Magnetic ground states of CrPS 4 and NiPS 3 monolayers from long-range exchange interactions

arXiv:2607.20631 · cond-mat.mtrl-sci · Submitted 2026-07-22 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: We're looking at the paper: "Magnetic ground states of CrPS 4 and NiPS 3 monolayers from long-range exchange interactions".

Mira: Here's the gist of the paper. We investigate the magnetic properties of monolayer CrPS4 and NiPS3 by combining firstprinciples calculations, second-principles spin models, and Monte Carlo simulations The resulting magnetic phase diagram contains multiple collinear and non-collinear phases that can be tuned by temperature and external magnetic fields The study employs a comprehensive approach to determine the magnetic exchange tensors directly from density functional theory using the LKAG formalism The researchers combine firstprinciples calculations, second-principles spin models, and Monte Carlo simulations to investigate the magnetic properties of monolayer CrPS4 and NiPS3.

Kai: They use siesta density functional theory (DFT) code with spin-orbit interaction included selfconsistently The exchange tensors are determined using the in-house arXiv:2607.20631v1 code, which implements the Liechtenstein–Katsnelson–Antropov–Gubanov (LKAG) formalism.

Mira: CrPS4 Magnetic Behavior In CrPS4, long-range exchange interactions destabilize the previously predicted ferromagnetic ground state and stabilize a spin-spiral phase This stabilization reduces the critical temperature to about 21 K, which is in agreement with available experiments.

Kai: The resulting magnetic phase diagram contains multiple collinear and non-collinear phases that can be tuned by temperature and external magnetic fields.

Paper summary: Kai: So, to recap what we just heard, this paper investigates the magnetic properties of monolayer CrPS4 and NiPS3 by combining first principles calculations, second-principles spin models, and Monte Carlo simulations. They are specifically looking at how including those long-range exchange interactions changes the magnetic behavior of both materials.

Mira: Exactly. The main argument they make is that when you include these long-range forces in their spin Hamiltonian, it qualitatively modifies the magnetic ground states for both systems.

Lev: And they aren't just looking at one material; they are testing this idea on CrPS4 and NiPS3 separately to see how the effect manifests in different systems.

Kai: In CrPS4, the long-range interactions destabilize what was thought to be a ferromagnetic ground state and instead stabilize a spin spiral phase.

Mira: And that spin spiral phase has a critical temperature reduced down to about twenty-one Kelvin, which aligns well with experimental data <ref:2607.20631#pg2>.

Lev: That's interesting because it shows that the long-range part of the physics is responsible for shifting the transition temperature in CrPS4 toward what we actually observe.

Kai: Then they look at NiPS3, where they find that the experimentally observed zigzag antiferromagnetic order only emerges when they include exchange interactions up to the fifth shell.

Mira: That requires those specific long-range interactions—the fifth-neighbor ones—to provide the necessary frustration to stabilize that zigzag order.

Lev: So, for NiPS3, it's not just about adding more shells; it's about needing that specific interaction strength to get the observed magnetic order.

Kai: The authors are essentially saying that the widely used three shell approximation is insufficient for describing these thiophosphate monolayers quantitatively.

Conclusion: Kai: So, looking at the title, "Magnetic ground states of CrPS4 and NiPS3 monolayers from long-range exchange interactions," it really highlights that these specific forces are key to understanding the magnetic reality here.

Mira: Right. The authors are essentially showing that they can extract the magnetic exchange tensors directly from density functional theory using the LKAG formalism, instead of just fitting things to total energies as conventional approaches do.

Lev: That direct extraction is a big methodological point because it means they aren't relying on approximations to get those interaction tensors, which is a strong thing when building models for real hardware.

Kai: And the implication for the field is that this work provides a predictive framework for accurately describing two dimensional van der Waals magnets in general.

Mira: It sets a standard by showing that reliable microscopic spin models for thiophosphate monolayers must include exchange interactions extending well beyond the third shell, which is what they found necessary.

Lev: So, if you're trying to build a simulator or an AI model for these materials, you can't just stop at three shells; you need to go longer to get the right physics.

Kai: It means that structural effects also play a role; for instance, in CrPS4, the structural dimerization is key because it stabilizes that spin spiral ground state.

Mira: And for NiPS3, that same structural distortion enhances the exchange frustration needed for its zigzag antiferromagnetic state.

Balıazs Nagyfalusi, Ritesh Das, Alvaro Bermejillo-Seco, Herre S. J. van der Zant, Yaroslav M. Blanter, Ámador Garc༺-Fuente, Á Jaime Ferrer

Departamento de Fnica, Universidad de Oviedo · Kavli Institute of Nanoscience, Delft University of Technology · Centro de Investigaci༺on en Nanomateriales y Nanotecnolog༺

cond-mat.mtrl-sci

Submitted: 2026-07-22

Updated: 2026-11-03

Comments: 11 pages, 12 figures

Code: https://github.com/magnopy/magnopy

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 83/100

The gist: We investigate the magnetic properties of monolayer CrPS4 and NiPS3 by combining firstprinciples calculations, second-principles spin models…

Key concepts

LKAG formalism
This is a method used to directly extract magnetic exchange tensors from density functional theory calculations. Instead of fitting energies, it allows researchers to obtain the actual magnetic coupling strengths between atoms by including interactions until numerical convergence is achieved.
Spin-spiral phase
In CrPS4, this is a non-collinear magnetic state where spins rotate continuously in space rather than aligning simply up or down. Long-range exchange interactions stabilize this state, reducing the critical temperature to about 21 K and matching experimental results.
Exchange interactions (Shells)
The paper emphasizes that conventional three-shell models are insufficient. To accurately describe these materials, long-range exchange interactions extending beyond the third shell are necessary. For NiPS3, including up to the fifth shell is needed to stabilize the observed zigzag magnetic order.

Terminology

Summary

We investigate the magnetic properties of monolayer CrPS4 and NiPS3 by combining firstprinciples calculations, second-principles spin models, and Monte Carlo simulations The resulting magnetic phase diagram contains multiple collinear and non-collinear phases that can be tuned by temperature and external magnetic fields

Methodology Overview

The study employs a comprehensive approach to determine the magnetic exchange tensors directly from density functional theory using the LKAG formalism The researchers combine firstprinciples calculations, second-principles spin models, and Monte Carlo simulations to investigate the magnetic properties of monolayer CrPS4 and NiPS3. They use siesta density functional theory (DFT) code with spin-orbit interaction included selfconsistently The exchange tensors are determined using the in-house arXiv:2607. p. 120631v1 code, which implements the Liechtenstein–Katsnelson–Antropov–Gubanov (LKAG) formalism. p. 1

CrPS4 Magnetic Behavior

In CrPS4, long-range exchange interactions destabilize the previously predicted ferromagnetic ground state and stabilize a spin-spiral phase This stabilization reduces the critical temperature to about 21 K, which is in agreement with available experiments. The resulting magnetic phase diagram contains multiple collinear and non-collinear phases that can be tuned by temperature and external magnetic fields. p. 0 The calculated spin-wave spectrum reproduces the main experimental features of the available neutron-scattering data. p. 8

NiPS3 Magnetic Behavior

In NiPS3, the experimentally observed zigzag antiferromagnetic order only emerges when exchange interactions up to the fifth shell are included. The fifth-neighbor exchange interaction provides the frustration required to stabilize the experimentally observed zigzag antiferromagnetic order. p. 8 The calculated spin-wave spectrum reproduces the main features of the available neutron-scattering data. p. 8 p. 8

Conclusion on Model Sufficiency

The results demonstrate that quantitatively predictive spin models for thiophosphate monolayers require long-range exchange interactions and provide a predictive framework for accurately describing two-dimensional van der Waals magnets The commonly adopted three-shell Heisenberg models are insufficient to describe transition-metal thiophosphate monolayers quantitatively. Reliable microscopic spin models require long-range exchange interactions obtained directly from first principles. p. 8

Role of Structural Effects

The structural dimerization plays a key role in stabilizing the spin-spiral ground state in CrPS4 As the dimerization is reduced, the wavelength of the spin spiral increases continuously. In NiPS3, this structural distortion enhances the exchange frustration responsible for the spin-spiral state. p. 3

Spin Dynamics and Field Response

The magnetic ground state of monolayer NiPS3 is remarkably insensitive to experimentally accessible magnetic fields The zigzag antiferromagnetic ground state remains robust throughout this range, with the critical temperature remaining essentially unchanged over the entire field range. The calculated spin-wave spectrum reproduces the main features of the available neutron-scattering data.

Final Synthesis

Taken together, these results demonstrate that reliable microscopic spin models require exchange interactions extending well beyond the third shell In CrPS4, long-range interactions destabilize the previously predicted ferromagnetic ground state in favor of a spin-spiral phase and reduce the critical temperature to approximately 21 K. In NiPS3, the fifth-neighbor exchange interaction provides the frustration required to stabilize the experimentally observed zigzag antiferromagnetic order. p. 8 The authors expect that the methodology presented here will be broadly applicable to the accurate description and prediction of magnetic phenomena in two-dimensional van der Waals materials. p. 8

--- Page 1 ---

Magnetic ground states of CrPS4 and NiPS3 monolayers from long-range exchange interactions Bal´azs Nagyfalusi, Ritesh Das, Alvaro Bermejillo-Seco, Herre S. J. van der Zant, Yaroslav M. p. 0 Blanter, Amador Garc´ıa-Fuente and Jaime Ferrer We investigate the magnetic properties of monolayer CrPS4 and NiPS3 by combining firstprinciples calculations, second-principles spin models, and Monte Carlo simulations Unlike conventional approaches that truncate exchange interactions after only a few shells and determine them by fitting total energies, we extract the magnetic exchange tensors directly from density functional theory using the LKAG formalism and include interactions until numerical convergence is achieved We show that long-range exchange interactions qualitatively modify the magnetic behavior of both materials In CrPS4, they destabilize the previously predicted ferromagnetic ground state and stabilize a spin-spiral phase, reducing the critical temperature to about 21 K, in agreement with available experiments The resulting magnetic phase diagram contains multiple collinear and non-collinear phases that can be tuned by temperature and external magnetic fields In NiPS3, the experimentally observed zigzag antiferromagnetic order only emerges when exchange interactions up to the fifth shell are included These results demonstrate that quantitatively predictive spin models for thiophosphate monolayers require long-range exchange interactions and provide a predictive framework for accurately describing two-dimensional van der Waals magnets

--- Page 2 ---

II. METHODS The electronic structure of the monolayers was calculated using the siesta density functional theory (DFT) code where the spin-orbit interaction was included selfconsistently In-house norm-conserving pseudopotentials were generated for Cr, Ni, P and S, where the lattice constant, magnetic moment and electronic bands were fitted to those obtained from the plane-wave, all-electrons code ELK A triple–ζ polarized (TZP) basis set was employed The exchange–correlation potential was treated within the generalized gradient approximation (GGA) using the Perdew-Burke-Ernzerhof (PBE) functional After careful convergence and accuracy testing, a 30 × 30 × 1 Monkhorst–Pack k-point grid and a real–space mesh cutoff of up to 4000 Ry were adopted to ensure a reliable self-consistent convergence The DFT Kohn–Sham Hamiltonian was mapped to the following classical Heisenberg model using the grogu code Here, Jij is the symmetric part of the full tensorial 3 × 3 exchange interaction, and Dij is the Dzyaloshinskii–Moriya (DM) interaction between magnetic sites i and j, whereas Ai is the single-ion magnetic anisotropy at site i The leading exchange interactions in both CrPS4 and NiPS3 are the isotropic couplings, here defined as Jij = 1/3 P α J αα ij, where α = x, y, z We also define the diagonal exchange anisotropy for a given ij pair as d αα ij = J αα ij − J zz ij for α = x, y, and for the whole system as d αα = Pij d αα ij To assess the convergence of the spin Hamiltonian with interaction range, exchange tensors were progressively truncated after increasing numbers of shells The magnetic ground state and thermodynamic properties were determined using Monte Carlo (MC) simulations An initial supercell grid of 64 × 64 unit cells was employed, with periodic boundary conditions applied in all directions The critical temperature was identified from the peak of the heat-capacity curve An external magnetic field B oriented along the z-axis was incorporated into the spin Hamiltonian in Eq. (1) via the Zeeman term: H = H0 − µ X i eiB, where µ is the magnetic moment obtained in the DFT calculation The Magnopy code was employed to verify the relative stability of the magnetic configurations obtained from Monte Carlo simulations and to compute the spinwave spectra of both materials

--- Page 3 ---

III. CRPS4 RESULTS CrPS4 crystallizes in a monoclinic lattice in which each Cr atom is coordinated by six S atoms forming a distorted octahedron, as shown in Fig. p. 2 1 The combined structural and magnetic optimizations were performed for Hubbard parameters in the range U = 0–2 eV Among these values, U = 0 provides the best agreement with the experimentally reported transition temperature of approximately 23 K for monolayer CrPS4 The optimized ground state is an out-of-plane ferromagnet with spins aligned along the z direction. p. 2 The lowest–energy zigzag antiferromagnetic configuration lies approximately 25 meV above the ground state, while all remaining magnetic configurations are separated by more than 100 meV The calculated electronic structure exhibits a direct band gap of 0. p. 272 eV at the Γ point, smaller than the ∼ 1. p. 24 eV direct gap reported experimentally for bulk CrPS4 The origin of the magnetic ground state can be understood from the calculated exchange interactions The first five shells are illustrated in Fig. p. 2 2 (a) Owing to the structural dimerization, the first-neighbor exchange interactions split into two inequivalent couplings, denoted by J1a and J1b The fourth–shell interaction, J4, is antiferromagnetic, in agreement with Ref. p. 2 [17], and its magnitude reaches approximately one quarter of the average first-neighbor coupling This competition frustrates the magnetic order and favors a continuous rotation of the spins instead of a collinear ferromagnetic arrangement The combined effect of these long-range interactions stabilizes a spin-spiral ground state propagating along the b direction Restricting the Hamiltonian to the first three shells yields a FM ground state with a critical temperature of approximately 39 K, in agreement with previous theoretical studies As additional exchange shells are incorporated, the FM state is progressively destabilized The magnetic ground state converges to a SS configuration once interactions up to the seventh shell are included This reduction reflects the increased magnetic frustration introduced by the competing long-range interactions, which weaken the stability of the ordered phase despite leaving the dominant exchange couplings essentially unchanged The resulting transition temperature of 21 K is in excellent agreement with the experimentally reported value for monolayer CrPS4

--- Page 4 ---

Fig. 3 (a) Critical temperature Tc of a CrPS4 monolayer as a function of the number of exchange shells Ns included in the spin Hamiltonian The transition from blue to red marks the change in the magnetic ground state upon extending the interaction range The yellow region corresponds to the spin-spiral (SS) phase (see Fig. p. 3 3 (b)) The green region denotes the canted spin-spiral (SS’) phase The red and blue regions correspond to the ferromagnetic (FM) and paramagnetic (PM) phases, respectively

--- Page 5 ---

FIG. 4 Magnetic phase diagram of a CrPS4 monolayer as a function of temperature and out-of-plane magnetic field The yellow region corresponds to the spin-spiral (SS) phase (see Fig. p. 4 3 (b)) The green region denotes the canted spin-spiral (SS’) phase The red and blue regions correspond to the ferromagnetic (FM) and paramagnetic (PM) phases, respectively

--- Page 6 ---

FIG. 5 Average out-of-plane magnetization Mz of a CrPS4 monolayer as a function of temperature for magnetic fields between 0 and 1. p. 45 T in steps of 0. p. 11 T In the absence of an external magnetic field, the average magnetization vanishes and the system remains in the SS phase below the paramagnetic transition Applying a finite magnetic field gives rise to three distinct regimes, which can be classified according to their zerotemperature magnetization For magnetic fields below 0. p. 445 T, the average magnetization remains zero and the system forms the SS phase discussed above, in which the spins rotate within the x–z plane Between 0. p. 445 T and 1.3 T, the magnetization becomes finite while remaining smaller than unity Inspection of the Monte Carlo snapshots reveals that the spiral continuously rotates toward the x–y plane while developing a finite homogeneous component along the z direction Finally, for magnetic fields exceeding 1. p. 43 T, the spiral is completely suppressed and the system undergoes a transition into a fully polarized ferromagnetic state

--- Page 7 ---

FIG. 7 Linear spin-wave spectrum of a CrPS4 monolayer The high-symmetry points of the Brillouin zone are C (1/2, 1/2), Γ (0, 0), and X (1/2, 0) Solid lines show the fully converged calculation (Ns = 14) The agreement between our calculated spectrum and experiment is remarkable The CrPS4 monolayer provides a clear example of the limitations of short-range spin models While models truncated after three shells predict a ferromagnetic ground state, including longer-range interactions stabilizes an exchange-driven spin spiral and reduces the critical temperature Furthermore, both the magnetic phase diagram and the magnon spectrum require the combined inclusion of structural dimerization and longrange exchange interactions

--- Page 8 ---

FIG. 8 Optimized crystal structure of a NiPS3 monolayer The magnetic ground state consists of ferromagnetic zigzag chains coupled antiferromagnetically (see Fig 10) To capture this symmetry breaking, we employ the enlarged rectangular unit cell shown in Fig. p. 5 8, which allows both the atomic positions and the magnetic structure to relax consistently The resulting energy splitting is extremely small: the two domains with propagation vectors at ±120◦ differ by less than 0. p. 62 µeV per Ni atom

--- Page 9 ---

FIG. 10 Symmetry-equivalent zigzag magnetic configurations in the ideal hexagonal lattice ((a)–(c)) and the corresponding ground states obtained after including the anisotropy induced by magnetostriction ((d)–(f)) The six nearly degenerate magnetic states—three propagation directions combined with two opposite spin orientations—therefore constitute an almost ideal realization of a weakly perturbed six-state clock model

--- Page 10 ---

The high-symmetry points of the Brillouin zone are C (1/2, 1/2), Γ (0, 0), X (1/2, 0), and Y (0, 1/2) The Hubbard parameter has a strong influence on the magnon bandwidth Calculations with U = 4 eV provide the best overall agreement with the experimental dispersion Some discrepancies remain, particularly in the magnon gaps, which are overestimated in our calculations These differences may originate from effects absent in the monolayer model, such as interlayer exchange interactions and magnetoelastic couplings

--- Page 11 ---

The resulting lattice constants are a = 5.91 Å and b = 10.24 Å The local magnetic moment is 1. p. 6 p. 613 µB per Ni atom

--- Page 8 ---

The resulting lattice constants are a = 5.91 Å and b = 10.24 Å The local magnetic moment is 1.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI system can predict magnetic ground states for two-dimensional van der Waals magnets by using first principles calculations combined with an LKAG formalism to extract exchange tensors directly from density functional theory. This allows the system to determine the magnetic exchange tensors directly from density functional theory using the LKAG formalism and include interactions until numerical convergence is achieved.

  2. Improved AI system can predict the critical temperature for monolayer CrPS4 by identifying that long-range interactions destabilize the previously predicted ferromagnetic ground state in favor of a spin-spiral phase and reduce the critical temperature to about 21 K, in agreement with available experiments.

  3. Improved AI system can predict the magnetic order for NiPS3 by recognizing that the fifth-neighbor exchange interaction provides the frustration required to stabilize the experimentally observed zigzag antiferromagnetic order.

  4. Improved AI system can construct a quantitative spin model for monolayer CrPS4 by using an interaction range extending well beyond the third shell to achieve accuracy, instead of relying on truncated models.

  5. Improved AI system can predict phase diagrams for CrPS4 by mapping the (B, T) phase diagram using Monte Carlo simulations, showing how the system transitions between ferromagnetic (FM), spin-spiral (SS), and canted spin-spiral (SS') phases.

Related papers