Bartel Comp Materials

The average voltage (V) at which a guest ion (Li+, Mg2+, Ca2+) intercalates into or de-intercalates from a cathode host structure. Computed from DFT total energies of the charged and discharged states. A key performance metric for battery cathode materials — higher voltage means higher energy density.

Mean absolute error (eV/Å) of an ML interatomic potential's predicted atomic forces compared to DFT reference calculations. The primary accuracy benchmark for neural network potentials. CHGNet achieves ~30 meV/Å on MPtrj test set; lower values indicate more accurate molecular dynamics trajectories.

The Gibbs free energy change (ΔG, kJ/mol or eV/atom) of a solid-state synthesis reaction at a given temperature. More negative values indicate stronger thermodynamic favorability. Used to rank candidate synthesis pathways and predict whether a target material can be made via a particular precursor combination.

The accuracy of automated X-ray diffraction pattern matching for identifying crystalline phases in synthesized materials. In autonomous synthesis workflows, this is the classification accuracy of ML models that interpret XRD patterns to determine whether the target phase was successfully produced.

The excess potential (V) required above the thermodynamic minimum to drive the oxygen evolution reaction (OER) on a catalyst surface. Lower overpotential indicates a more active catalyst. Typically measured at 10 mA/cm² current density. A key metric for evaluating electrocatalysts for water splitting.

A topological data analysis method that characterizes the topology of electron density distributions in crystalline materials. Computes persistence diagrams from sublevel sets of the electron density field, capturing features like connected components, loops, and voids at multiple scales. Used to classify crystal structure types and predict material properties.

The classical geometric tolerance factor t = (r_A + r_X) / sqrt(2)(r_B + r_X) proposed by Goldschmidt (1926) for predicting perovskite stability. Values near 1.0 favor cubic perovskite; t < 0.8 or t > 1.0 favor non-perovskite structures. Achieves ~74% accuracy, superseded by Bartel's revised tau.

The ratio of B-site cation radius to X-site anion radius (μ = r_B/r_X) in ABX3 perovskites. Values between 0.44 and 0.90 generally support stable octahedral coordination. Used alongside tolerance factors for perovskite stability prediction.

The DFT-calculated energy difference between a perovskite ABX3 compound and its most stable decomposition products (eV/atom). Negative values mean the perovskite is thermodynamically stable; positive values indicate it will spontaneously decompose. More physically rigorous than geometric tolerance factors.

The electronic band gap (eV) of halide perovskite materials (ABX3 where X = Cl, Br, I). Critical for optoelectronic applications — optimal solar cell performance requires band gaps of 1.1-1.7 eV. Tunable via composition engineering of A, B, and X sites.

The activation energy barrier (eV) for Li-ion migration between adjacent sites in a cathode crystal structure. Computed via nudged elastic band (NEB) calculations. Lower barriers enable faster charge/discharge rates. Typically 0.2-1.0 eV for intercalation cathodes.

A computed score ranking candidate precursor combinations for solid-state synthesis of a target material. Incorporates thermodynamic driving force, selectivity against competing products, and practical considerations (melting point, reactivity, commercial availability). Higher scores indicate more favorable precursor choices.

The temperature (°C or K) at which a solid-state synthesis reaction is conducted. Must be high enough for sufficient kinetics but low enough to avoid unwanted side reactions or decomposition. Typical range for ceramics: 600-1400°C. ML models increasingly predict optimal synthesis temperatures.

Binary assessment of whether a proposed metathesis (double exchange) reaction between two precursors will produce the desired product. Determined by comparing the thermodynamic driving force of the target reaction against all competing reaction pathways. A feasible metathesis reaction has negative ΔG and high selectivity.

The solar-to-fuel energy conversion efficiency (η) of a thermochemical water splitting cycle using metal oxide redox pairs. Depends on the reduction temperature, oxidation temperature, oxygen nonstoichiometry, and heat recovery. Typical values range from 5-40% depending on the oxide system and operating conditions.

The directional control in programmable catalytic loops where dynamic oscillations in catalyst composition or surface state drive net reaction progress in a preferred direction. A key concept in catalytic resonance theory — the frequency and amplitude of oscillations determine whether the catalytic cycle runs forward, backward, or reaches a dynamic steady state.

A similarity metric between two crystalline materials based on the comparison of their persistent homology descriptors (persistence diagrams). Computed using Wasserstein or bottleneck distance between persistence diagrams derived from electron density fields. Enables structure-agnostic comparison of materials across different space groups.

Mean absolute error (meV/atom) of an ML model's predicted formation energy or total energy compared to DFT reference calculations. The primary accuracy benchmark for ML property prediction models. State-of-the-art models achieve ~20-30 meV/atom on Materials Project datasets.

A graph-based encoding of crystal structures where atoms are nodes and bonds are edges, with node/edge features encoding atomic properties and interatomic distances. Used as input to graph neural network models (CGCNN, MEGNet, SchNet, DimeNet) for property prediction. The choice of graph construction (cutoff radius, edge features) significantly affects model performance.

The electronic band gap of inorganic halide double perovskites (A2BB'X6), which determines their suitability for optoelectronic applications. Computed via hybrid DFT (HSE06) or GW-BSE methods.

The energy required to dissociate an electron-hole pair (exciton) in a semiconductor. In double perovskites, large exciton binding energies indicate strong electron-hole coupling, relevant for light-emitting applications.

The activation energy for Mg2+ ion hopping between sites in a host crystal structure, typically measured in meV. Lower barriers enable faster Mg-ion diffusion and better cathode rate capability in rechargeable Mg batteries.

The average electrochemical potential for reversible Ca2+ insertion/extraction in a host cathode material for Ca-ion batteries. Higher voltages enable higher energy density.

The ionic conductivity of solid-state Mg2+ conductors, critical for enabling all-solid-state Mg batteries. Limited by the high migration barriers of divalent Mg2+ ions in most crystalline hosts.

A quantitative measure of how selectively a solid-state reaction produces the desired target phase versus competing byproduct phases, derived from thermodynamic reaction energies across all possible competing reactions in a chemical network.

The degree to which a solid-state synthesis pathway preferentially forms one crystal polymorph over thermodynamically competing polymorphs, governed by the interplay of reaction energy, surface energy, and nucleation barriers.

The process by which solid-state ceramic synthesis proceeds through a sequence of binary reactions at precursor particle interfaces, where the most thermodynamically favorable pairwise reaction occurs first and intermediate phases progressively react to form the target product.

The local cation ordering in disordered rocksalt (DRX) cathode materials that deviates from a perfectly random distribution. SRO affects Li percolation networks, voltage profiles, and cycling stability in DRX cathodes.

The extent of fluorine incorporation into disordered rocksalt cathode materials, which modifies voltage, capacity, and Li percolation. Achieving target fluorination requires careful precursor selection to avoid LiF formation.

The activation energy for Ca2+ migration in solid-state materials, critical for both Ca-ion battery cathodes and solid electrolytes. Ca2+ mobility is generally much lower than Li+ due to its larger size and higher charge density.

ML-based prediction of optimal solid-state synthesis conditions (temperature, time, atmosphere) from precursor properties and reaction features, trained on text-mined literature data.

A two-dimensional descriptor combining cation size and electronegativity differences to predict the thermodynamic stability of NASICON-structured materials (NaxMM'(PO4)3). Enables rapid screening of the large NASICON composition space.

A double-ion exchange synthesis strategy where two anions are simultaneously exchanged between precursors, enabling formation of complex oxides at substantially lower temperatures than conventional ceramic routes by providing large thermodynamic driving forces.

The electrochemical voltage of sulfide-based cathode materials for Li-ion or multivalent batteries. Sulfide cathodes generally operate at lower voltages than oxides but may offer advantages in ionic conductivity and interface compatibility.

A physically motivated descriptor identified via SISSO (sure independence screening and sparsifying operator) that predicts the Gibbs free energy of inorganic crystalline solids as a function of temperature, enabling temperature-dependent phase stability calculations with ~50 meV/atom accuracy.

The enthalpy change associated with a compound decomposing into competing phases (other compounds and/or elemental forms). Unlike formation enthalpy which measures stability relative to elements only, decomposition enthalpy captures the true thermodynamic competition that determines compound stability.

The degree to which machine learning synthesizability scores align with ground-truth thermodynamic metrics (convex hull energies, selectivity scores), measuring whether ML models correctly estimate the likelihood that a hypothetical material can be experimentally synthesized.

The thermodynamic driving force for a reaction between a specific pair of solid precursors at their interface, used to predict which intermediate phases form first during solid-state synthesis of multicomponent ceramics.

The equilibrium yield of ammonia from solar thermochemical synthesis cycles involving metal nitride/oxide redox pairs, determined by Gibbs energy minimization across hydrolysis, reduction, nitrogen fixation, and nitride reformation steps.

The performance gain achieved by dynamically modulating catalyst surface binding energies through external forcing (voltage, strain, temperature oscillation), quantified as the ratio of dynamic to static catalytic rates or the reduction in required overpotential.

The optimal oscillation frequency at which forced dynamic modulation of a programmable catalyst achieves maximum rate enhancement, determined by the match between external forcing period and intrinsic catalytic turnover timescales.

ML prediction of optimal heating temperature for solid-state synthesis, learned from text-mined synthesis recipes. Correlated with precursor stability metrics (melting points, formation energies) following extended Tamman's rule.

Quantitative metrics (primary and secondary competition) assessing the favorability of target phase formation versus impurity phase formation in solid-state reactions. Used to rank and select synthesis reactions that maximize target yield.

The ability to selectively synthesize a metastable crystal polymorph over the thermodynamically stable ground state through careful control of reaction energetics and surface energy contributions in solid-state synthesis.

AI models (diffusion models, variational autoencoders, large language models) that generate novel crystal structures. Benchmarked against baseline methods like random charge-balanced prototype enumeration and data-driven ion exchange of known compounds.

ML-derived or analytically derived tolerance factors (e.g., tau) that predict the stability and formability of perovskite-structured compounds based on ionic radii, oxidation states, and electronegativity.

Quantitative measures (primary and secondary competition) of the thermodynamic favorability of target phase formation vs impurity phase formation in solid-state reactions. Higher selectivity indicates reactions that preferentially produce the desired product.

The thermodynamic energetics of proton (H+) incorporation into perovskite and brownmillerite oxide structures, relevant for protonic ceramic fuel cells and electrochemical applications.

A mechanism for thermochemical water splitting where cation vacancies in spinel metal oxides (rather than conventional oxygen vacancies) mediate the redox cycling, enabled by cation site inversion that lowers vacancy formation energies.

A topological descriptor derived from persistent homology applied to electron density fields of crystalline solids, encoding bonding characteristics by tracking topological features (connected components, loops, voids) across varying density thresholds as a function of filtration parameter.

A structured dataset of synthesis procedures extracted from scientific literature using NLP and text mining, containing precursors, conditions, and outcomes that enable data-driven analysis of synthesis-structure-property relationships.

Descriptors derived from persistent homology (Betti curves) that compress electron density distributions into compact representations capturing bonding characteristics through components, cycles, and voids across electron density thresholds.

Application of machine learning interatomic potentials (MLIPs) to compute surface phase diagrams, surface energies, and surface reconstructions at a fraction of the DFT cost while maintaining near-DFT accuracy.

The thermodynamic driving force available in metathesis (double displacement) reactions for inorganic synthesis, which can dramatically alter the reaction landscape to enable rapid and selective formation of target phases that are otherwise difficult to synthesize via traditional routes.

Betti curve descriptors derived from persistent homology that compress electron densities of crystalline materials into compact representations, capturing bonding characteristics by encoding topological features (components, cycles, voids).

Machine-learned tolerance factor for NASICON-structured materials based on Na content, elemental radii, electronegativities, and Madelung energy. Classifies NASICON phases in terms of their synthetic accessibility.

Machine learning interatomic potentials that incorporate charge/oxidation state information (e.g., via magnetic moment prediction) to describe both atomic and electronic degrees of freedom, enabling modeling of redox-coupled phenomena in electrochemical systems.

Use of machine learning models to correct systematic errors in DFT-computed enthalpies of formation by learning error patterns from electronic structure features. Can reduce PBE errors from MAE ~195 meV/atom to ~80 meV/atom.

Structured synthesis datasets extracted from scientific literature using NLP/text mining methods. These datasets capture synthesis procedures, precursors, conditions (temperature, time, atmosphere), and outcomes for thousands of inorganic materials, enabling data-driven synthesis prediction.