Light‐Written Nonvolatile Polarization via Defect‐Engineered Charge Trapping

ABSTRACT

Defect engineering offers a practical route to control interfacial charge states in polymer‐inorganic composites, yet translating this control into optically writable and non‐volatile polarization in soft dielectrics remains difficult. Here, we introduce a Polydimethylsiloxane (PDMS)‐based composite that can be written by light to form an interfacial polarization state. The design relies on FeTiO₃ (FTO) nanoparticles with oxygen‐vacancy associated trap states, whose density is tuned by spark plasma sintering to create a trap‐rich polymer‐oxide interface. Under illumination, photocarriers promote interfacial charge transfer by reducing the effective barrier at the metal‐composite contact, producing a rapid rise in interfacial charge accumulation. After the light is removed, a large fraction of the photoexcited electrons becomes immobilized, leaving a residual polarization that relaxes only slowly. We quantify the write‐relax behavior using a triboelectric nanogenerator configuration as a sensitive probe of interfacial charge transfer, and we directly visualize the photo‐written electrostatic state and its retention by Kelvin probe force microscopy. These results present defect‐mediated charge trapping as a materials‐level mechanism for light‐programmable, long‐retention polarization in soft composites, enabling remotely addressable electrostatic interfaces for soft electronic systems.

A soft Polydimethylsiloxane/FeTiO₃ (PDMS/FTO) composite enables light‐written and long‐retention interfacial polarization without high‐voltage poling. Oxygen‐vacancy traps, tuned by spark plasma sintering, support rapid photo‐assisted charge transfer during illumination and persistent trapped‐charge polarization after light‐off. Triboelectric probing and Kelvin probe mapping together reveal a stable, optically programmable electrostatic interface for soft electronic systems.

1. Introduction

Interfacial charge states and polarization in soft dielectrics are central to many technologies, including electrets, electrostatic actuators, triboelectric energy harvesting, touch sensing, and charge‐assisted adhesion [1, 2, 3, 4, 5, 6]. Conventional polymer electrets can retain charges for extended periods, but they are commonly fabricated by high‐voltage poling processes such as corona charging [7, 8, 9]. These methods increase processing complexity and can be difficult to implement in large‐area devices or systems integrated with mechanically fragile soft components. Ferroelectric approaches can generate stronger polarization, however, they often rely on rigid ceramic materials, which compromise mechanical compliance and increase fabrication complexity [10, 11]. These limitations motivate the development of alternative polarization strategies that are easier to program while maintaining the advantages of soft materials.

Light provides an attractive route for polarization because it can be applied without electrical wiring, is spatially localized, and rapidly switched on and off. The major challenge, however, is charge retention. In most polymer systems, photoexcited charges rapidly recombine or become screened at surfaces, causing the induced potential to decay after illumination is removed [12, 13, 14]. In polymer‐inorganic composites, semiconducting fillers can generate photocarriers under illumination, but long‐lived polarization requires trap states capable of stabilizing these carriers [15, 16]. The retention behavior of photo‐written states is therefore strongly influenced by heterogeneous interfaces, where trap density, trap depth, and charge leakage govern carrier dynamics.

Defect engineering provides a practical route to stabilize photo‐written charges because defects can introduce electronic states that capture carriers and suppress recombination [17, 18, 19]. In oxide semiconductors, oxygen vacancy‐related defects are among the most common and tunable sources of trap states, and their concentration can be adjusted through thermal treatment [20, 21]. In soft polymer‐oxide composites, a suitable defect landscape may contribute in two ways. During illumination, photocarriers can transfer more efficiently across heterogeneous interfaces and promote interfacial charge accumulation. After illumination is removed, a fraction of these carriers can remain trapped in deeper vacancy‐related states, making the written charge distribution less susceptible to thermal fluctuations and surface screening.

Here, we present a soft Polydimethylsiloxane (PDMS)‐based composite in which interfacial polarization can be written by light and retained for extended periods without high‐voltage poling. FeTiO₃ (FTO) nanoparticles were treated by spark plasma sintering (SPS) to increase oxygen‐vacancy‐related trap states while preserving the crystal structure. After incorporating the SPS‐treated FTO into PDMS, light illumination enhances charge transfer and increases charge accumulation at the interface. Meanwhile, some photoexcited electrons are captured by vacancy‐related trap states, leading to a retained polarization after the light is removed. The writing and retention behavior were monitored using a triboelectric nanogenerator as a probe of charge transfer dynamics. Kelvin probe force microscopy was further used to map the photo‐written surface potential. This defect‐engineering approach provides a simple route for light‐controlled programming of electrostatic states in soft composites and may be useful for soft electronic systems requiring rewritable and long‐retention interfacial functionality.

2. Results and Discussion

2.1. Structural and Defect Engineering of FTO for Charge Trapping

We first establish a defect‐engineered strategy to regulate the electronic structure of the inorganic filler. To this end, an oxide material capable of sustaining a high density of stable defect states was selected as the inorganic component [22, 23]. Transition‐metal titanates are known to host abundant oxygen vacancy‐related electronic states and exhibit strong defect polarization coupling, making them particularly suitable for defect‐mediated charge storage and interfacial polarization [24, 25, 26]. Spark plasma sintering (SPS) was employed as a non‐equilibrium thermal process to engineer oxygen vacancies in a controlled manner. Unlike conventional annealing, SPS combines rapid heating and a reducing local environment, which effectively promotes the formation of oxygen vacancy defects while suppressing excessive grain growth [27, 28]. This approach enables precise modulation of defect density without introducing secondary phases or compromising structural integrity. Figure 1a illustrates the defect evolution in the FeTiO₃ (FTO) lattice induced by SPS annealing. In the pristine structure, the metal‐oxygen framework remains largely stoichiometric, with oxygen atoms occupying their regular lattice sites. After SPS treatment, the rapid heating and highly reducing local environment promote the partial removal of lattice oxygen, leading to the formation of oxygen vacancies without disrupting the overall crystal structure. These oxygen vacancies act as intrinsic point defects, locally perturbing the electronic environment around neighboring metal cations. Figure 1b depicts the modification of the electronic band structure upon SPS annealing. In the defect‐free state, the Fermi level resides close to the mid‐gap region. The introduction of oxygen vacancies generates defect‐related electronic states within the band gap, effectively increasing the density of localized states near the conduction band edge and shifting the Fermi level upward. As a result, the electronic landscape becomes more trap‐rich and accessible for charge storage, providing favorable conditions for charge trapping and polarization under external stimuli.

FIGURE 1.

Defect engineering and structural evolution of the inorganic phase (FTO) induced by SPS annealing. (a) Schematic illustration of the crystal structure evolution and oxygen vacancy formation in FTO after SPS annealing. (b) Schematics of band‐structure variation upon SPS annealing. (c) Temperature‐resolved XRD maps of characteristic diffraction peaks. (d) SEM images of FTO annealed at different temperatures. Scale bar is 20 µm. (e) EPR spectra of untreated and SPS‐treated (700°C) FTO, showing enhanced oxygen‐vacancy‐related defect signals after SPS treatment. (f) Total and projected density of states (DOS) of pristine FTO. (g) Total and projected density of states (DOS) of vacancy‐associated FTO, showing defect‐induced electronic states near the Fermi level. (h) Surface SEM images and elemental oxygen mapping of the samples without annealing and after SPS treatment at 700°C. (i, j) Deconvoluted O 1s XPS spectra of the unannealed sample and SPS‐annealed (700°C) sample, highlighting the increased oxygen‐vacancy‐related states.

Figure 1c shows the temperature‐dependent X‐ray diffraction (XRD) evolution of the FTO phase during thermal treatment. All diffraction patterns can be well indexed to the standard FeTiO₃ phase (PDF#29‐0733) (Figure S1) [29], confirming that the overall crystal structure is preserved without the formation of secondary phases across the entire annealing temperature range. With increasing annealing temperature, several characteristic diffraction peaks, including (012), (104), (110), (024), and (116), gradually shift toward lower 2θ values. According to Bragg's law, this systematic shift indicates an expansion of the lattice spacing rather than a phase transition. The peak displacement becomes most pronounced at 700°C, suggesting a maximum lattice expansion at this temperature. Such lattice expansion is often associated with defect‐induced local lattice distortion and oxygen deficiency in transition‐metal titanates [30]. However, since peak shifts alone cannot uniquely identify oxygen vacancies, additional defect characterization was performed. Electron paramagnetic resonance (EPR) measurements reveal a characteristic signal at g ≈ 2.004, attributed to oxygen vacancy‐associated defect centers. Notably, SPS‐treated FTO exhibits a significantly enhanced second integral intensity (0.814) compared with the untreated sample (0.605), indicating an increased concentration of vacancy‐related paramagnetic centers (Figure 1e; Figure S2). Furthermore, density‐of‐states calculations reveal the emergence of defect states near the Fermi level after oxygen‐vacancy introduction, further supporting SPS‐induced electronic structure modulation (Figure 1f,g; Figure S3). These newly introduced states mainly originate from Fe and Ti orbitals, indicating that oxygen vacancies modify the local electronic environment of transition‐metal sites. The additional states near E_F provide energetically accessible trapping sites for carrier occupation and stabilization, which is consistent with the proposed defect‐mediated charge trapping mechanism.

Figure 1d presents Scanning Electron Microscopy (SEM)V images of FTO samples annealed at different temperatures, revealing the evolution of particle morphology after SPS treatment. The sample treated at 500°C exhibits relatively irregular particles with rough surfaces, suggesting incomplete microstructural development. Increasing the SPS temperature to 700°C leads to more compact particle packing and a more continuous microstructure, indicating enhanced particle rearrangement and local densification. Further increasing the temperature to 800°C results in the appearance of larger plate‐like grains and smoother surfaces, implying grain growth. High‐magnification SEM images (Figure S4) further reveal local surface evolution and grain‐boundary characteristics at different SPS temperatures. Figure 1h compares the surface morphology and oxygen distribution of the FTO particles before and after SPS annealing at 700°C. The corresponding O elemental maps reveal a noticeable change in oxygen distribution after annealing. To further probe the chemical origin of these changes, X‐ray Photoelectron Spectroscopy (XPS) measurements were performed. The survey spectra (Figure S5) confirm the presence of Fe, Ti, and O elements in both samples, with no detectable impurity species introduced by SPS processing. High‐resolution O 1s spectra (Figure 1i,j) provide evidence for defect evolution induced by SPS annealing. For the untreated sample, the O 1s peak can be deconvoluted into three components, corresponding to lattice oxygen (O²⁻), oxygen vacancies, and surface hydroxyl groups (─OH) [31, 32]. Quantitative fitting reveals that the fraction of oxygen‐vacancy‐associated species increases significantly after SPS treatment, confirming the effective generation of oxygen vacancies.

2.2. Enhanced Dielectric Polarization and Interfacial Charge Transfer

Since Fe‐based oxides may respond to both optical and magnetic stimuli [33, 34], it is necessary to clarify whether the observed polarization enhancement originates from magnetic effects or purely from photo‐induced processes. As seen in the magnetic hysteresis (M‐H) curves of the FTO annealed at different temperatures (Figure S6). All samples exhibit nearly linear M‐H behavior without evident magnetic saturation, indicating a weak magnetic response rather than ferromagnetism [35]. Such weak magnetization is insufficient to induce magnetically driven charge transport, ruling out magnetization‐dominated mechanisms reported in triboelectric devices. The relative permittivity of composites at different temperatures (Figure S7) shows a typical dielectric dispersion behavior, with permittivity decreasing gradually as frequency increases, reflecting the suppression of interfacial and dipolar polarization at high frequencies. Notably, the composite with 700°C‐annealed filler exhibits the highest relative permittivity, consistent with its higher defect density. Based on the optimized annealing condition, Figure S8 compares the dielectric response of composites with different FTO loadings. As the FTO loading increases, the relative permittivity exhibits a non‐monotonic dependence on filler content; it rises from pristine PDMS to moderate loadings and reaches a maximum at 12 wt.%, but decreases at higher loadings (15 wt.%). The initial enhancement at low‐to‐moderate filler contents can be attributed to the introduction of polarizable ceramic inclusions and the strengthened interfacial polarization, where charge accumulation and dipole formation are more effective. When the filler content becomes excessive, the permittivity decreases, likely due to particle agglomeration and reduced interfacial efficiency.

Triboelectric charge transfer during contact electrification is highly sensitive to surface potential and interfacial polarization [36, 37, 38]. Therefore, the output of a triboelectric nanogenerator (TENG) can provide a quantitative means to probe the evolution of trapped and polarized charges in dielectric materials (Figure S9). Here, the TENG is not primarily designed for energy harvesting. Instead, it serves as a platform to monitor how the incorporation of FTO modifies interfacial charge transfer and dielectric polarization behaviors. Before constructing the final TENG devices, key material and device parameters were systematically optimized, including the FeTiO₃ mass fraction in PDMS and the thickness of the spin‐coated composite film. The performance of TENG shows optimized output at 12 wt.%. (Figure S10) and 0.18 mm composite film thickness (Figure S11). This optimization arises from the balance between enhanced dielectric polarization and charge trapping at moderate FeTiO₃ loading, while avoiding particle agglomeration and excessive leakage at higher concentrations. Meanwhile, an intermediate film thickness ensures sufficient charge storage capability without introducing large internal resistance that hinders effective charge extraction.

Figure 2a shows the transferred charge between pure PDMS and Cu during periodic contact‐separation. The schematic energy diagram illustrates that electrons are partially transferred from Cu to PDMS upon contact, forming a weak interfacial dipole layer. After incorporating FTO nanoparticles into PDMS (Figure 2b), the transferred charge increases. The corresponding schematic depicts a reduced interfacial potential barrier, which facilitates electron injection from Cu into the composite. This enhanced charge transfer is attributed to the presence of semiconducting FeTiO₃ domains, which introduce additional trapping sites and local polarization, thereby strengthening the triboelectric output. To verify the origin of the enhanced charge transfer, the dielectric permittivity of pure PDMS and PDMS@FTO is compared in Figure S12. The composite exhibits a consistently higher relative permittivity over the entire frequency range, indicating strengthened dipolar polarization and contributing to higher charge storage capability.

FIGURE 2.

Charge transfer and polarization characteristics of PDMS@FTO composite films. (a) Transferred charge of pure PDMS‐based TENG and corresponding interfacial energy barrier diagrams. (b) Transferred charge of PDMS@FTO based TENG and schematic illustration of reduced interfacial barrier and enhanced electron transfer. (c) Transferred charge of PDMS@FTO based TENG under different magnetic fields and corresponding interfacial charge transfer schematics. (d) Transferred charge of PDMS and PDMS@FTO under different magnetic intensities. (e) Comparison of transferred charge under initial state, applied magnetic field, and after field removal. (f) Magnetization (M‐H) curve of FTO annealing at 700°C. (g) Surface potential of pure PDMS and PDMS@FTO films after repeated contact‐separation under different magnetic fields.

Figure 2c,d presents the transferred charge of PDMS@FTO‐based TENG under different magnetic fields (Figure S13). Introducing an external magnetic field produces only a slight enhancement in output, and the charge transfer increases with field strength. After removal of the magnetic field, the output rapidly returns close to its initial level (Figure 2e). The quantitative comparison confirms that the residual change in transferred charge after removing the magnetic field is small (∼0.9 nC). This behavior indicates that FeTiO₃ in the composite is weakly paramagnetic, consistent with its low saturation magnetization (Figure 2f). The minor field‐induced increase in transferred charge can be ascribed to subtle magnetically assisted carrier transport, while the absence of persistent enhancement after field removal confirms that magnetic ordering does not play a dominant role in the charge trapping process. To examine whether magnetic field exposure alters the surface charge state after repeated contact‐separation cycles, surface potential was tested via a non‐contact Infrared Potential Measurement (Figure S14). The PDMS@FTO film shows a slightly higher surface potential after contact electrification under a magnetic field compared to without a field, whereas pure PDMS exhibits nearly identical potentials regardless of magnetic environment (Figure 2g). These results demonstrate that incorporating FTO into PDMS enhances dielectric polarization and introduces additional interfacial trapping sites, leading to increased triboelectric charge transfer, and the magnetic field effects are observable but negligible.

2.3. Photo‐Induced Non‐Volatile Charge Trapping and Long‐Term Retention

To verify whether the defect‐assisted polarization in PDMS@FTO can be actively written and stored by external stimuli, optical excitation was introduced as a controllable route to generate carriers and populate trap states (Figure S15). The optical absorption spectrum (Figure 3a) shows broadband absorption from the UV to the near‐infrared region, indicating that the FeTiO₃ phase can efficiently harvest visible light and produce photocarriers within the composite. Consistent with this, when the PDMS@FTO film is integrated into a TENG and operated under periodic contact‐separation, illumination leads to a clear enhancement of transferred charge compared with the dark state, increasing from 15.6 to 28.1 nC (Figure 3b). The output of open‐circuit voltage and short‐circuit current also validates the charge results (Figure S16). Since illumination may induce a temperature increase and potentially affect TENG output, control experiments were performed to exclude possible thermal contributions. Despite a moderate temperature rise during illumination, transferred charge measurements conducted at controlled temperatures (20°C–35°C) exhibit negligible variation (Figure 3c), indicating that the enhanced output mainly originates from photo‐induced charge trapping. The write‐erase characteristics of this light‐controlled polarization were further examined by alternating dark and illuminated conditions (Figure 3d). Upon illumination, the transferred charge rises rapidly and reaches a stable plateau, reflecting a fast “writing” process. After switching off the light, the signal decreases but does not return to the initial baseline, leaving a pronounced residual level. Such partial relaxation implies that the illumination does not merely induce transient surface charging, but instead creates a metastable charge configuration in the absence of continued photo excitation. Quantitative analysis (Figure 3e,f) reveals that approximately 81% of the photo‐induced enhanced output remains after illumination is removed. This demonstrates that a large population of photoexcited electrons becomes trapped in the composite film, forming a persistent polarization state.

FIGURE 3.

Photo‐induced charge trapping and non‐volatile polarization in PDMS@FTO composites. (a) UV–Vis–NIR absorption spectrum of the PDMS@FTO composite film. (b) Transferred charge of the PDMS@FTO‐based TENG under illumination during periodic contact‐separation. (c) Transferred charge measurements of PDMS@FTO at different temperatures (20°C–35°C). (d) Transferred charge under alternating dark/illuminated conditions, showing reversible writing and residual output after light‐off. (e) Comparison of transferred charge in the initial dark state, under illumination, and after returning to dark. (f) Charge retention and transferred charge contribution analysis after light removal. (g) Schematic illustration of photo‐carrier generation and defect‐assisted electron trapping in the composite film: (i) Schematic of contact electrification with a lowered interfacial barrier after illumination and the corresponding charge‐transfer enhancement. (ii) Electrostatic induction and charge redistribution induced by trapped electrons, leading to residual polarization and electrostatic equilibrium.

As an additional control to exclude a magnetic‐origin effect, we monitored the magnetic domain contrast of the PDMS@FTO film before and after illumination. The domain‐like patterns and their spatial distribution remain essentially unchanged after light exposure (Figure S17), indicating that photo excitation does not induce detectable domain reconfiguration. The underlying mechanism is schematically illustrated in Figure 3g. Light absorption in FeTiO₃ generates electron‐hole pairs. Electrons are more likely to be trapped by defect‐associated states, including oxygen‐vacancy‐related traps, whereas holes tend to recombine or dissipate more readily. Once electrons are trapped, the composite behaves as a charged dielectric with an internal field. During subsequent contact electrification with Cu, electrostatic induction drives charge redistribution to screen this internal field, which manifests as an increased transferred charge in the external circuit. After the light is turned off, a large portion of trapped electrons remains immobilized, maintaining the internal polarization and giving rise to the non‐volatile photo‐written state.

In the experiment, we varied the illumination intensity during periodic contact‐separation (Figure 4a). For PDMS@FTO, the transferred charge increases with increasing light power, consistent with more efficient photocarrier generation and trap filling under stronger illumination. In contrast, the output of the pure PDMS‐based TENG remains nearly unchanged. To evaluate the temporal stability of the written state, a complete write‐relax cycle was recorded by turning the light on and off while continuously monitoring the transferred charge (Figure 4b). Upon illumination, the signal rises rapidly and reaches a stable plateau, indicating fast writing of a polarized state. After switching off the light, the transferred charge decreases but quickly converges to an almost constant level, evidencing a long‐lived residual polarization. The enlarged views in Figure 4c resolve the dynamics in each stage, including rapid build‐up during illumination, partial relaxation immediately after light‐off, and a stable baseline in the dark. Kelvin probe force microscopy (KPFM) provides direct spatial evidence of photo‐written charges on the composite surface (Figure 4d,e). Before illumination, the surface potential is slightly positive (+0.23 V). After photo‐writing, it shifts to a strongly negative value (−1.88 V), confirming substantial net negative charge/polarization generated by optical excitation (Figure S18). As the placement duration increases, the magnitude of the negative potential gradually decreases (to −1.61 V at 10 min and −1.40 V at 30 min), indicating partial relaxation of the surface potential under ambient conditions.

FIGURE 4.

Light‐intensity dependence and surface‐potential evidence of a photo‐written electret state. (a) Transferred charge of PDMS and PDMS@FTO under different illumination intensities. (b) Transferred charge evolution during a write‐relax cycle (light on/off), showing stable polarization and retention. (c) Enlarged views of representative time windows: (i) dark‐to‐light writing, (ii) steady illuminated state, (iii) light‐off relaxation, and (iv) residual polarization in the dark. (d) KPFM surface‐potential maps of PDMS@FTO before illumination and at selected times after illumination. Scale bars: 10 µm. (e) Statistics of the mean surface potential extracted from the KPFM maps.

The partial decay of surface potential suggests gradual release of carriers stored in relatively shallow trap states. Consistent with this interpretation, PDMS@ pristine FTO exhibits much faster charge decay after light removal, retaining only ∼33% of the photo‐induced charge contribution, whereas SPS‐treated samples preserve ∼81% (Figure S19). This substantial difference indicates that SPS treatment significantly improves charge‐retention capability. We attribute this behavior to oxygen‐vacancy‐mediated trap states introduced after SPS treatment. Under illumination, photogenerated carriers can be captured by defect‐associated states. Charges trapped in relatively shallow states gradually de‐trap and recombine after illumination is removed, resulting in partial decay. In contrast, deeper oxygen vacancy‐related traps generated by SPS treatment can immobilize carriers more effectively and suppress charge dissipation, thereby maintaining long‐lived polarization. Therefore, further improving the density and distribution of deep trap states through defect engineering and interface optimization may provide an effective strategy for enhancing long‐term polarization stability. Importantly, KPFM measures the local surface potential, which is governed by surface‐layer processes and is therefore expected to relax faster than the bulk‐stored polarization. The observed decay can be mainly attributed to tip‐induced readout perturbation during repeated KPFM scanning and thermally assisted release of charges from shallow traps. These changes are mostly due to surface‐level screening and charge movement, rather than a real loss of the internal trapped charges. This is demonstrated by the fact that a significant potential remains even after a long period, suggesting that deeper traps keep the material polarized like an electret.

3. Conclusions

In summary, we developed a soft PDMS‐based composite in which interfacial polarization can be written by light and retained after illumination without high‐voltage poling. Oxygen‐vacancy‐related trap states were introduced into FeTiO₃ (FTO) nanoparticles through spark plasma sintering while preserving the crystal structure. After incorporating SPS‐treated FTO into PDMS, illumination increases charge transfer and charge accumulation at the interface. Meanwhile, some photoexcited electrons are captured by vacancy‐related and interfacial trap states, leading to a retained polarization after the light is removed. The writing and retention behavior were monitored using a triboelectric probe and further verified by Kelvin probe force microscopy. Experimental and theoretical results indicate that oxygen‐vacancy‐related defects play an important role in stabilizing trapped charges and maintaining long‐lived polarization. Compared with untreated samples, SPS‐treated composites show much stronger charge‐retention capability, highlighting the importance of defect engineering in controlling interfacial electrostatic states. This work provides a simple strategy for programming electrostatic states in soft composites through light and may be useful for soft electronic systems requiring controllable and retained interfacial functionality.

4. Experimental Section

Detailed experimental procedures for SPS treatment of FeTiO₃, fabrication of PDMS@FTO composite films and TENG devices, materials characterization (XRD, XPS, SEM, MFM, AFM/KPFM, EPR), DFT calculation, and electrical/surface‐potential measurements under magnetic and optical stimuli are provided in the Supporting Information.

Author Contributions

Boxun Liu: conceptualization, methodology, data curation, investigation, formal analysis, writing – original draft. Jiahui Jiang: methodology. Yangtao Yu: validation. Jiayu Li: investigation. Shiyuan Zhao: writing – review and editing, project administration, resources, supervision, conceptualization. Huizhong Zeng: data curation. Lingyu Wan: conceptualization, writing – review and editing, project administration, resources, supervision, investigation. Xinchun Wang: methodology. Guanlin Liu: conceptualization, methodology, writing – review and editing. Xiaolong Hu: investigation. Yuanjin Zheng: writing – review and editing, supervision. Yahui Li: writing – review and editing, conceptualization, data curation, validation, investigation, supervision, methodology, visualization.

Ethical Statement

The authors declare that this study does not involve any human participants, animal experiments, or data requiring ethical approval.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma73970‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.