arxiv: 2602.12109 · v3 · submitted 2026-02-12 · ❄️ cond-mat.mtrl-sci · physics.chem-ph· physics.comp-ph

Recognition: 2 theorem links

· Lean Theorem

A critical assessment of bonding descriptors for predicting materials properties

Aakash Ashok Naik , Nidal Dhamrait , Katharina Ueltzen , Christina Ertural , Philipp Benner , Gian-Marco Rignanese , Janine George

Authors on Pith no claims yet

Pith reviewed 2026-05-16 02:45 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci physics.chem-phphysics.comp-ph

keywords bonding descriptorsmachine learningmaterials propertiesquantum chemical bondingsymbolic regressionelastic propertieslattice thermal conductivityhigh-throughput database

0 comments

The pith

Quantum-chemical bonding descriptors improve machine learning models for predicting elastic, vibrational, and thermodynamic properties in solids.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper extends an existing quantum-chemical bonding database to roughly 13,000 materials and extracts a new set of bonding descriptors from it. These descriptors are added to standard composition and structure features in machine-learning models trained on elastic, vibrational, and thermodynamic properties. Statistical significance tests show measurable gains in predictive accuracy when the bonding descriptors are included. The same descriptors also support symbolic regression that recovers simple, interpretable expressions for quantities such as projected force constants and lattice thermal conductivity. The work therefore tests whether bonding concepts that have long been useful in chemistry can be turned into scalable numerical features for materials informatics.

Core claim

Incorporating quantum-chemical bonding descriptors derived from a systematically generated database of approximately 13,000 materials improves the performance of machine-learning models that otherwise rely only on composition and structure features, and these descriptors further enable symbolic regression to identify intuitive expressions for properties such as the projected force constant and lattice thermal conductivity.

What carries the argument

Quantum-chemical bonding descriptors extracted from the extended Quantum-Chemical Bonding Database for Solid-State Materials, which encode bonding information from established theoretical frameworks and are added as additional input features to existing machine-learning pipelines.

If this is right

Machine-learning models for bond-related properties achieve higher accuracy once bonding descriptors are included alongside composition and structure features.
Symbolic regression applied to the descriptors produces compact, human-readable formulas for the projected force constant and lattice thermal conductivity.
The performance benefit appears across elastic, vibrational, and thermodynamic target properties that are conventionally linked to chemical bonding.
High-throughput bonding analysis can be integrated into routine materials-informatics workflows without requiring new theoretical machinery.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the observed gains hold on larger or more chemically diverse datasets, bonding descriptors could become a default feature set in materials property prediction.
The ability to recover simple expressions via symbolic regression suggests that the descriptors capture physically meaningful combinations that might guide new analytical models.
Extending the same workflow to additional property classes, such as electronic or magnetic behavior, would test how broadly bonding information transfers.

Load-bearing premise

The statistical tests used to confirm performance gains are free from data leakage or post-selection bias, and the set of 13,000 materials is representative of the materials space for the properties being modeled.

What would settle it

Retraining the models on an independent collection of materials withheld from the original database and checking whether the reported accuracy improvements remain statistically significant.

Figures

Figures reproduced from arXiv: 2602.12109 by Aakash Ashok Naik, Christina Ertural, Gian-Marco Rignanese, Janine George, Katharina Ueltzen, Nidal Dhamrait, Philipp Benner.

**Figure 1.** Figure 1: Simplified overview of the methods employed for the bonding descriptor evaluation. The color (symbol) of the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Figure 2: Descriptor ranking based on ARFS scores for the maximum of bond-projected force constant (max [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Figure 3: Distance correlation between descriptor sets and targets for (a) Maximum of bond-projected force constant, [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 4.** Figure 4: Descriptor ranking as SHAP scores for the maximum of bond-projected force (max [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

**Figure 5.** Figure 5: Corrected resampling t-test based on per-fold mean absolute errors from 10-fold cross-validation for model com [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗

**Figure 6.** Figure 6: Parity plots showing correlation between (a) Last peak of phonon density of states ,“last [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗

**Figure 7.** Figure 7: Schematic depicting the model training pipeline. Both models, RF and MODNet, are trained and evaluated [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗

read the original abstract

Most machine learning models for materials science rely on descriptors based on materials compositions and structures, even though the chemical bond has been proven to be a valuable concept for predicting materials properties. Over the years, various theoretical frameworks have been developed to characterize bonding in solid-state materials. However, integrating bonding information from these frameworks into machine learning pipelines at scale has been limited by the lack of a systematically generated and validated database. Recent advances in high-throughput bonding analysis workflows have addressed this issue, and our previously computed Quantum-Chemical Bonding Database for Solid-State Materials was extended to include approximately 13,000 materials. This database is then used to derive a new set of quantum-chemical bonding descriptors. A systematic assessment is performed using statistical significance tests to evaluate how the inclusion of these descriptors influences the performance of machine-learning models that otherwise rely solely on structure- and composition-derived features. Models are built to predict elastic, vibrational, and thermodynamic properties typically associated with chemical bonding in materials. The results demonstrate that incorporating quantum-chemical bonding descriptors not only improves predictive performance but also helps identify intuitive expressions for properties such as the projected force constant and lattice thermal conductivity via symbolic regression.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Bonding descriptors give a measurable lift to ML models for materials properties, with the extended database and symbolic regression as the main additions.

read the letter

The key takeaway here is that adding quantum-chemical bonding descriptors to standard composition and structure features measurably improves predictions for elastic, vibrational, and thermodynamic properties, and the symbolic regression part turns some of those gains into simple, intuitive formulas. What stands out is the scale: they extended their earlier bonding database to roughly 13,000 materials and ran a systematic check with statistical significance tests. That lets them show the bonding features help beyond the baselines. The symbolic regression for things like projected force constants and lattice thermal conductivity is a nice touch because it gives expressions that chemists might actually recognize. The work builds on existing bonding analysis frameworks, so the core idea isn't revolutionary, but the database extension and the controlled comparison are solid incremental steps. Using proper stats tests instead of just eyeballing error bars is a plus. On the downside, the write-up leaves some details thin on how the train-test splits were handled and exactly how the descriptors were derived from the bonding data. Those are the usual places where small leaks or selection effects can creep in, so the full paper needs a close look there. The database representativeness for the property classes also matters for how far the conclusions travel. This paper is aimed at the materials informatics crowd, especially groups working on property prediction pipelines that could use more chemically grounded inputs. Someone already running ML on crystal structures would find the comparison useful to decide whether to add bonding features. I would put it through peer review. The empirical claims are testable and the setup is clear enough that referees can verify the gains without too much trouble.

Referee Report

2 major / 1 minor

Summary. The paper extends the Quantum-Chemical Bonding Database for Solid-State Materials to approximately 13,000 materials and derives a new set of quantum-chemical bonding descriptors from it. It conducts a systematic assessment using statistical significance tests to evaluate how these descriptors improve the performance of machine-learning models for predicting elastic, vibrational, and thermodynamic properties compared to models relying only on structure- and composition-derived features. Symbolic regression is also used to identify intuitive expressions for properties such as the projected force constant and lattice thermal conductivity.

Significance. If the reported performance gains hold after addressing reproducibility details, this work would provide empirical evidence that quantum-chemical bonding descriptors add value beyond standard features in materials ML, enabling more accurate predictions and interpretable models for bonding-related properties. The scale of the database and use of statistical tests strengthen the case for integrating bonding frameworks into high-throughput pipelines.

major comments (2)

[Methods] The manuscript lacks explicit details on the train-test splitting strategy, cross-validation method, and any measures to prevent data leakage in the ML evaluations (Methods section), which directly impacts the reliability of the statistical significance tests for performance improvements.
[Results] Descriptor derivation steps from the extended database and the precise architectures/hyperparameters of the machine learning models are insufficiently described (Results and Methods sections), hindering assessment of whether the gains are robust or due to specific implementation choices.

minor comments (1)

[Abstract] The abstract would benefit from specifying the exact number of properties evaluated and the types of baseline models used for comparison.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive assessment and the constructive comments that will improve the reproducibility of our work. We address each major comment below and will incorporate the requested details in the revised manuscript.

read point-by-point responses

Referee: [Methods] The manuscript lacks explicit details on the train-test splitting strategy, cross-validation method, and any measures to prevent data leakage in the ML evaluations (Methods section), which directly impacts the reliability of the statistical significance tests for performance improvements.

Authors: We agree that these details are critical for validating the statistical tests. In the revised manuscript, the Methods section will be expanded to explicitly state the train-test splitting strategy (random 80/20 split using unique material identifiers), the cross-validation approach (5-fold CV), and the data-leakage prevention measures (ensuring no compositional or structural overlap between sets via material IDs). These additions will directly support the reported performance gains and significance tests. revision: yes
Referee: [Results] Descriptor derivation steps from the extended database and the precise architectures/hyperparameters of the machine learning models are insufficiently described (Results and Methods sections), hindering assessment of whether the gains are robust or due to specific implementation choices.

Authors: We acknowledge this point. The revised manuscript will include a new subsection in Methods detailing the exact descriptor derivation workflow from the ~13,000-material database (including bonding analysis steps and feature extraction) and will specify all model architectures and hyperparameters (e.g., model type, number of estimators, learning rates, and any grid-search or default settings used). This will allow readers to assess robustness independently. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper extends a prior bonding database, derives descriptors from it, and evaluates their added value in ML models via statistical significance tests on external property targets (elastic, vibrational, thermodynamic). No equations or claims reduce reported improvements or symbolic-regression expressions to quantities defined by the paper's own fitted parameters or self-citations. The central results remain independent empirical findings.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The assessment rests on the assumption that existing quantum-chemical bonding workflows produce transferable descriptors across the expanded set of materials and that standard statistical tests can isolate the contribution of those descriptors without confounding effects from model choice or data partitioning.

axioms (1)

domain assumption Statistical significance tests reliably detect genuine improvements when bonding descriptors are added to ML models
Invoked to evaluate whether descriptor inclusion changes predictive performance.

pith-pipeline@v0.9.0 · 5532 in / 1218 out tokens · 73187 ms · 2026-05-16T02:45:59.523397+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean Jcost_pos_of_ne_one; washburn_uniqueness_aczel echoes

?

echoes
ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.

max pfc = 1.77−4.25 (bwdfat dist0 / distat neg bwdf0); log klat 300 =−0.41−2.9 (pairbwdf skew mean / DensityFeaturesvpa); strongest-bond ICOHP correlates with max pfc (Pearson r=−0.91)
IndisputableMonolith/Foundation/BranchSelection.lean branch_selection; RCLCombiner_isCoupling_iff echoes

?

echoes
ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.

bonding heterogeneity (skew, kurtosis of BWDF) and effective interaction number (EIN ICOHP) reduce lattice thermal conductivity; significant MAE gains for max pfc, klat, moduli

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Bond strengths in solids computed from a Wannier-type construction of local vibrational modes
cond-mat.mtrl-sci 2026-05 unverdicted novelty 7.0

A new Wannier-type construction of local vibrational modes in periodic systems yields interpretable bond strengths and shows phonon dispersion contributions to those strengths.

Reference graph

Works this paper leans on

115 extracted references · 115 canonical work pages · cited by 1 Pith paper · 1 internal anchor

[1]

Data-driven materials science: status, challenges, and perspectives

Lauri Himanen et al. “Data-driven materials science: status, challenges, and perspectives”. In:Advanced Science 6.21 (2019), p. 1900808.doi:10.1002/advs.201900808

work page doi:10.1002/advs.201900808 2019
[2]

Data-Driven Materials Science: Status, Challenges, and Perspectives

Lauri Himanen et al. “Data-Driven Materials Science: Status, Challenges, and Perspectives”. In:Advanced Science 7.2 (2020), p. 1903667.doi:10.1002/advs.201903667

work page doi:10.1002/advs.201903667 2020
[3]

Machine-enabled inverse design of inorganic solid materials: promises and challenges

Juhwan Noh et al. “Machine-enabled inverse design of inorganic solid materials: promises and challenges”. In: Chemical Science11.19 (2020), pp. 4871–4881.doi:10.1039/D0SC00594K

work page doi:10.1039/d0sc00594k 2020
[4]

Representation of compounds for machine-learning prediction of physical properties

Atsuto Seko et al. “Representation of compounds for machine-learning prediction of physical properties”. In:Phys. Rev. B95 (14 Apr. 2017), p. 144110.doi:10.1103/PhysRevB.95.144110

work page doi:10.1103/physrevb.95.144110 2017
[5]

Machine learning in materials informatics: recent applications and prospects

Rampi Ramprasad et al. “Machine learning in materials informatics: recent applications and prospects”. In:npj Computational Materials3.1 (2017), p. 54.doi:https://doi.org/10.1038/s41524-017-0056-5

work page doi:10.1038/s41524-017-0056-5 2017
[6]

Matminer: An open source toolkit for materials data mining

Logan Ward et al. “Matminer: An open source toolkit for materials data mining”. In:Computational Materials Science152 (2018), pp. 60–69.doi:10.1016/j.commatsci.2018.05.018

work page doi:10.1016/j.commatsci.2018.05.018 2018
[7]

Electronic structure based descriptor for characterizing local atomic environments

Jan Jenke et al. “Electronic structure based descriptor for characterizing local atomic environments”. In:Physical Review B98.14 (2018), p. 144102.doi:10.1103/PhysRevB.98.144102

work page doi:10.1103/physrevb.98.144102 2018
[8]

Naturally-meaningful and efficient descriptors: machine learning of material properties based on robust one-shot ab initio descriptors

Sherif Abdulkader Tawfik and Salvy P Russo. “Naturally-meaningful and efficient descriptors: machine learning of material properties based on robust one-shot ab initio descriptors”. In:Journal of Cheminformatics14.1 (2022), p. 78.doi:10.1186/s13321-022-00658-9

work page doi:10.1186/s13321-022-00658-9 2022
[9]

Machine learning for catalysis informatics: recent applications and prospects

Takashi Toyao et al. “Machine learning for catalysis informatics: recent applications and prospects”. In:Acs Cataly- sis10.3 (2019), pp. 2260–2297.doi:10.1021/acscatal.9b04186

work page doi:10.1021/acscatal.9b04186 2019
[10]

Machine learning identified materials descriptors for ferroelectricity

Jingjin He et al. “Machine learning identified materials descriptors for ferroelectricity”. In:Acta Materialia209 (2021), p. 116815.doi:10.1016/j.actamat.2021.116815

work page doi:10.1016/j.actamat.2021.116815 2021
[11]

Finding unprecedentedly low-thermal-conductivity half-Heusler semiconductors via high- throughput materials modeling

Jes´ us Carrete et al. “Finding unprecedentedly low-thermal-conductivity half-Heusler semiconductors via high- throughput materials modeling”. In:Physical Review X4.1 (2014), p. 011019.doi:10.1103/PhysRevX.4.011019

work page doi:10.1103/physrevx.4.011019 2014
[12]

Machine learning-driven materials discovery: Unlocking next- generation functional materials – A review

Dilshod Nematov and Mirabbos Hojamberdiev. “Machine learning-driven materials discovery: Unlocking next- generation functional materials – A review”. In:Computational Condensed Matter45 (2025), e01139.issn: 2352- 2143.doi:10.1016/j.cocom.2025.e01139

work page doi:10.1016/j.cocom.2025.e01139 2025
[13]

Stereochemistry of electrocyclic reactions

Robert B Woodward and Roald Hoffmann. “Stereochemistry of electrocyclic reactions”. In:Journal of the American Chemical Society87.2 (1965), pp. 395–397.doi:10.1021/ja01080a054

work page doi:10.1021/ja01080a054 1965
[14]

Goodenough.Magnetism and the Chemical Bond

J.B. Goodenough.Magnetism and the Chemical Bond. Inorganic Chemistry Section / Interscience monographs on chemistry. Interscience Publishers, 1963

work page 1963
[15]

Die gesetze der krystallochemie

Victor Moritz Goldschmidt. “Die gesetze der krystallochemie”. In:Naturwissenschaften14.21 (1926), pp. 477–485. doi:10.1007/BF01507527

work page doi:10.1007/bf01507527 1926
[16]

The principles determining the structure of complex ionic crystals

Linus Pauling. “The principles determining the structure of complex ionic crystals”. In:Journal of the american chemical society51.4 (1929), pp. 1010–1026.doi:10.1021/ja01379a006

work page doi:10.1021/ja01379a006 1929
[17]

Electronic population analysis on LCAO–MO molecular wave functions. I

Robert S Mulliken. “Electronic population analysis on LCAO–MO molecular wave functions. I”. In:The Journal of chemical physics23.10 (1955), pp. 1833–1840.doi:10.1063/1.1740588

work page doi:10.1063/1.1740588 1955
[18]

Generalization of natural bond orbital analysis to periodic systems: applications to solids and surfaces via plane-wave density functional theory

Benjamin D Dunnington and JR Schmidt. “Generalization of natural bond orbital analysis to periodic systems: applications to solids and surfaces via plane-wave density functional theory”. In:Journal of Chemical Theory and Computation8.6 (2012), pp. 1902–1911.doi:10.1021/ct300002t

work page doi:10.1021/ct300002t 2012
[19]

Solid state adaptive natural density partitioning: a tool for deciphering multi-center bond- ing in periodic systems

Timur R Galeev et al. “Solid state adaptive natural density partitioning: a tool for deciphering multi-center bond- ing in periodic systems”. In:Physical Chemistry Chemical Physics15.14 (2013), pp. 5022–5029.doi: 10 . 1039 / C3CP50350J

work page 2013
[20]

A quantum theory

Richard FW Bader and Atoms In Molecules. “A quantum theory”. In:Clarendon: Oxford, UK556 (1990)

work page 1990
[21]

Electron localization in solid-state structures of the elements: the diamond structure

Andreas Savin et al. “Electron localization in solid-state structures of the elements: the diamond structure”. In: Angewandte Chemie International Edition in English31.2 (1992), pp. 187–188.doi:10.1002/anie.199201871

work page doi:10.1002/anie.199201871 1992
[22]

Charge displacement analysis: A new general method to estimate atomic charges in molecules and clusters

PS Bagus et al. “Charge displacement analysis: A new general method to estimate atomic charges in molecules and clusters”. In:Journal of Molecular Catalysis A: Chemical119.1-3 (1997), pp. 3–10.doi: 10 . 1016 / S1381 - 1169(97)80041-6

work page 1997
[23]

A periodic energy decomposition analysis method for the investigation of chemi- cal bonding in extended systems

Marc Raupach and Ralf Tonner. “A periodic energy decomposition analysis method for the investigation of chemi- cal bonding in extended systems”. In:The Journal of chemical physics142.19 (2015).doi:10.1063/1.4919943. 10

work page doi:10.1063/1.4919943 2015
[24]

Energy decomposition analysis for excited states: an extension based on TDDFT

Florian Kreuter and Ralf Tonner-Zech. “Energy decomposition analysis for excited states: an extension based on TDDFT”. In:Physical Chemistry Chemical Physics27.9 (2025), pp. 4728–4745.doi:10.1039/D4CP04207G

work page doi:10.1039/d4cp04207g 2025
[25]

Schaefer, Beth Shapiro, and Richard E

Carl-Friedrich Sch¨ on et al. “Classification of properties and their relation to chemical bonding: Essential steps to- ward the inverse design of functional materials”. In:Science Advances8.47 (2022), eade0828.doi: 10.1126/sciadv. ade0828

work page doi:10.1126/sciadv 2022
[26]

Programming material properties by tuning intermolecular bond- ing

Upamanyu Ray, Zhenqian Pang, and Teng Li. “Programming material properties by tuning intermolecular bond- ing”. In:Journal of Applied Physics132.21 (2022).doi:10.1063/5.0123058

work page doi:10.1063/5.0123058 2022
[27]

Chemical Bonding Engineering: Insights into Physicochemical Performance Optimization for Energy-Storage/Conversion

Zhifang Zhou et al. “Chemical Bonding Engineering: Insights into Physicochemical Performance Optimization for Energy-Storage/Conversion”. In:Accounts of Materials Research5.12 (2024), pp. 1571–1582.doi: https://doi. org/10.1021/accountsmr.4c00243

work page doi:10.1021/accountsmr.4c00243 2024
[28]

Antibonding valence states induce low lattice thermal conductivity in metal halide semi- conductors

Mohammad Ubaid et al. “Antibonding valence states induce low lattice thermal conductivity in metal halide semi- conductors”. In:Applied Physics Reviews11.4 (2024).doi:10.1063/5.0227080

work page doi:10.1063/5.0227080 2024
[29]

Exploiting chemical bonding principles to design high-performance thermoelectric materi- als

Anthony V Powell et al. “Exploiting chemical bonding principles to design high-performance thermoelectric materi- als”. In:Nature Reviews Chemistry(2025), pp. 1–20.doi:10.1038/s41570-025-00695-6

work page doi:10.1038/s41570-025-00695-6 2025
[30]

Adaptive machine learning framework to accelerate ab initio molecular dynamics

Venkatesh Botu and Rampi Ramprasad. “Adaptive machine learning framework to accelerate ab initio molecular dynamics”. In:International journal of quantum chemistry115.16 (2015), pp. 1074–1083.doi:10.1002/qua.24836

work page doi:10.1002/qua.24836 2015
[31]

Statistical analysis of coordination environments in oxides

David Waroquiers et al. “Statistical analysis of coordination environments in oxides”. In:Chemistry of Materials 29.19 (2017), pp. 8346–8360.doi:10.1021/acs.chemmater.7b02766

work page doi:10.1021/acs.chemmater.7b02766 2017
[32]

ChemEnv: a fast and robust coordination environment identification tool

David Waroquiers et al. “ChemEnv: a fast and robust coordination environment identification tool”. In:Structural Science76.4 (2020), pp. 683–695.doi:10.1107/S2052520620007994

work page doi:10.1107/s2052520620007994 2020
[33]

Physics-inspired structural representations for molecules and materials

Felix Musil et al. “Physics-inspired structural representations for molecules and materials”. In:Chemical Reviews 121.16 (2021), pp. 9759–9815.doi:10.1021/acs.chemrev.1c00021

work page doi:10.1021/acs.chemrev.1c00021 2021
[34]

Automated bonding analysis with crystal orbital hamilton populations

Janine George et al. “Automated bonding analysis with crystal orbital hamilton populations”. In:ChemPlusChem 87.11 (2022), e202200123.doi:10.1002/cplu.202200123

work page doi:10.1002/cplu.202200123 2022
[35]

Atomate2: Modular workflows for materials science

Alex M Ganose et al. “Atomate2: Modular workflows for materials science”. In:Digital Discovery(2025).doi: 10. 1039/D5DD00019J

work page 2025
[36]

A quantum-chemical bonding database for solid-state materials

Aakash Ashok Naik et al. “A quantum-chemical bonding database for solid-state materials”. In:Scientific Data10.1 (2023), p. 610.doi:10.1038/s41597-023-02477-5

work page doi:10.1038/s41597-023-02477-5 2023
[37]

Version v1.0.0

Aakash Ashok Naik et al.Quantum-Chemical Bonding Database (Unprocessed data : Part 1). Version v1.0.0. Apr. 2023.doi:10.5281/zenodo.7852083

work page doi:10.5281/zenodo.7852083 2023
[38]

Commentary: The Materials Project: A materials genome approach to accelerating materials innovation

Anubhav Jain et al. “Commentary: The Materials Project: A materials genome approach to accelerating materials innovation”. In:APL materials1.1 (2013).doi:10.1063/1.4812323

work page doi:10.1063/1.4812323 2013
[39]

Crystal orbital Hamilton population (COHP) analysis as projected from plane-wave basis sets

Volker L Deringer, Andrei L Tchougr´ eeff, and Richard Dronskowski. “Crystal orbital Hamilton population (COHP) analysis as projected from plane-wave basis sets”. In:The journal of physical chemistry A115.21 (2011), pp. 5461– 5466.doi:https://doi.org/10.1021/jp202489s

work page doi:10.1021/jp202489s 2011
[40]

Analytic projection from plane-wave and PAW wavefunctions and application to chemical- bonding analysis in solids

Stefan Maintz et al. “Analytic projection from plane-wave and PAW wavefunctions and application to chemical- bonding analysis in solids”. In:Journal of computational chemistry34.29 (2013), pp. 2557–2567.doi: https://doi. org/10.1002/jcc.23424

work page doi:10.1002/jcc.23424 2013
[41]

LOBSTER: A Tool to Extract Chemical Bonding from Plane-Wave Based DFT

Stefan Maintz et al. “LOBSTER: A Tool to Extract Chemical Bonding from Plane-Wave Based DFT”. In:Journal of Computational Chemistry37 (2016), pp. 1030–1035.doi:10.1002/jcc.24300

work page doi:10.1002/jcc.24300 2016
[42]

LOBSTER: Local orbital projections, atomic charges, and chemical-bonding analysis from projector-augmented-wave-based density-functional theory

Ryky Nelson et al. “LOBSTER: Local orbital projections, atomic charges, and chemical-bonding analysis from projector-augmented-wave-based density-functional theory”. In:Journal of computational chemistry41.21 (2020), pp. 1931–1940.doi:10.1002/jcc.26353

work page doi:10.1002/jcc.26353 2020
[43]

Chains of trans-edge-sharing molybdenum octahedra: metal-metal bonding in extended systems

Timothy Hughbanks and Roald Hoffmann. “Chains of trans-edge-sharing molybdenum octahedra: metal-metal bonding in extended systems”. In:Journal of the American Chemical Society105.11 (1983), pp. 3528–3537.doi: 10.1021/ja00349a027

work page doi:10.1021/ja00349a027 1983
[44]

Crystal orbital Hamilton populations (COHP): energy-resolved vi- sualization of chemical bonding in solids based on density-functional calculations

Richard Dronskowski and Peter E Bloechl. “Crystal orbital Hamilton populations (COHP): energy-resolved vi- sualization of chemical bonding in solids based on density-functional calculations”. In:The Journal of Physical Chemistry97.33 (1993), pp. 8617–8624.doi:10.1021/j100135a014

work page doi:10.1021/j100135a014 1993
[45]

Crystal orbital bond index: covalent bond orders in solids

Peter C M¨ uller et al. “Crystal orbital bond index: covalent bond orders in solids”. In:The Journal of Physical Chemistry C125.14 (2021), pp. 7959–7970.doi:10.1021/acs.jpcc.1c00718

work page doi:10.1021/acs.jpcc.1c00718 2021
[46]

Development of a robust tool to extract Mulliken and L¨ owdin charges from plane waves and its application to solid-state materials

Christina Ertural, Simon Steinberg, and Richard Dronskowski. “Development of a robust tool to extract Mulliken and L¨ owdin charges from plane waves and its application to solid-state materials”. In:RSC advances9.51 (2019), pp. 29821–29830.doi:10.1039/C9RA05190B. 11

work page doi:10.1039/c9ra05190b 2019
[47]

Electrostatic, Madelung and cohesive energies for solids

MG Trefry, EN Maslen, and MA Spackman. “Electrostatic, Madelung and cohesive energies for solids”. In:Journal of Physics C: Solid State Physics20.1 (1987), p. 19.doi:https://doi.org/10.1088/0022-3719/20/1/006

work page doi:10.1088/0022-3719/20/1/006 1987
[48]

On the Madelung part of lattice energy: New pathways to use it as a tool in solid state chemistry (Part 1)

Rudolf Hoppe. “On the Madelung part of lattice energy: New pathways to use it as a tool in solid state chemistry (Part 1)”. In:Zeitschrift f¨ ur Naturforschung A50.6 (1995), pp. 555–567.doi:10.1515/zna-1995-0607

work page doi:10.1515/zna-1995-0607 1995
[49]

LobsterPy: A package to automatically analyze LOBSTER runs

Aakash Ashok Naik et al. “LobsterPy: A package to automatically analyze LOBSTER runs”. In:Journal of Open Source Software9.94 (2024), p. 6286.doi:10.21105/joss.06286

work page doi:10.21105/joss.06286 2024
[50]

Long-Range Forces in Rock-Salt-Type Tellurides and How they Mirror the Underlying Chemical Bonding

Jan Hempelmann et al. “Long-Range Forces in Rock-Salt-Type Tellurides and How they Mirror the Underlying Chemical Bonding”. In:Advanced Materials33.37 (2021), p. 2100163.doi:10.1002/adma.202100163

work page doi:10.1002/adma.202100163 2021
[51]

Identification of material dimensionality based on force constant analysis

Mohammad Bagheri, Ethan Berger, and Hannu-Pekka Komsa. “Identification of material dimensionality based on force constant analysis”. In:The Journal of Physical Chemistry Letters14.35 (2023), pp. 7840–7847.doi: 10.1021/ acs.jpclett.3c01635

work page 2023
[52]

Benchmarking materials property prediction methods: the Matbench test set and Automat- miner reference algorithm

Alexander Dunn et al. “Benchmarking materials property prediction methods: the Matbench test set and Automat- miner reference algorithm”. In:npj Computational Materials6.1 (2020), p. 138.doi: 10.1038/s41524-020-00406- 3

work page doi:10.1038/s41524-020-00406- 2020
[53]

Machine learning assisted prediction of the phonon cutoff frequency of ABO3 perovskite materi- als

Chen Gong et al. “Machine learning assisted prediction of the phonon cutoff frequency of ABO3 perovskite materi- als”. In:Computational Materials Science239 (2024), p. 112943.doi:10.1016/j.commatsci.2024.112943

work page doi:10.1016/j.commatsci.2024.112943 2024
[54]

Upper bounds on the highest phonon frequency and superconducting temperature from fundamental physical constants

Kostya Trachenko et al. “Upper bounds on the highest phonon frequency and superconducting temperature from fundamental physical constants”. In:Journal of Physics: Condensed Matter37.16 (2025), p. 165401.doi: 10.1088/ 1361-648X/adbc39

work page 2025
[55]

Thinking like a chemist: intuition in thermoelectric materials

Wolfgang G Zeier et al. “Thinking like a chemist: intuition in thermoelectric materials”. In:Angewandte Chemie International Edition55.24 (2016), pp. 6826–6841.doi:10.1002/anie.201508381

work page doi:10.1002/anie.201508381 2016
[56]

Accelerated discovery and design of ultralow lattice thermal conductivity materials using chemical bonding principles

Jiangang He et al. “Accelerated discovery and design of ultralow lattice thermal conductivity materials using chemical bonding principles”. In:Advanced Functional Materials32.14 (2022), p. 2108532.doi: 10 . 1002 / adfm . 202108532

work page 2022
[57]

Elastic moduli: a tool for understanding chemical bonding and thermal transport in thermo- electric materials

Eleonora Isotta et al. “Elastic moduli: a tool for understanding chemical bonding and thermal transport in thermo- electric materials”. In:Angewandte Chemie135.12 (2023), e202213649.doi:10.1002/ange.202213649

work page doi:10.1002/ange.202213649 2023
[58]

Materials property prediction for limited datasets enabled by feature selection and joint learning with MODNet

Pierre-Paul De Breuck, Geoffroy Hautier, and Gian-Marco Rignanese. “Materials property prediction for limited datasets enabled by feature selection and joint learning with MODNet”. In:npj computational materials7.1 (2021), p. 83.doi:10.1038/s41524-021-00552-2

work page doi:10.1038/s41524-021-00552-2 2021
[59]

Accurate experimental band gap predic- tions with multifidelity correction learning

Pierre-Paul De Breuck, Gr´ egoire Heymans, and Gian-Marco Rignanese. “Accurate experimental band gap predic- tions with multifidelity correction learning”. In:Journal of Materials Informatics2.3 (2022), N–A.doi: 10.20517/ jmi.2022.13

work page 2022
[60]

A unified approach to interpreting model predictions

Scott M Lundberg and Su-In Lee. “A unified approach to interpreting model predictions”. In:Advances in neural information processing systems30 (2017)

work page 2017
[61]

From local explanations to global understanding with explainable AI for trees

Scott M Lundberg et al. “From local explanations to global understanding with explainable AI for trees”. In:Na- ture machine intelligence2.1 (2020), pp. 56–67.doi:10.1038/s42256-019-0138-9

work page doi:10.1038/s42256-019-0138-9 2020
[62]

Permutation importance: a corrected feature importance measure

Andr´ e Altmann et al. “Permutation importance: a corrected feature importance measure”. In:Bioinformatics26.10 (2010), pp. 1340–1347.doi:https://doi.org/10.1093/bioinformatics/btq134

work page doi:10.1093/bioinformatics/btq134 2010
[63]

SISSO: A compressed-sensing method for identifying the best low-dimensional descrip- tor in an immensity of offered candidates

Runhai Ouyang et al. “SISSO: A compressed-sensing method for identifying the best low-dimensional descrip- tor in an immensity of offered candidates”. In:Physical Review Materials2.8 (2018), p. 083802.doi: 10 . 1103 / PhysRevMaterials.2.083802

work page 2018
[64]

SISSO++: A C++ Implementation of the Sure-Independence Screening and Sparsify- ing Operator Approach

Thomas A. r. Purcell et al. “SISSO++: A C++ Implementation of the Sure-Independence Screening and Sparsify- ing Operator Approach”. In:Journal of Open Source Software7.71 (2022), p. 3960.doi:10.21105/joss.03960

work page doi:10.21105/joss.03960 2022
[65]

Contribution of vi- brational excited molecular nitrogen to ammonia synthesis using an atmospheric-pressure plasma jet,

Thomas AR Purcell, Matthias Scheffler, and Luca M Ghiringhelli. “Recent advances in the SISSO method and their implementation in the SISSO++ code”. In:The Journal of Chemical Physics159.11 (2023).doi: 10.1063/5. 0156620

work page doi:10.1063/5 2023
[66]

Combining feature-based approaches with graph neural networks and symbolic regression for synergistic performance and interpretability

Rog´ erio Almeida Gouvˆ ea et al. “Combining feature-based approaches with graph neural networks and symbolic regression for synergistic performance and interpretability”. In:npj Computational Materials(2026).doi: 10.1038/ s41524-025-01938-2

work page 2026
[67]

Robust model benchmarking and bias- imbalance in data-driven materials science: a case study on MODNet

Pierre-Paul De Breuck, Matthew L Evans, and Gian-Marco Rignanese. “Robust model benchmarking and bias- imbalance in data-driven materials science: a case study on MODNet”. In:Journal of Physics: Condensed Matter 33.40 (2021), p. 404002.doi:10.1088/1361-648X/ac1280

work page doi:10.1088/1361-648x/ac1280 2021
[68]

Inference for the generalization error

Claude Nadeau and Yoshua Bengio. “Inference for the generalization error”. In:Advances in neural information processing systems12 (1999). 12

work page 1999
[69]

Approximate statistical tests for comparing supervised classification learning algorithms

Thomas G Dietterich. “Approximate statistical tests for comparing supervised classification learning algorithms”. In:Neural computation10.7 (1998), pp. 1895–1923.doi:10.1162/089976698300017197

work page doi:10.1162/089976698300017197 1998
[70]

Hierarchical symbolic regression for identifying key physical parameters correlated with bulk properties of perovskites

Lucas Foppa et al. “Hierarchical symbolic regression for identifying key physical parameters correlated with bulk properties of perovskites”. In:Physical Review Letters129.5 (2022), p. 055301.doi: 10.1103/PhysRevLett.129. 055301

work page doi:10.1103/physrevlett.129 2022
[71]

Symbolic regression as a feature engineering method for ma- chine and deep learning regression tasks

Assaf Shmuel, Oren Glickman, and Teddy Lazebnik. “Symbolic regression as a feature engineering method for ma- chine and deep learning regression tasks”. In:Machine Learning: Science and Technology5.2 (2024), p. 025065.doi: 10.1088/2632-2153/ad513a

work page doi:10.1088/2632-2153/ad513a 2024
[72]

Experimental and Theoretical Force Constants as Meaningful Indicator for Interatomic Bond- ing Characteristics and the Specific Case of Elemental Antimony

Franziska Zahn et al. “Experimental and Theoretical Force Constants as Meaningful Indicator for Interatomic Bond- ing Characteristics and the Specific Case of Elemental Antimony”. In:Advanced Materials37.7 (2025), p. 2416320. doi:10.1002/adma.202416320

work page doi:10.1002/adma.202416320 2025
[73]

Vibrational properties and bonding nature of Sb2Se3 and their implications for chalco- genide materials

Volker L Deringer et al. “Vibrational properties and bonding nature of Sb2Se3 and their implications for chalco- genide materials”. In:Chemical science6.9 (2015), pp. 5255–5262.doi:10.1039/C5SC00825E

work page doi:10.1039/c5sc00825e 2015
[74]

The Orbital Origins of Chemical Bonding in Ge- Sb- Te Phase-Change Materials

Jan Hempelmann et al. “The Orbital Origins of Chemical Bonding in Ge- Sb- Te Phase-Change Materials”. In: Angewandte Chemie International Edition61.17 (2022), e202115778.doi:10.1002/anie.202115778

work page doi:10.1002/anie.202115778 2022
[75]

Bonding hierarchy gives rise to high thermoelectric performance in layered zintl compound baau2p4

Koushik Pal, Jiangang He, and C Wolverton. “Bonding hierarchy gives rise to high thermoelectric performance in layered zintl compound baau2p4”. In:Chemistry of Materials30.21 (2018), pp. 7760–7768.doi: 10 . 1021 / acs . chemmater.8b03130

work page 2018
[76]

Bonding heterogeneity in mixed-anion compounds realizes ultralow lattice thermal conductivity

Naoki Sato et al. “Bonding heterogeneity in mixed-anion compounds realizes ultralow lattice thermal conductivity”. In:Journal of Materials Chemistry A9.39 (2021), pp. 22660–22669.doi:10.1039/D1TA04958E

work page doi:10.1039/d1ta04958e 2021
[77]

Autieri, G

Xin Jin et al. “Bonding heterogeneity leads to hierarchical and ultralow lattice thermal conductivity in sodium metavanadate”. In:The Journal of Physical Chemistry Letters13.48 (2022), pp. 11160–11168.doi: 10.1021/acs. jpclett.2c03061

work page doi:10.1021/acs 2022
[78]

Accelerated discovery of extreme lattice thermal conductivity by crystal graph atten- tion networks and chemical bonding

Mohammed Al-Fahdi et al. “Accelerated discovery of extreme lattice thermal conductivity by crystal graph atten- tion networks and chemical bonding”. In:npj Computational Materials(2025).doi:10.1038/s41524-025-01871-4

work page doi:10.1038/s41524-025-01871-4 2025
[79]

High Lattice Thermal Conductivity Solids

Donald T. Morelli and Glen A. Slack. “High Lattice Thermal Conductivity Solids”. In:High Thermal Conductivity Materials. Springer New York, 2006, pp. 37–68.isbn: 978-0-387-25100-4.doi:10.1007/0-387-25100-6_2

work page doi:10.1007/0-387-25100-6_2 2006
[80]

Identification of crystalline materials with ultra-low thermal conductivity based on machine learning study

Xinming Wang et al. “Identification of crystalline materials with ultra-low thermal conductivity based on machine learning study”. In:The Journal of Physical Chemistry C124.16 (2020), pp. 8488–8495.doi: 10.1021/acs.jpcc. 9b11610

work page doi:10.1021/acs.jpcc 2020

Showing first 80 references.