inorganic-rich component compared to the one formed with PVDF electrolyte. Intensities of C–O, C = O, and Li₂CO₃ in C 1s spectra of SEI matched with PVDF-BCTZ electrolyte are lower than those using PVDF electrolyte (Fig. 5e, f), suggesting that the decomposition of the DMF is greatly inhibited^51,52. This inhibition is attributed to the strong anchoring of DMF by the filler, as evidenced by the FTIR and DFT results above. In addition, the high intensity of N-Li in N 1s spectra in the SEI matched with PVDF-BCTZ electrolyte (Fig. 5g, h), indicating a deep FSI^– degradation⁵³.

Fig. 5. Characterization of cycled electrodes and solid-electrolyte interphases.

Characterization of cycled lithium metal in Li||Li systematic cells matching with (a, c, e, g) PVDF and (b, d, f, h) PVDF-BCTZ electrolytes after 50 times Li plating/striping at 0.1 mA cm^–2, 0.1 mA h cm^–2: a, b surface morphologies of cycled lithium metal, c, d quantified atomic composition ratios of the SEI obtained from XPS, e, f C 1s and g, h N 1s spectra of the SEI composition obtained from high-resolution XPS. TEM images of NCM811 particles in Li||NCM811 full cells matched with i PVDF electrolyte and j PVDF-BCTZ electrolyte after 50 cycles at 180 mA g^–1. Initial cycle curve and corresponding in situ XRD patterns of the NCM811 (003) and (101) planes for Li||NCM811 full cells using k PVDF and l PVDF-BCTZ electrolytes at 36 mA g^–1.

In addition, the influence of the piezoelectric PVDF-BCTZ electrolyte on the structural stability of NCM811 positive electrode in Li||NCM811 full cells was also investigated after 50 cycles at the specific current of 180 mA g⁻¹ within 2.8 to 4.3 V. The cathode-electrolyte interphase (CEI) layer formed on the surface of NCM811 particles matched with PVDF electrolyte (Fig. 5i) was thicker than that matched with PVDF-BCTZ PCE (Fig. 5j). Furthermore, the NCM811 particles matched with PVDF electrolyte have intergranular cracks, while the ones matched with PVDF-BCTZ PCE still maintain original morphology with no obvious damage (Fig. S36). The intergranular cracks of NCM811, which are induced by the nonuniform volume variation of particles, can expose new inner surfaces to electrolytes, promoting side reactions to thicken the CEI and accelerating cell degradation⁵⁴.

In situ XRD measurements were performed to track the phase transition behavior of NCM811 during cycling. The (003) diffraction peak reflects the c-axis lattice parameter variation responsible for particle cracking. Two phases coexist during H1-M and H2-H3 phase transitions as presented in the (003) diffraction peak of NCM811 matched with PVDF electrolyte (Fig. 5k), which results in low discharge capacity and weak structure stability⁵⁵. While three phase transitions (H1-M, M-H2, and H2-H3) all occur smoothly and completely for the one paired with PVDF-BCTZ PCE (Fig. 5l). Similarly, better phase transition is also presented in the (101) diffraction peak of NCM811 matched with PVDF-BCTZ PCE compared to PVDF electrolyte. In situ XRD result indicates that the NCM811 undergoes a fully reversible phase transition during cycling when paired with the piezoelectric PVDF-BCTZ PCE. Inhibition of DMF decomposition and alleviation of internal stress variation, driven by the anchoring effect of filler and the piezoelectric effect of PCE, might be the reason for the structural stability and the complete phase transition of NCM811^50,56. Overall, the piezoelectric PVDF-BCTZ electrolyte is able to maintain the interfacial stability of Li metal and the structural stability of the NCM811, which is necessary to achieve long-term stable cycling.

Discussion

In summary, we propose the piezo-assisted electromechanical coupling effect that actively converts the unavoidable volume fluctuations of lithium metal negative electrodes into a gradient piezo-field, effectively suppressing concentration polarization through selectively accelerating Li⁺ migration while impeding anion movement. Such a unique piezo-assisted electromechanical coupling effect enables the piezoelectric PVDF-BCTZ PCE to show flattened Li⁺ concentration gradient at the PCE/Li interface and low concentration polarization throughout the electrolyte. As a result, the PVDF-BCTZ PCE shows competitive CCD of 3.7 mA cm⁻², sustaining stable Li plating/stripping at high current density of 1 and 2 mA cm⁻². Remarkably, the solid-state NCM811|PVDF-BCTZ|Li cell withstands 2600 cycles at the specific current of 900 mA g⁻¹ within a potential window of 2.8–4.5 V. This work demonstrates a transformative strategy for self-adaptive concentration polarization suppression via piezo-electromechanical coupling, enabling SSLMBs to cycle at high current densities.

Methods

Preparation of BCTZ

0.5Ba(Zr_0.2Ti_0.8)O₃–0.5(Ba_0.7Ca_0.3)TiO₃ (BCTZ) was synthesized by electrospinning followed by annealing treatment. The amounts of compounds, barium acetate, calcium acetate monohydrate, titanium butoxide, and zirconium propoxide (Shanghai Aladdin Biochemical Technology Co., Ltd., 99.0%), are calculated according to the stoichiometric formula of BCTZ. First, solution A was prepared via adding barium acetate and calcium acetate monohydrate in glacial acetic acid (Shanghai Aladdin Biochemical Technology Co., Ltd., 99.8%) and stirred at 80 °C for 1 h. Meantime, solution B was obtained via adding titanium butoxide and zirconium propoxide in acetyl acetone (Shanghai Aladdin Biochemical Technology Co., Ltd., 99.5%), chelating at 25 ± 0.5 °C for 1 h. Then, 2-methoxyethanol (Shanghai Aladdin Biochemical Technology Co., Ltd., 99.8%) and polyvinylpyrrolidone (Shanghai Aladdin Biochemical Technology Co., Ltd., M_w = 1,300,000) were added into the mixture of solution A and B to prepare a precursor solution of BCTZ. A syringe loaded the precursor solution of BCTZ was placed on a syringe pump with a feed rate of 0.5 mL h⁻¹. Electrospinning process was run at 15 kV high voltage and 15 cm tip-to-collector distance. Next, the electrospun BCTZ mat was annealed at 1100 °C for 2 h. Finally, BCTZ nanofiber was grinded for 30 min twice.

Preparation of solid-state electrolytes

PVDF powder (Mw = 600,000, Kynar 761, 99%, Arkema), BCTZ, and BTO (from Macklin) were dried for 24 h in a vacuum oven prior to use. The precursor slurry of PVDF electrolyte was made by mixing 400 mg PVDF and 267 mg lithium bis(fluorosulfonyl)imide (LiFSI, Guangdong Canrd New Energy Technology Co., Ltd., 99.9%) with a 3/2 weight ratio (a commonly employed ratio reported in literatures^57–59) in 15 mL N,N-Dimethylformamide (DMF, Shanghai Aladdin Biochemical Technology Co., Ltd., 99.9%) under magnetic stirring for 6 h. The precursor slurry of BCTZ or BTO embedded electrolyte was prepared by adding fillers with certain weight percentage to the pure PVDF precursor slurry under magnetic stirring and ultrasonic vibration for 6 h alternately. These slurries were cast into 100 mm diameter glass dishes and dried for 23 h in a convection oven at 55 °C.

Preparation of cells

Single-crystalline LiNi_0.8Co_0.1Mn_0.1O₂ (NCM811, Guangdong Canrd New Energy Technology Co., Ltd.), carbon black (Guangdong Canrd New Energy Technology Co., Ltd.), and PVDF (M_w = 1,200,000, Kynar 5130, 99%, Arkema) were mixed with an 8:1:1 weight ratio in N-methyl-2-pyriolidone (NMP, Shanghai Aladdin Biochemical Technology Co., Ltd., 99.9%) solvent for 6 h magnetic stirring, followed by casting the slurry on single side of an aluminum foil using doctor blade holder. The mass loading of active NCM811 in positive electrode was adjusted by casting thickness. The CR2032 solid-state cells matching 1.1 mm thickness spring were assembled in an Ar-filled glove box (O₂ < 0.1 ppm, H₂O < 0.1 ppm). The symmetric Li||Li cell was assembled by placing electrolyte (19 mm diameter, around 100 μm thickness) between a large Li metal piece with diameter of 15.6 mm and a small Li metal piece with diameter of 12.5 mm. Li metal pieces with two specifications (450 μm thickness, China Energy Lithium Co., Ltd.) were directly used without any pre-activation or pre-treatment. A copper foil (14 mm diameter, 10 μm thickness, purchase from Canrd) on one side and a small Li metal were placed on the opposite sides of electrolyte for the Li||Cu cell. The Li||NCM811 full cell was assembled by placing electrolyte between a large Li metal piece and a positive electrode (diameter of 12.5 mm). For Li||NCM811 pouch cell, a positive electrode (3 × 3.5 cm) and a negative electrode (4 × 4.5 cm, thickness of 35 μm, China Energy Lithium) were stacked on the opposite sides of electrolyte (5 × 5.5 cm). During the fabrication and formation process at 25 ± 3 °C, a proper pressure of 700 kg (4.9 × 10⁶ Pa) was applied to ensure excellent interfacial contact between electrolyte and electrode.

Materials Characterizations

The X-ray diffraction (XRD) measurements of the samples were characterized on a Bruker D8 Advance with Cu-Kα radiation applied a scan rate of 5 deg min⁻¹ and step size of 0.0025 deg. The in situ XRD test at specific current of 36 mA g^–1 were performed at 25 °C and diffraction patterns were collected every 8 min. The morphology was examined by scanning electron microscope (SEM, HITACHI S4800) at 5 kV, coupled with energy dispersive X-ray spectroscopy. The dielectric value and electric displacement-electric field curve were measure via BDS (Novocontrol Concept 40) with an applied voltage of 1.0 V_rms (i.e., root-mean-square voltage) and Premiere II ferroelectric test system (Radiant Technologies, Inc., USA), and samples were evaporated with Au electrodes on both sides before test. The localized piezoelectric properties were characterized using piezoelectric force microscopy in contact mode on an atomic force microscope (AFM, Bruker Dimension Icon). A conductive Pt/Ir-coated Si cantilever (Bruker, SCM-PIT-V2, nominal tip radius: 25 nm, spring constant: 3.0 N m⁻¹) was used. Concurrently, the local electrical conductivity was mapped using the PeakForce Tunneling AFM (PeakForce TUNA) mode. The voltage response of film under dynamic impact forces were measured via a custom-built system, in which an electrometer recorded the voltage output. TGA was performed using a Netzsch STA 449F3 thermal analyzer from 25 to 600 °C at a heating rate of 10 °C min^–1 in N₂ atmosphere. Raman spectroscopy was conducted by a Micro-laser confocal Raman spectrometer (Horiba LabRAM HR800) at 25 ± 3 °C with a 532 nm laser. Tensile experiments were measured on a universal test instrument (EUT4103, SAAS), using a constant strain rate of 5 mm min^–1 until fracture, and the resulting stress‑strain curves were used to determine tensile strength and yield strength. Nano-indentation experiments were carried out at ambient temperature using a Berkovich diamond tip (NanoFlip, KLA iMicro), applying a load-controlled triangular waveform up to a maximum load of 0.2 mN. The stress variation of the pouch cell during the charging and discharging process was tested by battery pressure distribution measure system on the basis of controlling a constant gap. Lithium metal negative electrodes with SEI were disassembled from cycled cells in the argon-filled glove box. For XPS measurement, a 2 × 2 mm specimen was randomly sampled and attached to the test stage, then transferred to a PHI 5000 VersaProbe II instrument via a sealed transfer device to avoid air exposure. Depth profiling was performed at a rate of 1 nm per minute, and all spectra were energy-calibrated relative to the C–C peak at 284.8 eV. The Fourier transform infrared (FTIR) spectra were executed by VERTEX 70 spectrometer in an attenuated total reflection mode. Cold-field emission spherical aberration corrected transmission electron microscope (Thermo Fisher Scientific, Spectra 300) was applied for the SEI morphological analysis via observing the deposited Li (0.1 mA cm⁻², 0.5 mAh cm⁻²) on a Cu grid, which was loaded on a Gatan 698 cryo-transfer holder, at a stable temperature of −170 °C and operated voltage of 200 kV.

Electrochemical measurements

The ionic conductivities were tested through electrochemical impedance spectroscopy (EIS) tested from 7 MHz to 0.1 Hz with a 10 mV AC oscillation at a quasi-stationary potential condition on a VMP3 multichannel electrochemical station (Bio Logic Science Instruments, France). The test cells were assembled by sandwiching electrolytes (diameter of 19 mm) between two stainless-steel (SS) (thickness: 1 mm; diameter: 15.8 mm) blocking electrodes. A steady current under 1 V chronoamperometry (CA) was measured by electrochemical station to obtain the electronic conductivity. The Li||Li pouch cells were assembled and polarized by a voltage (ΔV = 10 mV) for 3500 s to test the lithium transference number t_Li+. Tafel plot were measured in symmetric Li||Li cell via electrochemical station with the scan rate of 1 mV s⁻¹ and the voltage window from −0.2 to 0.2 V. The LSV curves were conducted within a potential window of 0 to 6 V (vs. Li/Li⁺) at a scanning rate of 0.1 mV s^–1 using the electrochemical station. The cyclic voltammetry (CV) profiles of NCM811||Li full cells were recorded between 2.8 ~ 4.3 V for 5 cycles at 0.05 mV s^–1. The galvanostatic plating/stripping stability of Li||Li symmetric cells and the cycling performance of Li||NCM811 full cells were evaluated at 25 °C using a NEWARE CT4008T battery testing system. Reported results represent findings from individual cell tests.

Finite element simulation

A simplified battery model was simulated with several bars (denoting fillers) randomly distributing in a rectangle (denoting PVDF matrix) in a two-dimensional geometry. In the calculation process, the concentration field, electric field, and stress field are coupled together for calculation, and they will interact with each other. Applied equations are provided in SI. From the perspective of simplifying the simulation model, all possible side reactions are ignored in this simulation, and the concentration changes caused by the full battery during the charging and discharging process are replaced by simplified concentration boundary conditions. The use of simplified boundaries may introduce certain deviations, but the main purpose of this simulation is to study the influence of piezo-field on concentration diffusion. Based on this premise, ensuring that both models adopt the same boundary conditions can complete the qualitative research and analysis of concentration diffusion by piezoelectric materials. To simplify the calculation, the size of the computation domain is set to 170 × 100 μm. And the wideness of the negative electrode, the electrolyte, and the positive electrode is 50, 100, and 20 μm, respectively. The surface of negative electrode is applied with 0 V electric potential. The dielectric constants and conductivity are set based on the experimental results. The piezoelectric coefficient is set as 0.01 C/m² for BCTZ (The only material which has piezoelectric effects). The Young’s modulus is set to 2 GPa for PVDF and 100 GPa for fillers, and the Poisson ratio are set to 0.32 for all material. The initial Li⁺ concentration in the electrolyte is set to 1000 mol m^–3.

DFT calculation

The adsorption energies between DMF, PVDF, and BTO were calculated using first principles density-functional theory, performing on CP2K/Quickstep package⁶⁰. The calculations used the hybrid Gaussian and plane waves formalism together with the pseudopotential approximation, applying the Perdew–Burke–Ernzerhof generalized gradient approximation exchange-correlation functional⁶¹. Double-ζ MOLOPT basis sets (DZVP–MOLOPT–SR–GTH)⁶² and Goedecker−Teter−Hutter (GTH) pseudopotentials^63,64 have been used for all elements for the geometry optimization and energy calculation. A kinetic energy cutoff of 600 Ry and an electronic energy convergence criterion of 1 × 10⁻⁶ eV was applied. The long-range dispersion was accounted for using the Grimme’s DFT-D3 corrections⁶⁵. For the BTO-DMF adsorption system, a periodic slab model of the (110) surface was constructed and geometrically relaxed under fixed cell. The optimized adsorption models were visualized using VESTA software, as are shown in Fig. S32.

Author contributions

Y.-B.H. and Y.-F.H. conceived and supervised the project. J.-Y.Y., L.C., G.X., S.G, Y.M., and X.L. designed the experiments. J.-Y.Y. performed the experiments with the help of G.X. and X.A.D.Z. conducted the XPS analysis. All authors discussed the results in the manuscript. Y.-B.H., Y.-F.H., and J.-Y.Y. wrote and revised the manuscript, which was approved by all authors.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-72527-0.