Papers

Predicting the evolving microstructure of hydrating cement is essential for understanding and modeling its mechanical property development. Physics-based continuum approaches offer a rigorous framework for capturing the thermodynamics of dissolution and precipitation processes at the microstructural scale. In this work, we present an adapted Phase-Field (PF) model for cement hydration that resolves key physical inconsistencies in existing PF formulations by introducing a revised free-energy potential and distinct equilibrium constants for clinker dissolution and hydrate precipitation. The resulting PF framework reproduces microstructural evolution, yielding realistic porosity levels and continuous phase boundaries in close agreement with experimental observations. The predicted hydrated microstructures are subsequently used in a computational homogenization scheme to evaluate the elastic response of the material. The PF-derived mechanical properties show good agreement with experimental trends, supporting the ability of the proposed framework to consistently link hydration chemistry, microstructure formation, and the resulting mechanical response.

The pseudogap phase of the underdoped cuprate superconductors harbours diverse manifestations of different ordered electronic-states, and then these ordered electronic-states coexist or compete with superconductivity. Here starting from the microscopic electron propagator, the nature of the ordered electronic-states in the pseudogap phase is investigated within the $T$-matrix approach. This $T$-matrix is derived in terms of the inverse of matrix for various kinds of a single impurity, and then is used to evaluate the local density of states (LDOS) by the involvement of all the quasiparticle excitations and scattering processes. It is shown that a number of the anomalous properties in the underdoped cuprate superconductors is directly correlated to the opening of the normal-state pseudogap: (i) the structure of the microscopic octet scattering model generated by the normal-state pseudogap is essentially the same both in the superconducting (SC)-state and pseudogap phase, which naturally leads to that the quasiparticle scattering interference octet phenomenology observed in the SC-state exists in the pseudogap phase; (ii) however, the spectral weight at around the antinodal region in the SC-state is gapped out completely by both the SC gap and normal-state pseudogap, while it in the pseudogap phase is suppressed partially by the normal-state pseudogap, this directly leads to that the non-dispersive checkerboard charge ordering with a finite wave vector ${\bf Q}$ appears in the pseudogap phase only. The theory therefore also shows that the electronic-states affected by the normal-state pseudogap exhibit the LDOS modulation spectrum organization.

It is long established that overdoped cuprate superconductors experience a loss of superfluid density (SFD) with increasing doping, p, along with the decline in T_c. Such behavior is unconventional and suggests a depletion of the condensate by increasing pairbreaking or the growth of a second non-pairing channel. This led to a recent suggestion that the condensate arises from an incoherent charge channel which progressively gives way with overdoping to a second, coherent non-pairing channel. Contra these ideas, we report analysis of the field-dependent electronic specific heat of several cuprates from which we find no apparent loss of SFD with overdoping. The SFD per CuO_2 plaquette is found to rise progressively with overdoping from p towards (1+p), undiminished and much the same as the Hall number, thus implying that all available carriers contribute to the condensate. We suggest this could be the underlying intrinsic behavior for all cuprates. Our samples include (Y,Ca)Ba_2Cu_3O_{7-\delta}, Bi_2Sr_2CaCu_2O_{8+\delta}, La_{2-x}Sr_xCuO_4 and Tl_2Ba_2CuO_6, with the latter being the only exception. Our results signal a possible return to a more conventional picture.

Nonequilibrium spatiotemporal patterns have been extensively studied. However, a single oscillator or cyclic loop of states is typically employed at each site in theories and simulations. Here, we investigate how competition among multiple identical cyclic loops at each site alters patterns. We simulate active Potts models with standard Potts interactions between neighboring sites in two-dimensional square lattices. When multiple three-state cycles exist in state flips, such as in octahedral and square-antiprism networks, all types of spiral waves comprising the three states are formed simultaneously at high flip energies. However, at lower energies, only one or a few types emerge and switch stochastically into different types. At even lower energies, cyclic changes in single-state dominant homogeneous phases emerge [homogeneous cycling (HC) mode]. At intermediate flip energies, the spiral wave and HC modes temporally coexist in small systems but do not switch between each other in large systems. Conversely, when multiple four-state cycles exist in six-state and cubic networks, one state remains dominant for the entire range of flip energies, whereas the other states occasionally form domains at intermediate flip energies. Therefore, the number of spatially coexisting states can be controlled using flip networks and energies.

We present Josephson junctions with the metallic chiral helimagnet Cr$_{1/3}$NbS$_{2}$ as the weak link that exhibit magnetic diffraction patterns with asymmetry in both applied magnetic field and the critical (switching) current. The nonreciprocity between positive and negative critical currents (the Josephson diode effect) reaches efficiencies of $|\eta|>20\%$, where $\eta \equiv (I_{c+}-|I_{c-}|)/(I_{c+}+|I_{c-}|)$, and the effect persists even at zero applied field. As a chiral helimagnet, Cr$_{1/3}$NbS$_{2}$ lacks inversion symmetry and time reversal symmetry, providing a platform to explore the interplay of superconductivity with both symmetries broken. We propose that pinned Abrikosov vortices are a primary mechanism for the asymmetric magnetic-field response, while the nonzero spin chirality of Cr$_{1/3}$NbS$_{2}$ gives rise to the diode effect. Simulations of magnetic diffraction patterns with vortices show offsets from zero field consistent with observations, while simulations of chiral spin structures with out-of-plane canting show a diode effect.

Recent advances in optics and time-resolved techniques have facilitated the exploration of new states of matter under nonequilibrium conditions. Here, we predict that irradiated graphene can host two novel nonequilibrium steady states of matter with zero resistance when exposed to circularly polarized light: (i) Floquet Chern insulators and (ii) a radiation-induced zero-resistance state with spontaneous formation of an inhomogeneous current distribution. Specifically, we calculate nonequilibrium anomalous Hall and longitudinal conductivities to map the nonequilibrium phase diagram of irradiated graphene as a function of the driving frequency and the electric-field strength of circularly polarized light. As a result, Floquet Chern insulators are found to occur at high driving frequencies above the graphene band width. By contrast, at low driving frequencies below the graphene band width, the nonequilibrium anomalous Hall conductivity deviates from the expected quantized values, and the nonequilibrium longitudinal conductivity exhibits highly irregular behavior, including negative resistance. It is predicted that the thermodynamically unstable negative resistance will trigger a catastrophic breakdown, inducing a zero-resistance state with spontaneous formation of an inhomogeneous current distribution, similar to the radiation-induced zero-resistance state observed in quantum Hall systems.

We propose a modified swarmalator model that generates collective rotational currents in phase synchronization. Our approach builds on the original swarmalator model [4], introducing a key modification: the phase-dependent terms in the spatial dynamics are replaced with a simpler driving term that depends on both the phase and a specified origin. We investigate the dynamics of this model through extensive numerical simulations. When the origin is fixed, spiral motions of synchronized and clustered swarmalators emerge from a finite fraction of random initial conditions, resulting in collective currents. To prevent the unrealistic divergence of these spirals, we introduce a dynamic origin, defined as the center of the swarmalators' positions. With this dynamic origin, the system evolves into rotating collective currents, where synchronized swarmalators form stable circular patterns. In both the fixed and dynamic origin cases, we also observe no-current states, in which synchronized swarmalators aggregate near the origin. Finally, we find that the formation of collective currents can be facilitated by tuning the phase variables either at initialization or during the system's evolution.

The Sllod equations of motion enable modeling of homogeneous flow at the atomic scale, and are commonly used to predict fluid properties such as viscosity. However, few publicly available codes support such simulations, and those that do often do not implement a reversible numerical integration scheme or have other subtle problems. Here, we demonstrate a reversible and energy-conserving integration scheme for the Sllod equations of motion with error on the order of $\delta t^3$, in line with typical operator splitting integrators used in standard molecular dynamics simulations. We discuss various implementation details, and implement the scheme in LAMMPS where we find that our changes enable more accurate simulation of transient responses, mixed flows, and steady states, especially at high rates of flow. Importantly, we show that a lack of energy conservation can manifest as a systematic error in the direct ensemble average of the pressure tensor, leading to an error in the calculated viscosity which becomes significant at high flow rates.

Measurement-induced entanglement (MIE) captures how local measurements generate long-range quantum correlations and drive dynamical phase transitions in many-body systems. Yet estimating MIE experimentally remains challenging: direct evaluation requires extensive post-selection over measurement outcomes, raising the question of whether MIE is accessible with only polynomial resources. We address this challenge by reframing MIE detection as a data-driven learning problem that assumes no prior knowledge of state preparation. Using measurement records alone, we train a neural network in a self-supervised manner to predict the uncertainty metric for MIE–the gap between upper and lower bounds of the average post-measurement bipartite entanglement. Applied to random circuits with one-dimensional all-to-all connectivity, our method reveals a learnability transition with increasing circuit depth: below a threshold the MIE can be effectively learned with resources that grow only polynomially with system size, whereas above it the required resources grow exponentially. This computational phase transition coincides with the breakdown of efficient classical simulation of the underlying quantum state. We further observe signatures of this transition on current noisy quantum devices. These results highlight the power of data-driven approaches for learning MIE and delineate the practical limits of its classical learnability.

As a newly identified single-crystalline van der Waals dielectric with a high dielectric constant, Bi2SeO5 plays a pivotal role in advancing 2D electronic devices. In this work, we systematically investigate the defect properties of Bi2SeO5 using first-principles calculations based on a hybrid functional. Although Bi2SeO5 is a chemically ternary compound, each constituent element occupies several crystallographically nonequivalent sites, rendering its defect chemistry highly complex. Due to the anomalous +4 cationic valence state of Se, the defect formation energies of same main group anion antisite defects (SeO and OSe) are prohibitively high, and their concentrations can therefore be neglected. In contrast, the extraordinary cation-cation antisite defects BiSe and SeBi emerge as the dominant defects. The pronounced variability in the formation energies of the six types of VO defects demonstrates that identical defect types located on nonequivalent atomic sites can exhibit markedly different properties. Under O-rich and Se/Bi-poor conditions, Bi2SeO5 shows relatively robust p-type behavior. Conversely, under O-poor and Se/Bi-rich conditions, or at intermediate O, Se, and Bi partial pressures, Bi2SeO5 behaves as an intrinsic semiconductor or displays very weak n-type conductivity due to strong donor-acceptor compensation. This study provides theoretical insights to guide the design and development of high-performance Bi2SeO5-based electronic devices.

We propose a theoretical framework in which a one-dimensional (1D) tight-binding model incorporating unconventional magnetic order together with Rashba and Ising spin-orbit couplings are considered to realize two key phenomena in condensed matter systems: topological superconductivity and the superconducting diode effect (SDE). We first elucidate the underlying band topology of the normal-state Hamiltonian and subsequently introduce an on-site attractive Hubbard interaction. Performing a a self-consistent mean-field analysis, we establish superconducting order parameters in both the conventional Bardeen-Cooper-Schrieffer (BCS) and finite-momentum Fulde-Ferrell-Larkin-Ovchinnikov (FFLO) pairing channels. Intriguingly, both pairing states can support topological superconductivity, characterized by a nontrivial winding number, and lead to the emergence of four zero-energy Majorana modes localized at the ends of the 1D chain. The FFLO state further gives rise to an intrinsic field-free SDE, manifested as a nonreciprocal supercurrent and quantified by the diode efficiency $\eta$. Notably, our model yields a large diode efficiency $\eta \sim 65\%$, highlighting its potential for realising topological superconductivity and highly efficient superconducting devices.

We study the ferromagnetic instability in an SU($N$) Fermi-Hubbard model on the hypercubic lattice. Combining dynamical mean-field theory with continuous-time quantum Monte Carlo simulations, we find that, in the strong-coupling regime at low temperatures, ferromagnetically ordered (FM) states develop away from the commensurate fillings. In the particle-doped SU($3$) system near one-third filling, the FM state is accompanied by a spontaneous flavor-selective Mott state, where two of the three flavors are Mott insulating while the remaining flavor is metallic. Since particles in the metallic flavor can almost freely move on the lattice without correlation effects, the ordered state is stabilized by the kinetic-energy gain of the doped particles. This is similar to the generalized Nagaoka ferromagnetism discussed in the one-hole-doped system at one-third filling. In the SU($4$) case, we find that six distinct types of FM states appear as the particle density varies. The results uncover the nature of the FM state in the SU($N$) Fermi-Hubbard systems and highlight the rich magnetic behavior enabled by enlarged internal symmetries.

We study the van der Waals-London force, which is typically associated with fluctuating electric dipoles in atoms, in a mesoscopic circuit consisting of two inductively coupled superconducting loops. We investigate the inductive van der Waals-London interaction using both semiclassical and quantum electrodynamic (QED) approaches. The semiclassical model predicts a repulsive interaction due to anticorrelated current fluctuations. In contrast, the QED framework, which incorporates virtual photon exchange, reveals a predominantly attractive force. A key contribution comes from a state-independent two-photon exchange, which is absent in the semiclassical description and undetectable by spectroscopy. Our study introduces a theoretical framework for exploring the van der Waals force between individual artificial atoms via controlled mesoscopic circuits.

Intense terahertz (THz) pulses induce transient inversion-symmetry breaking in quantum paraelectric SrTiO$_3$, yet the underlying mechanism remains controversial. Using fields up to $\sim$1.1 MV/cm, we reveal spatially inhomogeneous THz-field-induced second harmonic generation (TFISH) governed by competing lattice and defect dynamics. Short-lived coherent antiferrodistortive (AFD) modes suppress dipole correlations within $\sim$5 ps, while heavily damped soft/AFD modes and a defect-induced low-frequency mode ($\sim$0.1-0.3 THz) jointly prevent long-range ferroelectric coherence in oxygen-vacancy-rich regions. Collective modes manifested by oscillatory TFISH components exhibit softening followed by hardening below a critical temperature $T^*\simeq$28 K, confirming transient ferroelectric order where defects are sparse. These results reconcile conflicting interpretations, establish defect-mediated competition as a central regulator of light-induced ferroelectricity, and open routes to ultrafast control of quantum materials.

We propose an experimentally accessible scheme for realizing tunable nonclassical light in cavity-coupled reconfigurable atomic arrays. By coherently controlling the collective interference phase, the system switches from single-photon blockade to high-purity multiphoton bundle emission, unveiling a hierarchical structure of photon correlations dictated by atom-number parity and cavity detuning. The scaling of photon population identifies the transition between superradiant and subradiant regimes, while parity- and phase-dependent spin correlations elucidate the microscopic interference processes enabling coherent multiphoton generation. This work establishes a unified framework connecting cooperative atomic interactions to controllable nonclassical photon statistics and introduces a distinct interference-enabled mechanism that provides a practical route toward high-fidelity multiphoton sources in scalable cavity QEDs.

This article discusses a failure of the widely used method of images to describe the time evolution of probability distributions in diffusive processes with memory. A walker in one-dimensional space draws a dead or a surviving path depending on whether it has touched a target during stochastic evolution. For the dead walker, we define its conjugate twin paired by paths spatially reflected at the first passage time. The probability distribution of the reflected dead path coincides with that of the free image walker in a physical domain, but not generally with that of the original dead walker for non-Markovian processes, which violates the method of images. For systems reducible to the generalized Langevin equation with the fluctuation-dissipation relation, we propose an energetic interpretation in terms of a path-memory force, where the path-probability ratio of the dead to the reflected path obeys an analogous relation to the fluctuation theorem associated with heat from the reflected to the original dead walker. This framework provides a quantitative basis as well as an intuitive picture of how and why the method of images breaks down for non-Markovian processes.

Altermagnets, characterized by spin-split bands without net magnetization, have recently emerged as a promising platform for spintronics. However, their microscopic mechanisms remain elusive, often relying on abstract group theory. In this work, we present an intuitive and pedagogical framework to understand the origin of spin splitting in altermagnets. We identify two essential ingredients: (1) alternating spin-polarized wavefunction localization on sublattices, and (2) broken translational symmetry caused by distortions in non-magnetic ion cages. We discuss a minimal model Hamiltonian based on an atomic exchange-driven spin splitting and anisotropic hopping that captures these effects and reproduces the hallmark features of altermagnetic band structures, including nodal spin degeneracies and large spin splittings. Our model is further validated by ab initio calculations on MnF2. By demystifying the microscopic origins of altermagnetism, our work bridges symmetry analysis and material realizations, shedding light on practical designs of altermagnetic spintronic devices.

Collective modes in superconductors, such as the Higgs mode, offer deep insights into the nature of condensates. Third-harmonic generation (THG) is a primary tool for probing the Higgs mode, but its signal competes with that of quasiparticle excitations depending on impurity scattering rates. In particular, in the clean regime the standard BCS theory generally predicts the dominance of quasiparticle contributions. Here, we propose and demonstrate that the quantum geometry of electronic bands can be a key mechanism governing this competition. By developing a formalism that explicitly incorporates the quantum metric, and applying it to a tunable model of a dispersive-band superconductor, we show that the quantum metric can dramatically amplify the nonlinear light-Higgs coupling by several orders of magnitude. Our results establish that a large quantum metric can cause the Higgs mode to dominate the THG response, resolving the puzzle of Higgs and quasiparticle competition in the clean regime and identifying band geometry as a crucial ingredient for designing and understanding the nonlinear response of superconductors.

A Mn dopant in a III-V semiconductor produces a highly-entangled, coherent triplet ground state not fully captured by single-determinant theories of electron structure. We directly construct an analytic form for its ground-state wavefunction, finding surprising spin-charge correlations not revealed by semiclassical calculations. Spin-correlated circulating currents associated with the dopant yield remarkably large magnetic fringe fields of $\sim$1$\,\mu$T at distances of $\sim 10$~nm from Mn in GaAs, potentially detectable by NV-diamond magnetometry while the dopant spin coherently precesses.

Geometric phases can manifest when a relativistic quantum particle moves cyclically along a loop in parameter space. The phase can be affected by the presence of a background field and can be accompanied by nontrivial topological features. The appearance of adiabatic geometric phases in momentum space is demonstrated for a Lorentz-violating Weyl fermion, where the role of the background is played by the coefficients for Lorentz violation. As explicit examples, the Berry curvature and the first Chern number are derived for two cases with large Lorentz violation, one incorporating CPT violation and one preserving CPT symmetry. Some alternative topological invariants are also obtained. In certain scenarios with large Lorentz violation, the physical vacuum is associated with a topological phase.

High temperature superconductors, especially YBa$_2$Cu$_3$O$_{7-\delta}$ (YBCO), are considered a key enabling technology towards a clean energy future. Hole doping in YBCO is a prerequisite for the emergence of its unchallenged superconducting properties. Up to now, research was focused on the under- and optimally doped region, due to practical limitations in reaching the overdoped state, despite being highly interesting from fundamental and applied aspects as competing orders vanish and critical current densities are expected to peak. Here, we deploy for the first time an electrochemical method to access the mostly uncharted overdoped region. We demonstrate precise control over the bulk oxygen concentration in YBCO thin films across the full off-stoichiometry window (0$\le{\delta}\le$1) using electrochemical oxidation combined with in situ XRD and electrical measurements. Resulting high doping states and critical current densities are confirmed using a multi modal approach, including x-ray diffraction, electrical, Hall and magnetic characterization. Thus, this work opens a promising pathway based on electrochemical oxidation towards electronically clean, oxygen overdoped cuprate superconductors and therefore will assist to further push the critical current density to its intrinsic limit.

We present a quantitative theory for simulating optically detected magnetic resonance (ODMR) measurements of optically-active spin centers using steady-state Lindblad equations. We apply the theory to an experimental ODMR spectrum associated with the negatively-charged silicon vacancy V2 center in 6H-SiC, showing that spin Hamiltonian parameters, optical transition rates, and even coherence times may be extracted, with values consistent with recent literature. Notably the $T_2$ spin coherence time is measurable, not just the $T_2^*$ dephasing time. Furthermore, we simulate the ODMR spectra of a V2 center in isotopically-purified 6H-SiC, and predict an order-of-magnitude narrowing of some, but not all spectral lines compared with natural abundance samples.

We develop a framework for the classification of invertible translation-invariant stabilizer codes modulo condensation and stabilization with simple codes. Our stabilizer codes generalize the Pauli codes: they are extracted from local commuting projector Hamiltonians whose terms are made of local alphabets more general than the Pauli operators. In particular, not all such codes can be entangled/disentangled with Clifford Quantum Cellular Automata. We compute their Witt-equivalence classes in any spatial dimension in terms of relative L-theory groups. In particular, we show that the group of equivalence classes of such codes in three spatial dimensions is isomorphic to the Witt group of abelian topological orders in two spatial dimensions. In four spatial dimensions we obtain a group of order two, conjecturally corresponding to the non-trivial invertible phase. Additionally, we propose a representing spectrum for the generalized cohomology theory corresponding to the invertible stabilizer codes.

The frequency structure of the superconducting correlations in cuprates gives insights on the pairing mechanism. Here we present an exhaustive study of this problem in the two-dimensional Hubbard model with cellular dynamical mean-field theory. To this end, we systematically quantify the dependence on doping $\delta$ and interaction strength $U$ of the superconducting gap, of the frequency scales where $d$-wave pairing occurs, and of their relative contribution to pairing. For all values of $U$ and $\delta$, we find pair-forming processes confined to frequencies set by the superexchange interaction and followed by pair-breaking processes, ruling out both pair-forming and pair-breaking processes on the scale of $U$. This suggests that at high frequencies, the effect of $U$ is eliminated by the $d$-wave paring, and that at small frequencies, $U$ generates the superexchange interaction that leads to low-frequency pair-forming processes providing the net contribution to pairing.

There is increasing evidence that numerous membrane proteins can assemble into aggregates that modulate their function and affect many cellular processes such as signal transduction and endocytosis. Here, we present a theoretical description of the instantaneous translational diffusion coefficients of transmembrane protein aggregates on free and supported lipid membranes using Kirkwood-Riseman theory. We find that hydrodynamic interactions within protein aggregates must be accounted for, as neglecting them yields several times lower diffusion coefficients. By deriving hydrodynamic radii for free and supported lipid membranes, we identify effective length scales that accurately characterize aggregate diffusivities in the presence of hydrodynamic interactions. These findings motivate the approximation of an aggregate by its outline and a random particle distribution inside it. We show that this approach provides a practical method to accurately determine aggregate diffusion coefficients when the particle locations cannot be resolved. The results presented in this article have immediate implications for the formation and function of membrane protein aggregates.

The role of additives such as FEC in extending the calendar life of silicon anodes beyond the cycling benefits is still not fully understood. Herein, the calendar life of high-loading Si (80 wt%) using baseline 1.2 M LiPF6 in EC-EMC electrolyte versus adding 10 wt% FEC is investigated over months. Over 8 days of aging, FEC leads to a 13-fold reduction in irreversible capacity loss in Si-LiFePO4 full cells. Cells without FEC are projected to fall below 80% of their initial capacity within approx. 22 days versus approx. 279 days with FEC. Symmetric Si-Si cells from harvested electrodes show greater increase in interphase resistance without FEC, whereby an increase of 10.81 Ohms is measured for 0 wt% FEC vs. only 3.37 Ohms for 10 wt% FEC over 2 months. Power law modeling of this long-term interphase resistance finds mixed transport-reaction growth behavior in FEC-free cells, suggesting significant dissolution, whereas cells with 10 wt% FEC added display a diffusion-controlled impedance growth behavior, suggesting a robust surface passivation film. Post-mortem FTIR and XPS confirm polycarbonate enrichment of the SEI, which was discovered to predominantly emerge from FEC self-polymerization during the idle aging. When the Si electrodes aged with and without FEC are harvested and reassembled into full cells with the same electrolytes used at aging, the first-cycle coulombic efficiency is 71% for 0 wt% FEC versus 97% for 10 wt% FEC. Subsequent cycling maintains over 99.7% CE with 10 wt% FEC, surpassing the pre-aging CE of 98.8%. This elevated CE indicates better passivation provided by the polymer fragments formed during aging compared to electrochemically formed SEI where no strong polymer FTIR signal is found. The self-polymerization during idle aging with additives such as FEC is therefore an opportune in situ mechanism to further engineer in extending the life of Si-based batteries.

Accelerated material discovery increasingly relies on artificial intelligence and machine learning, collectively termed "AI/ML". A key challenge in using AI is ensuring that human scientists trust the models are valid and reliable. Accordingly, we define a trustworthy AI framework GIFTERS for materials science and discovery to evaluate whether reported machine learning methods are generalizable, interpretable, fair, transparent, explainable, robust, and stable. Through a critical literature review, we highlight that these are the trustworthiness principles most valued by the materials discovery community. However, we also find that comprehensive approaches to trustworthiness are rarely reported; this is quantified by a median GIFTERS score of 5/7. We observe that Bayesian studies frequently omit fair data practices, while non-Bayesian studies most frequently omit interpretability. Finally, we identify approaches for improving trustworthiness methods in artificial intelligence and machine learning for materials science by considering work accomplished in other scientific disciplines such as healthcare, climate science, and natural language processing with an emphasis on methods that may transfer to materials discovery experiments. By combining these observations, we highlight the necessity of human-in-the-loop, and integrated approaches to bridge the gap between trustworthiness and uncertainty quantification for future directions of materials science research. This ensures that AI/ML methods not only accelerate discovery, but also meet ethical and scientific norms established by the materials discovery community. This work provides a road map for developing trustworthy artificial intelligence systems that will accurately and confidently enable material discovery.

First-principles DFT calculations on the hydrides Ca2NiH6, Sr2NiH6, and Ba2NiH6 reveal key thermodynamic properties. These compounds exhibit increasing entropy and heat capacity with temperature, and are thermodynamically stable at elevated temperatures due to negative free energies. The kinetics of hydrogen storage is influenced by entropy changes during hydrogen adsorption and desorption. Optically, Ba2NiH6 shows a high refractive index at low energies. Mechanical assessments indicate Sr2NiH6 is incompressible with moderate malleability, Ca2NiH6 has the highest resistance to deformation, while Ba2NiH6 is most compressible. Formation energies and hydrogen storage capacities (4.005 wt% for Ca2NiH6, 2.548 wt% for Sr2NiH6, and 1.750 wt% for Ba2NiH6) highlight Ca2NiH6 as the most promising candidate for hydrogen storage technology.

Machine learning interatomic potentials (MLIPs) are commonly trained and evaluated by their ability to reproduce first-principles energies and forces. Practical value, however, is often determined by downstream processes such as structural relaxation. Much of MLIP literature is motivated by the expectation that improving force accuracy will produce correspondingly better downstream behavior. Although this expectation is natural, it may not hold in practice due to model inaccuracies. Here, we demonstrate that force accuracy and relaxation accuracy can be driven in opposite directions. Starting from pretrained MLIPs, we differentiate through relaxation trajectories and optimize the the final structures. This intervention produces increases in force-prediction errors while substantially reducing structural errors. Importantly, this represents a trade-off between two aspects of physical fidelity: the learned forces become less faithful to first-principles forces, while the resulting relaxed structures become more faithful to first-principles predictions. Rather than observing a discrepancy between local quality and downstream performance, we demonstrate that this discrepancy can be induced through optimization. Our findings suggest that MLIPs need not be optimized as approximations of first-principles forces, but can instead learn task-specific vector fields that guide relaxation trajectories toward more accurate final structures. By propagating supervision through complete relaxation trajectories, our framework captures dependencies between successive predictions that are absent when structure–force pairs are treated independently. More broadly, these results demonstrate that conventional force metrics can provide an incomplete assessment of downstream performance and highlights the importance of evaluating scientific machine-learning models within computational workflows.

Living organisms can demonstrate highly adaptable and sophisticated responses using memory resulting from repeated exposure to external conditions or training. However, realizing similar adaptability in mechanical responses in inanimate, physical materials presents an outstanding challenge in several fields, including soft matter, materials science, and in the domain of soft robotics, to name a few. Our study focuses on disordered solids, which are model systems that resemble granular matter, foam and other disordered, soft solids. Here, combining bulk rheology, in-situ optical imaging, and numerical simulations, we demonstrate how training via cyclic shear can encode memories that tune the yield point in a unique way and over unprecedented ranges. Our study reveals that such tunability is intricately linked to the plasticity, non-affine deformations, and formation of shear bands. Remarkably, our numerical simulations illustrate that systems with identical internal energies, prepared via different protocols (mechanical or thermal), can display markedly different rheological responses, indicating that energy alone does not determine mechanical behavior. Moreover, while the yield strain increases with training amplitude, the material simultaneously softens, contrasting with the thermal case where both quantities increase monotonically with increasing annealing. Our results open up possibilities for memory-induced tuning of mechanical response in trainable amorphous matter, independently or in combination with thermal annealing, far beyond the material–feature space achievable via the latter alone.