Schottky engineering of GDYO@Pt to boost piezoelectric and oxidative stress modulation for accelerated cranial regeneration

Read post on nature.com

IntroductionPiezoelectric semiconductor-based electrical stimulation offers a non-invasive alternative by converting mechanical stress into electrical energy under frequency- and intensity-controlled ultrasound (US) irradiation, generating adjustable electric fields for therapeutic intervention1,2,3. The majority of piezoelectric semiconductors including BaTiO3, ZnO, and Bi2WO6 for bone repair exhibit excellent piezoelectric responses but have wide bandgaps (>2.0 eV)4,5,6,7,8. The piezoelectric potential generation is accompanied by the generation of carriers that undergo redox reactions with water, resulting in ROS which cause oxidative damage to healthy tissues9,10. The overproduction of endogenous ROS and the presence of a hypoxic microenvironment at a bone-defect site severely impede healing, prolonging the course of electrical stimulation11,12. So it is essential to develop narrow-bandgap piezoelectric materials delivering efficient electrical stimulation and regulating oxidative stress simultaneously.The absence of a centrosymmetric structure is a fundamental characteristic for achieving piezoelectric properties13,14,15. Strategies including interlayer stacking, elemental doping, and electric field stimulation would reduce or break the inversion symmetry to induce piezoelectric responses15,16,17,18,19. Monolayer graphene exhibits no intrinsic piezoelectricity but multilayer stacking configurations disrupt its overall inversion symmetry, enabling piezoelectric effects through interlayer coupling and intralayer electron transition competition16,17,20. Oxygen as a dopant into graphene oxide further breaks the intralayer inversion symmetry via “clamp–release” structural distortions, enhancing the piezoelectric effect15,21. Charge redistribution in graphene nanostructures is driven by adjacent charged molecules, permanent dipoles, or built-in electric fields. This interaction can disrupt the inversion symmetry to amplify the piezoelectric performance22,23,24. Built-in electric fields, which arise from band bending or chemical potential gradients at heterojunction interfaces, have been shown to universally induce polar structures and substantial piezoelectric effects in semiconductors when interfaced with noble metals, as demonstrated by Yang et al.18,25. These findings underscore the feasibility of engineering nonpiezoelectric 2D carbon materials into high-performance piezoelectric systems through strategic modifications.Piezoelectric effect-driven electrical stimulation not only directly modulates cellular electrophysiological activity, but also its accompanying piezocatalytic reactions can directionally generate redox-active species (e.g., ROS, H2, or O2) through ultrasound-induced charge separation, dynamically regulating local oxidative stress levels26,27,28,29,30. The underlying mechanism lies in the fact that the valence/conduction band positions of piezoelectric semiconductors determine interfacial redox potentials (e.g., the conduction band ought to be positioned below the$$2{{{{\rm{H}}}}}^{+}+2{{{{\rm{e}}}}}^{-}\to {{{{\rm{H}}}}}_{2}\uparrow$$(1)reduction potential to drive hydrogen production, while the valence band should exceed the$${{{{\rm{H}}}}}_{2}{{{\rm{O}}}}+{{{{\rm{h}}}}}^{+}\to \,\cdot {{{\rm{OH}}}}+{{{{\rm{H}}}}}^{+}$$(2)and$${{{{\rm{H}}}}}_{2}{{{{\rm{O}}}}}_{2}+2{{{{\rm{h}}}}}^{+}\to {{{{\rm{O}}}}}_{2}\uparrow+2{{{{\rm{H}}}}}^{+}$$(3)oxidation potentials to trigger hydroxyl radical generation and hypoxia-alleviating oxygen supply)28,29,31. This provides a solution for precisely regulating oxidative stress via band engineering. During oxidative stress modulation, molecular hydrogen (H2) serves as both a safe therapeutic agent and an anti-inflammatory mediator, which selectively neutralizes highly toxic •OH radicals through its antioxidant activity32,33,34. The exceptional tissue permeability of H2 allows its effective diffusion into ROS-generating organelles (e.g., mitochondria and nuclei)33. However, hypoxia is known to upregulate hypoxia-inducible factor 1α (HIF-1α) in infiltrating immune cells, thereby promoting ROS production35,36,37. Concurrently addressing the issue of hypoxia during ROS scavenging may offer a more efficient strategy for managing oxidative stress and electrical stimulation therapy.Graphdiyne (GDY), a carbon-based two-dimensional material, exhibits exceptional structural stability and functional tunability due to its unique configuration of sp-/sp²-hybridized carbon atoms38,39. The sp-hybridized carbon (C ≡ C) bonds endow GDY with superior reducibility and electron transfer capability40,41, while its narrow bandgap ( ~ 1.22 eV, comparable to silicon’s 1.11 eV) further establishes it as an ideal platform for synergistic regulation of piezoelectric effects and catalytic reactions42,43.Herein, this study focuses on the development of a narrow-bandgap piezoelectric semiconductor heterojunction (GDYO@Pt) as a dual-function platform for piezoelectric stimulation and oxidative stress regulation (Fig. 1). In this system, the charge distribution in the graphdiyne oxide (GDYO) nanosheets is reconfigured by the interfacial dipole-induced built-in electric field within the depletion region. This is proposed to promote breaking of the structural inversion symmetry and the formation of asymmetric polar configurations, aiming to significantly enhance the piezoelectric performance. Positively polarized charges are generated at the Schottky junction interface under ultrasound, which should induce downward band bending and barrier reduction. It is expected to facilitate electron transfer from GDYO to Pt, and promote sustained H2 production. Concurrently, it is proposed that the holes oxidize H2O2 to generate O2, while the GDYO@Pt acts as a nanozyme to catalytically decompose H2O2 into O2. This strategy effectively alleviates hypoxia and scavenges ROS, overcoming the limitations of oxidative stress on piezoelectric stimulation. Cellular validation confirms the GDYO@Pt system’s modulation of BMSCs membrane potential, calcium influx, H2/O2 levels, and ROS clearance, verifying piezoelectric–catalytic synergy. Finally, GDYO@Pt-loaded thermoresponsive hydrogels are applied under ultrasound activation to promote defect repair via osteogenic differentiation, angiogenesis, and immunomodulation.Fig. 1: Schematic illustration of GDYO@Pt synthesis and synergistic therapy for bone defects.Fabrication protocol of GDYO@Pt nanosheets and their hydrogel composite. Enhancement mechanism and workflow of piezoelectric activation and catalytic therapy.Full size imageResultsMaterials preparation and characterizationGDY was synthesized according to a previously reported method44,45. As shown in Fig. 1, GDYO nanosheets were prepared via a modified Hummer’s method using H2O2/H2SO4 as oxidants. GDYO@Pt was subsequently fabricated by in situ reduction of K2PtCl4 with ascorbic acid. During this process, the strong affinity between the C≡C bonds of GDYO and the Pt2+ species enabled efficient capture and reduction of the Pt ions. Transmission electron microscopy (TEM) and atomic force microscopy (AFM) revealed nanosheet structures for both GDYO and GDYO@Pt, with lateral dimensions of ~500 nm and thicknesses of ~3 and ~7 nm, respectively (Fig. 2a, e, f). Pt nanoparticles ( ~ 3 nm in diameter) were uniformly distributed on the GDYO surface (Supplementary Fig. 1). Energy-dispersive X-ray spectroscopy (EDS) mapping confirmed the homogeneous distribution of C, O, and Pt in the prepared GDYO@Pt (Fig. 2b). Raman spectroscopy yielded characteristic peaks at 1950 and 2159 cm−1 (Fig. 2c), while Fourier-transform infrared (FTIR) spectroscopy gave a single peak at 2050 cm−1 (Supplementary Fig. 2), thereby confirming the retention of C≡C bonds in GDYO and GDYO@Pt. The X-ray diffractometry (XRD) results showed that GDYO@Pt exhibited diffraction peaks corresponding to the Pt standard card (PDF#04-0802), consistent with the lattice spacing observed in the high-resolution TEM image (Fig. 2d and Supplementary Fig. 3), and confirming the presence of metallic Pt. X-ray photoelectron spectroscopy (XPS, Fig. 2g, h and Supplementary Fig. 4) was employed to further elucidate the chemical states of the prepared GDYO@Pt. The distinct Pt 4f peaks were observed, corresponding to metallic Pt0 (Pt 4f7/2: 71.12 eV; Pt 4f5/2: 74.38 eV) and Pt2+ (Pt2+ 4f7/2: 72.54 eV; Pt2+ 4f5/2: 76.16 eV), indicating the predominant reduction of Pt2+ to Pt0, but with some residual surface oxidation. The high-resolution C 1s spectra recorded for both GDYO and GDYO@Pt displayed characteristic peaks corresponding to C=C, C≡C, C=O, and C–O bonds, confirming the structural integrity of GDYO following the incorporation of Pt (Fig. 2g, h). All C 1s peaks in GDYO@Pt shifted to higher binding energies than those in GDYO, suggesting electron transfer from GDYO to Pt during GDYO@Pt formation.Fig. 2: Characterization of nanosheets and thermoresponsive gels.a TEM image of GDYO and GDYO@Pt. b The elements scanning mapping of GDYO@Pt. c The Raman spectra of GDYO and GDYO@Pt. d High-resolution TEM of GDYO@Pt. e AFM image (3D topography) of GDYO and GDYO@Pt. f AFM image (2D) and size distribution curve of GDYO@Pt nanosheets (inset). High-resolution XPS spectra of g Pt and h C. i SEM images at magnifications from (I) low to (III) high, j elements scanning mapping, k FT-IR spectra and rheological testing including l shear-thinning (orange line) and step-shear (cyan line), m step-strain-sweep, n oscillatory strain-sweep, o oscillatory time-sweep, p temperature-dependent-sweep of GDYO@Pt gel. q Digital photographs of GDYO@Pt gel at room temperature (25 °C) and physiological temperature (37 °C). Representative images of three replicates per group are shown in (a, d, and i), respectively.Full size imagePoloxamer 407 (P407), a PEO-PPO-PEO triblock copolymer (PEO: poly(ethylene oxide); PPO: poly(propylene oxide)), requires concentrations ≥18% (w/v) to form thermoresponsive hydrogels46,47. While higher P407 concentrations (20–24%) lead to increased mechanical strengths, they reduce the sol–gel transition temperature from 26.03 ± 0.21 to 22.07 ± 0.21 °C, rendering the resulting hydrogels unsuitable for injectable applications46,48. Thus, a P407 concentration of 18% was selected for the purpose of the current study. Tannic acid (TA), which is a natural polyphenol with bioadhesive properties, was incorporated at varying concentrations (0–2%) to optimize crosslinking48,49. As shown in Supplementary Figs. 5 and 6 and Fig. 2q, TA concentrations ≥1% prevented gelation at 37 °C, although the incorporation of GDYO@Pt (0.5 mg/mL) maintained the thermoresponsive nature of the hydrogel. Thus, a TA concentration of 0.7% was selected for subsequent experiments.Scanning electron microscopy (SEM) of the lyophilized gels revealed porous network structures for both the gel (P407/TA) and GDYO@Pt gel (Fig. 2i, І). The gel exhibited uniformly distributed surface micropores, whereas the GDYO@Pt gel showed markedly reduced porosity, indicating homogeneous GDYO@Pt dispersion within the matrix (Fig. 2i, ІІ和ІІІ). This spatial uniformity was confirmed by EDS mapping, revealing consistent distribution of C, O, and Pt elements across the composite (Fig. 2j). The FTIR spectra confirmed successful modification, with peaks corresponding to the C=O (1730.5 cm−1) stretching vibrations of TA, and the C–O (1110.8 cm−1) and C–H (2889.5 cm−1) vibrations of P407 dominating the spectra (Fig. 2k). Rheological studies using the GDYO@Pt gel demonstrated its shear-thinning behavior (Fig. 2l), self-healing capability under step strain/shear tests (Fig. 2l, m), and optimal viscoelastic properties during strain/frequency sweep tests (Fig. 2n and Supplementary Fig. 7). Time-dependent phase transition measurements and tube inversion tests confirmed a sol–gel transition time of 5 s and full gelation within 3 min at 37 °C (Fig. 2o and Supplementary Fig. 8). Temperature ramps revealed a phase transition temperature of 26.3 °C, enabling liquid-state storage at room temperature and rapid solidification upon injection into the tissues (Fig. 2p). The visual observations supported this temperature-triggered phase transition, thereby confirming successful fabrication of the GDYO@Pt gel (Fig. 2q).Piezoelectric response characterization of the nanosheetsThe piezoelectric properties of the nanosheets were investigated using piezoresponse force microscopy (PFM), a technique that is widely employed for the high-resolution characterization of piezoelectric materials. The vertical piezoresponse amplitudes and phase images of the GDYO and GDYO@Pt nanosheets were shown in Fig. 3a, b, e, f. Both materials exhibited distinct contrast variations, indicative of their piezoelectric activities, with GDYO@Pt demonstrating an enhanced responsiveness. Under a ±6 V or ±10 V ramped voltage, distinct butterfly-shaped amplitude loops were observed, demonstrating that a consistent strain variation was induced by the applied electric field. The corresponding local piezoelectric hysteresis loops revealed an ~180° phase switching behavior, confirming the intrinsic piezoelectric nature of these materials (Fig. 3c, g). Quantitative analysis of the amplitude loop slopes demonstrated that Pt incorporation significantly enhanced the piezoelectric response of GDYO (Fig. 3d). Kelvin probe force microscopy (KPFM) measurements showed a surface potential of 41 mV for GDYO@Pt, which was markedly higher than that of GDYO (22 mV), attributed to piezoelectricity-induced electrical polarization effects (Fig. 3i–k). Ultrasound propagation in the medium generates periodic stresses and cavitation effects with localized pressures of up to 108 Pa50,51,52. Ultrasound-induced strain behavior was simulated via COMSOL under 0–108 Pa cyclic loading (T1–T5). As depicted in Supplementary Fig. 9, the material strain fluctuated regularly with the pressure cycle, and the maximum strain generated by GDYO@Pt at T3 was 0.0291%, which was significantly higher than that of GDYO (0.00417%). The piezoelectric potential of GDYO at T3 was 5 V, while that of GDYO@Pt reached 9 V, confirming the critical role of Pt nanoparticle modification in amplifying the piezoelectric performance of GDYO (Fig. 3m, n).Fig. 3: Piezoelectric and mechanistic of nanosheets.PFM including a, e Amplitude, b, f phase mapping, c, g the butterfly amplitude loop and phase curve and d the slope of the hysteresis loop of GDYO and GDYO@Pt. h The charge difference distribution (2D slice) of GDYO@Pt interface, red and blue indicate charge accumulation and depletion, respectively. i, j The potential maps in KPFM mode and k their corresponding potential amplitude of GDYO and GDYO@Pt. l The calculated dipole moment of Pt, GDYO and GDYO@Pt. The simulated potential distribution of m GDYO and n GDYO@Pt. o The potential data extracted from the potential distribution between GDYO and Pt. p Schematic representation of the polarization mechanism of GDYO@Pt nanosheets. The data were shown as the mean ± SD in (d) (n = 3 independent experiments). Statistical significance was determined using a two-tailed unpaired t-test.Full size imageSubsequently, the piezoelectric output and carrier dynamics of the materials were evaluated (Supplementary Fig. 10). GDYO@Pt exhibited significant charge separation and electric field output capabilities under ultrasonic irradiation (0.1–0.7 W/cm²), with a clear ultrasonic power dependency (Supplementary Fig. 10a, b, e, f). The average electric field strength generated by GDYO@Pt reached 499 mV/mm (99.8 mV, 0.83 MHz) at 0.7 W/cm², representing a 228% enhancement over GDYO’s 152 mV/mm (30.4 mV, 0.83 MHz) (Supplementary Fig. 10c, d, g, h). This indicates that the Schottky junction formed between GDYO and Pt can effectively enhance piezoelectric and carrier separation performance. The endogenous electric field of bone tissue ranges from 40 to 500 mV/mm53. An electric field of ≤500 mV/mm or piezoelectric stimulation of 20–890 mV/1 MHz can elevate intracellular Ca2+ concentration by activating voltage-gated calcium channels, thereby promoting osteogenesis and angiogenesis54,55,56,57,58. The output parameters of GDYO@Pt (499 mV/mm, 99.8 mV, 0.83 MHz) fully cover the biological effective window, indicating its critical potential to promote bone regeneration.Density functional theory (DFT) and finite element analysis were employed to elucidate the mechanism of the enhanced piezoelectric effect. The differential charge density diagram clearly revealed the out-of-plane charge transfer behavior between the Pt nanoparticles and GDYO (Fig. 3h), indicating the formation of a strong interfacial electric field through electron delocalization during the tight bonding process between the Pt and GDYO. The process of charge transfer was also one of charge redistribution. As shown in Fig. 3o, the contact interface between the GDYO nanosheets and the Pt metal particle exhibited a potential difference of ~3.3 mV, with adjacent areas of the nanosheets showing polarization-affected potentials, which may be attributed to charge redistribution. Figure 3l showed that along the a- and b-axes, the unit cell dipole moments of GDYO@Pt significantly increased to 82.03 and −176.59 Debye (D), respectively, which were 1.4-fold greater than those of GDYO (i.e., 57.67 and −132.24 D). These values confirmed the substantial enhancement in the degree of polarization of GDYO following the loading of Pt nanoparticles. As shown in Supplementary Fig. 11, the increased relative dielectric constant of GDYO@Pt quantitatively reflected the enhanced polar intensity, consistent with the aforementioned results. These results indicate that in the Pt-GDYO Schottky junction, interface charge transfer triggered by work function differences forms a dipole layer, and the resulting built-in electric field further induces charge redistribution within GDYO nanosheets. This process disrupts the central symmetry of the semiconductor, enabling more efficient separation of polarization charges under stress and thereby significantly enhancing the material’s piezoelectric performance (Fig. 3p).Piezocatalytic properties of the GDYO@Pt nanosheetsConsidering the distinct piezoelectric properties of the GDYO@Pt nanosheets, the piezocatalytic reactions induced by the US-activated nanosheets were evaluated. Using electron spin resonance (ESR) spectroscopy to detect the generation of ROS (Supplementary Fig. 12), it was found that no ROS production occurred under US stimulation, perhaps due to the insufficient redox potential of the band structure for ROS generation. The piezocatalytic H2 production performance of the nanosheets was evaluated using a methylene blue (MB) probe to detect the catalytic H2 generation (Fig. 4a, b and Supplementary Fig. 13). The GDYO@Pt solution containing the MB probe exhibited a time-dependent decrease in absorbance at 664 nm, with a more significant reduction observed compared to the other groups, indicating that US-activated H2 generation occurred in the GDYO@Pt system. Gas chromatography (GC) was used to quantify H2 production (Fig. 4c), and it was found that the GDYO@Pt + US group exhibited the highest H2 yield. Since Pt nanoparticles are known to catalyze the generation of O2 from H2O2, the O2 generation capacity was also monitored (Fig. 4d and Supplementary Fig. 14). It was found that GDYO@Pt acted as a nanozyme for oxygen evolution, with the GDYO@Pt + H2O2 + US group exhibiting significantly enhanced O2 production compared to the GDYO + H2O2 + US and Pt + H2O2 + US groups. These phenomena were attributed to Schottky junction formation between the Pt metal and GDYO, which enhanced the piezoelectric performance and carrier separation efficiency, further promoting the catalytic performance in both H2 and O2 generation.Fig. 4: Piezocatalytic behavior and mechanisms of GDYO@Pt nanosheets.a UV–vis–NIR spectra of the mixture of GDYO@Pt and MB under US stimulation at different times. b The ratio of absorbance at 664 nm of GDYO or GDYO@Pt mixed with MB under US stimulation at different time points. c The amount of H2 produced in each group of solutions was measured by GC. d Oxygen generation by GDYO, Pt and GDYO@Pt under H₂O₂ and ultrasonic treatment (On: 5 min; Off: 5 min, repeated for 5 cycles). Work function of e GDYO nanosheets, f GDYO@Pt nanosheets and g Pt nanoparticles. h Mott-Schottky curves of GDYO@Pt nanosheets. i The work function and band structure of GDYO and Pt. The j Schottky junction construction and k piezocatalytic mechanism in GDYO@Pt (SBH: Schottky barrier height). l Mott-Schottky curves of GDYO@Pt nanosheets tested with Ultrasonic Vibration. m The electron density distribution at the GDYO/Pt interface (green: electron-enriched regions; yellow: electron-depleted zones). n Laterally averaged electron density difference along the z-direction. o Adsorption energy of water and p Gibbs free energy diagram of hydrogen evolution for Pt, GDYO, and GDYO@Pt.Full size imageThe macroscopic energy band structures of GDYO and GDYO@Pt were analyzed by Mott–Schottky measurements (Fig. 4h and Supplementary Fig. 15) and ultraviolet-visible diffuse reflectance spectroscopy (UV-vis DRS) (Supplementary Fig. 16). Mott-Schottky analysis confirmed the n-type semiconductor characteristics of GDYO and GDYO@Pt, manifested by positive slopes with flat-band potentials (Efb) of −0.83 V and −1.03 V (vs. Ag/AgCl, pH 7) (Fig. 4h and Supplementary Fig. 15), corresponding to −0.63 V and −0.83 V vs. NHE at pH 7, respectively. After calibration to pH 0, the CB potentials were determined to be −0.22 and −0.42 V (vs. NHE, pH 0), both of which were more negative than the$${{{{\rm{2H}}}}}^{+}{+{{{\rm{2e}}}}}^{-}\to {{{{\rm{H}}}}}_{2}\uparrow$$(4)reduction potential (i.e., 0 V vs. NHE, pH 0), indicating efficient electron-driven H2 evolution. With bandgaps of 1.36 eV (GDYO) and 1.49 eV (GDYO@Pt), both materials retain the narrow-bandgap characteristic of GDY (Supplementary Fig. 16). The calculated VB potentials (1.14 V and 1.07 V vs. NHE, pH 0) exceeded the$${{{{\rm{H}}}}}_{2}{{{{\rm{O}}}}}_{2}+2{{{{\rm{h}}}}}^{+}\to {{{{\rm{O}}}}}_{2}\uparrow+2{{{{\rm{H}}}}}^{+}$$(5)oxidation potential (0.69 V vs. NHE, pH 0), satisfying the thermodynamic requirements for the hole-mediated generation of O2 from H2O2. Collectively, the band positions of narrow-bandgap GDYO and GDYO@Pt theoretically satisfied both H2 and O2 evolution criteria.The mechanism of piezocatalysis was further investigated. Ultraviolet photoelectron spectroscopy (UPS) was employed to determine the macroscopic work functions (W) of GDYO, Pt, and GDYO@Pt (Fig. 4e–g). The cutoff energies (Ecutoff) for GDYO, Pt, and GDYO@Pt were determined to be 17.93, 16.80, and 17.52 eV, respectively. Using the equation$$W=21.22-|E_{{{\rm{cutoff}}}}-E_{{{\rm{F}}}}|,$$(6)the calculated W values for GDYO, Pt, and GDYO@Pt were obtained, i.e., 3.29, 4.42, and 3.70 eV (relative to the vacuum level), respectively. The Pt nanoparticles exhibited a higher W value than GDYO, indicating that when GDYO was in intimate contact with the Pt nanoparticles, electrons transferred from GDYO to Pt. The space-charge region formed on the GDYO side caused upward bending of the energy bands, establishing a Schottky barrier (Fig. 4i, j). Upon ultrasonic excitation, the valence band (VB) electrons of GDYO were excited to the conduction band (CB) and transferred to the Fermi level of the Pt nanoparticles through the Schottky barrier, whereas holes remained in the VB of GDYO (Fig. 4k). The Efb and CB potentials of ultrasound-activated GDYO@Pt were −0.85 and −0.24 V (vs. NHE, pH 0), respectively, which were positively shifted by 0.18 V compared to those obtained under static conditions (Fig. 4h, l). This suggested that the piezoelectric effect in GDYO@Pt induced downward band bending and lowered the Schottky barrier height (SBH), thereby facilitating electron migration from GDYO to Pt and enhancing the H2 and O2 production (Fig. 4j, k).DFT calculations were performed to validate the electron transfer process (Fig. 4m, n). As shown in Fig. 4m, electrons predominantly accumulate on the Pt side (green regions), whereas the yellow peripheral regions of GDYO indicate electron depletion. Corresponding to the differential charge distribution, charge displacement (Δρ) in the GDYO@Pt system was quantitatively calculated (Fig. 4n). The negative and positive Δρ values observed for the GDYO and Pt regions, respectively, confirmed that the primary electron migration pathway was from GDYO to Pt, which was consistent with the above results (Fig. 2h).As shown in Supplementary Fig. 17, DFT simulations were used to visualize the top and side views of the H2O adsorption configurations on different materials. Compared to GDYO (−0.508 eV) and Pt (−0.613 eV), GDYO@Pt (−0.847 eV) was found to exhibit a stronger adsorption energy, indicating an enhanced surface affinity for water molecules and more abundant reaction sites (Fig. 4o). The Gibbs free energy of H2 adsorption (|ΔGH*|) serves as a critical descriptor for determining the H2 precipitation activity of a system, with values closer to zero indicating a greater reaction probability. The calculated |ΔGH*| value for GDYO@Pt was 0.15 eV, which was significantly lower compared to GDYO (0.40 eV) and Pt (0.28 eV) (Fig. 4p and Supplementary Fig. 18), thus the Schottky junction formed by GDYO and Pt effectively lowered the activation energy for H2 precipitation. The improved H2O adsorption capacity of GDYO@Pt along with its reduced H* activation energy led to more abundant surface reduction reactions, ensuring efficient piezocatalytic H2 evolution.The non-uniform Pt distribution on GDYO may induce interfacial heterogeneity, causing localized Schottky barrier variations and irregular carrier transport. Interface states and Fermi-level pinning are also critical concerns. A high interface state density can significantly increase carrier recombination rates, shortening carrier lifetime59. At extreme densities, Fermi-level pinning can fix the Fermi level, preventing the Schottky barrier from adjusting with work function changes and impairing charge separation efficiency60,61. As shown in Fig. 4e–g, the macroscopic work function of GDYO@Pt (3.70 eV) did not approach Pt’s value (4.42 eV), indicating moderate interfacial coupling without significant Fermi-level pinning and suggesting low interface state density. Compared to GDYO, GDYO@Pt exhibited a reduced electrochemical impedance arc radius, demonstrating lower charge transfer resistance and enhanced interfacial charge transfer efficiency (Supplementary Fig. 19). This further confirmed that nanoparticle distribution variations, interface states, and Fermi-level pinning minimally hindered charge transfer in this composite. The enhanced hydrogen/oxygen evolution activity substantiated improved charge separation/transfer efficiency with limited interference from non-ideal factors. Therefore, the work function difference-induced band bending serves as the core thermodynamic driver for charge separation. The non-ideal factors (e.g., interface states, particle distribution variations, Fermi-level pinning) may slightly modify barrier height/shape but cannot alter the fundamental nature of work function difference-driven charge separation in this system.Ultrasound-activated GDYO@Pt nanosheets promote cell proliferationThe cytocompatibility of the GDYO and GDYO@Pt nanosheets was subsequently evaluated using the Cell Counting Kit-8 (CCK-8) assay. The obtained results demonstrated that neither GDYO nor GDYO@Pt exhibited observable toxicity toward BMSCs, even at a high concentration of 500 μg/mL (Supplementary Fig. 20). Live/dead cell staining confirmed the excellent cytocompatibility of the nanosheets (Supplementary Fig. 21), consistent with the aforementioned results.To simulate a damaged microenvironment, oxidative stress injury was induced in BMSCs through the addition of H2O2 (0.2 mM). The oxidative stress-injured BMSCs were exposed to US at different power densities (0.3, 0.5, 0.7, 1, and 1.5 W/cm2) for 10 min to determine the optimal conditions. Compared to the control group, the BMSCs viability significantly decreased by day 3 in the 1.0 and 1.5 W/cm2 groups, whereas the other groups remained unaffected. However, good cell viability was maintained when US was applied at a power density of 0.7 W/cm2 for 10 min, indicating that this was a suitable power density and duration for subsequent experiments (Fig. 5a and Supplementary Fig. 22).Fig. 5: Piezoelectric stimulation of BMSCs in physiological media.a Viability of BMSCs cultured for 3 days under varied ultrasonic power (1.0 MHz, 50% duty cycle). b The cellular survival rates of various treatment groups, including control, ROS, US, GDYO, GDYO@Pt, GDYO + US and GDYO@Pt + US (US: 0.7 W cm−2, 10 min). c The current generated by GDYO@Pt under ultrasound irradiation at varying power levels (5-minute excitation duration) were rigorously quantified. d Schematic illustration of the decoupled experimental setup for detecting electrical signals generated by ultrasound-activated materials and applying electrical stimulation to cells. e The current stability of ultrasound-triggered samples in PBS at pH 7.4 was systematically assessed over five repeated ultrasound irradiation on/off cycles. f Semi-quantitative statistics (F/F0 represents the ratio of post-ultrasound fluorescence intensity to pre-ultrasound fluorescence intensity, with the inset displaying the statistical analysis of fluorescence intensities across experimental groups) and g fluorescence images of BMSCs stained with Di-8-ANEPPS after different treatments. h The current stability of ultrasound-triggered samples in PBS at pH 4.5 was systematically assessed over five repeated ultrasound irradiation on/off cycles. i Semi-quantitative statistics (F/F0 represents the ratio of post-ultrasound fluorescence intensity to pre-ultrasound fluorescence intensity, with the inset displaying the statistical analysis of fluorescence intensities across experimental groups) and j fluorescence images of BMSCs stained with Fluo-4 after different treatments. The data were shown as the mean ± SD in (a) and (b) (n = 4 biologically independent replicates). Statistical significance was determined using a two-tailed unpaired t-test.Full size imageThe effects of the ultrasound-activated nanosheets on BMSCs proliferation were further assessed using CCK-8 assays and live/dead cell staining. Compared to the GDYO@Pt and GDYO + US groups, the GDYO@Pt + US group exhibited significantly enhanced proliferation levels on days 3, 5, and 9 (Fig. 5b and Supplementary Fig. 23). This enhancement was attributed to the synergistic effects of the Schottky junction-induced piezoelectric stimulation and oxidative stress regulation in GDYO@Pt. The GDYO@Pt group showed a slight improvement in cell viability compared with the control group. This was likely due to the nanozyme activity of GDYO@Pt, which consumes overexpressed H2O2 in damaged cells to generate O2, thereby enhancing the cell viability.Piezoelectric stimulation of BMSCsThe piezoelectric properties and cycle stabilities of the nanosheets were systematically evaluated in physiological environments through ultrasound activation in a physiological medium. As shown in Fig. 5e, h, compared to a neutral pH environment (pH 7.4), a pH of 4.5 (simulating the bone-defect microenvironment) significantly enhanced piezoelectric current generation in the nanosheets. During five on-off cycling tests, the GDYO@Pt nanosheets exhibited stable current–output characteristics. The current intensity of the GDYO@Pt system reached 1.7 μA (pH 4.5), which was ~2.1-fold higher than that of the GDYO nanosheets (0.8 μA), consistent with the piezoelectric characterization results (Fig. 3 and Supplementary Fig. 10). Further evaluations revealed that the piezoelectric current of US-activated GDYO@Pt displayed significant power-dependent characteristics (0–0.7 W/cm2) (Fig. 5c). In bone electrophysiology, it has previously been reported that microcurrents in the 0.1–10 μA range effectively promote bone tissue regeneration62,63,64. This study confirmed that the piezoelectric current intensity generated by GDYO@Pt fell precisely within this range. Furthermore, the observed environmental adaptability and output stability of this system met the requirements for bone-defect repair applications, highlighting its significant potential in bone regeneration therapy.After evaluating the piezoelectric properties, the effects of piezoelectric stimulation on the electrophysiological activity of BMSCs were investigated. For this purpose, a spatially separated experimental design was adopted, wherein BMSCs were seeded in the central area (1.0 × 1.0 cm2) of the sterilized conductive glass, while the nanosheet-coated region was fixed at the bottom. To precisely assess the direct impact of the piezoelectric effect of the material on the cellular electrophysiology, the nanosheet-loaded electrode area was submerged in a physiological medium (pH 4.5), whereas the BMSCs-seeded electrode area remained exposed (Fig. 5d). The voltage-sensitive fluorescent probe Di-8-ANEPPS was used to monitor changes in the cell membrane potential induced by US-activated piezoelectric stimulation (Fig. 5f, g). This probe operates via an electrochromic mechanism, wherein its excitation/emission spectra shift in response to changes in the membrane potential65,66. The slow internalization properties of Di-8-ANEPPS ensured specific localization on the cell membrane surface, enabling the accurate detection of membrane dipole potential alterations67,68. As shown in Fig. 5f, g, the GDYO@Pt group exhibited greater fluorescence intensity changes than the GDYO group, with a significant increase in the fluorescence intensity ratio observed after US stimulation. This phenomenon directly correlated with the enhanced piezoelectric current output of GDYO@Pt (1.7 vs. 0.8 μA, as shown above), confirming that US-activated GDYO@Pt can modulate the electrophysiological activity of BMSCs by altering the cell membrane potential through superior piezoelectric stimulation.As a secondary messenger that regulates osteogenesis, intracellular Ca2+ signaling is closely linked to the electrophysiological activity69,70. To investigate the possible activation of calcium signaling pathways under the influence of GDYO@Pt piezoelectric stimulation, a Fluo-4 fluorescent probe was employed with the same spatially separated design approach described above (Fig. 5d, i, j). The experimental results revealed that the GDYO@Pt group exhibited a significantly enhanced Ca2+ fluorescence signal intensity under US stimulation, with a 1.6-fold increase in the fluorescence ratio (post-/pre-stimulation) compared to the control group. This indicated that the piezoelectric electric field generated by the material effectively induced intracellular Ca2+ enrichment in BMSCs.As shown in Supplementary Fig. 24, to directly assess the influence of ultrasound-mediated mechanical effects on cellular electrophysiology, the BMSCs-seeded electrode region was immersed in a physiological medium (pH 4.5) and exposed to ultrasound (Mechanical Effect Group, Mech). Compared with piezoelectric material groups (GDYO/GDYO@Pt), significantly weaker alterations in membrane potential and Ca2+ concentration were observed in this group pre- and post-stimulation, indicating that the pure mechanical effect of ultrasound exerted limited impact on the electrophysiology of BMSCs. Combined with the above results (Fig. 5f, g, i, j), the findings indicate that the corresponding mechanism likely involves piezoelectric stimulation-induced membrane depolarization, which regulates the voltage-gated calcium channel activity to promote extracellular Ca2+ influx. The specific activation of Ca2+ signaling pathways provides a critical ionic environment for BMSCs proliferation, osteogenic differentiation, and extracellular matrix mineralization.Evaluation of oxidative stress levels in vitroThe effects of nanosheet-based piezocatalysis on the oxidative stress levels in BMSCs were evaluated. Initially, the GC and MB probe were used to analyze intracellular H2 release. As shown in Fig. 6a, US-activated GDYO@Pt-treated BMSCs exhibited a significant time-dependent increase in H2 generation, with a hydrogen evolution rate far exceeding that of other groups. Further validation via the MB probe (Fig. 6b) revealed near-complete fading (loss of blue color) in the GDYO@Pt + US group, whereas other groups retained distinct blue coloration, visually confirming its superior hydrogen production activity. These results demonstrated the universal mechanism by which the Schottky heterojunction (GDYO@Pt) enhanced piezocatalytic hydrogen production across diverse environments, spanning from in vitro systems to cellular contexts.Fig. 6: Assessment of oxidative stress levels at the cellular level.a Quantitative assessment of ultrasound-activated GDYO@Pt for H₂ production in BMSCs and b qualitative evaluation of H₂ production using MB staining in BMSCs under different treatments including control, US, GDYO, GDYO@Pt, GDYO + US, GDYO@Pt + US (US: 1.0 MHz, 50% duty cycle, 0.7 W cm−2, 10 min). c Fluorescence images of BMSCs stained with [Ru(dpp)3]Cl2 (red) and DAPI (blue) after different treatments. d Fluorescence microscopy images of BMSCs stained with DCFH-DA (green) and DAPI (blue) after different treatments. US: 1.0 MHz, 50% duty cycle, 0.7 W cm−2, 5 min for (c) and (d). Representative images of three replicates per group are shown in (b–d), respectively.Full size imageAlthough H2 is beneficial for scavenging highly toxic hydroxyl radicals (i.e., ROS) in damaged cells, a hypoxic environment can lead to the upregulated expression of HIF-1α in the infiltrating immune cells, further inducing ROS production and creating a vicious cycle of oxidative stress. Alleviating hypoxia while scavenging ROS may lead to the more efficient resolution of oxidative stress in a bone defect microenvironment. As shown in Fig. 6c, the hypoxia-sensitive fluorescent probe Ru(dpp)3Cl2 was employed to evaluate nanosheet-catalyzed O2 generation in the BMSCs. Compared with the ROS group (oxidative stress model group), hypoxia was significantly alleviated in the GDYO@Pt, GDYO + US, and GDYO@Pt + US groups, with the GDYO@Pt + US group exhibiting the strongest hypoxia mitigation capability. This was attributed to the synergistic catalytic decomposition of H2O2 into O2 by the GDYO@Pt nanosheets and holes, thereby providing a safeguard for ROS regulation.The indicator 2′,7′-dichlorofluorescin diacetate (DCFH-DA) was used to investigate the effects of different treatments on the intracellular ROS levels. DCFH from DCFH-DA is initially hydrolyzed by cellular esterases, and is subsequently oxidized by ROS to generate the strongly green-fluorescent 2′,7′-dichlorofluorescein (DCF). As shown in Fig. 6d, compared to the ROS group, the GDYO@Pt group exhibited weakened green fluorescence, indicating its intrinsic ROS-scavenging capability owing to the ability of the GDYO@Pt nanozyme to decompose H2O2. More importantly, DCF fluorescence was almost absent in the GDYO@Pt + US group, demonstrating a significantly enhanced ROS depletion level that far exceeded that in the GDYO + US group. This result stems from the synergistic effects of piezocatalytic H2 and O2 production during the continuous ROS-scavenging process.Promotion of osteogenic differentiationTo evaluate the effects of US-activated GDYO@Pt on the osteogenic differentiation of BMSCs, assessments of the alkaline phosphatase (ALP) activity and Alizarin Red S (ARS) staining were conducted. ALP is a critical intracellular enzyme that serves as a marker of early osteogenic differentiation. As shown in Fig. 7a, b, while the US group exhibited slight increases in both ALP-stained area and activity after 7 days compared with the ROS group, the GDYO@Pt + US group showed significantly higher ALP-stained area and activity than those in the US, GDYO@Pt, and GDYO + US groups. Furthermore, the late-stage (14 d) mineralization of BMSCs was assessed using Alizarin Red S (ARS) staining. As shown in Fig. 7a, c, the GDYO@Pt + US group displayed the highest number of mineralized nodules and cellular calcium deposits, indicating superior mineralization compared with the other groups. These results further confirmed that US-triggered GDYO@Pt facilitated the osteogenic differentiation of BMSCs.Fig. 7: Osteogenic differentiation at the cellular level.a, b ALP (7 days) and a, c ARS (14 days) staining images and semi-quantitative analysis of BMSCs in different treatments including control, ROS, US, GDYO, GDYO@Pt, GDYO + US, GDYO@Pt + US. Relative mRNA expression levels of osteogenic genes from BMSCs after different treatments, including d ALP, e OCN, f OPN, g COL-1 and h RUNX-2. US: 1.0 MHz, 50% duty cycle, 0.7 W cm−2, 10 min. The data were shown as the mean ± SD in (b–h) (n = 3 biologically independent replicates). Statistical significance was determined using a two-tailed unpaired t-test.Full size imageThe molecular mechanisms underlying this enhanced osteogenic differentiation were elucidated by analyzing gene expression levels using the quantitative reverse transcription polymerase chain reaction (RT-qPCR). Key osteogenic markers including ALP, osteocalcin (OCN), osteopontin (OPN), type I collagen (COL-1), and runt-related transcription factor 2 (RUNX-2) were examined. As shown in Fig. 7d–h, the GDYO@Pt + US group exhibited significantly higher expression levels of these osteogenic genes than the other groups after 14 d, indicating a markedly augmented osteogenic differentiation capacity. This can be attributed to the synergistic effects of the enhanced piezoelectric stimulation and down-regulated oxidative stress at the Schottky junction.In vivo skull regeneration and repairTo systematically evaluate the osteoinductive potential of US-activated GDYO@Pt nanosheets in vivo, a nanocomposite hydrogel system was developed by encapsulating the nanosheets within a TA-modified poloxamer thermosensitive gel to enhance their retention at bone-defect sites. This composite gel exhibited suitable biosafety and viscoelasticity, with the optimized formulation enabling liquid-state behavior at room temperature (25 °C) and semi-solid-state behavior at physiological temperature (37 °C) (Fig. 2l–q and Supplementary Figs. 7, 8 and 25). According to the therapeutic protocol shown in Fig. 8a, a 3 mm diameter cranial defect model was created in the parietal bone of mice using a precision drill. Subsequently, an aliquot (300 μL) of pre-chilled (4 °C) gel, GDYO gel, or GDYO@Pt gel was precisely injected into the defect cavity via a micro syringe and maintained at 37 °C for 10 min to ensure in situ gelation. The surgical incisions were closed in layers, followed by postoperative analgesia and anti-infection treatments. Twenty-four mice were randomized into six groups (control, gel + US, GDYO gel, GDYO@Pt gel, GDYO gel + US, and GDYO@Pt gel + US; n = 4 per group) across two timepoints (6 and 12 weeks). The US parameters (1 MHz, 0.7 W/cm2, 10 min/d) matched those used in the in vitro studies, and treatment was applied for 4 weeks.Fig. 8: In vivo bone regeneration assessment.a Therapeutic protocol for calvarial defect repair in C57BL/6 mice. b Representative microscopic CT images of the cranial defect site at 6 and 12 weeks after implantation in different treatment groups. Scale bars: 0.5 mm. White circles represent defect boundaries. c The ratio of new bone to total area, d BMD and e BV/TV of cranial defect region in control, gel + US, GDYO gel, GDYO@Pt gel, GDYO gel + US and GDYO@Pt gel + US at different times. Representative f HE and g Masson’s trichrome staining images of the cranial defect areas across various groups at the 12-week post-implantation period (CT connective tissues, NB new bone). Typical immunohistochemical staining patterns demonstrating h OCN, i RUNX-2, and j CD31 expression in neoformed calvarial regions at 12 weeks post-implantation (black arrows: positive staining regions). k Immunofluorescence micrographs showing CD206 (red) and CD86 (green) expression in the defect region. US: 1.0 MHz, 50% duty cycle, 0.7 W cm−2, 10 min. The data were shown as the mean ± SD in (c–e) (n = 4 biologically independent replicates). Statistical significance was determined using a two-tailed unpaired t-test.Full size imageMicro-computed tomography (micro-CT) imaging and reconstruction revealed that the GDYO@Pt gel + US group achieved the highest degree of bone regeneration and repair at both 6 and 12 weeks compared to the other groups (Fig. 8b). Semi-quantitative analysis (Fig. 8c–e) indicated poor bone regeneration in the control, gel + US, and GDYO gel groups, with low new bone area ratios (%), bone volume/total volume ratios (BV/TV), and bone mineral densities (BMD) observed in each case. In contrast, the GDYO@Pt gel + US group exhibited superior repair metrics, reaching a new bone area of 75.37%, a BV/TV ratio of 69.00%, and a BMD of 0.658 g/cm3 at 12 weeks, significantly surpassing the other groups. Histological analysis using hematoxylin and eosin (H&E) and Masson’s trichrome staining confirmed these findings. More specifically, from H&E staining, the GDYO@Pt gel + US group displayed a continuous, well-organized neobone with abundant vascularization (Fig. 8f), whereas Masson’s staining highlighted mature osteoid deposition within the defect (Fig. 8g); the other groups showed limited discontinuous bone formation. Immunohistochemical (IHC) staining for osteogenic markers (OCN and RUNX-2; Fig. 8h, i) showed the strongest degree of staining in the GDYO@Pt gel + US group, consistent with the micro-CT and histological results. CD31 immunostaining (angiogenesis marker) revealed superior vascular protein expression in this group (Fig. 8j). These results were attributed to the synergistic effects of Schottky junction-enhanced piezoelectric stimulation and oxidative stress modulation.Immunofluorescence (IF) staining was subsequently performed to investigate the effects of the different treatment conditions on the inflammatory microenvironment at the defect site. M1 (CD86) and M2 (CD206) macrophages, representing pro-inflammatory and anti-inflammatory phenotypes respectively, play critical roles in regulating the osteogenic microenvironment. Although short-term inflammation facilitates the recruitment of endogenous stem cells, prolonged inflammation impedes bone repair. As shown in Fig. 8k, compared to the other groups, the GDYO@Pt gel + US group exhibited a significantly higher proportion of M2-type (CD206, red) macrophages and a lower proportion of M1-type (CD86, green) macrophages at the cranial defect site. This polarization toward an anti-inflammatory phenotype is conducive to bone tissue repair, thereby indicating that the US-triggered GDYO@Pt gel effectively orchestrated the osteogenic, angiogenic, and immune microenvironments to promote bone regeneration.Assessment of oxidative stress levels in bone defect tissuesConsidering the aforementioned in vivo bone repair outcomes, the mechanism underlying the US-activated GDYO@Pt-mediated remodeling of redox homeostasis in the bone-defect microenvironment was further elucidated. Initially, a dihydroethidium (DHE) fluorescent probe was used to evaluate the ROS levels at defect sites. As shown in Fig. 9a, b, the GDYO@Pt gel + US group displayed substantially attenuated red fluorescence compared with the other groups, confirming its superior ROS-scavenging capacity. HIF-1α immunofluorescence analysis demonstrated significantly reduced HIF-1α expression in this group (Fig. 9c, d), indicating that the GDYO@Pt system-mediated elevation of the O2 partial pressure alleviated tissue hypoxia, thereby disrupting the hypoxia → HIF-1α activation → ROS accumulation cycle.Fig. 9: Oxidative stress levels in vivo.a Fluorescent photographs and b semi-quantitative analysis of ROS levels in different treatment groups. c Immunofluorescence staining and d semi-quantitative analysis of HIF-1α in bone defect area. e Characteristic immunohistochemical staining profiles of NRF2 following various therapeutic interventions at 12 weeks post-treatment (black arrows: positive staining regions). US: 1.0 MHz, 50% duty cycle, 0.7 W cm−2, 10 min. The data were shown as the mean ± SD in (b) and (d) (n = 4 biologically independent replicates). Statistical significance was determined using a two-tailed unpaired t-test.Full size imageTo further explore the regulatory pathways associated with oxidative stress, immunohistochemical staining was performed to analyze the expression of nuclear factor erythroid 2-related factor 2 (NRF2). As shown in Fig. 9e, the GDYO@Pt gel + US group exhibited a significantly higher proportion of NRF2-positive cells in the newly formed bone tissue than the other groups. These findings demonstrated‌‌ that this multidimensional antioxidant strategy not only effectively reduced local ROS levels, but also enhanced endogenous antioxidant defenses via the NRF2 pathway, establishing a redox-balanced microenvironment conducive to electrical stimulation therapy for bone defect repair.H&E staining of tissue sections from all treatment groups post-therapy demonstrated a preserved structural integrity in each case, in addition to the absence of pathological abnormalities in the heart, liver, spleen, lungs, and kidneys (Supplementary Fig. 25). These results indicated the favorable biocompatibility of GDYO@Pt gel at therapeutic doses in vivo.DiscussionTraditional piezoelectric ceramics primarily generate piezoelectric potentials under stress (e.g., ultrasonic stimulation), which drive the directional migration of ions (e.g., Ca2+, PO₄3-, H+, OH-), indirectly regulating ion distribution and concentration in the microenvironment. However, physiological electrolyte ions (Na⁺, K⁺, Cl-, etc.) rapidly migrate to the material surfaces to screen the electric field, significantly limiting its effective range and duration. This screening effect severely restricts the ability to regulate the microenvironment. In contrast, the piezoelectric semiconductor GDYO@Pt not only generates piezoelectric potentials but also excites free electrons (e-) and holes (h⁺), thereby initiating catalytic reactions that drive local redox processes and enable regulation of the bone defect microenvironment.In this study, the platform that integrates efficient piezoelectric stimulation and multimodal antioxidant stress functions is developed based on a Schottky heterojunction formed between GDYO nanosheets and Pt nanoparticles. The interfacial dipole effect at the heterojunction induces a built-in electric field, driving charge rearrangement in GDYO nanosheets to break the semiconductor’s inversion symmetry, enhance polarity, and significantly boost piezoelectric performance. In addition, the US-triggered Schottky interfaces are found to generate positive polarization charges that induce downward band bending, lowering the energy barriers, and promoting electron migration from GDYO to Pt for enhanced H2 evolution. In this process, the consumption of H⁺ may improve the acidic microenvironment at bone defect sites through localized pH elevation. The generated H2 can further scavenge highly oxidative •OH. Concurrently, the hole-mediated oxidation of H2O2 generates O2, synergizing with the nanozyme activity of GDYO@Pt to achieve the multimodal scavenging of ROS. Ca²⁺ combines with PO₄3- to form an amorphous calcium phosphate (ACP) precursor, providing a matrix foundation for bone tissue mineralization. When the local microenvironmental pH increases, the solubility of hydroxyapatite (HA) decreases, driving the transformation of ACP into crystalline HA71,72. Meanwhile, the downregulation of oxidative stress may prevent ROS from damaging the collagen template, ensuring the integrity of the periodic structure of collagen fibers73,74. This guides the oriented deposition of HA crystals along the long axis of collagen, ultimately forming a biomimetic hierarchical mineralized structure.In vitro and vivo experiments demonstrate that the US-activated GDYO@Pt nanosheets promote osteogenic differentiation, angiogenesis, and immunomodulation by remodeling the electrophysiological microenvironments and downregulating oxidative stress, ultimately achieving a superior cranial bone defect repair efficacy. This piezoelectric-catalytic dual-function platform synchronizes efficient electrical stimulation with oxidative stress regulation, offering a promising solution for bone defect repair. Through band engineering, it precisely controls interfacial redox kinetics during piezoelectric stimulation, avoiding toxic ROS generation while directionally producing therapeutic H2 and O2 to establish a low-oxidative-stress microenvironment conducive to osteogenesis.This mechanism can be extended to diverse regenerative applications, including articular cartilage repair, neural axon regeneration, and skin wound healing, demonstrating universal applicability. Thus, elucidating the spatiotemporal regulatory principles of electro-chemical coupling on cell fate and establishing quantitative structure-activity relationships among band structures, interfacial reaction kinetics, and biological responses will critically guide the development of next-generation intelligent bone repair materials.MethodsMaterialsGraphite was sourced from Beijing Gaoke New Materials (China). Potassium tetrachloroplatinate (II) (K2PtCl4), H2O2 (30%), H2SO4 (98%) and ethanol (99%) were purchased from Shanghai Chemical Reagent (China). Kolliphor® P 407 was obtained from BASF (Ludwigshafen, Germany). Tannic acid (USP, ≥99%) and Methylene blue ( ≥ 98%) were obtained from Macklin (China). The water in the experiments was deionized. All reagents were used directly without further purification.Synthesis of GDYOThe graphdiyne (GDY) was first synthesized on a copper surface via cross-coupling reactions using hexaethynylbenzene as the precursor. 50 mL hexaethynylbenzene acetone solution (2 mg/mL) was added slowly into the three-necked flask containing ten pieces of copper foils (1.5 cm × 5 cm) and 100 mL pyridine under an argon atmosphere at 55 °C. After a 72-hour reaction, the GDY powder was collected and washed thoroughly by acetone, deionized water, 1 M hydrochloric acid, and deionized water, respectively. Subsequently, GDY powder (50 mg) was homogenized with H2SO4 (2.5 mL) under ice-bath conditions, followed by slow dropwise addition of 30% hydrogen peroxide solution (1 mL) under vigorous stirring for 2 hours. Deionized water (20 mL) was then added to the mixture, and the suspension was ultrasonicated for 1 hour, followed by centrifugation (3577.6 × g, 10 min) and repeated washing with deionized water until the supernatant reached neutral pH.To further exfoliate stacked GDYO, concentrated hydrochloric acid (20 μL) was added to 20 mL of GDYO dispersion (0.5 mg/mL), followed by ice-bath ultrasonication for 6 hours (150 W power, 40 kHz frequency). The dispersion was then centrifuged (3577.6 × g, 10 min) and washed twice with deionized water. Finally, unexfoliated residues and large aggregated nanosheets were removed via low-speed centrifugation (503.1 × g, 5 min), yielding a stable exfoliated GDYO nanosheet dispersion, which was stored at 4 °C for subsequent experiments.Fabrication of GDYO@PtThe heterojunction GDYO@Pt nanosheets were synthesized via an in situ reduction method. First, GDYO (60.0 mg) and K2PtCl₄ (30.0 mg) were dispersed in deionized water (120 mL) under ultrasonication. The mixture was then stirred for 30 minutes at 600 rpm. Subsequently, 12 mL of ascorbic acid solution (0.1 M) was added dropwise to the above solution, followed by stirring at 60 °C for 3 hours. The precipitate was collected by centrifugation (3577.6 × g, 10 min) and washed three times with deionized water. Finally, the GDYO@Pt nanosheets were obtained as the final product through freeze-drying.Fabrication of GDYO@Pt gelFirst, the blank thermosensitive gel was prepared via a low-temperature swelling method. Specifically, 0.07 g of TA (tannic acid) was added to 10 mL of deionized water and stirred continuously at room temperature until fully dissolved. Subsequently, 1.8 g of P407 (poloxamer 407) was slowly added to the TA solution, followed by stirring for 30 minutes. The mixture was then transferred to a 4 °C refrigerator and allowed to stand for 24 hours to obtain the blank thermosensitive gel solution. Next, GDYO@Pt nanosheets were incorporated into the blank gel and thoroughly mixed under ice-bath conditions to prepare the thermosensitive GDYO@Pt gel (0.5 mg/mL).Gelation and rheological properties of the thermosensitive gelThe gelation behavior of various hydrogel formulations was assessed via the tube inversion method. Specifically, glass vials containing gel-1 (18% P407, 0% TA), gel-2 (18% P407, 0.5% TA), gel-3 (18% P407, 0.7% TA), gel-4 (18% P407, 1% TA), gel-5 (18% P407, 2% TA), and GDYO@Pt gel formulations (0.5 mg/mL GDYO@Pt with 18% P407 and 0%, 0.5%, 0.7%, 1%, or 2% TA) were immersed in a 37 °C water bath for 10 minutes. After thermal equilibration, the vials were inverted to confirm the sol-gel transition.Additionally, rheological experiments were conducted using a parallel plate (40 mm diameter, 1 mm gap) on a rheometer to characterize GDYO@Pt gel. The storage modulus (G′) and loss modulus (G″) of GDYO@Pt gel were quantified under varying conditions. Oscillatory strain sweeps were performed at 37 °C and 1 Hz to measure G′ and G″ as a function of strain amplitude. Furthermore, time- and frequency-dependent measurements of G′ and G″ were carried out at 37 °C with a fixed strain amplitude of 0.5%. Step-strain measurements were conducted under high (45%) and low (0.1%) strain conditions at 37 °C and 1 Hz. Steady-state shear viscosity was analyzed at 37 °C by applying shear rate sweeps, while step-shear measurements employed low (1 s-1) and high (100 s-1) shear rates.CharacterizationThe comprehensive characterization of the materials was systematically conducted using multiple analytical techniques. Morphological features were examined by transmission electron microscopy (TEM, JEOL JEM-2100F, Japan), scanning electron microscopy (SEM, ZEISS MERLIN Compact, Germany), and atomic force microscopy (AFM, Bruker Dimension ICON, Germany). Crystallographic and chemical analyses were performed via X-ray diffraction (XRD, Bruker D8 Advance, Germany), X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, USA), and Fourier transform infrared spectroscopy (FT-IR, Thermo Scientific Nicolet iS50, USA). Optical properties were assessed through ultraviolet-visible-near infrared (UV-Vis-NIR) spectroscopy (Shimadzu UV-1900 and Hitachi UH4150, Japan), while reactive oxygen species (ROS) signals were detected via electron spin resonance (ESR, Bruker EMXplus, Germany). Functional evaluations included rheological measurements (Anton Paar MCR 702e, Austria), piezoelectric/piezocatalytic activation using an ultrasonic therapeutic device (Chattanooga Intelect 2776, USA), and quantitative hydrogen gas analysis via gas chromatography (Agilent GC7890, USA). The finite element simulations were carried out using COMSOL Multiphysics 6.2 software, which has the main advantage of adapting to multiphysics field coupled simulations, making it easier to perform iterative calculations between different physics fields. Specifically, the simulations were carried out through the solid mechanics module and the electrostatics module in COMSOL, and the use of the stress-charge form was determined according to the material and the requirements. The imposed boundary conditions were a fixed constraint on the left side and stresses in the range of 0 to 108 Pa on the right side. Details of the DFT calculations were provided in Supplementary Note 1. Microscopic images were analyzed by using ImageJ (version 1.49 v). Statistical data analyses were performed using Origin 2017 (version 94E).Piezoelectric output under ultrasonic stimulationGDYO and GDYO@Pt were dispersed in a deionized water solution containing 50% ethanol. This dispersion was coated onto an ITO conductive glass substrate and dried to form a uniform thin film (1.0 cm × 3.0 cm × 0.02 cm3). A platinum (Pt) top electrode layer was then spray-coated onto the film surface. For electrical measurements, one test electrode was connected to the Pt layer, and the other to the exposed ITO region without the film, establishing a closed-loop circuit. The output voltage signal under continuous ultrasonic stimulation was monitored using a digital oscilloscope (Tektronix, 2 GHz). The estimated average electric field was calculated as the output voltage divided by the film thickness. Ultrasonic stimulation parameters: power density 0.1–0.7 W/cm², frequency 1 MHz, Probe diameter 2.5 cm (vertically aligned 1–2 cm above the entire device).In vitro hydrogen detectionGDYO and GDYO@Pt were uniformly dispersed in PBS-containing headspace vials (0.5 mg/mL) and sealed. The samples were then exposed to ultrasound (1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density) in the dark. At different time points (0, 10, 20, 30, and 40 minutes), 1 mL of gas from the vial headspace was extracted and analyzed for hydrogen using a gas chromatograph (GC) equipped with a thermal conductivity detector (TCD) and nitrogen as the carrier gas. Under the catalytic action of platinum (Pt) nanoparticles, H2 reacts with methylene blue (MB) to form colorless reduced leucomethylene blue (MBH2). Thus, MB can also serve as an indicator for H2 detection. Experimental details are as follows: 2 mL of PBS, GDYO, or GDYO@Pt dispersion (200 µg/mL) was added to quartz Petri dishes (3 cm diameter), followed by 8 µL of MB solution (1 mg/mL). Equal amounts of Pt nanoparticles were added to the GDYO and PBS groups as catalysts. To eliminate interference from •OH radicals, methanol was added as a quencher to all groups. Ultrasound treatment (1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density) was applied in the dark at fixed intervals. The dynamic variations in MB levels were tracked by ultraviolet-visible spectrophotometry through absorbance detection at 664 nm. The normalized absorbance was calculated as:$${{{{\rm{A}}}}={{{\rm{A}}}}}_{{{{\rm{t}}}}\,\min }{/{{{\rm{A}}}}}_{0\,\min }\times 100\%$$(7)where A0 min and At min represent the absorbance values at 664 nm before and after ultrasound treatment, respectively.In vitro oxygen detectionOxygen production was monitored using a dissolved oxygen meter (OHAUS ST 300D, USA). Specifically, GDYO, Pt, and GDYO@Pt (200 μg/mL) were dispersed in 30 mL of 10 mM hydrogen peroxide (H2O2) solution. Ultrasound stimulation was then applied to or withheld from each group, and dissolved oxygen levels were monitored in real time.Piezoelectric and electrochemical characterizationPiezoelectric current measurements were conducted using an electrochemical workstation (Metrohm VIONIC, Switzerland) with a standard two-electrode system, where indium tin oxide (ITO) conductive glass (1.0 × 4.0 cm²) served as both counter and working electrodes. The working electrode was modified with the material as follows: 100 μL of sample dispersion (1.5 mg/mL in H2O) was drop-casted onto one end of the ITO glass (1.0 × 1.0 cm²) and dried at 60°C for 24 hours. The piezoelectric behavior of the material was evaluated by real-time monitoring of current responses under varying ultrasound power densities (0, 0.3, 0.5, and 0.7 W/cm²) and five on/off cycles in PBS electrolytes at pH 4.5 and pH 7.4.Mott-Schottky measurements were performed in a 0.5 M Na2SO₄ electrolyte using a three-electrode system. The working electrode (prepared as described above), a Pt plate counter electrode, and an Ag/AgCl reference electrode were employed. Scanning potentials from 1 V to 0.1 V were applied at frequencies of 1000 Hz, 1316 Hz, and 1732 Hz to obtain Mott-Schottky curves before and after ultrasound stimulation.Electrochemical impedance was measured in a mixed electrolyte containing 1 mM potassium ferricyanide (K3[Fe(CN)₆]), 1 mM potassium ferrocyanide (K₄[Fe(CN)₆]), and 0.5 M KCl. The working electrode was prepared by drop-casting 10 μL of sample dispersion (1.0 mg/mL) onto a glassy carbon electrode. The counter and reference electrodes were identical to those used in the Mott-Schottky analysis. AC impedance spectra of different samples were acquired using an electrochemical workstation within a selected frequency range.Cell cultureBMSCs were purchased from Yuchi Biotechnology Co., Ltd. (Shanghai, China) and cultured in complete medium (Gibco, Thermo Fisher Scientific, USA; composition: α-MEM medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin and 100 μg/mL streptomycin). Cells were maintained in a humidified incubator (Thermo Fisher Scientific, USA) at 37 °C with 5% CO2.Cell viability assessment with nanosheets at varying concentrationsCells were seeded in 24-well plates at a density of 5 × 10³ cells per well and co-cultured with nanosheets at varying concentrations (n = 4). After the designated incubation period, the medium (with or without nanosheets) was removed, and 10% CCK-8 solution (Beyotime, C0038) was added. The plates were then incubated at 37 °C for 1 hour. Subsequently, 100 μL of solution from each well was transferred to a 96-well plate, and absorbance values were recorded at 450 nm. The relative cell viability was calculated as follows:$${{{\rm{Cell}}}}\,{{{\rm{viability}}}}\, (\%)=\frac{{{{\rm{Absorption}}}}\,{{{\rm{of}}}}\,{{{\rm{different}}}}\,{{{\rm{treatment}}}}\,{{{\rm{groups}}}}}{{{{\rm{Mean}}}}\,{{{\rm{absorption}}}}\,{{{\rm{of}}}}\,{{{\rm{control}}}}}\times 100\%$$(8)Effects of ultrasound-activated nanosheets on cell viabilityFirst, oxidative stress injury was induced in BMSCs using H2O2. Specifically, BMSCs (2 × 104 cells/well) were seeded in 6-well plates and cultured for 24 hours. The medium was then replaced with H2O2-containing medium (0.2 mM) for 6 hours, followed by fresh complete medium.To evaluate the impact of ultrasound treatment on cell viability, Oxidative stress-injured BMSCs (2 × 104 cells/well in 6-well plates) were exposed to ultrasound at varying power densities (0.3, 0.5, 0.7, 1.0, and 1.5 W/cm²) and durations (0, 5, 10, 15, and 20 minutes) to determine the threshold of ultrasound parameters affecting cell survival (n = 4).Then, GDYO or GDYO@Pt (200 µg/mL) was incubated with oxidative stress-damaged BMSCs; ultrasound stimulation (1.0 MHz frequency, 0.7 W/cm² power density, 50% duty cycle, 10 min per session) was applied or withheld on days 1, 3, and 7. Cell viability was assessed for all groups on days 3, 5, and 9 (n = 4).Effects of ultrasound-induced piezoelectricity on cell membrane potentialThe materials were spin-coated and immobilized on one end of the ITO conductive glass as described in the piezoelectric and electrochemical characterization section. BMSCs were seeded at a density of 1 × 104 cells onto the central region of sterilized ITO glass (seeding area: 1.0 × 1.0 cm²) and incubated in complete medium for 24 hours. For the ultrasonic mechanical effect group, BMSCs were cultured on one end of the piezoelectric-free ITO glass (equivalent seeding density/area to piezoelectric material groups) to assess ultrasound-mediated membrane potential alterations independent of piezoelectric phenomena. The medium was then replaced with complete medium containing Di-8-ANEPPS (2 μM, Macklin, D881943), followed by 30 minutes of incubation in the dark. Cells were gently rinsed three times with dye-free medium to remove excess probe.Subsequently, the material-coated end of the ITO glass was immersed in a PBS-filled electrochemical cell (pH 4.5). After stabilizing the dual-electrode system in the electrolyte for 30 seconds, ultrasound (0.7 W/cm², 5 minutes) was applied to activate the piezoelectric signals of the material. The BMSCs-seeded end of the piezoelectric-free ITO glass was subjected to the same procedure as described above. Fluorescence microscopy (Olympus CKX41SF, Japan) was employed to detect cellular fluorescence on the ITO glass (excitation: 488 nm, emission: 605 nm, red fluorescence).Effects of US-induced piezoelectricity on intracellular calcium levelsFollowing the aforementioned protocol, 1 × 104 BMSCs were seeded onto the central region of sterilized ITO conductive glass pre-coated with materials (seeding area: 1.0 × 1.0 cm2) and incubated in complete medium for 24 hours. For the ultrasonic mechanical effect group, BMSCs were cultured on one end of the piezoelectric-free ITO glass (equivalent seeding density/area to piezoelectric material groups) to assess ultrasound-mediated Ca²⁺ concentration alterations independent of piezoelectric phenomena. The medium was then replaced with Hanks’ Balanced Salt Solution (HBSS) containing Fluo-4 (4 μM, Solarbio, F8500), followed by 30 minutes of incubation in the dark. Cells were gently rinsed three times with HEPES-buffered saline to remove excess dye.The material-coated end of the ITO glass was subsequently immersed in a PBS-filled electrochemical cell (pH 4.5). After stabilizing the dual-electrode system in the electrolyte for 30 seconds, ultrasound (0.7 W/cm², 5 minutes) was applied to activate the piezoelectric signals of the material. The BMSCs-seeded end of the piezoelectric-free ITO glass was subjected to the same procedure as described above. Intracellular calcium dynamics were monitored via fluorescence microscopy by detecting green fluorescence signals (excitation: 494 nm; emission: 516 nm) from Fluo-4-loaded cells.In vitro detection of intracellular hydrogenAdherent BMSCs were co-cultured with GDYO or GDYO@Pt (200 μg/mL) for 12 hours. Following trypsin digestion, PBS was used to prepare cell suspensions (density: 1 × 106 cells/mL). 1 mL of suspension was transferred into sealed headspace vials under gas-tight conditions. After ultrasonic irradiation for 0, 10, 20, 30, and 40 min respectively, 1 mL of headspace gas was collected for gas chromatography quantification of H2 production.BMSCs were seeded in 6-well plates at a density of 2 × 104 cells per well and cultured for 24 hours. Subsequently, the cells were incubated with GDYO or GDYO@Pt (200 µg/mL) for an additional 12 hours. Equal amounts of Pt nanoparticles were added to the GDYO and PBS groups as catalysts. After incubation, the cells were gently rinsed with PBS and fixed with 4% paraformaldehyde (PFA) for 20 minutes. The fixed cells were then co-mixed with methylene blue (MB, 100 µM) in PBS for 20 minutes. Afterwards, the medium was removed and the cells were washed three times with PBS. Ultrasound irradiation was applied to the cells with or without the following parameters: 1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density, and irradiation durations of 10 min. Bright-field images were acquired using fluorescence microscopy to analyze changes in the blue intensity of MB.In vitro detection of intracellular oxygenThe Ru(dpp)3Cl2 (RDPP, Solarbio, YS172601) probe was employed to monitor O2 generation. RDPP is a red-fluorescent oxygen-sensitive indicator (excitation/emission wavelengths: 488 nm/620 nm), whose fluorescence is quenched upon interaction with O2 via energy transfer. Thus, changes in RDPP fluorescence intensity inversely correlate with intracellular oxygen levels.Specifically, BMSCs were seeded in 6-well plates at a density of 2 × 104 cells/well and allowed to adhere. The medium was then replaced with H2O2-containing medium (0.2 mM) for 6 hours. The oxidatively stressed cells were then incubated with complete medium containing GDYO or GDYO@Pt (200 µg/mL) for 24 hours. Subsequently, the cells were incubated in complete medium supplemented with Ru(dpp)3Cl2 (30 μM) in the dark for 40 minutes. After incubation, the medium was removed, and the cells were gently washed three times with PBS. Ultrasound irradiation (parameters: 1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density, 5-minute duration) was applied or withheld. Fluorescence microscopy was used to capture images and analyze changes in the red fluorescence intensity of Ru(bpy)3Cl2.In vitro detection of intracellular ROSIntracellular ROS levels were measured using a ROS assay kit (DCFH-DA fluorescent probe; Beyotime, S0033M). Briefly, BMSCs were seeded in 6-well plates at a density of 2 × 104 cells per well and cultured for 24 hours. The medium was then replaced with H2O2-containing medium (0.2 mM) for 6 hours, followed by fresh complete medium. Experimental groups were co-incubated with 200 µg/mL GDYO@Pt or GDYO nanosheets for 24 hours, followed by ultrasound treatment (1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density, 5-minute irradiation) or no treatment.After washing with PBS, all groups were stained with the DCFH-DA probe under dark conditions for 30 minutes and washed again with PBS to remove unbound probes. DCF fluorescence signals (excitation: blue light; emission: green light) in BMSCs were detected using an inverted fluorescence microscope.ALP staining assayBMSCs were seeded in 6-well plates at a density of 2 × 104 cells/well and allowed to adhere. The medium was then replaced with H2O2-containing medium (0.2 mM) for 6 hours. The oxidatively stressed cells were then incubated with complete medium containing GDYO or GDYO@Pt (200 µg/mL) for 7 days, during which ultrasound was applied/not applied every two days (1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density, irradiation time of 10 min). For the ultrasonic mechanical effect group, BMSCs without the materials were subjected to the same procedures as described above. Subsequently, the cells were washed with PBS, fixed with 4% PFA for 30 min, and an appropriate amount of ALP staining solution (Solarbio, G1480) was added to cover the cells uniformly, and then stained for 30 min at room temperature and protected from light. After the staining was completed, the cells were washed thoroughly with PBS, and the ALP-active areas were observed and photographed under a microscope to record the ALP activity areas to assess the early osteogenic differentiation ability of BMSCs. For quantitative analysis, ALP activity was measured using an ALP assay kit (Beyotime, P0321M) according to the manufacturer’s instructions (n = 3).Alizarin Red S staining assayBMSCs were seeded in 6-well plates at a density of 2 × 104 cells/well and allowed to adhere. The medium was then replaced with H2O2-containing medium (0.2 mM) for 6 hours. The oxidatively stressed cells were then incubated with complete medium containing GDYO or GDYO@Pt (200 µg/mL) for 14 days, during which ultrasound was applied/not applied every two days (1.0 MHz frequency, 50% duty cycle, 0.7 W/cm² power density, irradiation time of 10 min). For the ultrasonic mechanical effect group, BMSCs without the materials were subjected to the same procedures as described above. After incubation, cells were washed with PBS and fixed with 4% PFA for 30 minutes. A 0.2% Alizarin Red S solution (Solarbio, G1450) was added to completely cover the cells, followed by 30 minutes of staining at room temperature. Unbound dye was removed by thorough PBS rinsing. Calcium nodule formation, indicated by orange-red mineralized deposits, was visualized and photographed under a microscope to evaluate late-stage osteogenic differentiation and mineralization capacity. For quantitative analysis, calcium deposits were dissolved in 2% cetylpyridinium chloride (Sigma-Aldrich, C0732), and absorbance was measured at 570 nm using a microplate reader (BioTek Cytation 3, USA) (n = 3).RT-qPCR analysis of osteogenic differentiation genes in BMSCsTo further investigate the osteogenic effects of ultrasound-activated nanosheets, total mRNA was extracted from all experimental groups using Trizol reagent (Invitrogen, 15596018CN). First-strand cDNA was synthesized with the PrimeScript RT reagent kit (TaKaRa, RR047A) following the manufacturer’s protocol. Real-time quantitative PCR amplification was performed on a QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific, USA) using SYBR Green fluorescent dye (Roche, KK4602). Gene-specific primers for osteogenic markers – alkaline phosphatase (ALP), osteocalcin (OCN), osteopontin (OPN), type I collagen (COL-1), and Runt-related transcription factor 2 (RUNX-2) – were designed (primer sequences provided in Supplementary Table 1). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as the internal reference gene for data normalization. Relative mRNA expression levels were calculated using the 2-ΔΔCt method to quantify ultrasound-enhanced osteogenic differentiation (n = 3).In vivo experimental studyAll experimental procedures were conducted in strict compliance with the National Standards for the Care and Use of Laboratory Animals in China. Eight-week-old wild-type C57BL/6 mice (purchased from the National Institutes for Food and Drug Control, NIFDC, China) were housed in specific pathogen-free (SPF) facilities pre- and post-operatively, with ad libitum access to water and a standard rodent diet. The ambient temperature was maintained at approximately 25 °C, humidity was controlled, and a 12-hour light-dark cycle was employed. Surgical protocols and perioperative management followed the guidelines approved by the Animal Care and Use Committee of Beihang University (License No. BM20230009).Establishment of cranial defect model and in situ injection of gelMice were group-housed for one week preoperatively for environmental acclimatization. Anesthesia was induced and maintained via isoflurane inhalation. After incising the scalp, a 3.0 mm diameter full-thickness circular bone defect was created in the parietal bone using a dental trephine drill. The defect area was irrigated with saline, followed by in situ injection of thermosensitive gel formulations into the defect cavity according to experimental groups. After complete gelation (37°C, 10 minutes), the wound was closed in layers. Postoperatively, ibuprofen (10 mg/kg, Macklin, I821809) was administered in drinking water for 24 hours to alleviate pain. To prevent infection, sulfamethoxazole (15 mg/kg, Macklin, S861458) and trimethoprim (30 mg/kg, Macklin, T832374) were added to drinking water for one week.Therapeutic procedure for murine cranial defect repairCranial defect-bearing mice were randomly allocated into six experimental groups (n = 4): (1) Control (no treatment), (2) gel + US (thermosensitive gel with ultrasound stimulation), (3) GDYO gel (gel containing 0.5 mg/mL GDYO nanosheets without ultrasound), (4) GDYO@Pt gel (gel containing 0.5 mg/mL GDYO@Pt nanosheets without ultrasound), (5) GDYO gel + US (GDYO-loaded gel with ultrasound), and (6) GDYO@Pt gel + US (GDYO@Pt-loaded gel with ultrasound). All interventions were standardized to ensure uniform nanosheet concentrations and treatment protocols across groups. Ultrasound stimulation (1.0 MHz, 50% duty cycle, 0.7 W/cm²) was administered once daily (10 minutes per session), for 4 consecutive weeks. Cranial bone samples were harvested at 6 and 12 weeks post-treatment and analyzed for bone regeneration outcomes using micro-CT and histological assessments.Micro-CT analysisThe extracted cranial bone samples post-treatment were scanned using a Micro-CT system (Skyscan 1276, Bruker, Belgium) under 55 kV voltage and 200 μA current. Three-dimensional image reconstruction and analysis of the bone defect region were performed using NRecon and CTvox software. Bone regeneration parameters, including BV/TV and BMD, were quantified with CTAn software.Histological evaluationFollowing micro-CT analysis, samples from the 12-week treatment groups were decalcified in 15% ethylenediaminetetraacetic acid (EDTA) solution and subsequently embedded in paraffin. Tissue sections (4 μm thickness) were prepared, deparaffinized with xylene, and rehydrated through a graded ethanol series (70% to 100%). Each sample underwent hematoxylin and eosin (H&E, Servicebio, G1076) staining and Masson’s trichrome staining (Servicebio, G1006). Histomorphological structures were examined using a digital tissue section scanner (Pannoramic MIDI, Hungary).IHC and IF staining protocolIHC staining: paraffin-embedded sections were dewaxed and blocked, then incubated overnight at 4 °C with the following primary antibodies: anti-OCN antibody (Servicebio, GB11233), anti-RUNX-2 antibody (Servicebio, GB115631), anti-CD31 antibody (Servicebio, GB113151) and anti-NRF2 antibody (Servicebio, GB113808). All primary antibodies were diluted at 1:100 using 3% (w/v) bovine serum albumin (BSA, Sigma-Aldrich, USA). After PBS washing, samples were incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (Servicebio, GB23303, 1:200). Staining signals were visualized using the digital tissue section scanner.IF staining: dewaxed paraffin sections were blocked with 3% BSA to prevent nonspecific binding, followed by overnight incubation at 4 °C with primary antibody (anti-CD86 (Servicebio, GB115630, 1:300)/anti-CD206 (Servicebio, GB125273, 1:300) blend and anti-HIF-1α (Beyotime, AH339, 1:100)). After PBS washing, sections were incubated with Alexa Fluor 488-labeled (Servicebio, GB25301, 1:400 for CD86) or CY3-labeled (Servicebio, GB21303, 1:300 for CD206; Servicebio, GB21301, 1:300 for HIF-1α) secondary antibody at room temperature for 1 hour, then counterstained with DAPI for nuclear visualization. Final images were captured using the digital tissue section scanner.Statistical analysisStatistical differences were calculated using a two-tailed unpaired t-test in IBM SPSS Statistics 25 software. Replicate experimental data were presented as mean ± standard deviation (mean ± SD). P-values  0.05 were considered statistically non-significant.Reporting summaryFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.