Disorder-Mediated Magnetism in Transition-Metal Dichalcogenide Monolayers: XS $_2$ and XSe $_2$ (X = Ti, V, Co, Mo, and Ru)

Chi-Cheng Lee, Yukiko Yamada-Takamura, Taisuke Ozaki

Oct 21, 2026arXiv:2610.29235v1

cond-mat.mtrl-sci

Scorecard· 16/16

4.5/10 impact

Competent, clearly framed DFT study offering a plausible but speculative explanation for the VSe2 FM puzzle; limited by GGA-only treatment, sub-meV energy claims, and no experimental link.

Abstract

Magnetism in transition-metal dichalcogenide monolayers has attracted considerable attention, yet its nature remains elusive, particularly in monolayer 1T-VSe $_{2}$ . Our first-principles calculations show that Stoner instability can drive ferromagnetism in XS $_{2}$ and XSe $_{2}$ monolayers (X = Ti, V, Co, Mo, and Ru), associated with a reduced density of states at the Fermi level. We demonstrate that mixing nonmagnetic Ti, Co, Mo, and Ru disulfides or diselenides can induce magnetism. Conversely, mixing magnetic VS $_{2}$ or VSe $_{2}$ with other compounds can completely suppress magnetism. Both scenarios can lower the total energy. Moreover, substituting a small number of V atoms with Co defects can also disrupt ferromagnetism in 1T-V(S,Se) $_{2}$ . Even in pristine 1T-VSe $_{2}$ , we find ferrimagnetic configurations with reversed local V moments that are nearly degenerate with the ferromagnetic state. Their competition may contribute to the suppression of long-range ferromagnetic order. This reduced energy penalty for magnetic and structural disorder arises from a suppressed density of states at the Fermi level, resembling the mechanism underlying the Peierls transition. Electronically driven energy lowering near the Fermi level may therefore be widespread among transition-metal dichalcogenide monolayers in various forms.

AI Impact Assessments

(1 model)

Core Contribution

This is a DFT-GGA study (OpenMX) of magnetism in 1H and 1T monolayers of XS2/XSe2 (X = Ti, V, Co, Mo, Ru). Its central claims are:

Ferromagnetism in these compounds follows a Stoner picture. Magnetism appears when chalcogen height tunes a large unpolarized DOS peak at E_F.

Mixing nonmagnetic compounds (Ti/Co/Mo/Ru) can induce magnetism, with the Stoner criterion met locally.

Mixing magnetic VS2/VSe2 with other compounds, or adding about 11% Co impurities, can suppress magnetism. Mixing energies are negative.

Pristine 1T-VSe2 has dynamically stable ferrimagnetic configurations (6×6 and 3√3×√7 supercells) that lie within 0.2 meV/atom of the FM state. The authors propose that this competition suppresses long-range FM order, which could help explain the experimental absence of FM despite the GGA prediction.

The reduced DOS at E_F in these states is likened to a Peierls mechanism, and the authors suggest that electronically driven energy lowering near E_F is widespread in TMD monolayers.

The framing of "disorder-mediated magnetism" is conceptually interesting. Disorder lowers energy by reducing the DOS at E_F, and magnetic and chemical disorder are treated on the same footing as CDW formation. It extends the authors' earlier work on mixing to stabilize CDWs (PRB 2025).

Methodological Rigor

The approach is standard spin-polarized GGA with no Hubbard U or hybrid functional check, and no spin-orbit coupling. The authors themselves note that the 1T-VSe2 prediction is functional-dependent. Results for a system whose magnetism is known to be sensitive to correlation and functional are therefore only partially probative.

Specific concerns:

The key claim is a 0.2 meV/atom energy difference between FM and ferrimagnetic states. This is below the typical accuracy of the method and of the basis set (the authors admit 0.1 meV/atom is within their uncertainty in another case). Near-degeneracy is plausible, but it is not a sharp prediction.

Only two ferrimagnetic cells were tried. There is no exchange-coupling (Heisenberg) mapping, no Monte Carlo, and no finite-temperature analysis. The step from "near-degenerate states" to "suppression of long-range order" is speculative, and in 2D the Mermin-Wagner theorem and magnetic anisotropy matter.

The "Peierls-like" link relies on qualitative DOS comparisons. There is no quantitative Stoner parameter analysis, no susceptibility calculation, and no analysis of the Fermi-surface nesting that would make the Peierls analogy concrete.

The mixed-compound results rely on one or a few special quasi-random-style configurations (4-component 6×6 cells), with no configurational sampling or statistics. Negative mixing energies are reported relative to unmixed compounds in the same structure. The "relative energies" against the lower-energy structure are mostly positive for the 1H mixed phases, so the stability claim is more modest than the abstract implies.

Mixed-phase CDW competition and magnetism are only partly separated. The reference energies are the relaxed (6×6) CDW structures, whose configuration space is not exhaustively searched. Phonon stability checks are a nice addition.

The claim that 1T-VSe2 is more readily observed because impurities favor it is explicitly labeled speculation.

No comparison with experimental magnetization or XMCD is made beyond a qualitative narrative.

Potential Impact

The paper offers a plausible, qualitative mechanism for the long-standing discrepancy between DFT (FM) and experiments (no long-range FM) in monolayer VSe2. This puzzle is a well-known topic, so the paper will be cited by the VSe2 and 2D magnet communities. The suggestion of magnetic and compositional design, with alloys that switch magnetism on or off, is of interest for 2D spintronics, but it remains purely computational and has no experimental pathway. The broader "electronic energy lowering near E_F" idea is suggestive but vague and unlikely to change practice. Overall impact is likely moderate and confined to computational 2D magnetism.

Timeliness

2D magnets and TMD alloys are active areas, and the VSe2 controversy remains unresolved. The study is timely, but it is a short, incremental contribution in a crowded literature. Many prior works have already discussed near-degenerate magnetic states and CDW-magnetism competition in VSe2 (Yin et al. 2022, Esters et al. 2017, Fumega et al. 2019).

Strengths

It treats a systematic set of ten compounds in both phases.

It links chalcogen height, DOS at E_F, and Stoner instability transparently.

It includes phonon-stability checks and large supercells.

It offers an interesting inverse viewpoint in which disorder lowers energy.

It is concise and logically organized.

Limitations

No beyond-GGA validation (DFT+U, hybrid functionals) for a system where correlation is known to matter.

Energy differences are at or below numerical uncertainty.

Sparse configurational sampling, with no statistics.

No Heisenberg or Monte Carlo analysis connecting near-degeneracy to loss of order.

The Peierls analogy is qualitative.

No experimental comparison or testable prediction, such as specific alloy compositions for synthesis or XMCD signatures.

Supplemental details are needed for reproducibility, and code is not released, though OpenMX is public.

The generalization claim ("widespread among TMD monolayers") goes beyond the evidence.

Other Observations

Reproducibility is moderate. OpenMX is open, and the structures and settings are described in the supplement, but the specific mixed configurations are not fully specified in the main text. Resource demands are modest, since 6×6 supercell DFT is feasible for a small group with a cluster. The paper is a solid but incremental computational contribution. It is likely to gain modest citations from the VSe2 and 2D alloy communities.

Rating:4.5/ 10

Significance 4.5Rigor 4Novelty 5Clarity 6.5

Generated Oct 22, 2026

Comparison History (0)

No comparisons yet.