pith. sign in

arxiv: 2605.27888 · v1 · pith:XQTGXQS3new · submitted 2026-05-27 · ❄️ cond-mat.mtrl-sci

Machine-learning-accelerated discovery of synthesizable high-temperature altermagnets with giant spin splitting

Yi-Fei Jiang , Jia-Xuan Guo , Zhen Zhang , Xin-Wei Yi , Jing-Yang You This is my paper

Pith reviewed 2026-06-29 11:42 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci
keywords altermagnetsspin splittingmachine learninghigh-throughput screeningtetragonal compoundsantiferromagnetsspintronicsNéel temperature
0
0 comments X

The pith

Machine learning identifies 34 tetragonal compounds as altermagnets with spin splittings exceeding 1.5 eV.

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

The paper develops a machine-learning framework to search for altermagnets in the tetragonal AB2C2D family. It screens thousands of candidates and uses an XGBoost model to select those with large spin splitting and low formation energies. One example, RbMn2Te2O, shows 1.88 eV splitting and a Néel temperature around 390 K. This matters because altermagnets could enable spintronic devices without stray magnetic fields. The work shows these materials can be stable at high temperatures and tunable by pressure or structure changes.

Core claim

Screening 8640 AB2C2D variants yields 1347 altermagnetic candidates, from which an XGBoost model trained on first-principles data selects 34 low-hull-energy compounds with giant non-relativistic spin splittings over 1.5 eV. Calculations on RbMn2Te2O confirm 1.88 eV maximum splitting, dynamical stability, and Néel temperature of 390 K, arising from symmetry-locked Mn exchange fields amplified by Mn-d/Te-p hybridization. The altermagnetic splitting persists through a soft-mode structural transition in SrMn2Te2O and can be modulated by hydrostatic pressure.

What carries the argument

The interpretable XGBoost model trained on first-principles spin-splitting data, which ranks candidates by hull energy to isolate synthesizable high-splitting altermagnets.

Load-bearing premise

The XGBoost model trained on first-principles spin-splitting data generalizes accurately to the screened compounds and correctly ranks them by hull energy without significant false positives from overfitting or incomplete training data.

What would settle it

Experimental synthesis of RbMn2Te2O followed by ARPES or transport measurement showing maximum spin splitting below 1 eV or Néel temperature far below 390 K would falsify the central discovery.

read the original abstract

Altermagnets offer a route to spin-polarized electronic states without macroscopic magnetization, because compensated magnetic order can generate momentum-dependent spin splitting through crystal-symmetry-controlled exchange fields. However, experimentally viable altermagnets combining large spin splitting, thermodynamic stability and high magnetic ordering temperatures remain scarce. Here, we develop a machine-learning-accelerated high-throughput framework to explore the tetragonal AB$_2$C$_2$D compounds. Screening 8640 variants identifies 1347 compensated antiferromagnetic candidates satisfying altermagnetic symmetry. An interpretable XGBoost model trained on first-principles spin-splitting data then isolates 34 low-hull-energy candidates,including four previously reported, with giant non-relativistic spin splittings exceeding 1.5 eV near the Fermi level. Detailed first-principles calculations of the representative RbMn$_2$Te$_2$O confirm a maximum spin splitting of $\sim$1.88 eV with dynamical stability and an estimated N\'eel temperature of $\sim$390 K. The giant splitting originates from symmetry-locked Mn-sublattice exchange fields amplified by directional Mn-d/Te-p hybridization. Furthermore, we uncover a profound soft-mode-driven structural transition associated with an interlayer dimensionality crossover in SrMn$_2$Te$_2$O, yet the unfolded electronic structure demonstrates that the altermagnetic spin splitting remains robust after lattice reconstruction. Hydrostatic pressure provides an additional tuning route, producing non-monotonic modulation of the spin-split Fermi surface governed by local coordination and orbital hybridization. These results establish tetragonal AB$_2$C$_2$D compounds as a tunable materials platform for stray-field-free spintronic devices and provide a general data-driven strategy for discovering robust giant-splitting altermagnets.

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.

Referee Report

3 major / 2 minor

Summary. The manuscript develops a machine-learning-accelerated high-throughput workflow to discover altermagnets in the tetragonal AB₂C₂D family. Screening 8640 variants yields 1347 compensated antiferromagnetic candidates satisfying altermagnetic symmetry; an interpretable XGBoost model trained on first-principles spin-splitting data then isolates 34 low-hull-energy compounds (including four previously reported) with giant non-relativistic spin splittings >1.5 eV near the Fermi level. Detailed DFT calculations on the representative RbMn₂Te₂O confirm a maximum spin splitting of ∼1.88 eV, dynamical stability, and an estimated Néel temperature of ∼390 K. The work further examines the microscopic origin of the splitting, a soft-mode-driven structural transition in SrMn₂Te₂O that preserves the altermagnetic splitting, and hydrostatic-pressure tuning of the spin-split Fermi surface.

Significance. If the central ML predictions prove robust, the work would constitute a meaningful advance by establishing a scalable, data-driven route to high-temperature, synthesizable altermagnets with large spin splittings suitable for stray-field-free spintronics. The explicit identification of RbMn₂Te₂O as a dynamically stable candidate with TN∼390 K and the demonstration that the splitting survives a dimensionality-crossover transition are concrete, testable outputs that could guide experimental synthesis efforts.

major comments (3)
  1. [Abstract and ML pipeline description] The XGBoost model is the load-bearing step that reduces 1347 altermagnetic candidates to the final list of 34 low-hull-energy compounds; however, the manuscript reports neither cross-validation scores, feature-ablation results, nor performance on any held-out set of known altermagnets. Without these metrics it is impossible to quantify generalization error or the risk of false positives arising from overfitting or incomplete chemical-space coverage.
  2. [Results section on RbMn₂Te₂O] The headline numerical results for RbMn₂Te₂O (maximum spin splitting ∼1.88 eV and Néel temperature ∼390 K) are presented without accompanying error bars, convergence tests, or uncertainty estimates from the underlying first-principles calculations, which weakens the claim that these values are quantitatively reliable.
  3. [Candidate selection and hull-energy discussion] The claim that the 34 candidates are “synthesizable” rests on low hull energies, yet no explicit comparison is made to the hull energies of the four previously reported compounds or to any experimental synthesis thresholds, leaving the practical viability of the new candidates unquantified.
minor comments (2)
  1. [Abstract] The abstract states that the XGBoost model is “interpretable” but does not list the most important features or provide SHAP-value analysis in the main text; adding this would strengthen the methodological transparency.
  2. [Results] Notation for the chemical formula (AB₂C₂D) is used consistently, but the manuscript would benefit from an explicit table listing the 34 final candidates with their predicted and (where available) computed spin-splitting and hull-energy values.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major point below and will revise the manuscript to incorporate additional validation, quantitative details, and comparisons as suggested.

read point-by-point responses

  1. Referee: [Abstract and ML pipeline description] The XGBoost model is the load-bearing step that reduces 1347 altermagnetic candidates to the final list of 34 low-hull-energy compounds; however, the manuscript reports neither cross-validation scores, feature-ablation results, nor performance on any held-out set of known altermagnets. Without these metrics it is impossible to quantify generalization error or the risk of false positives arising from overfitting or incomplete chemical-space coverage.

    Authors: We agree that explicit validation metrics are needed to assess model reliability. In the revised manuscript we will report 5-fold cross-validation scores, feature-ablation results, and performance metrics on a held-out set of known altermagnets. These additions will quantify generalization error and address concerns about overfitting or chemical-space coverage. revision: yes

  2. Referee: [Results section on RbMn₂Te₂O] The headline numerical results for RbMn₂Te₂O (maximum spin splitting ∼1.88 eV and Néel temperature ∼390 K) are presented without accompanying error bars, convergence tests, or uncertainty estimates from the underlying first-principles calculations, which weakens the claim that these values are quantitatively reliable.

    Authors: We acknowledge the value of convergence tests and uncertainty estimates. The revised manuscript will include k-point and energy-cutoff convergence data, together with estimated uncertainties for the reported spin-splitting and Néel-temperature values obtained from the DFT calculations. revision: yes

  3. Referee: [Candidate selection and hull-energy discussion] The claim that the 34 candidates are “synthesizable” rests on low hull energies, yet no explicit comparison is made to the hull energies of the four previously reported compounds or to any experimental synthesis thresholds, leaving the practical viability of the new candidates unquantified.

    Authors: We will add an explicit comparison of hull energies for the 34 candidates versus the four previously reported compounds, and will reference typical experimental synthesis thresholds from the literature to better quantify practical viability. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's central claims rest on direct first-principles DFT calculations for the representative compound RbMn₂Te₂O, which independently yield the reported ~1.88 eV spin splitting, dynamical stability, and Néel temperature estimate. The XGBoost model, trained on separate first-principles spin-splitting data, functions only as a screening filter to identify candidates from the 8640 variants; its outputs are not used to define or substitute for the verified physical quantities. No equations reduce by construction to fitted inputs, no self-citations provide load-bearing uniqueness theorems, and no ansatzes or renamings are smuggled in. The derivation chain is therefore self-contained against external DFT benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities are stated. The workflow implicitly assumes standard DFT accuracy and ML generalization, but these are not itemized.

pith-pipeline@v0.9.1-grok · 5865 in / 1272 out tokens · 36690 ms · 2026-06-29T11:42:49.241715+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

52 extracted references · 3 canonical work pages · 2 internal anchors

  1. [1]

    Wolf, S. A. et al. Spintronics: A Spin-Based Electronics Vision for the Future. Science 294, 1488– 1495 (2001)

    2001

  2. [2]

    & Das Sarma, S

    Žutić, I., Fabian, J. & Das Sarma, S. Spintronics: Fundamentals and applications. Rev. Mod. Phys. 76, 323–410 (2004)

    2004

  3. [3]

    Baltz, V . et al. Antiferromagnetic spintronics. Rev. Mod. Phys. 90, 015005 (2018)

    2018

  4. [4]

    & Wunderlich, J

    Jungwirth, T., Marti, X., Wadley, P. & Wunderlich, J. Antiferromagnetic spintronics. Nat. Nanotechnol. 11, 231–241 (2016)

    2016

  5. [5]

    Guo, Z. et al. Spin‐Polarized Antiferromagnets for Spintronics. Adv. Mater. 37, 2505779 (2025)

    2025

  6. [6]

    & Jungwirth, T

    Šmejkal, L., Sinova, J. & Jungwirth, T. Beyond Conventional Ferromagnetism and Antiferromagnetism: A Phase with Nonrelativistic Spin and Crystal Rotation Symmetry. Phys. Rev. X 12, 031042 (2022)

    2022

  7. [7]

    & Jungwirth, T

    Šmejkal, L., Sinova, J. & Jungwirth, T. Emerging Research Landscape of Altermagnetism. Phys. Rev. X 12, 040501 (2022)

    2022

  8. [8]

    Krempaský, J. et al. Altermagnetic lifting of Kramers spin degeneracy. Nature 626, 517–522 (2024)

    2024

  9. [9]

    González-Hernández, R. et al. Efficient Electrical Spin Splitter Based on Nonrelativistic Collinear Antiferromagnetism. Phys. Rev. Lett. 126, 127701 (2021)

    2021

  10. [10]

    & Liu, Z

    Yan, H., Zhou, X., Qin, P. & Liu, Z. Review on spin-split antiferromagnetic spintronics. Appl. Phys. Lett. 124, 030503 (2024)

    2024

  11. [11]

    Bai, L. et al. Altermagnetism: Exploring New Frontiers in Magnetism and Spintronics. Adv. Funct. Mater. 34, 2409327 (2024)

    2024

  12. [12]

    Song, C. et al. Altermagnets as a new class of functional materials. Nat. Rev. Mater. 10, 473–485 (2025)

    2025

  13. [13]

    B., González-Hernández, R., Sinova, J

    Šmejkal, L., Hellenes, A. B., González-Hernández, R., Sinova, J. & Jungwirth, T. Giant and Tunneling Magnetoresistance in Unconventional Collinear Antiferromagnets with Nonrelativistic Spin-Momentum Coupling. Phys. Rev. X 12, 011028 (2022)

    2022

  14. [14]

    Superexchange interaction and symmetry properties of electron orbitals

    Kanamori, J. Superexchange interaction and symmetry properties of electron orbitals. J. Phys. Chem. Solids 10, 87–98 (1959)

    1959

  15. [15]

    Jiang, B. et al. A metallic room-temperature d-wave altermagnet. Nat. Phys. 21, 754–759 (2025)

    2025

  16. [16]

    Zhang, F. et al. Crystal-symmetry-paired spin–valley locking in a layered room-temperature metallic altermagnet candidate. Nat. Phys. 21, 760–767 (2025)

    2025

  17. [17]

    Jain, A. et al. A high-throughput infrastructure for density functional theory calculations. Comput. Mater. Sci. 50, 2295–2310 (2011)

    2011

  18. [18]

    T., Davies, D

    Butler, K. T., Davies, D. W., Cartwright, H., Isayev, O. & Walsh, A. Machine learning for molecular and materials science. Nature 559, 547–555 (2018)

    2018

  19. [19]

    Abduljabbar, Z., A

    Arif Ali, Z., H. Abduljabbar, Z., A. Tahir, H., Bibo Sallow, A. & Almufti, S. M. eXtreme Gradient Boosting Algorithm with Machine Learning: a Review. Acad. J. Nawroz Univ. 12, 320–334 (2023)

    2023

  20. [20]

    & Simon, R

    Varma, S. & Simon, R. Bias in error estimation when using cross-validation for model selection. BMC Bioinformatics 7, 91 (2006)

    2006

  21. [21]

    A Unified Approach to Interpreting Model Predictions

    Lundberg, S. & Lee, S.-I. A Unified Approach to Interpreting Model Predictions. Preprint at https://doi.org/10.48550/arXiv.1705.07874 (2017)

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1705.07874 2017

  22. [22]

    Lundberg, S. M. et al. From local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell. 2, 56–67 (2020)

    2020

  23. [23]

    Allen, L. C. Electronegativity is the average one-electron energy of the valence-shell electrons in ground-state free atoms. J. Am. Chem. Soc. 111, 9003–9014 (1989)

    1989

  24. [24]

    Liu, C.-C. et al. Physical properties and first-principles calculations of an altermagnet candidate Cs1 − δV2Te2O. Phys. Rev. B 112, 224439 (2025)

    2025

  25. [25]

    Lin, H. et al. Structure and physical properties of CsV2Se2−xO and V2Se2O. Phys. Rev. B 98, 075132 (2018)

    2018

  26. [28]

    Trip-bench: A benchmark for long-horizon interactive agents in real-world scenarios.CoRR, abs/2602.01675, 2026

    Cheng, X., Gao, Y . & Liu, J. P. J. Realistic tight-binding model for V2Se2O-family altermagnets. Preprint at https://doi.org/10.48550/ARXIV .2602.09465 (2026)

    work page internal anchor Pith review doi:10.48550/arxiv 2026

  27. [29]

    S., Arita, R

    Bahramy, M. S., Arita, R. & Nagaosa, N. Origin of giant bulk Rashba splitting: Application to BiTeI. Phys. Rev. B 84, 041202 (2011)

    2011

  28. [30]

    W., Rosenbluth, M

    Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H. & Teller, E. Equation of State Calculations by Fast Computing Machines. J. Chem. Phys. 21, 1087–1092 (1953)

    1953

  29. [31]

    Evans, R. F. L. et al. Atomistic spin model simulations of magnetic nanomaterials. J. Phys. Condens. Matter 26, 103202 (2014)

    2014

  30. [32]

    The dynamics of charge-density waves

    Grüner, G. The dynamics of charge-density waves. Rev. Mod. Phys. 60, 1129–1181 (1988)

    1988

  31. [33]

    Johannes, M. D. & Mazin, I. I. Fermi surface nesting and the origin of charge density waves in metals. Phys. Rev. B 77, 165135 (2008)

    2008

  32. [34]

    Zhu, X., Cao, Y ., Zhang, J., Plummer, E. W. & Guo, J. Classification of charge density waves based on their nature. Proc. Natl. Acad. Sci. 112, 2367–2371 (2015)

    2015

  33. [35]

    O’Brien, M. C. M. & Chancey, C. C. The Jahn–Teller effect: An introduction and current review. Am. J. Phys. 61, 688–697 (1993)

    1993

  34. [36]

    & Zunger, A

    Popescu, V . & Zunger, A. Extracting E versus k ⃗ effective band structure from supercell calculations on alloys and impurities. Phys. Rev. B 85, 085201 (2012)

    2012

  35. [37]

    & Geng, W.-T

    Wang, V ., Xu, N., Liu, J.-C., Tang, G. & Geng, W.-T. V ASPKIT: A user-friendly interface facilitating high-throughput computing and analysis using V ASP code. Comput. Phys. Commun. 267, 108033 (2021)

    2021

  36. [38]

    Tunneling between ferromagnetic films

    Julliere, M. Tunneling between ferromagnetic films. Phys. Lett. A 54, 225–226 (1975)

    1975

  37. [39]

    & Furthmüller, J

    Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186 (1996)

    1996

  38. [40]

    Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953–17979 (1994)

    1994

  39. [41]

    P., Burke, K

    Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 77, 3865–3868 (1996)

    1996

  40. [42]

    Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations. Phys. Rev. B 13, 5188–5192 (1976)

    1976

  41. [43]

    Horton, M. K. et al. Accelerated data-driven materials science with the Materials Project. Nat. Mater. 24, 1522–1532 (2025)

    2025

  42. [44]

    Jain, A. et al. Commentary: The Materials Project: A materials genome approach to accelerating materials innovation. APL Mater. 1, 011002 (2013)

    2013

  43. [45]

    L., Botton, G

    Dudarev, S. L., Botton, G. A., Savrasov, S. Y ., Humphreys, C. J. & Sutton, A. P. Electron-energy- loss spectra and the structural stability of nickel oxide: An LSDA+U study. Phys. Rev. B 57, 1505– 1509 (1998)

    1998

  44. [46]

    Jain, A. et al. Formation enthalpies by mixing GGA and GGA + U calculations. Phys. Rev. B 84, 045115 (2011)

    2011

  45. [47]

    Ong, S. P. et al. Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis. Comput. Mater. Sci. 68, 314–319 (2013)

    2013

  46. [48]

    Parlinski, K., Li, Z. Q. & Kawazoe, Y . First-Principles Determination of the Soft Mode in Cubic ZrO 2. Phys. Rev. Lett. 78, 4063–4066 (1997)

    1997

  47. [49]

    & Tanaka, I

    Togo, A., Chaput, L., Tadano, T. & Tanaka, I. Implementation strategies in phonopy and phono3py. J. Phys. Condens. Matter 35, 353001 (2023)

    2023

  48. [50]

    First-principles Phonon Calculations with Phonopy and Phono3py

    Togo, A. First-principles Phonon Calculations with Phonopy and Phono3py. J. Phys. Soc. Jpn. 92, 012001 (2023)

    2023

  49. [51]

    Ward, L. et al. Matminer: An open source toolkit for materials data mining. Comput. Mater. Sci. 152, 60–69 (2018)

    2018

  50. [52]

    Ketkaew, R. et al. OctaDist: a tool for calculating distortion parameters in spin crossover and coordination complexes. Dalton Trans. 50, 1086–1096 (2021)

    2021

  51. [53]

    Pedregosa, F. et al. Scikit-learn: Machine Learning in Python. Mach. Learn. PYTHON

  52. [54]

    & Koyama, M

    Akiba, T., Sano, S., Yanase, T., Ohta, T. & Koyama, M. Optuna: A Next-generation Hyperparameter Optimization Framework. in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining 2623–2631 (ACM, Anchorage AK USA, 2019). doi:10.1145/3292500.3330701. Fig. 1 | Machine -learning-accelerated high -throughput discovery o...

    work page doi:10.1145/3292500.3330701 2019