Pith.

Materials Science · arXiv:2611.16308 v1 · submitted 12 Nov 2026 · 57 references

ORCHID-SAGE: Fully Reproducible CALPHAD Assessment from Evidence to Phase Diagram

Raymundo Arróyave

The paper says

“Knowledge of phase stability is the first requirement for materials design, and predicting the conditions under which new materials can be synthesized has become a key capability of self-driving laboratories (SDLs).”

more of the abstract

For inorganic materials, notably alloys, these predictions come from thermodynamic databases built with the CALPHAD method. Building a database, however, requires expert judgment at many steps, and because these decisions are rarely recorded, others can use a database but cannot regenerate or update it. Machines can already collect the data, build a first model with machine-learned interatomic potentials and refine it against experiments. The assessment itself, the sequence of judgments that turns these parts into a reliable database, is still carried out by human experts. We argue that it can be automated reproducibly only if the expert knowledge behind an assessment is encoded in formal protocols, so that an assessment delivers a record of its construction and not only its parameters. We introduce ORCHID-SAGE, a standard that organizes this record into reproducible and automatable protocols. In this first attempt, applied to Cr-Ni, the database was regenerated from its record by a script on a second machine, by an AI agent working only through the Model Context Protocol (MCP) and by a refit from public sources only, with the same predictions. Recorded in this way, a database need no longer be fixed at publication but can be updated by AI agents as SDLs and other sources provide new evidence, while the judgment remains with people and on the record.

from the abstract

Pith says

Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded · not checked by a person

It can be wrong. Challenge this review or reply to it.

The regenerated-fit demonstration is real and well-supported, but the title's 'from evidence to phase diagram' runs one stage ahead of what was actually executed, since the one-shot MLIP stage was not regenerated — and the authors say so themselves. This line is wrong

the referee pass did not run

How often objections from reviews like this one are right: not yet measured

Are you an author? What you can do.

Not checked: figures, code, what the 57 references say.

What would settle it

Perform a third regeneration on a second GPU machine in which the one-shot model is recomputed from its raw inputs (SGTE unary descriptions, MLIP checkpoint, SQS files, liquid MD inputs) using the recorded one-shot run cards, and the fitted one-shot TDB is rebuilt from the regenerated energies. Then run the PhaseForge+ fit from the specific protocol using those regenerated one-shot observations in place of the recorded ones, and compare the final TDB against the original using the same 13 checks and the same declared tolerances. If all checks pass, the full evidence-to-phase-diagram claim is supported; if the regenerated one-shot inputs shift any check, the record is not yet shown to be sufficient for the full pipeline.

from the skeptic's reading

How the review went

3 readings by deepseek-v4-flash · 13 Nov 2026

  1. Reader holds up, with conditions confidence high
  2. Skeptic partly agrees 1 min 33 s · 34,903 tokens read, 12,694 written
  3. Editor wrote the verdict above 23 s · 36,557 tokens read, 609 written

Not run on this paper: the referee, the simulated authors' defence.

Pith did not check this review's objections against the paper.

The review in full

The review in brief, machine-written

Read the ORCHID-SAGE paper; it's worth your time even though the subfield is narrow. The central demonstration is sound: the Cr–Ni database was regenerated from its recorded protocol twice, once by script on a DGX Spark against the Mac Studio original and once by an AI agent working through the MCP server, and both reproductions hold within the declared tolerances. Deterministic checks identical, numerical checks inside tolerance, eutectic and CrNi2-top at 1614 K and 860 K in all three runs. The refit from public sources alone also lands on the same solution. That is a real, new demonstration of a genuinely useful capability, and the artifact provision is about as complete as I have seen in this field: RO-Crate, SHA-256 hashes, pinned commits, run cards, decision log, dashboard.

The genuinely novel piece is the decision log. Recording not just data and code but the judgments — which phases to exclude, which uncertainties to assign, which fit among competing starts to pick — and making those judgments part of the reproducible record directly addresses a dysfunction the author states plainly: even the original assessor usually cannot reproduce a fit after a few years. That is the contribution.

The soft spots, in proportion. The sharpest one is exactly what the stress test flags. The executed regeneration starts from the evidence bundle and the one-shot energetics; the PhaseForge MLIP stage — SQS relaxations (including the bcc/fcc structures that transformed and entered as midpoints of their relaxation intervals) and the liquid MD — was only checked on a small sample of structures on a second GPU. Section 3.3 says this outright. Since the one-shot energies enter the fit as observations at 9.7 kJ/mol, an upstream shift could propagate to the final database even though the downstream fit is perfectly reproducible. The two-DGX-Spark check of the liquid MD, agreeing within that tolerance, is encouraging but does not cover the solids. So the title claim 'from evidence to phase diagram' exceeds what was executed; the authors acknowledge it, which makes it a revision point, not a hidden flaw.

Two lesser items, both stated: the pair-entropy estimator is explicitly provisional and unreviewed, and part of the evidence is single-reader pending a second pass. Single binary system. None of this is concealed. The paper deserves a serious referee and, after a title-scope edit and ideally one more system, publication. Send it out.

The strongest objection

The scripted DGX Spark rerun and the MCP-agent rerun in Section 3.1 are strong evidence that, once the evidence bundle and the one-shot model are fixed, the specific protocol is sufficient to regenerate the fitted database. The load-bearing gap is the step before that: Section 2.2 states that the evidence bundle contains 'the observations derived from the one-shot model constructed using PhaseForge,' and Section 3.3 states that 'the reproduction reported here, finally, starts from the one-shot energetics,' with the MLIP calculations 'only checked for a small sample of structures on a second GPU.' The authors acknowledge that a full regeneration is possible but was not performed. The central title claim, 'from evidence to phase diagram,' therefore exceeds what has actually been executed: the raw-evidence-to-one-shot stage, including SQS relaxations of transforming structures, liquid molecular dynamics, and the provisional pair-entropy estimator, has not been regenerated from the record. Section 2.3 shows that this stage contains genuine judgment: several bcc and fcc structures transformed during relaxation and were entered with the midpoint of their relaxation interval as the value and half the interval as the uncertainty. That rule is recorded, but whether following it on a second machine reproduces sufficiently close one-shot observations is untested. Since the one-shot energies enter the fit as observations with 9.7 kJ/mol uncertainty, a shift at this stage could propagate to the final database even though the downstream fit is perfectly reproducible. This is an acknowledged limitation, but it is the sharpest unresolved point bearing on the central claim.

The reader's strongest claim is well supported for the part of the pipeline that was actually re-run: the DGX Spark script reproduced all 13 checks, and the AI agent reproduced them on the original machine, while the mode A/A2 analysis shows that the undetermined parameters have little effect on predictions. My concern is narrower and is acknowledged by the authors at the end of Section 3.3: the regeneration starts from the one-shot energetics, so the link from raw evidence and MLIP calculations to the one-shot model is asserted but not executed. That link contains the most machine-dependent and judgment-heavy steps, including the treatment of transforming SQS structures and the provisional liquid entropy estimator. A full regeneration on a second GPU would settle whether the record is sufficient for the entire pipeline or only for the fit stage. This reinforces the CONDITIONAL verdict rather than overturning it: I see no basis for rejection, and the evidence is not yet strong enough for unconditional acceptance. The reader's weakest assumption about record completeness is the right theme, and the unexecuted upstream stage is the sharpest unverified part of that completeness.

Pith's reading of the paper (a machine's summary, not the authors')

The paper introduces ORCHID-SAGE, a three-level protocol standard for recording CALPHAD thermodynamic assessments so that the resulting database can be regenerated from the record alone, without any additional information. The central claim is that if an assessment is documented as a general protocol (the method), a specific protocol (the system's decisions and inputs), and run cards (execution details), then the database can be reproduced on a different machine, by a script, or by an AI agent. The authors test this on the Cr–Ni binary system: the database was regenerated on a DGX Spark (scripted) and on a Mac Studio (by an AI agent via the Model Context Protocol), with all 13 declared checks passed—five deterministic checks byte-identical and eight numerical checks within tolerances. The paper argues that this standard enables 'living thermodynamic databases' that can be updated as new evidence emerges, essential for self-driving laboratories that use phase-stability predictions to guide synthesis and then refine models from the results.

What carries the argument. The three-level protocol hierarchy: (1) general protocol (45 steps, versioned and hashed) describing the method applicable to any binary system; (2) specific protocol listing the system's inputs, executed runs, and all decisions with justifications; (3) run cards documenting individual executions with software versions, commands, hashed inputs/outputs, and optimizer trajectories. The record is packaged as an RO-Crate with a dashboard. Reproduction is verified by comparing predictions (not parameters) within declared tolerances, since the CALPHAD objective landscape is nearly flat in some directions.

Core claim. ORCHID-SAGE is a standard for recording CALPHAD assessments at three levels—general protocol, specific protocol, and run cards—such that the thermodynamic database can be fully regenerated from the record without any external information. The standard is validated on the Cr–Ni binary system: a scripted rerun on a different machine (DGX Spark) and an AI agent rerun (Claude Opus 5.5 via MCP server) both reproduced the original predictions within declared tolerances, including the eutectic temperature (1614 K) and the top of CrNi2 (860 K). The paper further demonstrates that the record captures not only data and code but also the assessor's decisions, enabling reproducibility, error detection, and future updates by AI agents.

The premise it rests on. The three-level record (general protocol, specific protocol, run cards) together with the evidence bundle contains all the information needed to reproduce the database, and comparing predictions within declared tolerances is a faithful criterion for reproduction.

If the paper is right (Pith's inferences, not the authors')

A CALPHAD database can be regenerated from its record without contacting the authors, enabling verification, correction, and extension.

The standard allows databases to become 'living'—updated as new experimental or computational evidence becomes available, with each version traceable to its evidence and decisions.

AI agents can carry out the assessment and regeneration, reducing human effort while keeping judgment steps recorded and auditable.

The approach identifies undetermined parameters (e.g., two modes in Cr–Ni with identical predictions) that would go unnoticed in conventional assessments.

The Cr–Ni assessment itself points to a specific region (Cr-rich bcc boundary) where new measurements would be most useful.

If ORCHID-SAGE is adopted, the CALPHAD community could build a library of 'living' databases that evolve with new data, potentially transforming alloy design from static reference databases to dynamic, continuously improving resources.

The requirement to record every decision explicitly may reveal implicit assumptions and biases in current assessment practices, improving the rigor and transparency of the field.

The standard's reliance on AI agents for extraction, fitting, and regeneration suggests a future where human experts focus on judgment steps (phase selection, model choice) while machines handle execution, potentially accelerating database construction for complex multicomponent systems.

The identified metastable bcc miscibility gap in Cr–Ni, which could not be resolved without the full record, illustrates how the standard enables deeper physical analysis that would be impossible from parameters alone.

Check this review

Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.

How often reviews like this one are right

Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.

How often objections from reviews like this one are right: not yet measured

Every value above, in the review's public record (JSON)

Reviews by people

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review

How Pith checks papers · The full audit of this paper

The paper

Abstract

Knowledge of phase stability is the first requirement for materials design, and predicting the conditions under which new materials can be synthesized has become a key capability of self-driving laboratories (SDLs). For inorganic materials, notably alloys, these predictions come from thermodynamic databases built with the CALPHAD method. Building a database, however, requires expert judgment at many steps, and because these decisions are rarely recorded, others can use a database but cannot regenerate or update it. Machines can already collect the data, build a first model with machine-learned interatomic potentials and refine it against experiments. The assessment itself, the sequence of judgments that turns these parts into a reliable database, is still carried out by human experts. We argue that it can be automated reproducibly only if the expert knowledge behind an assessment is encoded in formal protocols, so that an assessment delivers a record of its construction and not only its parameters. We introduce ORCHID-SAGE, a standard that organizes this record into reproducible and automatable protocols. In this first attempt, applied to Cr-Ni, the database was regenerated from its record by a script on a second machine, by an AI agent working only through the Model Context Protocol (MCP) and by a refit from public sources only, with the same predictions. Recorded in this way, a database need no longer be fixed at publication but can be updated by AI agents as SDLs and other sources provide new evidence, while the judgment remains with people and on the record.

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.

abstract compiled · 2026-11-13

The authors' main claim, as the classifier picked it out 0.71

“We argue that it can be automated reproducibly only if the expert knowledge behind an assessment is encoded in formal protocols, so that an assessment delivers a record of its construction and not only its parameters.”

How the main claim is stated: hedged possibility 0.52

What the abstract itself states, as probabilities: claims first 0.35 · states limitation 0.83 · human participants 0.20 · controlled comparison 0.11 · randomized assignment 0.01 · states quantitative result 0.07 · compares with prior results 0.35 · mentions code or data availability 0.35

The abstract, sentence by sentence

This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.

SentenceRoleThe authors' own workStates the main conclusionStrength
Knowledge of phase stability is the first requirement for materials design, and predicting the conditions under which new materials can be synthesized has become a key capability of self-driving laboratories (SDLs). background 1.00 0.05 0.03 asserted 1.00
For inorganic materials, notably alloys, these predictions come from thermodynamic databases built with the CALPHAD method. background 1.00 0.04 0.02 asserted 1.00
Building a database, however, requires expert judgment at many steps, and because these decisions are rarely recorded, others can use a database but cannot regenerate or update it. gap 0.82 0.08 0.04 asserted 1.00
Machines can already collect the data, build a first model with machine-learned interatomic potentials and refine it against experiments. background 0.93 0.09 0.03 asserted 0.91
The assessment itself, the sequence of judgments that turns these parts into a reliable database, is still carried out by human experts. gap 0.81 0.09 0.05 asserted 1.00
We argue that it can be automated reproducibly only if the expert knowledge behind an assessment is encoded in formal protocols, so that an assessment delivers a record of its construction and not only its parameters. gap 0.42 0.17 0.68 hedged possibility 0.52
We introduce ORCHID-SAGE, a standard that organizes this record into reproducible and automatable protocols. contribution 1.00 0.90 0.22 asserted 0.99
In this first attempt, applied to Cr-Ni, the database was regenerated from its record by a script on a second machine, by an AI agent working only through the Model Context Protocol (MCP) and by a refit from public sources only, with the same predictions. result 0.86 0.94 0.16 asserted 1.00
Recorded in this way, a database need no longer be fixed at publication but can be updated by AI agents as SDLs and other sources provide new evidence, while the judgment remains with people and on the record. interpretation 0.90 0.41 0.35 hedged possibility 0.58

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from simulation or computation 0.80

What was examined, and what was not

Sources read

  • arXiv listing abstract v1

Not examined

  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers

Levels run

  • P abstract

Not run

  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

How the labels were made

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-13: 1 request, 7,046 input tokens.

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON

Figures (12)
Figure 1 of the paper
Figure 1. ORCHID-SAGE, shown for Cr–Ni. The record (general protocol, specific protocol with its decisions, run cards)
Figure 2 of the paper
Figure 2. An example of an automated workflow for CALPHAD model refinement:
Figure 3 of the paper
Figure 3. The ORCHID-SAGE standard, shown for Cr–Ni.
Figure 4 of the paper
Figure 4. Gates and ranking in the general protocol. (a) The
Figure 6 of the paper
Figure 6. Energetics of the one-shot model of Cr–Ni, in
Figure 7 of the paper
Figure 7. Excess entropy of the Cr–Ni liquid from molecu
Figure 8 of the paper
Figure 8. Verification of the specific protocol and of its reruns.
Figure 9 of the paper
Figure 9. Regeneration of the Cr–Ni database from its record. (a) Phase diagrams calculated from the database of the original
Figure 11 of the paper
Figure 11. The Cr–Ni fit sequence, from the one-shot model
Figure 12 of the paper
Figure 12. Physics and uncertainty of the final fit. (a) Stable diagram (purple) with the metastable miscibility gap of BCC_A2:

Showing 10 of the paper's 12 figures. Figures are the authors', reproduced from the paper.

References (57)
  1. [1] Hans Leo Lukas, Suzana G. Fries, and Bo Sund- man.Computational Thermodynamics: The Cal- phad Method. Cambridge University Press, 2007. doi:10.1017/CBO9780511804137 on Pithdoi:10.1017/cbo9780511804137
  2. [2] Richard Otis and Zi-Kui Liu. pycalphad: CALPHAD-based computational thermodynam- ics in Python.Journal of Open Research Software, 5 (1):1, 2017. doi:10.5334/jors.140 on Pithdoi:10.5334/jors.140
  3. [3] Raymundo Arróyave. Phase stability through machine learning.Journal of Phase Equilibria and Diffusion, 43(6):606–628, 2022. doi:10.1007/ s11669-022-01009-9 on Pith
  4. [4] Brandon Bocklund, Richard Otis, Aleksei Egorov, Abdulmonem Obaied, Irina Roslyakova, and Zi- Kui Liu. Espei for efficient thermodynamic database development, modification, and uncer- tainty quantification: application to cu–mg.MRS Communications, 9(2):618–627, 2019 on Pith
  5. [5] Raymundo Arróyave, James E. Saal, and Dong- won Shin. ORCHID: an open, AI-ready resource of CALPHAD literature data, calculated phase di- agrams, phase identities and DFT references, 2026 on Pith
  6. [6] Felix Mölder, Kim Philipp Jablonski, Brice Letcher, Michael B. Hall, Christopher H. Tomkins- Tinch, Vanessa Sochat, Jan Forster, Soohyun Lee, Sven O. Twardziok, Alexander Kanitz, Andreas Wilm, Manuel Holtgrewe, Sven Rahmann, Sven Nahnsen, and Johannes Köster. Sustainable data ... on Pith
  7. [7] Paolo Di Tommaso, Maria Chatzou, Evan W. Flo- den, Pablo Prieto Barja, Emilio Palumbo, and Cedric Notredame. Nextflow enables reproducible computational workflows.Nature Biotechnology, 35(4):316–319, 2017. doi:10.1038/nbt.3820 on Pithdoi:10.1038/nbt.3820
  8. [8] Michael R. Crusoe, Sanne Abeln, Alexandru Io- sup, Peter Amstutz, John Chilton, Nebojša Ti- jani´ c, Hervé Ménager, Stian Soiland-Reyes, Bog- dan Gavrilovi´ c, Carole Goble, and The CWL Com- munity. Methods included: Standardizing compu- tational reuse and portability with the... on Pithdoi:10.1145/3486897
  9. [9] Sebastiaan P . Huber, Spyros Zoupanos, Martin Uhrin, Leopold Talirz, Leonid Kahle, Rico Häusel- mann, Dominik Gresch, Tiziano Müller, Aliak- sandr V . Yakutovich, Casper W. Andersen, Fran- cisco F. Ramirez, Carl S. Adorf, Fernando Gargiulo, Snehal Kumbhar, Elsa Passaro, Conrad... on Pith
  10. [10] Claudia Draxl and Matthias Scheffler. The NO- MAD laboratory: from data sharing to artificial in- telligence.Journal of Physics: Materials, 2(3):036001, on Pith
  11. [11] Timothy Lebo, Satya Sahoo, and Deborah McGuin- ness. PROV-O: The PROV ontology. W3C Recom- mendation, 30 April 2013, 2013. URL https:// www.w3.org/TR/2013/REC-prov-o-20130430/ on Pith
  12. [12] Stian Soiland-Reyes, Peter Sefton, Mercè Crosas, Leyla Jael Castro, Frederik Coppens, José M. Fer- nández, Daniel Garijo, Björn Grüning, Marco La Rosa, Simone Leo, Eoghan Ó Carragáin, Marc Portier, Ana Trisovic, RO-Crate Community, Paul Groth, and Carole Goble. Packaging resea... on Pith
  13. [13] Simone Leo, Michael R. Crusoe, Laura Rodríguez- Navas, Raül Sirvent, Alexander Kanitz, et al. Recording provenance of workflow runs with RO- Crate.PLOS ONE, 19(9):e0309210, 2024. doi:10. 1371/journal.pone.0309210 on Pith
  14. [14] Milad Abolhasani and Eugenia Kumacheva. The rise of self-driving labs in chemical and materi- als sciences.Nature Synthesis, 2(6):483–492, 2023. doi:10.1038/s44160-022-00231-0 on Pithdoi:10.1038/s44160-022-00231-0
  15. [15] Nathan J. Szymanski, Bernardus Rendy, Yuxing Fei, Rishi E. Kumar, Tanjin He, David Milsted, et al. An autonomous laboratory for the accelerated syn- thesis of inorganic materials.Nature, 624(7990): 86–91, 2023. doi:10.1038/s41586-023-06734-w on Pithdoi:10.1038/s41586-023-06734-w
  16. [16] Pejman Honarmandi, Thien Chi Duong, S Fath- eme Ghoreishi, Douglas Allaire, and Raymundo Arroyave. Bayesian uncertainty quantification and information fusion in calphad-based thermo- dynamic modeling.Acta Materialia, 164:636–647, 2019 on Pith
  17. [17] Siya Zhu, Do˘ guhan Sarıtürk, and Raymundo Ar- róyave. Accelerating CALPHAD-based phase di- agram predictions in complex alloys using uni- versal machine learning potentials: Opportunities and challenges.Acta Materialia, page 120747, 2025 on Pith
  18. [18] Siya Zhu, Do˘ guhan Sarıtürk, and Raymundo Ar- róyave. Machine learning potentials for alloys: A detailed workflow to predict phase diagrams and benchmark accuracy, 2025. arXiv:2506.16771 on PitharXiv
  19. [19] Courtney Kunselman, Siya Zhu, Do˘ guhan Sarıtürk, and Raymundo Arróyave. Construction and tuning of CALPHAD models using machine- learned interatomic potentials and experimental data: A case study of the Pt–W system.Journal of Phase Equilibria and Diffusion, 2025. doi:10.1007/... on Pith
  20. [20] Courtney Kunselman, Brandon Bocklund, Axel van de Walle, Richard Otis, and Raymundo Ar- róyave. Analytically differentiable metrics for phase stability.Calphad, 86:102705, 2024. 22 on Pith
show 37 more
  1. [21] Courtney Kunselman, Brandon Bocklund, Richard Otis, and Raymundo Arroyave. Analytical gradient-based optimization of CALPHAD model parameters.arXiv preprint arXiv:2505.01585, 2025. doi:10.48550/arXiv.2505.01585. URL https:// arxiv.org/abs/2505.01585. arXiv:2505.01585 [cond-mat... on PithPith's reviewarXivdoi:10.48550/arxiv.2505.01585
  2. [22] Do˘ guhan Sarıtürk, Courtney Kunselman, and Siya Zhu. PhaseForgePlus, 2025. Software, version 0.1.0 on Pith
  3. [23] Dictionary.com. sage (noun). Dictio- nary.com Unabridged, based on the Random House Unabridged Dictionary, https://www.dictionary.com/browse/sage, on Pith
  4. [24] Anthropic. Model context protocol specification. Version 2025-06-18, 2025. URL https://modelcontextprotocol.io/ specification/2025-06-18 on Pith
  5. [25] Byeong-Joo Lee. On the stability of Cr car- bides.Calphad, 16(2):121–149, 1992. doi:10.1016/ 0364-5916(92)90002-f on Pith
  6. [26] Liangyan Hao, Andrei Ruban, and Wei Xiong. CALPHAD modeling based on Gibbs energy func- tions from zero kevin and improved magnetic model: A case study on the Cr–Ni system.Cal- phad, 73:102268, 2021. doi:10.1016/j.calphad.2021. 102268 on Pithdoi:10.1016/j.calphad.2021
  7. [27] Courtney Kunselman, Brandon Bocklund, Richard Otis, and Raymundo Arróyave. Gradient- informed Hamiltonian Monte Carlo for multicom- ponent CALPHAD model optimization and uncer- tainty quantification.Acta Materialia, 315:122412, on Pith
  8. [28] Qi-Jun Hong and Zi-Kui Liu. Generalized ap- proach for rapid entropy calculation of liquids and solids.Physical Review Research, 7:L012030, 2025. doi:10.1103/PhysRevResearch.7.L012030 on Pithdoi:10.1103/physrevresearch.7.l012030
  9. [29] Schema.org. HowTo, HowToSection and How- ToStep types. Schema.org vocabulary, 2026. URL https://schema.org/HowTo. Accessed 7 Novem- ber 2026 on Pith
  10. [30] L. A. Pugliese and G. R. Fitterer. Activities and phase boundaries in the Cr-Ni system using a solid electrolyte technique.Metallurgical Transactions, 1: 1997–2002, 1970. doi:10.1007/BF02642800 on Pithdoi:10.1007/bf02642800
  11. [31] A. T. Dinsdale. SGTE data for pure ele- ments.Calphad, 15(4):317–425, 1991. doi:10.1016/ 0364-5916(91)90030-N on Pith
  12. [32] Alex Zunger, S-H Wei, Luiz G Ferreira, and James E Bernard. Special quasirandom structures. Physical review letters, 65(3):353, 1990 on Pith
  13. [33] Raymundo Arróyave. m_scraps: Multi-sublattice short-range-order and cluster-correlation target- ing, 2026. Software, version 0.6.0 on Pith
  14. [34] Rahul Singh, Aayush Sharma, Prashant Singh, Ganesh Balasubramanian, and Duane D. John- son. Accelerating computational modeling and design of high-entropy alloys.Nature Com- putational Science, 1:54–61, 2021. doi:10.1038/ s43588-020-00006-7 on Pith
  15. [35] Anton Bochkarev, Yury Lysogorskiy, and Ralf Drautz. Graph atomic cluster expansion for semilocal interactions beyond equivariant mes- sage passing.Physical Review X, 14:021036, 2024. doi:10.1103/PhysRevX.14.021036 on Pithdoi:10.1103/physrevx.14.021036
  16. [36] Xuan L. Liu, Greta Lindwall, Thomas Gheno, and Zi-Kui Liu. Thermodynamic modeling of Al–Co– Cr, Al–Co–Ni, Co–Cr–Ni ternary systems towards a description for Al–Co–Cr–Ni.Calphad, 52:125– 142, 2016. doi:10.1016/j.calphad.2015.12.007 on Pithdoi:10.1016/j.calphad.2015.12.007
  17. [37] Guillaume Bracq, Jean-Claude Crivello, Mathilde Laurent-Brocq, Jean-Marc Joubert, and Ivan Guil- lot. What is the enthalpy contribution to the sta- bilization of the Co–Cr–Fe–Mn–Ni face-centered cubic solid solution?Journal of Phase Equilib- ria and Diffusion, 42:561–570, 2021... on Pith
  18. [38] Scott Kirklin, James E. Saal, Bryce Meredig, Alex Thompson, Jeff W. Doak, Muratahan Aykol, Stephan Rühl, and Chris Wolverton. The Open Quantum Materials Database (OQMD): Assessing the accuracy of DFT formation energies.npj Com- putational Materials, 1:15010, 2015. doi:10.1038/... on Pith
  19. [39] David J. Wales and Jonathan P . K. Doye. Global optimization by basin-hopping and the lowest energy structures of Lennard-Jones clusters con- taining up to 110 atoms.The Journal of Physical Chemistry A, 101(28):5111–5116, 1997. doi:10.1021/ jp970984n on Pith
  20. [40] Y. Xiang, D. Y. Sun, W. Fan, and X. G. Gong. Gen- eralized simulated annealing algorithm and its ap- plication to the Thomson model.Physics Letters A, 233(3):216–220, 1997. doi:10.1016/S0375-9601(97) 00474-X on Pithdoi:10.1016/s0375-9601(97
  21. [41] Pauli Virtanen, Ralf Gommers, Travis E. Oliphant, Matt Haberland, Tyler Reddy, et al. SciPy 1.0: Fundamental algorithms for scientific computing in Python.Nature Methods, 17(3):261–272, 2020. doi:10.1038/s41592-019-0686-2 on Pithdoi:10.1038/s41592-019-0686-2
  22. [42] John W. Cahn. On spinodal decomposition.Acta Metallurgica, 9(9):795–801, 1961. doi:10.1016/ 0001-6160(61)90182-1 on Pith
  23. [43] Luke Tierney and Joseph B. Kadane. Accurate ap- proximations for posterior moments and marginal densities.Journal of the American Statistical Associ- ation, 81(393):82–86, 1986. doi:10.1080/01621459. 1986.10478240. 23 on Pithdoi:10.1080/01621459
  24. [44] V . N. Svechnikov and V . M. Pan. Characteristics of the equilibrium diagram and processes of solution and precipitation in the Cr–Ni system.Sbornik Nauchnykh Rabot Instituta Metallofiziki Akademii Nauk Ukrainskoi SSR, 15:164–178, 1962 on Pith
  25. [45] W. A. Dench. Adiabatic high-temperature calorimeter for the measurement of heats of al- loying.Transactions of the Faraday Society, 59:1279– 1292, 1963. doi:10.1039/TF9635901279 on Pithdoi:10.1039/tf9635901279
  26. [46] M. J. Collins. Electron optic determination of solid phase boundaries in Ni–Cr system.Ma- terials Science and Technology, 4(6):560–561, 1988. doi:10.1179/mst.1988.4.6.560 on Pithdoi:10.1179/mst.1988.4.6.560
  27. [47] E. R. Jette, V . H. Nordstrom, B. Queneau, and F. Foote. X-ray studies on the nickel-chromium sys- tem.Transactions of the AIME, 111:361–373, 1934 on Pith
  28. [48] E. A. Guggenheim. Statistical thermodynamics of mixtures with zero energies of mixing.Proceedings of the Royal Society of London. Series A, 183:203–212, on Pith
  29. [49] Ryoichi Kikuchi. A theory of cooperative phenom- ena.Physical Review, 81:988–1003, 1951. doi:10. 1103/PhysRev.81.988 on Pith
  30. [50] Raymundo Arróyave. pair-entropy: Pair (quasi- chemical) estimator of the excess entropy of liq- uids from molecular dynamics. https://github. com/ArroyaveLab/pair-entropy, 2026. 24 on Pith
  31. [1944] doi:10.1098/rspa.1944.0032 on Pithdoi:10.1098/rspa.1944.0032
  32. [2019] doi:10.1088/2515-7639/ab13bb on Pithdoi:10.1088/2515-7639/ab13bb
  33. [2020] doi:10.1038/s41597-020-00638-4 on Pithdoi:10.1038/s41597-020-00638-4
  34. [2021] doi:10.12688/f1000research.29032.2 on Pithdoi:10.12688/f1000research.29032.2
  35. [2022] doi:10.3233/DS-210053 on Pithdoi:10.3233/ds-210053
  36. [2023] Accessed 11 November 2026 on Pith
  37. [2026] doi:10.1016/j.actamat.2026.122412 on Pithdoi:10.1016/j.actamat.2026.122412

Challenges and discussion

Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.

No comments yet.

Sign in to comment