The microstructural evolution and mechanical responses under multi-degree-of-freedom reciprocating torsion-compression deformations remain to be fully elucidated, particularly regarding the Swift and inverse Swift effects and their physical mechanisms, which constrain the design and formability of textured Mg alloys. Therefore, the multi-degree-of-freedom reciprocating pre-torsional-compressive loadings along extrusion direction (ED) were specifically designed. The twinning behaviours and the radial distribution of twin structures were systematically analysed. The driving mechanisms of the Swift and subsequent inverse Swift effects were discussed. Results demonstrated that free end torsion (FET) deformation induced radially linear-gradient twinning structure, while reverse FET (RFET) loading triggered FET twins detwinning and extensive {10$ \bar{1} $2} tensile twin activation within the basal textures, driving the reverse FET twin texture further tend towards ED aggregation. FET twins inhibited the nucleation of RFET twins, resulting in the formation of a distinctive inverse-gradient twinning structure. 65% of the Swift-effect strain under low-strain FET (γ < 0.12) was coordinated by dislocation slip, whereas the misfit strain induced by FET twins accommodated more than 85% of the axial shortening during γ_FET = 0.38. RFET-stage axial elongation was governed by detwinning, with subsequent axial shortening attributable to large-scale RFET twin activation. {10$ \bar{1} $2} tensile twins dominated initial free rotational compression (FRC) strain, the interactions between the release of residual shear stress and the reverse shear strain component induced by FRC twins results in the circumferential motionlessness during initial FRC. Proliferation of FRC twins promoted cumulative circumferential shear strain component, inducing further macroscopic reverse spontaneous rotation.

Lanthanide-doped ultrasmall upconversion nanocrystals exhibit unique optical properties that differ significantly from bulk materials, making them highly promising for diverse applications. However, their widespread applications have long been impeded by extremely weak photoluminescence due to severe surface quenching effect. Herein, we present a Li⁺-doping-induced phase engineering strategy to achieve significantly enhanced upconversion luminescence in ultrasmall Cs₂ZrF₆:Yb/Er nanocrystals (NCs). While initial Li⁺ doping improves crystallinity, yielding up to a 63-fold emission increase, further doping triggers a phase transition to form heterophase nanocrystals consisting of trigonal Cs₂ZrF₆ and tetragonal LiYbF₄. The synergistic effect of improved crystallinity and reduced local symmetry around lanthanide ions results in a 302-fold boost in emission intensity, even as the particle size decreases to ~6.1 nm. Moreover, these ultrasmall nanocrystals display anomalous anti-thermal quenching behavior, with luminescence intensity increasing as temperature rises from 303 to 483 K, driven by defect-mediated energy repopulation. This work not only offers a robust approach for fabricating ultrasmall, high-brightness NCs but also establishes phase engineering as a pivotal mechanism for modulating local crystal field, paving the way for high-performance, thermally stable ratiometric nanothermometry.

Transition metal nitride coatings often face a trade-off between thermal stability and oxidation resistance at high temperatures. Here, we address this challenge using a mechanism-guided high-throughput combinatorial strategy, implemented through multi-target co-deposition, to rapidly probe the Al-Cr-Ti-Si-N compositional space. An optimal composition window was identified (Al 13.3-24.1 at.%, Cr 7.6-14.6 at.%, Ti 10.4-23.9 at.%, Si 1.3-3.2 at.%). Coatings within this window exhibit high hardness (> 32 GPa), maintain a face-centered cubic (fcc) matrix up to 1,000 °C, and form oxide scales thinner than 0.5 μm after oxidation at 1,000 °C. Multi-scale characterization reveals that these outstanding properties originate from nanoscale synergistic mechanisms. Specifically, spinodal decomposition generates Ti-rich and Al/Cr-rich domains. These domains, coupled with dense dislocation networks and pronounced lattice distortion, collectively underpin mechanical robustness and structural stability. The inner SiO₂ and intermediate (Cr,Al)₂O₃ layers act together to block oxygen diffusion and metal ion migration, resulting in exceptional oxidation resistance. High-speed dry cutting tests validate the engineering relevance of these coatings, showing a tool life up to 2.4 times longer than that of commercial AlTiN coatings. This work resolves the stability-oxidation trade-off and establishes a generalizable pathway for the rational design of multicomponent ceramic coatings.

Electric current induced structural change is a critical reliability issue in nanoelectronics. So far, microstructure evolution models have considered the electromigration (EM) force caused by high current density to be uniform everywhere in the conductor. In this work, the EM force acting on atoms near Σ5 grain boundaries (GBs) of aluminum is calculated for the first time using a quantum transport method. Although the EM forces exhibit a complex distribution near the GB, the regional averaged EM force from GB atoms points toward the opposite direction of the bulk EM force. This striking result leads to GB motion in the direction of the electron current, whereas previous structural evolution models considering uniform EM forces predict GB motion in the opposite direction. This result highlights the necessity to consider the current-induced grain structure evolution at quantum mechanical level since the local EM force on atoms depends on the electron scattering behavior of the GB structure.

Perovskite solar cells (PSCs) are widely recognized as one of the most promising candidates for next-generation photovoltaics (PV), attributed to their outstanding power conversion efficiency and easily adjustable bandgap. Nevertheless, their practical commercialization remains challenging due to inherent material instabilities and rapid degradation triggered by exposure to moisture, oxygen, ultraviolet light, and thermal stress. In response, atomic layer deposition (ALD) has gained prominence as an advanced thin-film deposition technique, offering atomic-level precision and enabling the fabrication of uniform, conformal, and defect-minimized metal oxide (MO) layers. The ALD MOs serve versatile roles as charge transport layers, interfacial passivation coatings, and encapsulation barriers, collectively enhancing device performance, mechanical integrity, and operational durability. This review highlights the working mechanism of ALD, structural features of PSCs, and key strategies that leverage ALD-grown MOs to address efficiency loss and instability issues. Specific focus is given to defect passivation and protection against environmental factors. Moreover, challenges including precursor optimization, process compatibility with PSC architectures, and cost considerations are examined, along with future perspectives for industrial translation. Finally, the integration of ALD MOs holds strong potential to deliver the stability, efficiency, and scalability essential for the industrial application of PSCs in sustainable PV technologies.

This review highlights an innovative co-delivery system using hydrogel microneedles (MNs) to encapsulate exosomes and drugs, presenting a transformative strategy for treating interconnected bone and brain disorders. By offering potential strategies to address the blood-brain barrier, MNs enable minimally invasive delivery of exosomes, natural vesicles rich in microRNAs and cytokines, facilitating spatiotemporal control over therapeutic release. This system enhances regenerative outcomes in diabetic wounds, myocardial infarction, osteochondral defects, and neurological conditions via multi-mechanistic actions including immunomodulation, angiogenesis, and tissue repair. However, clinical translation remains challenging due to issues in exosome standardization, microneedle mechanical properties, and long-term biosafety.

The non-equilibrium dendrite growth of Mg-6wt.%Al alloy during quasi-rapid solidification is studied by combining phase-field simulations and comparative experiments (furnace/air/water cooling: 0.07/2.9/181 K/s). The kinetic behavior of the solid-liquid interface is characterized, and the solute trapping-drag competition is emphasized. The effects of undercooling (28-36 K), cooling rate, and orientation angle (0-π/6) on dendrite morphological evolution are systematically explored, and the laws governing the morphological transition of the interface front are analyzed according to the condition criterion of interface transition. The experimental primary dendrite arm spacing decreases from ~45 ± 4.2 μm (0.07 K/s) to ~5.5 ± 0.7 μm (181 K/s), matching the phase-field simulations (relative error less than 9%). The solid-phase Al concentration rises from 2.6 ± 0.22 wt.% to 5.4 ± 0.26 wt.% experimentally, consistent with the simulated trends. The simulated critical value for planar-to-cellular transition (1.2 × 10⁹ K·s/m²) is lower than the theoretical value (1.48 × 10¹⁰ K·s/m²) due to solute drag. The analysis of the partial drag condition is extended by integrating thermodynamics and kinetics. The limitation of the current model in capturing the intermediate partial drag state and potential future direction to address this are discussed. Through combining with the experimental results under different cooling rates, the simulation results are further interpreted and validated in both quantitative and qualitative way. This research provides theoretical basis for the regulation of magnesium alloy microstructures under quasi-rapid solidification such as industrial die-casting process.

The microstructure is an important key factor for the reliability of deep through silicon via (TSV) electroplated copper. In the present study, the influence of current density on the surface physical field and microstructure in the electroplated copper were discussed using a combined method of numerical simulation and experiment. The results showed that the optimal electroplated parameters with the defect-free filling were confirmed to be at 0.1-0.17 Ampere per Square Decimeter (ASD) with an accelerator-to-suppressor ratio of 1:10 when the size of deep TSV was Φ20 μm × 200 μm with an aspect ratio of 10:1. The microstructure of TSVs exhibited a distinct distribution feature: fine grains along sidewalls and at the mouth, large columnar grains in mid-regions, and equiaxed grains at the bottom. The grain size and the quantity of Σ3 twin boundaries first decreased and then increased with increasing current density due to the competing effect of nucleation rate and grain growth by governing the polarization and suppressor desorption. Moreover, the intermediate 9R structure between the matrix and the twin was first observed in the electroplated copper, which provided a new way for strain accommodation in deep TSV electroplated copper, and the matrix→9R→twin pathway was proposed through the slip of Shockley partial dislocations. These findings served as a valuable reference for modeling microstructure evolution and laid a foundation for both microstructure prediction and control in deep TSV electroplated copper.

Freestanding oxide membranes, which are free from epitaxial constraint and substrate clamping unlike complex oxide thin-film heterostructures, provide a fascinating platform for realizing multi-functional devices via heterogeneous integration of freestanding membranes with different physical properties. For further applications of freestanding membranes to actual devices, single-crystalline nanomembranes are highly essential. Herein, we demonstrate the fabrication of freestanding BaZrO₃ membranes with high crystallinity via BaZrO₃/SrCuO₂ bilayer thin films epitaxially grown on SrTiO₃ (001) substrates, followed by selective etching of the sacrificial SrCuO₂ layer. The exfoliated freestanding BaZrO₃ membranes serve as a robust template layer for the epitaxial growth of ferroelectric BaTiO₃, producing freestanding BaTiO₃/BaZrO₃ membrane heterostructures. The crystallinity and epitaxy of the as-fabricated freestanding heterojunctions is characterized by X-ray diffraction analyses. Raman spectroscopy also reveals that the upper layers of ferroelectric BaTiO₃ are mainly in-plane polarized probably due to the biaxial in-plane tensile strain imposed by the lower BaZrO₃ layer with cubic symmetry, although the initial tensile strain is progressively mitigated with the increasing thickness of the topmost BaTiO₃ layer. The perovskite BaZrO₃ membrane templates are of potential interest for designing strain-tuned heterojunctions of freestanding oxide membranes, where the associated physical properties can be manipulated compared with the bulk counterparts and/or epitaxial oxide thin-film heterostructures.

To elucidate the correlation between the catalytic properties of metal nanoparticles used in polymer electrolyte fuel cells and the atomic arrangements of such nanoparticles, high-energy X-ray diffraction measurements were conducted. Using the measured data as a reference, Reverse Monte Carlo was performed on isolated finite-size spherical cluster models to visualize the atomic arrangements of Pt and Pt₃Co nanoparticles. In the Pt₃Co nanoparticles, the local composition of the nanoparticle center and the surface region differed. The tendency of Co atoms to be located closer to the surface rather than the center was confirmed. Furthermore, the atomic arrangements in the center of the Pt and Pt₃Co nanoparticles were more disordered than those in the surface region. Thus, we clarified the difference in the distribution of atoms in the center and surface regions by creating a three-dimensional model of the atomic arrangement of the nanoparticles. The visualized structural model can contribute to the development and performance enhancement of nanoparticle catalysts.

The biomineralization process is initiated when bacteria recruit dissolved mineral ions, leading to the spontaneous formation of biomineral layers on metal surfaces, which can substantially influence the service life of the metal. Herein, we investigated Pseudoalteromonas spongiae (P. spongiae)-induced biomineralization layer formation on Q235B carbon steel and 2507 duplex stainless-steel in bacterial suspension, and we evaluated the anti-corrosion and antifouling performance of mature biomineralized layers in artificial seawater (ASW). On Q235B, the higher surface reactivity promoted a uniform, dark-gray, superhydrophilic biomineralized layer with minor iron oxides. This layer increased interfacial (charge-transfer) resistance, suppressed corrosion, and reduced the attachment of Phaeodactylum tricornutum. On 2507, the native passive barrier limited nucleation, producing Ca–Mg carbonate deposits with locally exposed metal and faster charge transfer. Nonetheless, the hydrophilic protective layer still reduced biofouling in ASW. Overall, P. spongiae–mediated biomineralization was substrate-dependent, exhibiting dual anti-corrosion and antifouling capacity on carbon steel. On 2507, the heterogeneous biomineralized layer, while reducing biofouling, locally compromised the integrity of the passive film, resulting in pitting corrosion.

Enhancing the fatigue resistance of metals remains a significant challenge in materials engineering. This study demonstrates that titanium exhibits remarkable fatigue resistance when heterostructures are introduced via additive manufacturing. Compared to homogeneously structured titanium, heterostructured titanium shows a remarkable 141% improvement in fatigue strength and a 53% enhancement in fatigue ratio. The heterostructure promotes the formation of high-density geometrically necessary dislocations, leading to hetero-deformation-induced strengthening under cyclic loading. This process enhances structural stability, suppressing fatigue crack initiation and propagation, thus improving fatigue resistance. These findings suggest that heterogeneity is a promising strategy for enhancing fatigue resistance across various alloy systems.

Grain boundary segregation plays a critical role in determining the properties of polycrystalline materials, yet its influence on piezoelectric performance remains underexplored. In this work, bismuth layer-structured piezoceramic W⁶⁺-doped CaBi₂Nb₂O₉ (WCBN) was chosen to investigate the effect of grain boundary segregation on the piezoelectric properties through multiscale structural characterization and phase-field simulations. The results reveal that improper grain boundary segregation can induce internal stress fields that restrict domain switching dynamics, leading to deterioration of the piezoelectric response. Therefore, a novel poling process was developed, which effectively alleviated the segregation-induced stress constraints and enhanced the piezoelectric coefficients by 180%. More importantly, optimizing the preparation process significantly enhances the mechanical properties, particularly increasing the fracture toughness of WCBN ceramics to 2.73 MPa m^1/2, which is more than twice that of traditional Pb(Zr, Ti)O₃ piezoceramics. These findings establish direct correlations between grain boundary segregation, internal stress, and domain switching behavior, providing fundamental insights for the design of piezoelectric materials that integrate both high piezoelectric and mechanical properties, which could be greatly beneficial to long-term stable operation in harsh environments with high temperatures and complex vibrations for bismuth layer-structured piezoceramics.

Electrocaloric (EC) cooling represents a promising solid-state approach for next-generation thermal management. However, achieving substantial temperature modulation remains a challenge due to intrinsic material limitations and inefficient energy conversion. Herein, we focus on microstructure regulation to enhance thermal conductivity and EC performance. A hydroxyl-functionalized Ba_0.63Sr_0.37Zr_0.01(Ti_0.999Mn_0.001)_0.99O₃ (BSZMT-OH) and h-BNNS-OH composite was designed, with enhanced interfacial hydrogen bonding to optimize electric field response in a P(VDF-TrFE-CFE) matrix. Finite element analysis (FEA) and piezo response force microscopy (PFM) reveal strengthened interfacial coupling, which facilitates rapid domain switching kinetics by amplifying tetragonal P4mm phase responses. These enhancements yield a peak EC temperature change (ΔT) of 1.59 K under a low electric field of 40 MV/m in the optimized 6% BSZMT-OH@4 h-BNNS-OH (6@4BNNS) composite. Integrated into a double-layer four-section electrocaloric cooling (DL 4 EC) device, it cools from 70 to 23 °C in 23 s, outperforming water cooling. Our findings offer insights into EC mechanisms and present a high-performance thermal management strategy.

To meet the escalating demands for high-performance piezoelectric materials in fields such as modern medical diagnostics, precision manufacturing, etc., we developed a ternary 0.555Pb(Ni_1/3Nb_2/3)O₃-0.145PbZrO₃-0.30PbTiO₃ (PNN-PZ-PT) piezoelectric ceramic located at the tricritical point of rhombohedral, tetragonal and pseudocubic phases. This ceramic, with coexistence of multiple ferroelectric phases, demonstrates an ultrahigh piezoelectric coefficient d₃₃ of 1190 picocoulombs per newton (pC/N) and a large relative dielectric constant ε_r of 9900. Analysis of the domain structure reveals irregular maze-like nanodomains with strong local disorder. These nanodomain structures exhibit excellent local piezoelectric response and polarization switching characteristics, thereby enhancing the alignment of polarization vectors during poling and ensuring the high stability after being poled. The tricritical point composition with disordered nanodomains can significantly enhance the contribution of polarization to macroscopic piezoelectric properties, offering a promising approach for developing ceramics with superior piezoelectricity.

With the ongoing miniaturization of multilayer ceramic capacitors (MLCCs), there is an increasing demand for dielectric materials that simultaneously exhibit high dielectric constant, excellent DC-bias stability, and high reliability. To address this challenge, B-site Ca doping was employed to regulate the polar structure of BaTiO₃-based ceramics. In this study, we systematically investigated the effects of Ca doping on the crystal structure, defects, and microstructure by varying the dopant concentration and occupancy behavior. The B-site Ca-doped BaTiO₃ ceramics exhibit a pseudo-cubic structure, characterized by the coexistence of tetragonal and cubic phases. Ca²⁺ substitution for Ti⁴⁺ disrupts the long-range ferroelectric order, leading to the formation of polar nanoregions (PNRs) interconnected and embedded within a non-polar matrix. Defect analysis and studies on reducing atmosphere sintering reveal that oxygen vacancies are effectively localized by cation defects, thereby suppressing long-range conduction. These structural features synergistically result in a high dielectric constant, superior DC-bias stability, enhanced insulation resistance, and strong non-reducibility. This work provides fundamental insights into the microstructural design of BaTiO₃-based ceramics and highlights their potential for high-reliability MLCC applications.

Nonlinear photonic crystals (NPCs) have garnered significant attention due to their capability to manipulate and enhance optical interactions via quasi-phase-matching. The periodic domain inversion structures inside NPCs, corresponding to periodically modulated second-order nonlinear coefficients χ⁽²⁾, enable the generation of light at new frequencies. However, due to the internal stress and defects, forming a structurally perfect superlattice in naturally grown crystals is challenging, and artificial engineering often entails high costs and complex procedures. In this study, we successfully fabricated a unique three-dimensional (3D) potassium tantalate niobate (KTN) NPC using a direct current poling treatment. Nonlinear Cherenkov diffraction occurs when the fundamental light propagates parallel to the poling direction, whereas nonlinear Bragg diffraction is observed when the light is incident perpendicular to the poling direction. By analyzing both the linear and nonlinear optical responses of the stably poled KTN crystal, we reveal the 3D distribution of ferroelectric domains, which advances the understanding of ferroelectric domain dynamics during the poling process and opens new avenues for research in anisotropic nonlinear optics and optoelectronic applications.

Metamaterials have become an important strategy for enhancing electromagnetic wave transmission, and the realization of transmission tunability has attracted considerable attention in current scientific research. In this study, we propose a transmissive metamaterial consisting of periodically arranged SrTiO₃ ceramic particles array and metallic grid array. Taking advantage of the Mie resonance of the dielectric particle array, the dispersion characteristics at the interface are precisely tailored, resulting in high transmission in the microwave frequency range. Furthermore, the temperature-sensitive permittivity of SrTiO₃ enables dynamic tuning of the transparency window toward higher frequencies through electrically induced thermal modulation. Both the selective high-transmissive performance and electrical tunability are validated through numerical simulations and experimental measurements. This work provides a convenient route to design tunable metamaterials, offering fascinating possibilities for the development of active microwave windows.

Nonlocal metasurfaces exhibit significant potential for advanced all-optical image processing by leveraging their exceptional capability to regulate spatial dispersion through precise tailoring of optical transfer functions (OTFs). However, the inverse design of specific OTFs remains challenging due to the inherently complex and highly nonlinear relationship between metasurface structural parameters and angular-dependent optical responses, which conventional empirical trial-and-error approaches struggle to address. To overcome this limitation, we propose an automated inverse design framework integrating a deep neural network acting as a forward predictor with Bayesian optimization. This framework enables automated OTF tailoring by optimizing metasurface structural parameters for targeted image processing operations at desired wavelengths within the 1,200-1,400 nm range. We validate the framework by designing nine dedicated silicon hollow brick metasurfaces: for each operational wavelength (1,250, 1,300, and 1,350 nm), three distinct devices are engineered to separately execute 2D second-order differentiation, 2D fourth-order differentiation, and 2D Gaussian high-pass filtering in transmission mode through targeted OTF engineering. These inversely designed nonlocal metasurfaces achieve a numerical aperture close to 0.4 and serve as fundamental components for edge detection and image sharpening. This intelligent, automated design paradigm dramatically accelerates the design process and significantly expands the scope of achievable functionalities for optical computing metasurfaces, paving the way for more sophisticated all-optical information processing systems.

Non-antibiotic antibacterial agents can significantly reduce the risk of bacterial drug resistance, but they face limitations such as poor tissue penetration, insufficient in vivo stability, and significant side effects. Metal-organic frameworks (MOFs), as a novel class of porous functional materials, exhibit broad application prospects in the antibacterial field due to their highly tunable structures, large specific surface areas, and precisely controllable pore systems. This review introduces the antibacterial mechanisms of MOFs, explores their responsive antibacterial strategies within the bacterial infection microenvironment, and summarizes their applications in biofilm eradication, immuno-synergistic antibacterial therapy, and implant surface modification. Finally, it discusses the biocompatibility and biosafety of MOFs, addresses the existing deficiencies in this field, and outlines future directions.

In this study, we synthesized a novel heterostructure comprising two-dimensional tungsten sulfide nanoflowers and chiral silver nanoparticles. Leveraging the chiral-induced spin selection (CISS) effect, our WS₂@L-His/Ag(chiral) system achieved a remarkable 77% enhancement in photocatalytic degradation efficiency over the WS₂@Ag(achiral) counterpart. The chiral nanostructures selectively facilitate the transport of electrons with defined spin orientations, thereby suppressing the recombination of electrons and holes and augmenting the generation of hydroxyl radicals, which significantly boosts degradation efficiency. Our findings underscore the potential of electron spin manipulation and herald novel avenues for chiral material applications in spin-dependent photoelectric chemical processes.

In this paper, the effect of natural convection on the stability of the dendritic tip during directional solidification under various gravitational field conditions is numerically investigated using the phase-field lattice-Boltzmann method. The coupled equations were implemented for parallel computing on multi-graphics processing units using in-house code written in modular Compute Unified Device Architecture. In the single-crystal case, downward buoyancy transports solute from the dendrite roots toward the tips, leading to solute enrichment near the tips. This enrichment reduces the local undercooling, slows dendrite growth, and eventually triggers tip splitting at high convection intensities. In contrast, the upward buoyancy moves rejected solutes from tips to interdendritic regions, barely affecting solute distribution along the crystal symmetry axis and stabilizing tips. In bi-crystalline systems, downward convection induces tip splitting and plume formation in converging grain boundaries, and drives solute flow to promote sidebranching in diverging grain boundaries, while upward convection has a negligible impact on grain boundaries. This work offers quantitative insights into the dynamic mechanism by which natural convection regulates dendritic tip stability, thereby elucidating the role of natural convection in microstructure evolution during directional solidification.

Silicon (Si) anodes have emerged as promising candidates for next-generation lithium-ion batteries owing to their high theoretical capacity. However, their practical application is hindered by severe volume expansion and unstable electrode interfaces during cycling. Polymer binders play a critical role in mitigating these issues by maintaining electrode integrity and enhancing interfacial stability. This review provides a systematic classification of polymer binders for Si-based anodes according to their binding mechanisms, including covalent, hydrogen-bonding, and supramolecular interactions. The discussion emphasizes how structural configurations, such as linear, branched, and cross-linked architectures, affect mechanical resilience, adhesion strength, and compatibility with high-loading electrodes. Recent advances in multifunctional and dynamic cross-linked binders are highlighted, with particular focus on strategies to accommodate large volume changes and suppress interfacial degradation. In addition, a comparative analysis of advantages and limitations for various binder systems is provided, along with perspectives on future development trends. This work aims to guide the rational design of polymer binders for achieving stable, high-energy-density Si-based anodes.

Microstructure evolution in service significantly influences the properties of advanced materials. Numerical simulation can effectively capture microstructure development and provide abundant high-fidelity data. However, effective 3D microstructure-informed property prediction methods are lacking due to the complexity and richness of 3D microstructural data. In this work, a novel approach combining phase-field simulation and a 3D convolutional neural network is proposed to explore the composition-process-structure-property relationship in Ti_1-xAl_xN coatings. A large dataset of 4,962 simulated 3D microstructures under various heat treatment conditions was first generated using phase-field simulations. Then, a reconstructible feature extraction model was trained to compress each 48 × 48 × 48-grid microstructure into a 128-dimensional latent vector with a reconstruction accuracy of up to 99%. Using the extracted features, a microstructure-based hardness prediction model was constructed, achieving a low prediction error of 1.6 GPa (ca. 5.3% error for an average hardness of 30.8 GPa). The results demonstrate the effectiveness of 3D microstructure-informed deep learning for accurate property prediction, providing a promising tool for the data-driven design of high-performance materials.