B-site engineered medium-entropy perovskite as a dual-purpose material enabling piezoelectric energy harvester and supercapacitor electrode applications

Abstract

This work focused on the synthesis of B-site-engineered medium-entropy perovskite oxides (MEPOs) with a general equimolar composition of La(Ti_0.25Mn_0.25Fe_0.25Co_0.25)O₃ (LTMFC), along with non-equimolar compositions in the B-site to explore the effect of cation disorder on structural and electrochemical performance. X-ray diffraction studies confirm the formation of a single-phase orthorhombic perovskite structure with the pbnm space group for both equimolar and non-equimolar compositions. The non-equimolar La(Ti_0.15Mn_0.15Fe_0.35Co_0.35)O₃ (TM-0.15) exhibits lattice expansion induced by oxygen vacancies, accompanied by partial reduction of Ti⁴⁺ to Ti³⁺, as verified by X-ray photoelectron spectroscopy. Compared with the equimolar LTMFC, non-equimolar TM-0.15 exhibits higher conductivity and faster ion diffusion, resulting in a specific capacitance of 526 F g⁻¹ at 1 A g⁻¹, exceeding that of the equimolar LTMFC (486 F g⁻¹) under identical conditions. Binder-free electrophoretic deposition enables uniform coating over Ni foam with excellent interfacial contact. TM-0.15 was incorporated into poly(vinylidene fluoride) (PVDF) to fabricate a piezoelectric nanogenerator (PENG) that charges the charge-storing devices. 5 wt% of TM-0.15 in PVDF exhibits a higher piezoelectric output voltage of 5 V at 1.5 kgf and is capable of charging a 2.2 µF capacitor to 1.45 V within 25 s. This study suggests that non-equimolar B-site engineering in MEPOs offers a viable strategy to outperform equimolar counterparts, enabling synergistic energy storage and harvesting.

---

1. Introduction

The rapid population growth has led to an increasing demand for energy, but conventional energy sources, particularly fossil fuels, face issues such as environmental pollution, greenhouse gas emissions, low energy efficiency, and limited reserves, prompting a shift towards renewable energy sources such as solar, geothermal, and hydropower. A viable solution involves converting and storing renewable energy as electricity, which emphasizes the need for efficient electrochemical energy storage devices.^1,2 Among the prevailing energy storage technologies, batteries, fuel cells, and supercapacitors, each presents distinct advantages and constraints. Batteries offer high energy density but are constrained by low power density, limited discharge rates, and limited lifespan. In contrast, supercapacitors bridge the functional gap between conventional capacitors and batteries, combining exceptional power density, rapid charge–discharge kinetics, and long cycling stability, making them indispensable in electric transportation, grid balancing, portable electronics, aerospace, and emergency power systems.^3–5 Parallel to advances in energy storage, piezoelectric materials known for their ability to transduce mechanical deformation into electrical energy have gained prominence in the field of self-powered devices. Their ability to harvest biomechanical or environmental vibration has positioned them as pivotal in wearable and autonomous electronics. This allows for the integration of a piezoelectric material with an energy storage unit, combining energy harvesting and storage in a single architecture.^6,7

The concept of a piezoelectric nanogenerator (PENG) originated with Wang et al. in 2006, who used ZnO nanowire arrays.⁸ Subsequent milestones include the hybridization of PENG with supercapacitors,⁹ and the first all-in-one self-charging power system developed by Xue et al. in 2012, which integrated PENG and an energy-storing lithium-ion battery within a single coin-type cell.¹⁰ Self-charging piezoelectric supercapacitors (SCPSCs) utilize PENG to enhance energy storage through piezoelectric principles. Conventional supercapacitors (SCs) store energy electrochemically by forming charge layers at electrode interfaces when a voltage is applied. In SCPSCs, piezoelectric separators enable energy generation from external pressure, facilitating energy harvesting from movements and vibrations and providing a power source for wearable devices.^11,12 Over the past few decades, extensive research has focused on developing advanced electrode materials for supercapacitors. A wide range of materials, transition metal oxides, sulphides, nitrides, and carbonaceous composites have been explored to achieve high specific capacitance and stability. Among these, transition-metal oxides stand out for their multivalent redox activity, robust structural tunability, environmental benignity, and cost efficiency.¹³ In particular, perovskite oxides (ABO₃) have garnered increasing attention due to their ferroelectricity, electronic flexibility, and tunable band structure, making them promising for both electrochemical and piezoelectric applications.⁷

Owing to their piezoelectric properties, perovskite oxides with an ABO₃ structure are promising electrode materials for supercapacitors, as their electronic structure can be tuned by manipulating cations and oxygen anions. Research, work by Kang et al., on LaMnO₃ synthesis,¹⁴ and work by Muhammed Shafi et al., on La_0.7Sr_0.3Mn_0.5Fe_0.5O₃ synthesis,¹⁵ highlights cation substitution effectiveness, which enhanced the electrochemical performance by improving ionic and electronic conductivity through valence state changes in transition metals. High-entropy oxides (HEOs) gained attention for energy applications due to their high intrinsic conductivity and stability compared to traditional metal compounds.¹⁶ Configuration entropy (S_config) distinguishes high-entropy systems, calculated based on component mole fractions,

where R is the gas constant; N, M, and P are the number of ions on the A, B, and O sites of perovskite, respectively; and A_i, B_j, and O_q are the mole fractions of the ith, jth, and qth ions, respectively.¹⁷ Materials with ΔS_config ≥ 1.5 R, materials with 1.0 R ≤ ΔS_config ≤ 1.5 R, and ΔS_config ≤ 1.0 R are considered high, medium, and low entropy, respectively.

The first high-entropy ceramic, (Mg_0.2Co_0.2Ni_0.2Cu_0.2Zn_0.2)O, featuring a regular Fm3̄m rock-salt structure, was introduced by Rost et al. in 2015. Mixing different cations into a single lattice site creates a new class of entropy-stabilized oxides with unique stability and transformation behavior.¹⁸ Jiang et al. synthesised 13 perovskite-type high-entropy oxides (HEPOs) for the first time in 2018, out of which 6 compositions formed a homogeneous single solid-solution phase and concluded that, instead of cation-size mismatch, the Goldschmidt tolerance factor is the key predictor for a stable phase at high temperature.¹⁹ Subsequently, various high-entropy ceramics have been developed, including those with oxide structures like fluorite, perovskite, and spinel. Perovskite-structured HEO exhibits remarkable properties, making it suitable as a cathode material.^20,21 Recent studies have highlighted the efficiency of entropy engineering in tailoring the electrochemical and catalytic properties of perovskites. Guo, Meng, et al. synthesized Al-doped HEPO La(CoCrFeMnNiAl_0.5)_1/5.5O_3−δ by the co-precipitation method, delivering a capacitance of 353.65 F g⁻¹ at 1 A g⁻¹.²² Wang, Yuan, et al. synthesized La(Mn_0.25Fe_0.25Co_0.25Ni_0.25Cu_0.25)O₃ using a facile metal organic framework gel method, exhibiting a capacitance of 609.3 F g⁻¹.²³ The versatile composition of high-entropy ceramics enables precise adjustments to electronic states through selected elements.²⁴ This work focused on B-site medium entropy and altering the electronic state of the element by changing the non-equimolar ratio in B-site elements.

Herein, we synthesized a B-site medium-entropy perovskite oxide (MEPO) with lanthanum (La) at the A-site and Ti, Mn, Fe, and Co at the B-site. Because of their variable oxidation states, they are selected as the main elements in most of the cathode materials. Elements in the B-site have similar ionic radius, thus equimolar and non-equimolar compositions of La(Ti_0.25Mn_0.25Fe_0.25Co_0.25)O₃ (LTMFC) are easy to synthesize. A non-equimolar ratio with less Ti and Mn creates an oxygen vacancy, and it reduces some of the Ti⁴⁺ to Ti³⁺, which is observed in the XRD peak shift due to a change in the ionic radii of those ions. Conventional ball milling is employed to synthesize the material, which can be easily scaled up for industrial applications. The entropy of the LTMFC is calculated to be S_config = 1.39 R, and for non-equimolar La(Ti_0.15Mn_0.15Fe_0.35Co_0.35)O₃ (TM-0.15) S_config = 1.30 R. Working electrodes were prepared by binder-free electrophoretic deposition. It is observed that non-equimolar compositions TM-0.15 exhibit a higher capacitance of 526 F g⁻¹ compared to equimolar (486 F g⁻¹) compositions. TM-0.15 is then blended with PVDF to make a piezoelectric nanogenerator that delivers 5 V under mechanical stress, which is then used to charge the capacitor. Synthesized materials are analysed by X-ray diffraction (XRD), scanning electron microscopy (SEM), and X-ray Photoelectron Spectroscopy (XPS). Electrochemical properties were measured in an electrochemical workstation. This study demonstrates enhanced charge storage capabilities and effective piezoelectric generation in MEPOs by employing B-site medium entropy engineering and oxygen-vacancy tuning, thereby paving the way for the expanded applications of MEPOs in future energy harvesting and charge storage systems.

2. Experimental section

2.1. Materials

Lanthanum oxide (La₂O₃) was purchased from Uranus Chemicals Co., Ltd. Titanium dioxide (TiO₂) and manganese dioxide (MnO₂) were obtained from Shimakyu Co., Ltd. Iron(III) oxide (Fe₂O₃) was procured from Choneye Pure Chemicals Co., Ltd, while cobalt(II, III) oxide (Co₃O₄) was procured from Nippon Shiyaku Kogyo Co., Ltd. The above materials were used for the synthesis of MEPOs as received without further purification.

For the fabrication of poly(vinylidene fluoride) (PVDF) films, PVDF powder (MW: 250–450 kDa) was supplied by Thermo Scientific, and laboratory-grade N,N-dimethylformamide (DMF) by Macron fine chemicals. DMF was used as the solvent to dissolve PVDF.

2.2. Synthesis of MEPO

MEPOs were synthesized via high-energy ball milling in a Pulverisette-6 planetary ball mill. Stoichiometric amounts of the respective oxide precursors were accurately weighed to the desired composition and loaded into a zirconia jar along with zirconia balls, maintaining a ball-to-powder ratio of 15 : 1. Ethanol was employed as a milling medium to ensure homogeneous mixing. The mixture was milled for 2 h at 400 rpm, dried at 60 °C for 12 h, ground and calcined at 900 °C for 3 h. After cooling to room temperature, the calcined samples were ground and sintered at 1100 °C for 4 h. The sintered samples were allowed to cool naturally, ground, and used for further studies. The same procedure was followed to prepare the following compositions: LTMFC, TM-0.15, La(Ti_0.15Mn_0.35Fe_0.15Co_0.35)O₃ (TF-0.15), and La(Ti_0.15Mn_0.35Fe_0.35Co_0.15)O₃ (TC-0.15). The overall synthesis is illustrated in Scheme 1.

Scheme 1 Synthesis of MEPO.

2.3. Preparation of PVDF films

The PVDF films were fabricated by solution casting. TM-0.15 was dispersed in DMF by ultrasonication for 2 h, 10 wt% of PVDF was dissolved in the above solution under continuous stirring at 60 °C, and stirring was continued at room temperature for 12 h to obtain a homogeneous solution. The solution was cast in a Petri dish and dried in a vacuum at 70 °C for 6 h. The above procedure was followed for different weight percentages of TM-0.15 (0, 2.5, 5, and 7.5 wt%) relative to the weight of PVDF. For pure PVDF film, the same procedure was followed without TM-0.15. The overall procedure is illustrated in Scheme 2.

Scheme 2 Preparation of the PVDF film and fabrication of the PENG.

2.4. Fabrication of the piezoelectric nano-generator (PENG)

The prepared PVDF film was cut into 3 cm × 3 cm pieces and used as the active layer, which had a thickness of 0.1 mm. Each film was sandwiched between two commercially available 0.02 mm-thick aluminium sheets, with electrodes of identical dimensions on one side and 2.5 cm × 2.5 cm on the other to prevent a short circuit. Copper tape was attached to provide electrical contact. The assembled device was laminated to protect it from the external environment; the thickness of the lamination sheet was 0.08 mm + 0.08 mm. The final thickness of the PENG device was 0.28 mm.

2.5. Material characterization

An X-ray diffractometer (XRD, Rigaku D/Max-II; Japan) was used to obtain diffraction patterns for analysing the phase purity of the synthesized materials. The measurement was conducted using Cu Kα radiation (λ = 1.54 Å) at 30 kV and 50 mA, over the 2θ range of 20° to 80° at a scan rate of 3° min⁻¹. X-ray photoelectron spectroscopy (XPS; VGS Thermo Scientific K-Alpha, USA) was performed using Al Kα radiation as the excitation source to determine the chemical composition. Surface morphology study and elemental analysis were performed using a Hitachi S-3400N scanning electron microscope (SEM) coupled with energy-dispersive X-ray spectroscopy (EDS; Oxford Instruments). Fourier-transform infrared (FTIR) spectra were recorded using a Bruker spectrometer with attenuated transform reflectance (ATR) mode.

2.6. Electrode and PENG

Binder-free electrophoretic deposition was used to prepare the working electrodes. 50 mg of the material was dispersed in 40 ml of acetone–iodine solution, 8 mg (0.2 mg ml⁻¹) of I₂ was dissolved, and the mixture was sonicated for 30 min to ensure complete dispersion. Iodine was used as a charging agent, which positively charged the medium-entropy particles, leading to coating on the negative-potential electrode. Ni foam was used as both the positive and negative electrode. Deposition was performed at 15 V for 5 min, with a 2 cm electrode spacing. Residual iodine was removed by drying the sample at 200 °C under Ar for 30 min.²⁵

3. Results and discussion

3.1. Structure and morphology analysis

Room-temperature X-ray diffraction (XRD) was performed on the synthesized MEPOs to examine their crystal structures. The diffraction patterns confirmed that all characteristic peaks correspond to an orthorhombic perovskite structure indexed to the Pbnm space group, consistent with COD ID: 2310587. The identified diffraction peaks at specific 2θ values include (110), (111), (112), (120), (202), (220), (221), (222), (312), (224), (314), and (240) planes. The interplanar spacing (d) was calculated using Bragg's law:where λ represents the wavelength of X-ray (0.154 nm), θ denotes Bragg's angle, d represents interplanar spacing, and n is the order of diffraction. The d-spacing calculated from the main diffraction peak for LTMFC (32.36°) and TM-0.15 (32.29°) was 2.76 Å and 2.77 Å, respectively. This shows that the lattice expansion happened with TM-0.15. Fig. 1(b) shows the shift in TM-0.15 to lower 2θ, confirming the lattice expansion in TM-0.15.²⁶ The morphological features captured by SEM are presented in Fig. 1(c and d), and Fig. S1(a–c) of TM-0.15 and LTMFC at different magnifications show that both samples possess similar structural morphologies. A larger agglomeration of particles will degrade the material's charge storage performance by blocking available active sites. Therefore, from the SEM image, it is observed that the material is densely packed into clusters with minimal agglomeration. Elemental mapping in Fig. 2(a–g) shows the presence of all the input metals, which were evenly distributed in the sample.

Fig. 1 (a) XRD of ball mill synthesised MEPO, (b) enlarged high intensity peak between 32˚ and 33° 2θ, and (c and d) SEM image of TM-0.15 at different magnifications.

Fig. 2 (a) Elemental mapping of the selected area of TM-0.15, (b–g) mapping image of elements La, Ti, Mn, Fe, Co, and O, (h) EDS spectrum of TM-0.15, inset showing the atomic percentage of elements.

Fig. 2(h) presents the EDS spectrum of TM-0.15. The inset table shows the atomic percentage of the elements present. Fe and Co are present at higher atomic percentages than Ti and Mn, indicating that the desired composition of TM-0.15 is achieved. Fig. S1(k) presents the EDS spectrum of LTMFC; the inset table shows the atomic percentages of elements present. The atomic percentage of Ti and Mn is higher than that of TM-0.15, which confirms the difference in the molar ratio between LTMFC (equimolar) and TM-0.15 (non-equimolar). Collectively, the compositional analysis verifies that the synthesised medium entropy material adheres to the designed equimolar and non-equimolar stoichiometries, confirming precise elemental incorporation.

Fig. 3, S2, S3, and S4(a) represent the X-ray photoelectron spectroscopy (XPS) survey spectra of TM-0.15, LTMFC, TF-0.15, and TC-0.15, respectively, with distinct peaks at 834 eV, 458 eV, 642 eV, 711 eV, 780 eV, and 529 eV corresponding to La, Ti, Mn, Fe, Co, and O, respectively.^27,28 The presence of all these elements confirms the successful formation of the designed medium-entropy oxide, consistent with the elemental composition obtained from EDS. Fig. 3(b) presents the deconvoluted Ti 2p XPS spectrum of TM-0.15, showing the presence of Ti³⁺ along with Ti⁴⁺. Fig. S2, S3, and S4(b) present the deconvoluted Ti 2p spectrum of LTMFC, TF-0.15, and TC-0.15, which shows that Ti⁴⁺ is present. The reduction of Ti⁴⁺ to Ti³⁺ in TM-0.15 arises from the formation of oxygen vacancies. The creation of an oxygen vacancy involves the removal of an O²⁻ ion from the lattice, generating two excess electrons that are subsequently transferred to nearby atoms to maintain local charge neutrality.²⁹ The reduction mechanism of Ti is illustrated with eqn (3) and (4).

2Ti⁴⁺ + 2e⁻ → 2Ti³⁺ (4)

Fig. 3 XPS spectrum of TM-0.15 (a) survey spectrum, (b) Ti, (c) Mn, (d) Fe, (e) Co, (f) O 1s.

The reduction of Ti⁴⁺ → Ti³⁺ increases the ionic radius of Ti, which in turn expands the local lattice spacing in TM-0.15. This lattice expansion is consistent with the observed shift in the diffraction peaks towards lower 2θ, which is shown in Fig. 1(b). The increase in lattice space facilitates ion intercalation and enhances electrical conductivity by lowering the diffusional resistance within the crystal structure.³⁰ An increase in oxygen vacancies in the system increased the number of active sites for faradaic reactions, boosting electrochemical activity.³¹ Additionally, the presence of these vacancies increased the unit cell volume.³²Fig. 3(c) shows the Mn 2p_3/2 spectra, which exhibit two peaks at 641.2 eV and 642.9 eV, corresponding to Mn⁴⁺ and Mn³⁺, respectively. Changes in the molar ratio of B-site elements alter the valencies of elements in the same site. Fig. 3(d) represents the spectra of Fe 2p with Fe²⁺, and Fe³⁺, at 710.2 eV and 712.2 eV, respectively. The ratio of +3 valence is increased in TM-0.15, and the increase in the ratio of higher oxidation number facilitates higher redox performance.³³Fig. 3(e) presents the spectra of Co, which exists in two oxidation states, Co⁺³ and Co⁺², at 779.83 eV and 782.28 eV, respectively.³⁴ From the above XPS results, it is confirmed that the elements are present in multiple oxidation states, supporting multiple redox processes and enhancing electrochemical performance. Fig. 3(f) shows that the O 1s spectra of TM-0.15 are split into (O_I) lattice oxygen at 529.1 eV, (O_II) oxygen vacancy at 530.2 eV, and (O_III) surface adsorbed oxygen at 531.6 eV.^33,35 Oxygen vacancies play a crucial role in modulating the material's electrical conductivity. A high concentration of oxygen vacancies not only promotes OH⁻ adsorption but also accelerates surface redox reactions, thereby boosting the overall electrochemical activity of the material.

Reports demonstrate that the O_II/O_I ratio is a reliable metric for estimating oxygen vacancy density.^36,37Table 1 presents the percentages of O_I, O_II, and O_III of all the prepared samples. LTMFC, TM-0.15, TF-0.15, and TC-0.15 show oxygen vacancy concentrations of 14.1%, 33.0%, 21.7%, and 16.4%, respectively. This indicates that O_I and O_III decreased in TM-0.15, suggesting that oxygen is less abundant throughout the system. Conversely, the O_II percentage increased in TM-0.15 relative to the other prepared samples due to the absence of oxygen in the lattice. This result is consistent with the atomic percentage of elements obtained from the EDS spectrum. Oxygen vacancies disrupt the crystal lattice, increasing electron mobility and conductivity. This facilitates faster charge transfer during redox reactions, creates active sites for ion adsorption and electron exchange, and ultimately improves overall performance.³⁸

Table 1 The percentage of oxygen vacancies in the prepared samples was determined using XPS

Material	% OI	% OII	% OIII
LTMFC	48.5	14.1	37.3
TM-0.15	44.6	33.0	22.4
TF-0.15	58.7	21.7	19.5
TC-0.15	56.2	16.4	27.4

---

3.2. Electrochemical performance of the MEPO

The electrochemical performance of the prepared material was evaluated using a three-electrode setup. 3 M KOH was used as the electrolyte, and a material-coated Ni foam was used as the working electrode. Fig. 4(a) presents the electrochemical data of the prepared MEPOs. The CV potential window is fixed to 0–0.5 V. CV is performed at 5 mV s⁻¹. The clear, distinct oxidation and reduction peaks at 0.33 V and 0.15 V, respectively, indicate pseudocapacitive behaviour and a reversible redox process in all prepared MEPOs. Fig. 4(b) presents the CV of TM-0.15 at various scan rates from 5 mV s⁻¹ to 100 mV s⁻¹. As the scan rate increased, both the current and area under the CV curve increased, indicating the enhanced charge–discharge kinetics of the electrode. However, the specific capacitance (C_sp) decreased at higher scan rates due to the limited time available for ion diffusion. Under these conditions, OH⁻ ions could not penetrate deeply into the interior active sites of the electrode, leading to incomplete charge storage. In contrast, at lower scan rates, K⁺ ions had sufficient time to diffuse into the porous structure, enabling full utilization of the internal electroactive sites and resulting in higher C_sp.³⁹ The distinct redox peak at higher potentials indicates the contribution of Faradaic pseudo-capacitance to charge storage. With increasing scan rate, the anodic and cathodic peaks shift slightly towards the more positive side and negative side, respectively, reflecting enhanced polarization effects.

Fig. 4 Electrochemical analysis of LTMFC and TM-0.15: (a) CV @ 5 mV, (b) CV of TM-0.15 @ different scan rates from 5 mV to 100 mV, (c) discharge curve @ 1 A g⁻¹, (d) discharge curve of TM-0.15 from 1 A g⁻¹ to 10 A g⁻¹, (e) capacity difference at different current densities, (f) EIS (inset fitted circuit for TM-0.15).

Despite this shift, the overall CV retains its characteristic shape, indicating fast, reversible redox kinetics and excellent electrochemical stability.^36,40 The oxidation of transition metals in the B-site from M⁺²/M⁺³ → M⁺³/M⁺⁴ shows a peak between 0.3 and 0.4 mV, and the reduction peak between 0.2 and 0.1 mV is caused by the reduction of M⁺⁴/M⁺³ → M⁺³/M⁺².^41,42 The specific capacitance (C_sp) value is calculated from eqn (5).

where i, Δt, m, and ΔV represent discharge current (A), discharge time (s), active mass loading (g), and potential window (V), respectively.

Discharge curves of MEPOs displayed in Fig. 4(c) from which the capacitance is calculated, TM-0.15 has the highest capacitance of 526 F g⁻¹; meanwhile, LTMFC, TF-0.15, and TC-0.15 show 486 F g⁻¹, 415 F g⁻¹, and 420 F g⁻¹, respectively. Fig. 4(d) shows the GCD discharge curve of TM-0.15, indicating that capacitance decreases with increasing current density from 1 A g⁻¹ to 10 A g⁻¹. It shows a higher capacitance of 526 F g⁻¹ at 1 A g⁻¹, dropping to 442.6 F g⁻¹ at 2 A g⁻¹ and 200 F g⁻¹ at 10 A g⁻¹. The results are consistent with CV. Table 2 compares the performance with previous reported results and shows that TM-0.15 exhibits higher performance than most of them. Fig S3(a–c) presents the CV and S3(d and e) presents the discharge curve of LTMFC, TF-0.15, and TC-0.15 at different scan rates and different current densities, respectively. These figures show behaviour similar to TM-0.15, with lower capacitance, which might be due to a change in oxygen vacancy concentration. Fig. 4(e) illustrates the reduction in specific capacitance with increasing current density for all prepared MEPOs.

Table 2 The energy-storage comparison between this work and previously reported results

Material	Synthesis method	Electrolyte	Capacitance	Ref.
(CoCrFeMnNi)3O4	Reverse co-precipitation	2 M KOH	239 F g^−1 at 0.5 A g^−1	47
(Co0.2Cr0.2Fe0.2Mn0.2Ni0.2)3O4	PAA gel	6 M KOH	384 F g^−1 at 1 A g^−1	48
Porous (FeCoCrMnNi)3O4	Self-combustion	1 M KOH	204 F g^−1 at 0.5 A g^−1	49
(FeCoCrMnNi)3O4	Sol–gel	1 M KOH	332.2 F g^−1 at 0.3 A g^−1	50
(Mg0.21Cr0.21Mn0.21Fe0.21Cu0.16)3O4	Sol–gel	1 M KOH	241 F g^−1 at 1 A g^−1	41
La1−xSrx(Co0.2Mn0.2Fe0.2Ni0.2Cu0.2)O3	Sol–gel	3 M KOH	200.92 F g^−1 at 1 A g^−1	51
La(CoCrFeMnNiAlx)1/(5+x)O3	Coprecipitation	2 M KOH	353.6 F g^−1 at 1 A g^−1	52
HEO@-40Ni	Freeze casting and self-combustion	1 M KOH	942 F g^−1 at 1 A g^−1	53
La(Ti 0.15 Mn 0.15 Fe 0.35 Co 0.35 )O 3	Ball-milling	3 M KOH	526 F g ^−1 at 1 A g ^−1	This work

---

The increased capacitance in TM-0.15 stems from oxygen vacancies that promote strong OH⁻ adsorption and vacancy-coupled faradaic reactions, as detailed in Scheme 3. During charging, OH⁻/H₂O from the electrolyte adsorbs at oxygen vacancies, facilitating H⁺ insertion and triggering the proton-coupled electron transfer reaction among B-site cations. These reversible transitions, supported by a disordered multi-cation environment, enable a reversible pseudocapacitive cycle during discharge. The abundance of oxygen vacancies boosts active-site density, ion accessibility, and redox kinetics, resulting in improved capacitance.⁴³ In addition to the oxygen vacancy, the partial reduction of Ti⁴⁺ is accompanied by defect formation. Such mixed-valence Ti states are known to introduce redox-active centers and enhance electrical conductivity. In this context, Lopez, Maria C., et al.,⁴⁴ demonstrated that Ti³⁺ ions occupying defect or distorted coordination sites actively participate in reversible Ti⁴⁺/Ti³⁺ redox processes, leading to improved electrochemical performance in Ti-based frameworks. Therefore, the outperforming property of TM-0.15 is compared to other materials in the prepared series.

Scheme 3 Illustration of an oxygen vacancy acting as an active site for OH⁻ adsorption, promoting redox reactions and enhancing ion diffusion, resulting in improved capacitance with TM-0.15.

Electrochemical impedance spectroscopy was performed to determine the resistance and ion-transport mechanisms of the prepared medium-entropy perovskite oxides. The data were recorded at open-circuit potential over the frequency range of 10⁶ to 10⁻¹ Hz. The high-frequency intercept corresponds to R_s, which arises from the intrinsic resistances of the electrolyte, the current collector, and the contact interfaces. The semicircle observed in the mid-frequency region represents the charge transfer process at the electrode electrolyte interface, characterized by R_ct and C_dl. The low-frequency tail at approximately 45° reflects the Warburg impedance (R_W) associated with diffusion of electrolyte ions through the porous structure of the electrode.⁴⁵ The Nyquist plot of TM-0.15 and LTMFC is shown in Fig. 4(f), with the corresponding circuit of TM-0.15 illustrated in the inset. For all samples, the impedance spectra were fitted using an equivalent circuit comprising the solution resistance (R_s), charge transfer resistance (R_ct), double layer capacitance (C_dl), and Warburg resistance (R_W). Among the investigated composites TM-0.15 exhibited the lowest charge transfer resistance of 1.133 Ω, compared with LTMFC (1.483 Ω), TF-0.15 (1.638 Ω), and TC-0.15 (1.157 Ω). The reduced R_ct value for TM-0.15 indicates improved electrical conductivity and faster interfacial charge transfer kinetics relative to the equimolar composition.

To analyze the kinetics of the redox reaction on the electrode, CV at various scanning rates was evaluated using Conway's method. The quantitative contribution of pseudocapacitance is expressed through⁴⁶

i = av^b (6)

where i is the peak current, v is the scan rate, a and b are the constants; the above equation can be transformed into

log(i) = b log(v) + log(a) (7)

The log fitting of log i vs. log v yields a b-value of 0.168 as shown in Fig. 5(a). This value is significantly lower than 0.5, confirming that the charge storage process is predominantly diffusion-controlled, rather than surface capacitive.

Fig. 5 (a) Log current vs. log scan rates, (b) charge distribution at various scan rates, (c) deconvoluted CV at 50 mV s⁻¹ diffusion contribution, and (d) surface contribution, both shaded in red.

To evaluate the contribution ratio of pseudocapacitance, Dunn's method was employed.⁴⁶

where k₁ and k₂ are constants; k₁v and represent the capacitance behaviour and diffusion behaviour contribution, respectively. The above equation can be transformed into

Fig. 5(b) shows the surface and diffusion contribution of TM-0.15 at different scan rates. At a low scan rate, charge storage is dominated by diffusion 86% and a surface contribution 14% at 5 mV s⁻¹. At 60 mV s⁻¹, the diffusion and surface contributions change to 64% and 36%, respectively, due to the lack of time for ion intercalation into the electrode material at higher scan rates. It is clearly observed that the diffusion contribution decreases with increasing scan rate, while the surface contribution increases in the charge storage mechanism. Fig. 5(c) and (d) show the deconvoluted CV profile at 50 mV s⁻¹, confirming the dominance of diffusion and surface contribution, respectively, as shaded in red. From the electrochemical studies, it is observed that the prepared material can be used for charge storage purposes in the supercapacitors.

3.3. Characterization and energy-harvesting performance of PVDF films

The β-phase of PVDF is crucial for enhanced electroactive performance. Recent studies highlight that the incorporation of nanofillers into the PVDF matrix effectively promotes β-phase formation due to the ability to serve as nucleation sites. Pure PVDF primarily exhibits the α-phase, necessitating high electrical poling to align dipoles, which subsequently experience depolarization after the electric field is removed. Conversely, nanofillers can interact with polymer dipoles and ions, facilitating self-orientation and creating a net dipole moment in the PVDF film without requiring an external electric field.⁵⁴ PVDF powder is mixed with TM-0.15 at different weight percentages and checked for the piezoelectric nanogenerator (PENG) performance. TM-0.15, as shown in the above studies, has more oxygen vacancies, which in turn lead to a greater positive surface charge.⁴⁶ When TM-0.15 is introduced into the PVDF matrix, the surface positive charges on TM-0.15 and the –CF₂ group interact electrostatically, attracting the electronegative fluorine group toward TM-0.15 and causing chain alignment in a more trans conformation. This leads to β-phases, which are more electroactive phases in PVDF.⁵⁵ Different weight % of TM-0.15 were added to PVDF, and the voltage was measured using an oscilloscope.

IR spectroscopy is an effective method for evaluating the formation of β-phases in PVDF, as shown in Fig. 6(a). The percentages of the α and β phases were calculated using eqn (10), with absorbance peaks at 766 cm⁻¹ and 840 cm⁻¹ for A_α and A_β.⁵⁶

where K_α = 6.1 × 10⁴ for α-766 cm⁻¹ and K_β = 7.7 × 10⁴ for β-840 cm⁻¹. According to the above equation, a higher β + γ content was observed for the 5-TM (79.8%), whereas pure PVDF, 2.5-TM, and 7.5-TM were 76.5%, 76.4%, and 76.6%, respectively. The observed change can be attributed to the nucleating effect of TM-0.15, which promotes polymer-chain alignment and enhances β-phase formation through electrostatic interaction between the interface of PVDF and TM-0.15. Fig. 6(b) presents the XRD pattern of pristine PVDF and TM-0.15 incorporated PVDF films. Pristine PVDF exhibits a broad diffraction feature at 18.7°, characteristic of the α-phase with a largely amorphous nature, along with a peak at 20.7° corresponding to the (110)/(200) planes of the β-phase. With increasing TM-0.15, the α-phase related to the broad peak is suppressed, while the β-phase becomes sharper and more intense. Visually, 5-TM shows the disappearance of the α-phase and a dominant, narrow β-phase peak at 20.7°, consistent with FTIR results that indicate a higher β-phase fraction and enhanced crystallinity. The reduced FWHM and absence of amorphous halo confirm higher crystalline ordering in 5-TM compared to PVDF and other compositions. This arises from the trans (TTTT) conformation of the β-phase, which promotes denser chain packing and larger crystallites.^57,58Fig. 6(c) shows the open circuit voltage of the prepared PVDF film with varying weight % of TM-0.15. The alternative voltage produced by the PENG is measured with an oscilloscope; pure PVDF shows a maximum of 1 V at 0.5 kgf at 5 Hz. At 5% TM-0.15, the maximum voltage reached 2 V under the same conditions as PVDF; with 7.5%, it decreased to 1 V, indicating that the optimized percentage for maximum voltage is 5%. These results indicate the formation of the β-phase in the prepared films, and the higher voltage at 5% TM-0.15 is consistent with the higher β-phase observed in the FTIR spectrum. The enhancement in piezoelectric output is due to the improved interfacial interaction between the PVDF matrix and TM-0.15 at a moderate loading of 5 wt%, and the uniform dispersion increases charged interfaces that promote the dipole alignment in PVDF, as shown in Fig. 7(a). When it increased to 7.5 wt% the voltage decreased, which may be attributed to the aggregation of TM-0.15 in the PVDF matrix, as shown in Fig. 7(b). Such aggregation reduces the effective interfacial interaction between PVDF and the filler, which suppresses dipole alignment; consequently, piezoelectric output and dielectric constant decrease.⁵⁹Fig. 6(d) shows that with an increase in applied force, the charge separation inside the material is maximized at 1.5 kgf, corresponding to an open-circuit voltage of 5 V.

Fig. 6 (a) FTIR spectrum of PVDF with TM-0.15, (b) XRD of PVDF with TM-0.15, (c) open-circuit voltage of different wt% of TM-0.15 added PENG, and (d) open-circuit voltage of 5%TM-0.15 PENG at different forces (inset image of the PENG device).

Fig. 7 Schematic representation of the polarization of PVDF with TM-0.15 (a) at 5 wt% and (b) 7.5 wt%.

Dielectric properties are fundamental for understanding PVDF's piezoelectric properties, as measured using a Microtest 6623 precision impedance analyser. A dielectric polymer with a higher electric dipole moment yields improved piezoelectric response under external force, necessitating a substantial number of aligned polar crystal domains.^60,61 Dielectric constant values of the prepared films are shown in Fig. 8(a), representing the dielectric constant, and Fig. 8(b) represents the tangent loss. Compared with bare PVDF 2.5-TM, 5-TM, and 7.5-TM films, the films exhibit higher dielectric constants. PVDF has a dielectric constant of 14; at 2.5% dielectric, the value increases to 64, to 102 at 5%, and to 81 at 7.5%. The results show that, for piezo voltage, a higher open-circuit voltage is observed with 5-TM. This behaviour indicates that the nanofiller significantly affects the composite's homogeneity and dielectric constant due to strong interactions between the nanofiller's functional groups and the PVDF matrix, thereby enhancing interfacial interactions between the polymer and the filler.

Fig. 8 (a) Variation of dielectric constant and (b) tangent loss for all filler-added PVDF films.

Fig. S7 shows the SEM image of the prepared PVDF films. Fig. S7(a) shows a pure PVDF film with a plain, smooth surface and no pores. Fig S7(b) shows that the adhesion of 5 wt% TM-0.15 roughens the surface, and pores appear in the film, which is responsible for enhancing the piezoelectric response by promoting higher effective mechanical deformation and local stress concentrations in the PVDF film. This leads to improved polarization switching and more efficient charge generation.⁶²

The above results confirm that 5 wt% is the optimized percentage of TM-0.15 in PVDF for the effective piezo voltage. To evaluate the energy harvesting performance of 5-TM, the generated output was stored in commercial capacitors via a bridge rectifier, enabling continuous power delivery to small electronic devices. Fig. 9(a) illustrates the charging profiles of commercial capacitors with PENG. Upon repeated mechanical tapping, a 2.2 µF capacitor reaches its saturated voltage of 1.45 V in 25 seconds. With an increase in the capacitor's charge-storing ability, the time required increases, and PENG can charge the 4.7 µF capacitor to 0.9 V and the 10 µF capacitor to 0.4 V in the same 25 seconds. This shows that effective charging can be done with low-power systems. Fig. S6 presents the dependence of voltage, current, and areal power density on external load resistance (10 kΩ–30 MΩ). The maximum power density of 250 µW m⁻² is obtained at 5 MΩ, indicating effective impedance matching between the device and the external load. The comparatively low power density can be associated with the high internal resistance of the material. Fig. 9(b) shows the repeated charging and discharging of a 2.2 µF capacitor to demonstrate the repeated and steady power generation of 5-TM. From the OCP, the capacitor is rapidly charged to 1.6 V in less than 30 seconds, then allowed to discharge on its own without an external load to 0.6 V, and charged again with 5-TM. This shows that the prepared PENG can be repeatedly charged without performance loss. Fig. 9(c) shows the circuit connections used for charging the capacitor via a bridge rectifier to convert the AC voltage produced by the PENG to a DC voltage that can be used to charge the capacitors. The above results confirm the energy harvesting performance of 5-TM, and it can charge the capacitors.

Fig. 9 (a) Charging of different commercial capacitors with 5-TM PENG, (b) repeated charging and discharging of a 2.2 µF capacitor, (c) circuit connection for charging a capacitor with PENG (the inset image shows a symbolised replica of the real circuit).

4. Conclusion

We have successfully synthesized MEPOs at different molar ratios using the conventional ball-milling technique. From the XPS, it is found that a non-equimolar ratio alters the percentage of lattice oxygen, increases oxygen vacancy, and alters the valence state of the B-site metals. SEM shows that all prepared MEPOs have a similar structure, and the added elements are evenly distributed throughout the structure without agglomeration at a single site. The presence of oxygen vacancy in TM-0.15 reduces the fraction of Ti⁴⁺ to Ti³⁺, which alters the lattice size and is observed in XRD by a peak shift in TM-0.15, which is not observed in the rest of the prepared MEPOs. TM-0.15 shows a higher capacitance of 526 F g⁻¹ at 1 A g⁻¹, which is 7.5% higher than that of the equimolar LTMFC and also higher than other molar ratios of the prepared MEPOs. This justifies that the formation of oxygen vacancies increases the capacitance of TM-0.15. The prepared TM-0.15 was successfully employed as a nucleating agent in PVDF to enhance the β-phase formation. The SEM image shows that the surface becomes rough with the addition of TM-0.15, which increases polarisation switching and efficient charge generation. The addition of 5% TM-0.15 with PVDF results in a higher open-circuit voltage of 4.5 V. Repeated charging of the capacitor is demonstrated with 5-TM, demonstrating the prepared PENG's ability to charge repeatedly. The above study confirms that the prepared MEPO La(Ti_0.15Mn_0.15Fe_0.35Co_0.35)O₃ can be utilized as an electrode for supercapacitors and as an energy-harvesting device for future energy-harvesting and storage applications.

Author contributions

Ezhilarasan Murugesan: conceptualization, data curation, investigation, writing – original draft. Sanath Kumar: investigation, writing – review & editing. Yen-Pei Fu: investigation, funding acquisition, project administration, supervision, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: characterisation details and experimental data. See DOI: https://doi.org/10.1039/d5ta10024k.

Acknowledgements

The authors would like to thank the National Science and Technology Council of Taiwan for financially supporting this research under contract numbers NSTC 112-2221-E-259-002-MY3 and NSTC 114-2221-E-259-010. The authors thank Professor Yu-Shyan Lin, Department of Materials and Engineering, National Dong Hwa University, for providing the oscilloscope. The authors thank Bo-Han Li and Chen-Yen Kuo for their assistance with the experiments and reactions.

References

1. C. Hou, W. Yang, H. Kimura, X. Xie, X. Zhang, X. Sun, Z. Yu, X. Yang, Y. Zhang, B. Wang, B. Bin Xu, D. Sridhar, H. Algadi, Z. Guo and W. Du, J. Mater. Sci. Technol., 2023, 142, 185–195.
2. L. Wang, X. Gao, D. Wang, H. Shang, Y. Zhao and B. Zhang, Fuel, 2025, 381, 133639.
3. M. Islam, M. S. Hossain, B. Adak, M. M. Rahman, K. kubra Moni, A. S. M. Nur, H. Hong, H. Younes and S. Mukhopadhyay, J. Energy Storage, 2025, 107, 114838.
4. Reenu, Sonia, L. Phor, A. Kumar and S. Chahal, J. Energy Storage, 2024, 84, 110698.
5. X. Zhu, Y. Zeng, X. Zhao, D. Liu, W. Lei and S. Lu, EcoEnergy, 2025, 3, e70000.
6. S. A. Beknalkar, A. M. Teli, V. V. Satale, T. S. Bhat, R. U. Amate, P. J. Morankar and J. C. Shin, J. Alloys Compd., 2025, 1024, 180169.
7. K. Krishnamoorthy, P. Pazhamalai, S. Manoharan, N. U. H. Liyakath Ali and S. J. Kim, Carbon Energy, 2022, 4, 833–855.
8. Z. L. Wang and J. Song, Science, 2006, 312, 242–246.
9. J. Bae, Y. J. Park, M. Lee, S. N. Cha, Y. J. Choi, C. S. Lee, J. M. Kim and Z. L. Wang, Adv. Mater., 2011, 23, 3446–3449.
10. X. Xue, S. Wang, W. Guo, Y. Zhang and Z. L. Wang, Nano Lett., 2012, 12, 5048–5054.
11. S. Meng, N. Wang and X. Cao, Materials, 2023, 16, 6916.
12. K. Shu, W. Li, Q. Wu, Y. Zong, C. Zhao, Y. Zhang and C. Wang, J. Mater. Chem. A, 2024, 12, 22372–22395.
13. M. B. Askari, P. Salarizadeh, M. T. Tourchi Moghadam, S. Azizi and M. H. Ramezan zadeh, J. Alloys Compd., 2025, 1027, 180573.
14. T. Kang, P. Nakhanivej, K. J. Wang, Y. Chen, Y. G. Chung and H. S. Park, J. Energy Chem., 2024, 94, 646–655.
15. P. M. Shafi, D. Mohapatra, V. P. Reddy, G. Dhakal, D. R. Kumar, D. Tuma, T. Brousse and J. J. Shim, Energy Storage Mater., 2022, 45, 119–129.
16. W. Bao, H. Shen, Y. Zhang, C. Qian, G. Zeng, K. Jing, D. Cui, J. Xia, H. Liu, C. Guo, F. Yu, K. Sun and J. Li, J. Mater. Chem. A, 2024, 12, 23179–23201.
17. J. Ma, T. Liu, W. Ye, Q. He and K. Chen, J. Energy Storage, 2024, 90, 111890.
18. C. M. Rost, E. Sachet, T. Borman, A. Moballegh, E. C. Dickey, D. Hou, J. L. Jones, S. Curtarolo and J. P. Maria, Nat. Commun., 2015, 6, 8485.
19. S. Jiang, T. Hu, J. Gild, N. Zhou, J. Nie, M. Qin, T. Harrington, K. Vecchio and J. Luo, Scr. Mater., 2018, 142, 116–120.
20. Y. Ning, Y. Pu, C. Wu, Z. Chen, X. Zhang, L. Zhang and B. Wang, J. Eur. Ceram. Soc., 2024, 44, 4831–4843.
21. J. Baek, M. D. Hossain, P. Mukherjee, J. Lee, K. T. Winther, J. Leem, Y. Jiang, W. C. Chueh, M. Bajdich and X. Zheng, Nat. Commun., 2023, 14, 5936.
22. M. Guo, Y. Liu, F. Zhang, F. Cheng, C. Cheng, Y. Miao, F. Gao and J. Yu, J. Adv. Ceram., 2022, 11, 742–753.
23. Y. Wang, J. Liu, H. Cao, J. Ding, N. Wang, C. He and Y. Zhang, Chem. Eng. J., 2025, 509, 161246.
24. J.-X. Li, W.-B. Zhang, B. Chen, J. Feng, B. Ashkar, J.-J. Jiang, J. Pang, H. Gou, N.-S. Zhou and X.-J. Ma, Dalton Trans., 2025, 54, 12417–12431.
25. A. Mittal, H. Verma, S. Kumar, P. Kumar, Y. P. Fu and S. Upadhyay, Chem. Eng. J., 2025, 522, 167342.
26. L. Zouridi, D. Totnios, L. Papoutsakis, E. Daskalos, G. Karagiannakis, G. E. Marnellos and V. Binas, ACS Appl. Nano Mater., 2025, 8, 3389–3401.
27. Z. Pan, J. Chen, X. Jiang, Z. Lin, L. Zhang, L. Fan, Y. Rong, L. Hu, H. Liu, Y. Ren, X. Kuang and X. Xing, Inorg. Chem., 2017, 56, 2589–2595.
28. B. Bharti, S. Kumar, H. N. Lee and R. Kumar, Sci. Rep., 2016, 6, 32355.
29. C. Qiu, Y. Odarchenko, Q. Meng, S. Xu, I. Lezcano-Gonzalez, P. Olalde-Velasco, F. Maccherozzi, L. Zanetti-Domingues, M. Martin-Fernandez and A. M. Beale, ACS Catal., 2022, 12, 9125–9134.
30. G. P. Sharma, P. K. Gupta, S. K. Sharma, R. G. S. Pala and S. Sivakumar, ACS Appl. Energy Mater., 2021, 4, 4758–4771.
31. H. Fu, M. Wang, Q. Ma, M. Wang, X. Ma and Y. Ye, Nanoscale Adv., 2022, 4, 2704–2712.
32. L. Shen, Z. Du, Y. Zhang, X. Dong and H. Zhao, Appl. Catal., B, 2021, 295, 120264.
33. T. X. Nguyen, Y. C. Liao, C. C. Lin, Y. H. Su and J. M. Ting, Adv. Funct. Mater., 2021, 31, 2101632.
34. Y. Pan, X. Xu, Y. Zhong, L. Ge, Y. Chen, J. P. M. Veder, D. Guan, R. O'Hayre, M. Li, G. Wang, H. Wang, W. Zhou and Z. Shao, Nat. Commun., 2020, 11, 2002.
35. Y. Liu, J. Hou, C. Cheng, F. Cheng, T. Su, Y. Miao, C. Ma and X. Wang, Ceram. Int., 2023, 49, 21546–21554.
36. X. Guo, X. Sun, H. Ju, L. Jin, L. Wang and S. Dong, J. Energy Storage, 2025, 130, 117444.
37. M. Christy, Z. A. Sheikh, S. A. Patil, J. Kwon, S. Choi, Y. B. Kim, U. Paik and T. Song, Adv. Sustainable Syst., 2025, 9, e00788.
38. S. Khaja Hussain and J. H. Bang, Phys. Chem. Chem. Phys., 2023, 25, 11892–11907.
39. M. Shahzad, F. Ahmad, M. Ibraheem, A. Shakoor, S. M. Ramay, M. R. Raza and S. Atiq, RSC Adv., 2025, 15, 6308–6323.
40. S. Sahoo, A. Sood, R. Kumar, A. Milton, S. M. Choi, Y. J. Jo, I. Hussain and S. S. Han, J. Energy Storage, 2025, 110, 115368.
41. K. Chillal, A. Halder, S. K. Jha, R. Patel and J. Valand, Ceram. Int., 2025, 51, 3890–3898.
42. B. Talluri, M. L. Aparna, N. Sreenivasulu, S. S. Bhattacharya and T. Thomas, J. Energy Storage, 2021, 42, 103004.
43. J. T. Mefford, W. G. Hardin, S. Dai, K. P. Johnston and K. J. Stevenson, Nat. Mater., 2014, 13, 726–732.
44. M. C. López, G. F. Ortiz, P. Lavela, J. L. Tirado, R. Stoyanova and E. Zhecheva, Chem. Mater., 2013, 25, 4025–4035.
45. A. C. Lazanas and M. I. Prodromidis, Am. Chem. Soc., 2023, 3, 162–193.
46. G. M. Tomboc, X. Zhang, S. Choi, D. Kim, L. Y. S. Lee and K. Lee, Adv. Funct. Mater., 2022, 32, 2205142.
47. B. Talluri, M. L. Aparna, N. Sreenivasulu, S. S. Bhattacharya and T. Thomas, J. Energy Storage, 2021, 42, 103004.
48. L. Li, P. Ji, C. Geng, Y. Li, L. Meng, B. Zhou, J. Liang, J. Peng and X. Su, J. Energy Storage, 2023, 73, 109182.
49. Y. Yin, W. Bin Zhang, B. Chen, J. Feng, J. L. Yang, Z. Q. Yang, J. J. Li and X. J. Ma, Ceram. Int., 2024, 50, 10292–10304.
50. Y. Yin, W. Bin Zhang, X. L. Zhang, M. M. Theint, J. L. Yang, Z. Q. Yang, J. J. Li, S. Liang and X. J. Ma, Dalton Trans., 2023, 52, 9005–9016.
51. Z. Y. Wan, J. B. Lv and Y. X. Zhang, Ionics, 2025, 31, 3671–3682.
52. M. Guo, Y. Liu, F. Zhang, F. Cheng, C. Cheng, Y. Miao, F. Gao and J. Yu, J. Adv. Ceram., 2022, 11, 742–753.
53. L. Z. Wei, S. J. Shi, M. Y. Guo, R. F. Guo and P. Shen, J. Energy Storage, 2025, 110, 115351.
54. A. Chatterjee, A. Das, K. Saha and U. Jeong, Adv. Sens. Res., 2023, 2, 2200080.
55. B. Amrutha, A. Anand Prabu and M. Pathak, Heliyon, 2024, 10, e29192.
56. M. Yasar, P. Hassett, N. Murphy and A. Ivankovic, ACS Omega, 2024, 9, 26020–26029.
57. Y. Zhou, W. Liu, B. Tan, C. Zhu, Y. Ni, L. Fang, C. Lu and Z. Xu, Polymers, 2021, 13, 998.
58. T. Nishiyama, T. Sumihara, Y. Sasaki, E. Sato, M. Yamato and H. Horibe, Polym. J., 2016, 48, 1035–1038.
59. X. He, W. Liu, H. Fu, C. R. Bowen, J. Liu, X. Li, A. R. Akbar, J. Zheng and W. Yang, ACS Appl. Mater. Interfaces, 2025, 17, 11144–11153.
60. N. Ahbab, S. Naz, T. B. Xu and S. Zhang, Micromachines, 2025, 16, 386.
61. S. Wang and Q. Li, IET Nanodielectr., 2018, 1, 80–91.
62. N. Shehata, R. Nair, A. Jain, M. Gamal, A. Hassanin, S. Noman, I. Shyha, K. Kruczała, M. Saad and I. Kandas, Sci. Rep., 2025, 15, 14073.

---