Small inspection robots are highly desirable for inspecting complex machinery and detecting damage in confined spaces. However, common climbing robots that rely on vacuum suction or bioinspired dry adhesion often suffer from bulky sizes or slow locomotion speeds. Developing compact yet intelligent wall-climbing robots that mimic the agility and payload capacity of geckos remains an important challenge. In this work, we design a 20-g, 10-cm artificial intelligence (AI)-integrated robot capable of carrying a 70-g payload while climbing on vertical and inverted surfaces at a speed of 70 mm/s. Acoustic adhesion is generated by vibrating a flexible annular disk on smooth surfaces, where air is periodically absorbed and expelled, resulting in negative pressure. The thin air layer with negative pressure indicates anisotropic performance, characterized by strong normal adhesion and negligible tangential resistance, making it highly suitable for designing small, yet strong, climbing robots. The theoretical model and laser surface morphology measurements reveal the thickness-dependent adhesion of a thin air layer beneath the disk. A servo-spring system is designed to meet the stringent requirements of a thin air layer thickness, yielding robust normal adhesion. Resonance analysis and the use of proper spring material stiffness further enhance adhesion performance. Therefore, combining this innovative acoustic adhesion with optimized structural design, our robot achieves gecko-like mobility and payload capacity. Additionally, integrated AI techniques simplify robot control, allowing voice-commanded operation and autonomous task execution. We demonstrate the functions of these climbing robots through agile inspections in a 3-dimensional maze and retired aircraft engines. This work presents the design of small, strong, and agile climbing robots that utilize anisotropic acoustic adhesions, demonstrating agile mobility across gaps, right corners, and discontinuous curved surfaces. It offers potential solutions for in situ damage detection in aero-engines and other complex equipment cavities.

Magnets are essential for mobile consumer electronics, electric motors that power industry and the future of transportation as well as generating and transforming most electric power. Strong magnets reduce the size and weight of motors and generators as well as improve efficiency. The most powerful Nd 2 Fe ${\rm Nd}_2{\rm Fe}$ 14 B $_{14}{\rm B}$ -based magnets have a complex structure like millions that are expected to exist but have not been made and characterized. With the recent developments of AI materials discovery techniques, which enable data-driven and machine-learning-assisted screening, together with computational approaches that can accurately predict intrinsic magnetic properties of a given structure, and high-throughput autonomous labs, the discovery of new, ultra-powerful magnet materials with saturation magnetization greater than 2.5 Tesla or magnetic energy density ( B H m a x $BH_{max}$ ) greater than 800 kJ/ m 3 ${\rm m}^{3}$ is now quite possible.

Thermoelectric materials require precise doping to adjust the Fermi level in order to fully realize their thermoelectric potential. The effective mass (EM) model is commonly used to predict the maximum thermoelectric figure-of-merit (zT) and optimal carrier concentration, but its application is limited by the need for numerical Fermi integral solutions and Hall effect measurements. Since the thermopower (magnitude of the Seebeck coefficient, |S|) is effectively a measure of the Fermi level, it can be used as a direct descriptor of doping level in heavily doped semiconductors such as good thermoelectric materials. Here, we present a simple method to analyze thermoelectric transport using only the typical thermoelectric measurements: Seebeck coefficient, electrical conductivity and thermal conductivity. This enables evaluation of weighted mobility, quality factor (B), theoretical maximum zT, optimal thermopower, identification of anomalous scattering behavior; as well as a full prediction of zT, as a function of |S| given by

The ideal electronics packaging for next-generation wearable devices and integrated circuits would be a flexible material with high thermal conductivity of a metal. We report a liquid-metal polyurethane composite (LiMPuC) with a thermal conductivity of 23.42 W m-1 K-1 while retaining extreme stretchability. The performance arises from interfacial-chemistry-guided ordering of the polyurethane matrix and improved liquid-metal wetting. By modulating hydrogen-bond donor/acceptor densities in the thermoplastic polyurethane and grafting -NH2 groups onto eutectic gallium-indium (EGaIn) droplets, we create anchored, reconfigurable interfaces that (i) increase matrix chain alignment and intrinsic heat transport and (ii) promote stable, strain-tolerant thermal near-percolation of the liquid metal. Under large tensile strain, these coupled effects preserve high conductivity and deliver a flexibility figure of merit > 100. This chemistry-to-microstructure pathway, linking hydrogen-bond engineering with LM surface functionalization, provides a general strategy for designing flexible, high- κ composites for advanced thermal management in emerging electronics.

Mechanical interfaces-in which two materials are simply pressed together-are common in electronic systems, but microscopic surface roughness limits the area that is actually in contact, creating an impedance to electron or phonon transport. This can lead to Joule heating in electronic circuits and poor heat dissipation in power electronics. Here we show that mechanical metainterfaces, which are inspired by traditional carpentry joints such as mortise and tenon and finger joint, can enhance interfacial electronic and thermal transport. The interfaces use geometry-driven contact force augmentation in combination with conversion of interfacial stress from compressive to shear, which removes surface dielectric barriers. As an example for electrical interfaces, we create a mortise-tenon joint used in a plug-in connector for electric vehicles that exhibits one-eighth the area-normalized electrical resistance of its commercial counterparts. For thermal interfaces, we create a finger-joint interface used in electronic cooling that exhibits a thermal resistance as low as 2.3 K mm2 W-1, resulting in an additional chip temperature drop of 44 degrees C compared with current solutions when used between a light-emitting diode chip and a copper heat sink. The metainterfaces can be made with regular machining as the patterning technique and are readily scalable.

Grain-boundary (GB) dynamics control the stability, mechanical, and functional response of nanocrystalline materials, but direct experimental access to their slow non-equilibrium motion has been limited. Here we establish X-ray photon correlation spectroscopy (XPCS), combined with domain-adaptive machine learning, as a quantitative probe of GB dynamics. Temperature- and grain-size-dependent two-time XPCS measurements in nanocrystalline silicon reveal pronounced departures from time-translation invariance, showing that GB relaxation can remain far from equilibrium over experimental timescales. However, direct extraction of quantitative physical information from these high-dimensional, noisy fluctuation maps faces a significant challenge. To overcome this barrier, we develop a semi-supervised learning framework that transfers physical parameter labels from continuum simulations to unlabeled experimental XPCS maps through domain-adaptive representation alignment. This AI-augmented approach enables the extraction of key kinetic parameters, including bulk diffusivity, GB stiffness, and effective GB concentration, directly from experimental XPCS measurements. Our results show how machine learning can transform indirect fluctuation signals into quantitative materials dynamics, providing a general route to study non-equilibrium defect motion in solids.

Although numerous high-zT GeTe systems have been reported, the correlation between zT and conversion efficiency η remains poorly defined. Reporting peak or even average zT often leads to an overestimation of η as these metrics assume temperature-independent transport properties-an unrealistic simplification. In 2017, the thermoelectric device ZT for a single material (making either n- or p-type leg) was introduced to evaluate the maximum efficiency of a thermoelectric leg with temperature-dependent properties. This work provides an experimental validation of Device ZT as a practical metric for predicting thermoelectric conversion efficiency. By combining calibrated single-leg efficiency measurements, heat-flow analysis, and finite-element simulations, we establish a direct correlation between material transport properties, Device zT, and experimentally measured device performance. The CuSb-GeTe leg achieves an efficiency of 5.4% at ΔT = 400 K, while conventional zTavg-based estimates overpredict performance by nearly twofold. In contrast, Device ZT accurately reproduces the measured efficiency. These findings provide an experimentally validated framework for translating material-level thermoelectric properties into realistic device performance and establish Device ZT as a practical metric for evaluating thermoelectric devices.

Abstract Defects play a vital role in understanding and elucidating the structure-property relationship in materials science. Here we report the existence of a new kind of planar defect, filling fault, in half-Heusler (HH) compounds—a structurally and functionally diverse family of materials. The ideal HH structure is composed of three occupied and one vacant sublattices. The two-dimensional filling fault layer tends to form with the originally vacant 4d sites occupied, fully or partly, which results in a vacancy swap occupation between the 4c and 4d sublattices. It is found that the filling faults are apt to form fully in Ni-based ZrNiSn and ScNiSb, moderately in Co-based TiCoSb, and hardly in Fe-based NbFeSb and VFeSb. The unique defect configurations and the formation rule clarify the long-standing structure puzzles in HH compounds. Furthermore, we find that the filling faults and associated vacancy swap occupation can significantly enhance the piezoelectricity and change the temperature dependence of electrical conductivity. These results deepen the understanding of precise crystallographic structures and defect-property relationships in solids, and facilitate the design of advanced materials and functional devices.

Ga-In co-doped ZnO ceramics were synthesized by a solid-state reaction to elucidate the role of defect chemistry and dopant solubility in governing thermoelectric transport. Compared with pristine ZnO and Ga singly doped ZnO, the In-Ga co-doped sample (IGZO) exhibits a distinct evolution in microstructure and defect configuration. A minor fraction of Ga2O3(ZnO)9 spinel phase is present, contributing to phonon scattering as part of a composite-like microstructure. Structural analyses reveal that In incorporation suppresses the formation of Ga2O3(ZnO)9 spinel clusters, indicating an expansion of Ga solubility in the ZnO lattice. This solubility-assisted effect leads to a transition from secondary-phase-dominated microstructures toward point-defect-dominated configurations. The competition between spinel clusters and point defects plays a key role in determining phonon scattering behavior, where atomic-scale defects become dominant in the IGZO ceramic. At the same time, the enhanced substitution of Ga, assisted by In, modifies the electronic structure through increased carrier concentration and density-of-states effective mass. As a result, the IGZO ceramic exhibits reduced lattice thermal conductivity and modified charge transport behavior under conventional bulk processing conditions, with a ZT of 0.182 at 1073 K. These findings highlight that solubility-assisted defect engineering provides an effective pathway to tune defect hierarchy and transport mechanisms in ZnO-based thermoelectric materials.

ABSTRACT Grain boundary (GB) scattering of charge carriers plays an important role in the electrical properties of materials. For thermoelectrics, this scattering can significantly reduce the carrier mobility and output power. Various GB engineering strategies, such as increasing the grain size and decreasing the GB potential barrier height, have been demonstrated in n‐type Mg 3 Sb 2 ‐based thermoelectrics. Yet, similar effects in p‐type Mg 3 Sb 2 have been less reported, and effective methods to modify the GB potential are elusive. Here, we reveal that Cd can reduce the GB barrier height via segregation to the GB. The enrichment of Cd suppresses the formation of hole‐killer defects (Sb Mg + ) and thus reduces the GB resistance, leading to significant improvements in the electrical conductivity and power factor. A two‐phase model shows a progressive decrease in the GB barrier height with increasing Cd content. Simultaneously, the GB Cd segregation and grain‐interior Cd alloying strengthen the phonon scattering and reduce the sound velocity, substantially reducing the thermal conductivity. Consequently, a maximum ZT of 0.84 is achieved in Mg 2.5 Cd 0.5 Sb 2– 1 at% NaF at 773 K. These findings unravel the hidden role of Cd in tuning the GB characteristics and the transport properties of Mg 3 Sb 2 , revisiting the functionality of conventional dopants in materials design.

Thermoelectric materials, despite their desirable heat-to-electricity conversion properties, have seen limited commercial viability due to their labor-intensive and high-cost conventional synthesis and assembly processes. Here, we demonstrate an ink printing/sintering approach which unlocks cost-effective fabrication of high-performing, geometrically-complex La3-xTe4 thermoelectric legs with high relative densities and phase purity. LaTe1.47 legs are created by ink-extrusion printing of pre-alloyed powders, followed by debinding and sintering at high temperature; they achieve a high figure of merit (zT = 1.49 ± 0.24 at 1250 K), on par with state-of-the-art LaTe1.46 synthesized via traditional hot pressing. Furthermore, the ink printing methodology enables printing and diffusion bonding of non-flat interfaces between a LaTe1.47 leg and a Ni electrode, which are designed to achieve high mechanical interlocking with minimal chemical interdiffusion. After measuring creep properties on dense La3-xTe4, we perform simulations of the thermomechanical stress at these LaTe1.47/Ni interlocking interfaces during operation; this demonstrates the importance of creep in relaxing the stress state of both phases and the potential for designed interfaces to mitigate crack propagation at the thermoelectric-metal junction. This additive approach addresses the key challenges with thermoelectric device fabrication, enabling thermoelectric devices which are economical, scalable and thermomechanically-robust at very high temperatures.

The thermoelectric properties of undoped ZnO (ZnO), Mg-doped ZnO (MZO), Al-doped ZnO (AZO), and Ga-doped ZnO (GZO) films were addressed in terms of the electronegativity difference (Delta chi). A smaller Delta chi not only increases bond covalency and larger orbital overlap but also reduces intrinsic defects, leading to higher electrical conduction and larger carrier mobility observed in AZO and GZO films. In contrast, the MZO films with the largest Delta chi, which enhance intrinsic defects (oxygen vacancies, etc.) or potential barriers at grain boundaries, reduce bond covalency, or increase bond ionicity, result in the lowest electrical conductivity and carrier mobility. The power factor PF values are 93, 27, 49, and 148 mu W m-1 K-2 for the ZnO, MZO, AZO, and GZO films, respectively. Films having atoms with smaller Delta chi values achieve the larger PF value. As a result, replacing Zn sites with dopants of larger electronegativity or small electronegativity difference is an indicator for enhancing the thermoelectric properties of ZnO materials.

Triboelectric X-ray generators (TXGs), which convert mechanical energy from tribological interactions into soft X-ray output, represent an approach that offers advantages including structural simplicity, portability, and low fabrication cost. In this work, tribo-generated high-voltage electric fields (similar to 30 kV) accelerate triboelectrification electrons, and bremsstrahlung X-ray radiation occurs. This is attributed to the deceleration of high-energy electrons after their collision with adhesive tape, generating nanosecond X-ray pulses. The radiation pulse was quantified by a semiconductor X-ray detector, and imaging and defect detection were demonstrated using dental X-ray film. We investigated the optimization of the TXG energy conversion process, focusing on enhancing X-ray output through dual-level modulation of electric fields and the electron source. By regulating the inter-surface electric field and free electrons, the X-ray output intensity is increased by over 6-fold compared to the unoptimized TXG.

As the demand for advanced thermal management materials continues to grow, high thermal conductivity materials, especially those with superior chemical stability and mechanical strength, are increasingly sought after from both theoretical and applied perspectives. Two-dimensional transition metal carbides (MXenes) meet these criteria and have been theoretically predicted to possess high thermal conductivity. However, experimental validation of their intrinsic transport properties—particularly the effects of surface terminations, lateral size, and thickness—remains limited due to the difficulty of obtaining clean, large-area samples. Here, we experimentally investigate single-flake Ti3C2Tx with thicknesses ranging from ~9 to 45 nm to reveal its thickness-dependent thermal behavior. Using the energy transport state-resolved Raman technique, we measured thermal conductivities ranging from 30.6 W m−1 K−1 to 105 W m−1 K−1, consistent with theoretical predictions. These results offer direct insight into intrinsic thermal transport in MXenes and underscore their promise for efficient thermal management applications. Experimental validation of the intrinsic transport properties of 2D MXenes is challenging due to difficulty in obtaining clean, large-area samples. This paper investigates single-flake Ti3C2Tx with thicknesses ranging from ~9 to 45 nm to reveal its thickness-dependent thermal behavior.

Pushing the intrinsic lattice thermal conductivity (LTC) in crystalline materials to lower bounds is crucial for fundamental materials research towards emerging technologies including thermoelectric energy conversion and thermal management in both hypersonic aircraft and next-generation turbine systems. However, in the ultralow LTC regime (< 1 Wm(-1)K(-1)), the competition between propagative (particle-like) and coherent phonons-arising from off-diagonal components-poses a significant challenge in further reducing LTC. We perform quantitative analysis of 4700 materials using density functional theory (DFT), spanning all crystallographic groups, to elucidate the interplay between diagonal and off-diagonal phonon contributions. We identify a critical balance between these transport mechanisms, where intermediate phonon lifetimes (similar to 1 ps) and slow group velocities (similar to 1 km/s) collectively suppress both contributions, enabling ultralow LTC. Results from a large dataset of 31,058 structures by machine learning models strongly resemble the DFT trends of two-channel phonon transport. Leveraging these models, we screen 25,882 additional materials and confirm their properties with DFT, identifying 12 candidates with ultralow room-temperature LTC-including a record-low value of 0.132 Wm(-1)K(-1). Our large-scale analysis reveals fundamental insights into dual-channel phonon transport, enabling rational design of ultralow LTC materials and accelerating the discovery of advanced phononic crystals with tailored thermal transport properties.

Bi2SeO2 is a promising n-type semiconductor to pair with p-type BiCuSeO in a thermoelectric (TE) device. The TE figure of merit zT and, therefore, the device efficiency must be optimized by tuning the carrier concentration. However, electron concentrations in self-doped n-type Bi2SeO2 span several orders of magnitude, even in samples with same nominal compositions. Such unsystematic variations in the electron concentration has a thermodynamic origin related to the variations in native defect concentrations. In this study, we use first-principles calculations to show that the selenium vacancy, which is the source of n-type conductivity in Bi2SeO2, varies by 1-2 orders of magnitude depending on the thermodynamic conditions. We predict that the electron concentration can be enhanced by synthesizing under more Se-poor conditions and/or at higher solid-state reaction temperatures (T_SSR), which promote the formation of selenium vacancies without introducing extrinsic dopants. We validate our computational predictions through solid-state synthesis of Bi2SeO2. We observe more than two orders of magnitude increase in the electron concentration simply by adjusting the synthesis conditions. Additionally, we reveal the significant effect of grain boundary scattering on electron mobility in Bi2SeO2, which can also be controlled by adjusting T_SSR. By simultaneously optimizing the electron concentration and mobility, we achieve a zT of ~0.2 at 773 K for self-doped n-type Bi2SeO2. Our study highlights the need for careful control of thermodynamic growth conditions and demonstrates TE performance improvement by varying synthesis parameters according to thermodynamic guidelines.

Complex shapes are created from Yb14MnSb11, a high-temperature thermoelectric Zintl phase, via a two-step process: i) layer-by-layer 3D-extrusion of ink containing partially-reacted powders which are ball-milled from a blend of Yb, MnSb, and Sb powders; ii) heat treatment to synthesize the ternary compound Yb14MnSb11 and densify the extruded powders. A high phase purity for Yb14MnSb11 (83-94%) is achieved in both cast and 3D-extruded ink specimens via a solid-state reaction between Yb, MnSb, and Yb4Sb3 during reactive sintering. Pressure-free sintering at temperatures of 1200-1400 degrees C densifies the powders to 82% relative density but can also induce the decomposition of the Yb14MnSb11 phase due to Yb sublimation. A process window with optimized sintering temperature and time is identified, achieving both low porosity and high phase purity and reaching a maximum zT = 0.61 at 1000 degrees C, about half of the maximum zT value for bulk Yb14MnSb11 made via conventional processes (pressure sintering of precursor powders). The present approach - direct ink writing of ball-milled powders, combined with reactive sintering - is a scalable and affordable method to fabricate thermoelectric legs with intricate 3D shapes, for enhanced performances in high-temperature thermoelectric applications.

Conformal shaping of thermoelectric (TE) materials is essential for integrating energy harvesting devices, thermal management, and temperature sensing onto irregular surfaces in flexible or wearable electronics. Processing and figure of merit (zT) measurement of conformal TE materials remains a challenge. We introduce a novel mechanical hammering technique to fabricate high-performance Ag₂Se TE materials into complex conformal shapes while preserving their excellent TE properties. A specialized Harman measurement system was developed to accurately characterize the zT curved geometries, overcoming limitations of conventional TE measurement approaches. We successfully produced Ag₂Se films with curvature radii of 5 mm, 10 mm, and 15 mm, as well as a sophisticated saddle-shaped structure incorporating both positive and negative curvatures. The hammering process induces strong grain boundaries, leading to enhanced charge transport properties without compromising the material's low thermal conductivity. The curved Ag₂Se samples exhibit a remarkable peak zT of ~1.0 at room temperature, matching the performance of traditional bulk forms. This work establishes a scalable mechanical processing route for creating flexible, high-efficiency TE materials in diverse conformal configurations, opening new possibilities for their integration into wearable electronics, curved-surface energy harvesters, and advanced thermal management systems.

A new strategy to trace the chemical bonding origins of band structure explains silicon's unusual low-symmetry conduction band minimum.

In this study, we systematically investigate the thermoelectric properties of (NbxW1-x)18O49 compounds (0 <= x <= 0.20). Unlike conventional doping approaches that aim to introduce additional charge carriers to enhance electrical conductivity, our non-conventional method focuses on tuning the electronic structure of the W18O49 matrix by substituting W5 + with Nb5+ cations. This substitution reduces the maximum number of donating electrons from 108 to approximately 104, approaching the 98 electrons accepted by the anions. This electronic tuning results in a significant increase in the Seebeck coefficient to-133.717 mu V/K and achieves a remarkable zT value of approximately 0.318 at 800 degrees C. Additionally, Nb doping induces notable structural distortions, leading to modifications in morphology and a reduction in lattice thermal conductivity. Despite a decrease in electrical conductivity with higher Nb content, the overall thermoelectric performance is enhanced, highlighting the potential of (NbxW1-x)18O49 Pentacolumn structures for high-temperature thermoelectric applications.