Browse Articles
Discover research articles across all indexed journals
Ionic liquid induced structural transformation in a copper-based MOF synthesis: Experimental and DFT investigations
The development of advanced materials for carbon management has driven significant interest in hybrid systems based on metal–organic frameworks (MOFs) and ionic liquids (ILs), owing to their tunable structures and enhanced affinity for CO2. In this study, we demonstrate that the IL [BMIm][BF4], employed as an ionothermal medium, plays a dual role as both solvent and structure-directing agent, enabling the formation of a two-dimensional copper–terephthalate network (Cu-BDC-2), which differs from the material obtained via conventional solvothermal synthesis (Cu-BDC-1). Structural and spectroscopic analyses reveal that IL-derived species remain associated with the framework surface, directly influencing its physicochemical properties. CO2 adsorption measurements at 85 °C show that Cu-BDC-2 exhibits enhanced interactions with CO2, attributed to the effect of IL-derived species. In contrast, theoretical calculations indicate that, in Cu-BDC-1, van der Waals interactions between CO2 and the aromatic linkers dominate the adsorption process. Furthermore, the GFN1 − xTB method shows good agreement with DFT + U results, highlighting its potential as a computationally efficient approach for screening such systems. These findings demonstrate that ionothermal synthesis is an effective strategy for modulating both the structure and surface functionality of MOF-based materials, opening new perspectives for their applications.
Anisotropic magneto-transport properties of semimetallic LaNiSb3
Single crystals of LaNiSb3 were synthesized using the Sn-flux method. Structural characterization confirms that LaNiSb3 crystallizes in the orthorhombic Pbcm space group with lattice parameters a = 13.0970(2) Å, b = 6.1400(4) Å, and c = 12.1270(4) Å. Electrical resistivity measurements demonstrate metallic behavior over the entire temperature range of 3–300 K. The magnetoresistance (MR) exhibits a positive anisotropic response, attaining a maximum of 8% for H∥b, with a pronounced crossover from quadratic to nearly linear field dependence. Angular-dependent MR measurements reveal a pronounced twofold symmetry upon magnetic-field rotation within both the ab and ac crystallographic planes up to 50 K, indicating anisotropic charge transport. Hall resistivity measurements show predominantly electron-type conduction at high temperatures, with an increasing hole contribution upon cooling. The multiband character is further corroborated by the violation of Kohler’s scaling and is well described within a semiclassical two-band framework. Collectively, these results suggest LaNiSb3 has anisotropic multiband electronic transport and could be a compelling candidate to explore structure–property correlation as a topological semimetal.
N-component free energy lattice Boltzmann method with reduction consistency and global momentum conservation
We present a free energy lattice Boltzmann model capable, in principle, of simulating fluid systems with an arbitrary number of immiscible components. Our method is strictly reduction consistent, ensuring that absent fluid components do not spontaneously nucleate. We introduce a novel discretization of the surface tension force that globally conserves momentum to machine precision, and we enforce reduction consistency through a flux correction that is independent of the mobility. The method is benchmarked with a range of static and dynamic problems, including liquid lenses, Janus droplets, quaternary phase separation, and six-component layered Poiseuille flow, and we obtain excellent agreement with theoretical predictions throughout. Finally, we demonstrate the applicability of the proposed method through patterned liquid surfaces and microfluidic emulsion droplet generation.
Electronic transport in order–disorder coupled bilayer graphene
Within the tight-binding framework, we investigate the electronic transport properties of AA- and AB-stacked ordered–disordered bilayer graphene using matrix diagonalization combined with the quantum diffusion theory, in the framework of a realistic materialized tight-binding Hamiltonian with long-range hopping. It is found that interlayer compression opens a bandgap, with the critical interlayer distance for gap opening being influenced by the disorder strength. As disorder increases, the electronic bands broaden, the density of states in the band tails remains low, while the band-center density stays high and stable, approaching the monolayer graphene band structure under strong disorder. Participation-number analysis indicates that all electronic states are nonlocalized under weak disorder, whereas mobility edges emerge under strong disorder, separating nonlocalized states in the band center from localized states in the band tails, and gradually blur as the interlayer distance decreases. Importantly, a universal anomalous quantum diffusion behavior is observed, regardless of interlayer compression, the electronic mobility undergoes a transition from decrease to increase with increasing disorder strength. By jointly tuning disorder strength and interlayer distance, the electronic diffusion can be effectively controlled between superdiffusive and subdiffusive regimes. Ordered–disordered bilayer graphene quantum films may be realized via doping, micro-/nano-fabrication, or substrate engineering. These results provide new theoretical insights into bilayer graphene and guidance for the design of novel electronic devices.
Why projection-based WF-in-DFT cannot be exact, even with the exact exchange–correlation functional. Formal and practical sources of errors
We establish theoretical foundations for embedding a correlated wavefunction in an environment formed by Kohn–Sham orbitals. We show that introducing an approximation that equates two, in principle, distinct kinetic-energy functionals yields an embedding functional identical to the projection-based wavefunction-in-density functional theory (DFT) formulation of Miller and co-workers. We demonstrate that this functional is inherently nonvariational: its minimum is not guaranteed to coincide with the exact ground-state energy and remains bounded from above by it. Building on this formal framework, we analyze the dominant sources of error in projection-based density matrix renormalization group-in-DFT embedding with approximate exchange–correlation (xc) functionals. Using molecules with dissociating covalent bonds as a diagnostic example, we demonstrate that the primary source of error is the nonadditive exchange–correlation energy describing the nonclassical coupling between the active subsystem and its environment. Eliminating the fractional-spin error by employing a pair-density xc functional (pair-density functional theory) instead of a semilocal generalized gradient approximation (GGA) does not remedy this deficiency, because the inaccuracy stems from self-interaction effects at the subsystem–environment interface.
Comparison of terahertz radiation and hot-electron dynamics in CdTe and GaAs photoconductive antennas excited at 520 nm
Photoconductive antennas (PCAs) remain a cornerstone for efficient terahertz (THz) generation. Given the recent proliferation of ytterbium (Yb) laser technology, adapting PCAs to these specific laser characteristics is essential. In this work, we investigate the THz emission of typical dipole-type semi-insulating gallium arsenide (SI GaAs) and cadmium telluride (CdTe) PCAs with a 30 μm gap, driven by the second harmonic (520 nm) of an Yb-based femtosecond oscillator laser. We demonstrate that, with optimized bias and optical excitation, CdTe PCA can generate a peak-to-peak THz field 53% larger than that of GaAs. While GaAs initially exhibits higher emission at low fields, the ability of the CdTe PCA to sustain higher bias fields at high optical power without catastrophic failure is an advantage that leads to the generation of higher THz fields. Furthermore, we highlight fundamental differences in hot-electron dynamics between the two semiconductors. Disparities in intervalley scattering rates, which govern carrier mobility and the resulting space-charge screening regime, lead to markedly different evolutions of the THz waveforms in GaAs vs CdTe. These findings establish CdTe as a robust alternative for high-power THz applications driven by Yb-based laser systems.
Exploration of rotational and vibrational stimulated Raman scattering in oxygen-filled fiber
Stimulated Raman scattering (SRS) is performed in pressurized oxygen (O2) within a nested antiresonant fiber (NARF) to investigate the interplay between rotational and vibrational SRS. In this work, the O2-filled NARF is pumped by 500 ps pulses at λ = 1.064 μm with peak powers of up to 60 kW. The pump generates cascaded vibrational Stokes lines spaced by ∼1550 cm−1, and each of these generates cascaded Stokes and anti-Stokes rotational SRS. The extremely spectrally broad guidance of the NARF enables, for the first time, the measurement of the third vibrational Stokes order, which is beyond 2.1 μm with our pump wavelength. We report on the threshold energies and conversion efficiencies of each vibrational SRS order. In addition, we observe the pressure-dependent absorption of O2 in the near-infrared and discuss the ramifications. We also collect pump polarization-dependent spectra while varying O2 pressure and pump power to determine the gain competition between rotational and vibrational SRS. This work explores the spectroscopic complexities of O2 and presents a new multispectral fiber laser source.
Displacement damage induced disorder and thermal transport degradation in AlGaN alloys: A molecular dynamics study
High electron mobility transistors based on AlGaN are widely deployed in radiation rich environments, where displacement damage can degrade both material integrity and heat dissipation. Yet, how displacement induced disorder couples to thermal transport degradation in AlxGa1−xN alloys across composition and temperature remains unclear. Molecular dynamics simulations were used to quantify radiation-induced structural disorder and thermal transport degradation in AlxGa1−xN random alloys. Defect accumulation was tracked by Frenkel pair analysis, while structural decoherence was quantified by a tetrahedra based local order parameter distribution and an amorphous-like atomic fraction defined from its low order weight. The composition response is nonmonotonic: Al0.5Ga0.5N exhibits the strongest defect retention and the largest growth of amorphous-like environments, whereas Al0.25Ga0.75N shows the weakest overall damage signature under the same dose. For a fixed composition, higher temperature systematically increases residual damage and raises the amorphous-like fraction, indicating that thermally assisted migration promotes persistent clustered and reconstructed states rather than restoring crystalline order. Radial distribution functions confirm the inferred disorder through peak broadening and rapid attenuation of medium-range oscillations, supporting the reliability of the tetrahedral metric. Equilibrium molecular dynamics calculations show that irradiation amplifies compositional sensitivity of thermal conductivity, with Al0.5Ga0.5N remaining the lowest after damage. The largest dimensionless thermal resistance increase coincides with the highest amorphous-like fraction in Al0.5Ga0.5N, while the temperature dependence of the relative resistance change reflects joint control by the pristine temperature-dependent transport baseline and the disorder level.
Broadband dielectric measurements of primary linear monohydroxy alcohols with microwave microfluidic spectroscopy
Dielectric spectroscopy can provide information about the relaxation dynamics and intermolecular interactions in fluids. The past century of research has resulted in many hypotheses about intermolecular interactions in alcohols, including hydrogen bonding dynamics. In spite of extensive efforts to measure the dielectric properties of alcohols, there is a range of specificity of results in the literature. The range of reported values in existing data makes it challenging to validate models of relaxation dynamics and examine trends. Here, we report the dielectric spectra of eight primary linear monohydroxy alcohols, with accompanying uncertainties, measured at 25 °C from 100 MHz to 110 GHz using microwave microfluidic spectroscopy. The spectra of alcohols were fitted with a model containing two relaxations: a dominant Debye relaxation and a higher frequency non-Debye relaxation. We validated our fitting approach by comparing the residuals to the uncertainties. We found that the Debye relaxation changes systematically as a function of the number of carbons in the alkyl chain, decreasing in magnitude and increasing to higher time constants. The non-Debye relaxation also shifts downward in frequency as the chain length increases. Overall, these results are a comprehensive set of broadband dielectric spectra that can be used to test and validate models for intermolecular interactions in alcohols.
Monte Carlo simulation of transverse-field-induced spin reorientation in Dy2Fe14B under atomic-scale thermal fluctuations
Rare-earth permanent magnets are key materials for modern energy and transportation technologies, and understanding their magnetic behavior under applied magnetic fields is of both fundamental and technological importance. Recent advances in ultrahigh magnetic-field experiments are revealing previously inaccessible magnetization regimes in these materials. In this study, we investigate the transverse-field magnetization process of Dy2Fe14B over a wide magnetic-field range at different temperatures using an atomistic modeling approach. This recently developed framework enables us to study the microscopic mechanisms of magnetization processes, including the effects of atomic-scale thermal fluctuations. We show that, qualitatively, different mechanisms of field-induced spin reorientation operate in the low- and high-temperature regimes and identify the characteristic magnetic fields associated with these processes. Our results demonstrate how the competition among anisotropy, exchange, and Zeeman energies governs field-induced spin reorientation in heavy rare-earth magnets, providing microscopic insight into their high-field magnetic behavior and thermal stability.
Prospects for detecting UF6 hydrolysis intermediates by mass spectrometry, resonance Raman, and NQR spectroscopy
Simulations are performed to consider several spectrometric and spectroscopic candidates for elucidating the mechanism of the hydrolysis of uranium hexafluoride (UF6). This study is among the first to benchmark the def-mTZVP basis sets for actinide-containing molecules, and it is shown to be a suitable basis set for surveying geometrical structures and vibrational spectra when used in conjunction with density functional theory. An experiment is proposed coupling mass spectrometric ion selection with vibrational spectroscopy, and supporting infrared and Raman spectral simulations demonstrate that ionization blue-shifts bands and can change their qualitative features. Ultraviolet resonance Raman is shown to have good prospects for discriminating U–O–U bridged intermediates, evidence for which was observed recently [L. E. McNamara et al., J. Phys. Chem. A 130, 775–786 (2026)]. As a more speculative approach, we also consider the prospect of addressing 233U or 235U nuclei by quadrupole resonance (NQR) spectroscopy for assigning early stage intermediate complexes. In doing so, we provide a first order-of-magnitude estimate for the collision-induced NQR signal for the UF6 dimer, which is of interest for the interpretation of a fast decoherence time observed in liquid-phase nuclear magnetic resonance. Having provided critical insights into the formation and stability of UF6 hydrolysis intermediates in previous studies, this computational spectroscopy survey is expected to help guide and expedite future laboratory campaigns.
Noise-aware machine learning accelerates development of wide-hysteresis NiTi-based shape memory alloys
NiTi-based shape memory alloys (SMAs) with wide thermal hysteresis show great potential in engineering applications such as pipe couplings. However, traditional trial-and-error methods are costly and time-consuming, hindering the development of wide-hysteresis alloys. Although machine learning enables efficient exploration of NiTi-SMA compositions, most studies overlook experimental noise. To address this, we propose a noise-aware Kriging model that achieves high predictive accuracy (R2 = 0.91, RMSE = 6.02) for rapidly screening alloys with wide hysteresis. Using this approach, we designed novel NiTiNbTa alloys tailored to specific processing and storage requirements. Among them, Ni49.5Ti44.5Nb4.5Ta1.5 and Ni49.5Ti44.5Nb5.5Ta0.5 can be directly processed after low-temperature storage, while Ni49Ti45.5Nb4Ta1.5 remains stable under ambient conditions. All three exhibit a thermal hysteresis over 70 K without post-processing and a shape memory recovery rate above 90% under 400–900 MPa stress. This work offers a valuable strategy for designing high-performance wide-hysteresis SMAs.
Spectral leakage and masking effects in the measurement of hyperuniformity
Disordered hyperuniformity is a recently discovered novel state of matter, characterized by a complete suppression of normalized infinite-wavelength density fluctuations as in perfect crystals and lack of conventional long-range order nor broken symmetry as in glasses. The detection of hyperuniformity relies critically on accurate characterization of the small-wavenumber behavior of the static structure factor of the system. In practice, however, measurements are performed on finite subsystems or through incomplete observations that effectively mask portions of the underlying configuration. Inspired by a recent numerical study [Liu et al., J. Chem. Phys. 164, 094102 (2026)], we develop a unified theoretical framework that quantifies how finite windows and spatially correlated binary masks modify the observed structure factor. We show that the measured structure factor Sobs(k) is the convolution of the intrinsic structure factor with the spectral density of the observation function, whether it is a compact window or an extended random mask. For generic hyperuniform systems with small-k scaling S(k) ∼ kα, finite observation window induces a universal quadratic leakage term at sufficiently small wavenumbers (i.e., k ≲ 1/L), leading to an apparent k2 scaling independent of the true exponent. The true hyperuniform exponent α can only be measured in the intermediate regime 1/L ≪ k ≪ qc. In stealthy hyperuniform systems, where the intrinsic structure factor possesses a spectral gap, all observed small-k power arises entirely from this convolution mechanism. For spatially correlated masks, we derive the corresponding convolution relation in terms of the mask spectral density and identify conditions under which hyperuniform signatures are suppressed, preserved, or distorted. Our results establish quantitative criteria for reliably extracting intrinsic scaling exponents and distinguishing genuine hyperuniform order from measurement-induced artifacts.
Review: Structural properties of binary metal nitrides
Thin films of binary metal nitrides with cubic or hexagonal crystal structures have a wide range of industrial applications. Their potential and that of their alloys range from metallic protective coatings for cutting tools and quantum wells in light-emitting devices to current-carrying channels in high-frequency and power transistors, as well as piezoelectric actuators and resonators in microelectromechanical and acoustic components. For the epitaxy of the required films and layer structures as well as for the design of metal nitride-based piezoelectric, electronic, and optoelectronic devices, precise knowledge of their structural properties is particularly important. It is, therefore, surprising that simulated and experimental data on lattice parameters, even for binary nitrides that have been studied for many years, differ too widely or are not sufficiently consistent for targeted design, e.g., for piezoelectric resonators with applications in future mobile communication systems. This review combines two complementary approaches to select and improve the accuracy and consistency of the physical data. First, based on a search of published data and supplementary simulations, two cubic (rock-salt and zinc-blende) and two hexagonal crystal structures (layered hexagonal and wurtzite) are described for each binary metal nitride, regardless of whether the respective crystal lattices are mechanically and thermodynamically stable under normal conditions. Second, all metal nitrides present in these crystal lattices, from the metal atomic number of 5 for BN to 81 for TlN, are comparatively examined. The systematic comparison of the structural properties of related cubic and hexagonal crystal structures of each binary metal nitride, combined with comparing different metal nitrides with each other by varying the atomic number of the metal atom over a wide range, reveals a clear dependence of the lattice parameters, bond lengths, nearest neighbor metal–metal distances, average volumes required per atom, and mass densities on the electron configuration of the metal atoms. In particular, the systematic changes in the structural properties of binary metal nitrides, whose metal atoms belong to the same period or group in the periodic table of elements, are presented here and are used to determine improved data sets of structural parameters.
Pause and change point detection in single-molecule motor trajectories by BIC-based model selection
Detecting pause events in single-molecule trajectories is essential for extracting kinetic information but remains challenging in noisy measurements. Here, we present a pause-detection method based on model selection using the Bayesian Information Criterion. The approach operates directly on raw trajectories, requires no preprocessing, and does not rely on user-defined parameter tuning. We provide a rigorous analysis of detection sensitivity that quantifies the probability of missing events under given experimental conditions, and the method naturally accommodates missing data in trajectories. Benchmarking shows that the approach outperforms commonly used pause-detection and change-point-detection strategies across simulated and experimental datasets. The method provides a robust and general framework for pause and change-point detection in single-molecule trajectories and is broadly applicable to single-molecule measurements obtained with magnetic tweezers, optical tweezers, and single-particle tracking.
Stable low-resistance ohmic contact on GaN via an oxynitride interlayer over 20–200 °C: Mechanisms and performance
Achieving reliable, low-resistance ohmic contacts in GaN-based devices is fundamentally limited by interfaces with high interface-state densities, which are a direct result of essential fabrication steps. For example, etching to expose the n-GaN layer in laser diodes introduces surface defects, while regrown n+-GaN surfaces in selective-area regrowth (SAG) HEMTs exhibit high interface states due to growth kinetics. These high-frequency and high-power devices typically operate at elevated junction temperatures (100–200 °C), where high interface-state density exacerbates Fermi-level pinning (FLP) effect, leading to unstable contact resistance and reliability concerns. To address this issue, we propose and validate a universal interface engineering strategy: inserting an oxynitride interlayer. We systematically study Ti-based ohmic contacts on SAG n+-GaN from 20 to 200 °C, comparing untreated interfaces with those modified by TiOxNy or GaOxNy interlayers. Through correlated microstructural analysis, temperature-dependent electrical measurements, and conductance-based interface-state characterization, we show that the interlayers effectively suppress interface-state density, mitigate FLP, and stabilize the Schottky barrier height. While untreated contacts display non-monotonic and degrading resistivity with temperature, interlayer-modified contacts exhibit a stable, monotonic decrease in specific contact resistivity. The GaOxNy interlayer yields the best performance, achieving ∼10−8 Ω cm2 order above 100 °C. This work elucidates the critical role of interface states in contact thermal instability and provides a viable materials solution for stable ohmic contacts in GaN-based devices that operate at high temperatures with damaged or low-quality contact interface.
Quantum dynamics of photoexcited charge transfer in bridge-mediated porphyrin–naphthalenetetracarboxylic dianhydride donor–acceptor system
Photoinduced charge transfer (CT) in donor–bridge–acceptor (D-B-A) systems is governed by the interplay between electronic coupling and vibronic interactions. Here, we investigated ultrafast CT dynamics in Zn-porphyrin–NTCDA complexes connected through either an extended 1,4-diethynylphenylene (DEP) bridge or an ethynyl linker. The excited-state dynamics are described using a linear vibronic coupling Hamiltonian combined with quantum dynamical wavepacket propagation, thereby allowing explicit treatment of multidimensional nuclear motions. Electronic-structure analysis shows that the conjugated DEP bridge provides stronger donor–acceptor communication and more favorable energetic alignment between locally excited and CT states. As a result, population transfer to the CT configuration occurs on an ultrafast timescale of ∼9 fs. In contrast, modifying the bridge reduces π-conjugation and electronic interaction, leading to slower CT dynamics of ∼25 fs. Systematic inclusion of increasing numbers of vibrational modes reveals that cooperative multimode coupling enhances vibronic dephasing, stabilizes the charge-separated population, and suppresses coherent recurrences in the electronic dynamics. The simulations, therefore, identify bridge-mediated conjugation and multidimensional vibronic interactions as key factors governing charge-separation efficiency in these architectures. These findings provide molecular-level insights into how bridge structure and vibrational motions cooperatively regulate CT dynamics in D-B-A systems.
Front-end bandwidth and observability limits in confined electronic systems
The progressive miniaturization of solid-state electronic systems drives device operation toward regimes characterized by increasingly rapid electronic dynamics, particularly in reduced-dimensional structures subject to quantum confinement. In this regime, the central issue is not only the intrinsic speed of electronic processes, but also the extent to which their temporal structure remains experimentally accessible through conventional readout chains. This work develops a phenomenological framework for the mismatch between intrinsic confined-state dynamics and the finite temporal resolution of the measurement front-end. The physical ingredients involved—quantum confinement, finite bandwidth, and detector-mediated filtering—are individually well established; the contribution here is to combine them into an explicit observability-based description with operational regime descriptors. To this end, we introduce a dimensionless observability parameter that identifies the crossover between observable and observability-limited regimes, together with an observability transfer efficiency that quantifies the fraction of intrinsic dynamical content preserved by the readout chain. A minimal analytical model shows that the same control parameter governs retained spectral content, signal attenuation, and peak delay, thereby linking spectral filtering to measurable time-domain distortions. Within this formulation, the measurable electrical response is treated as a property of the coupled system formed by the electronic states and the front-end, rather than of the confined system in isolation. The resulting framework provides a compact basis for analyzing observability limits in nanometric electronic systems and for identifying experimentally testable crossover behavior when intrinsic and instrumental timescales become comparable.
Nonequilibrium vibrational distribution and kinetics of OH chemiluminescence in the hydrogen–oxygen reaction
This paper describes a spectroscopic approach for studying the kinetics of the reaction H + O + M → OH(A2Σ+) + M (R1). Vibrational distributions of OH(A) produced during high-temperature reactions of H2/O2 mixtures diluted with Ar have been measured by observing the emission spectra of the OH A2Σ+–X2Π transition. The experiments were performed behind reflected shock waves at temperatures of 1600–2200 K and pressures of 1.4–5.2 bar in a high-repetition-rate shock tube. The observed emissions are attributed to the chemiluminescence of OH(A) formed in (R1). The recorded spectra were fitted with the simulated ones to determine the vibrational distributions of OH(A) for vibrational levels of υ = 0–4. The distributions were significantly nonthermal due to competition between vibrational energy transfer and other involved processes. Then, in order to model the kinetics of (R1), a master equation for the vibrational states of OH(A) is constructed based on the obtained vibrational distributions, as well as available data on predissociation, collisional quenching, spontaneous emission, and vibrational energy transfer of OH(A). The rate constants for (R1) predicted for certain mixtures are found to be fairly consistent with those reported previously based on OH(A) concentration measurements. Meanwhile, substantially smaller rate constants are predicted for highly diluted mixtures. This result is supported by the variation seen in the relative emission rates observed for different mixtures. It is suggested that the rate constant for (R1) depends significantly on third-body molecules due to the dependency of the vibrational energy transfer rate on collision partners.
Machine-learning-assisted multi-objective optimization of thermal conductivity, thermal expansion coefficient, and fracture toughness in rare-earth zirconate ceramics
Although rare-earth zirconates (RE2Zr2O7) are promising top-coat candidates for thermal barrier coatings, their thermal conductivity (TC) still requires further reduction, and their limited thermal expansion coefficient (TEC) and fracture toughness (KIC) restrict further performance improvement. However, trial-and-error experiments and first-principles calculations are not well suited for rapid screening of the large RE2Zr2O7 compositional space, while empirical models usually have limited transferability to complex multi-component systems. To address these challenges, this study develops an interpretable machine learning (ML) framework that integrates multi-property prediction, feature interpretation, and multi-objective composition optimization for RE2Zr2O7 ceramics. Experimental data sets are selected from the literature and handbooks, including 1167, 660, and 309 samples for TC, TEC, and KIC, respectively. Composition-derived features related to atomic mass, configurational entropy, electronegativity, ionic radius, and ionic volume are generated, and a four-step feature selection strategy is applied to obtain compact feature subsets. Multiple ML algorithms are evaluated, and random forest is found to provide the best performance for TC, while extra trees perform best for TEC and KIC. The selected models show strong agreement between predicted and experimental values on both test sets and independent compositions. SHapley Additive exPlanations are further employed to interpret the role of temperature, mass disorder, electronegativity dispersion, and ionic-radius mismatch in governing the predictions. Finally, the trained ML predictors are integrated with non-dominated sorting genetic algorithm II to perform multi-objective optimization in the (Gd–Yb–Sc)2Zr2O7 compositional system, enabling the identification of composition regions with coordinated thermophysical and mechanical performance.