{"id":"8720f602-a44e-48cd-b469-6f6f89f3b65f","arxiv_id":"2607.20631","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"In monolayers of CrPS4 and NiPS3, exchange interactions beyond the third neighbor shell qualitatively change the magnetic ground state and critical temperature, matching experiments.","lead":"For two atomically thin magnetic crystals, long-range magnetic couplings change the predicted ground state: CrPS4 becomes a spin spiral instead of a ferromagnet, and NiPS3's zigzag order appears only when distant neighbor couplings are kept. The work warns that standard short-range spin models can produce the wrong magnetic state for 2D magnets.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CrPS4 spin-spiral claim rests on exchange tensors from a ferromagnetically relaxed lattice; if the spiral's equilibrium dimerization differs, the central result may not survive.","rationale":"The reader's verdict is already CONDITIONAL, and the concern I identify is the same one the reader highlighted: the CrPS4 spiral is computed on a ferromagnetically relaxed lattice. This is a genuine internal inconsistency, not merely a disagreement with consensus. However, it is testable, and the paper is otherwise careful and transparent—it provides shell-by-shell convergence, Monte Carlo parameters, and acknowledges the limitation. No equation-level circularity was found. The NiPS3 result (zigzag requires J5) is not affected by the FM-relaxation concern to the same degree because the structural relaxation there was performed in the zigzag cell that can accommodate the predicted order. Thus the overall verdict should remain CONDITIONAL pending the self-consistent spiral relaxation test. I agree with the reader's weakest_assumption.","tokens_in":14810,"tokens_out":4240,"duration_ms":37356,"concrete_test":"Perform a full structural relaxation of monolayer CrPS4 in a supercell of at least 7 unit cells along b (accommodating the predicted q≈0.124 Å^-1 spiral) with spin-orbit coupling and starting from the spiral order; recompute LKAG J_N and Monte Carlo Tc. If the optimized structure retains the FM-like dimerization (d12−d34 ≈ 0.15 Å) and the spiral persists with q≈0.124 Å^-1 and Tc≈21 K, the central claim survives. If the dimerization is reduced and the spiral unwinds or q shifts significantly, the FM-relaxed exchange tensors are not representative of the true ground state.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing vulnerability is in Sec. III: the CrPS4 exchange tensors are extracted from a DFT calculation whose structural relaxation was restricted to a four-Cr ferromagnetic cell. The authors explicitly state this cell 'cannot accommodate the long-wavelength spin-spiral state identified below.' Consequently, the lattice parameters and the Cr–Cr dimerization (d12=3.75 Å vs d34=3.60 Å) are FM-state values, and the LKAG J_N are computed in that FM reference. The model built from these J_N then predicts a spin spiral with q=0.124 Å^-1 and Tc≈21 K. The inconsistency is that the reference magnetic state used to generate the Hamiltonian is not the ground state of that Hamiltonian. This matters quantitatively because the authors' own interpolation (Eq. 3) shows the spiral is stabilized by the dimerization: reducing the J1a/J1b splitting causes the spiral to unwind into FM domains. If the true spiral state has a different magnetostrictive equilibrium, then J1a, J1b, J4 and the higher-shell couplings all change; q and Tc could move, and the spiral could disappear. The authors acknowledge the limitation but do not self-consistently relax the spiral, so the central spin-spiral claim is not yet established on a lattice consistent with its own ground state. A secondary concern is that U=0 was selected for CrPS4 to match the experimental Tc, making the 'agreement' partly a fitting outcome; robustness of the spiral to U∈{1,2} eV is not reported. The structural issue is the more fundamental of the two.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates monolayer CrPS4 and NiPS3 using DFT with the LKAG formalism to extract tensorial exchange interactions, then uses Monte Carlo to determine ground states and critical temperatures as a function of the number of exchange shells. For CrPS4, the authors claim that including long-range exchange (beyond third shell) destabilizes the previously predicted ferromagnetic state and stabilizes a spin spiral, reducing Tc to ~21 K, in agreement with experiments. For NiPS3, they claim that the experimentally observed zigzag antiferromagnetic order only emerges when fifth-shell interactions are included. The central methodological claim is that conventional three-shell Heisenberg models fitted to total energies are insufficient for these thiophosphate monolayers, and that LKAG-derived exchange tensors with convergence in the number of shells are required.","tokens_in":15165,"tokens_out":2432,"duration_ms":22541,"significance":"If correct, the results would establish that short-range Heisenberg models are qualitatively inadequate for two important 2D magnetic semiconductors, providing a broader lesson for first-principles spin-model construction. The use of LKAG without total-energy fitting, systematic shell-number convergence checks, and Monte Carlo with demonstrated finite-size control are explicit strengths. The NiPS3 result—fifth-shell exchange providing the frustration for zigzag order—appears robust and is a valuable contribution. The CrPS4 spin-spiral claim, however, rests on an acknowledged but unresolved structural inconsistency: the exchange tensors are extracted from a ferromagnetically relaxed unit cell that cannot accommodate the spiral, and the Hubbard U is chosen to match the very Tc later quoted as agreement. These issues do not invalidate the methodology but leave the most dramatic claim not yet established.","major_comments":[{"comment":"The CrPS4 exchange tensors are obtained from a DFT calculation whose structural relaxation was restricted to a four-Cr ferromagnetic unit cell, which the authors state 'cannot accommodate the long-wavelength spin-spiral state identified below.' Since the spiral is stabilized by the dimerization (reducing J1a/J1b splitting unwinds the spiral, per Eq. (3)), the FM-relaxed lattice may not be representative of the spiral ground state. If the true spiral has a different magnetostrictive distortion, J1a, J1b, J4, and higher shells all change, potentially altering q and Tc or destroying the spiral. The authors acknowledge the limitation but do not perform a self-consistent relaxation of the spiral state or a robustness check with respect to the lattice. This is load-bearing for the central claim; I request either a spiral-state relaxation in a sufficiently large cell or a controlled test showin","section":"Sec. III, Eq. (3)"},{"comment":"The Hubbard U for CrPS4 was selected as U=0 because it reproduces the experimental Tc~23 K, and then the computed Tc≈21 K is quoted as 'excellent agreement.' This is partially circular: the agreement is built into the parameter choice. The paper does not report how the exchange interactions, spiral wavevector, or Tc vary with U∈{1,2} eV, even though the authors state they tested U=0–2 eV. A robustness scan over U (or a U determined from an independent observable such as the magnon spectrum or band gap) is needed to support the claim that the long-range-exchange mechanism is not an artifact of the U selection.","section":"Sec. III, 'U=0 provides the best agreement...'"},{"comment":"The transition from FM to spin-spiral is stated to occur when interactions up to the seventh shell are included, and Tc converges only at the tenth shell. However, the figure appears to show the ground-state switch at N_s=7 with Tc still decreasing substantially beyond that. Since the central claim involves both the ground-state symmetry and the quantitative Tc, the convergence criterion for 'numerical convergence' should be specified explicitly (e.g., change in Tc below a threshold) rather than inferred from the figure. This is a presentation issue, but it bears on the reproducibility of the claimed convergence.","section":"Sec. III, Fig. 3(a)"}],"minor_comments":[{"comment":"The definition of J_ij as the isotropic coupling (1/3 trace) should be stated more prominently; currently it appears only in text after the Hamiltonian. Also, the single-ion anisotropy tensor A_i is not defined in the equation explicitly (it is clear from context).","section":"Sec. II, Eq. (1)"},{"comment":"The phase diagram labels SS and SS' but the reader must infer the field/temperature ranges from the text. Adding dashed lines or annotations to mark the 0.45 T and 1.3 T boundaries would improve readability.","section":"Sec. III, Fig. 4"},{"comment":"The Tc values for N_s=3 and N_s=5 are close (68 K vs 61 K), and the figure error bars are not defined. Please indicate whether these differences are within Monte Carlo uncertainty, especially since 'the fifth-neighbor interaction stabilizes the zigzag phase' but Tc changes modestly.","section":"Sec. IV, Fig. 11"},{"comment":"The phrase 'parameter-free exchange tensors' in Sec. III is overstated: the LKAG extraction avoids fitting total energies, but the results still depend on the choice of XC functional and Hubbard U. Suggest rephrasing to 'fit-free' or 'not fitted to total energies.'","section":"General"},{"comment":"Ref. [45] is dated 2026 and Ref. [46] is dated 2025; please verify these are published or give preprint DOIs. Several references to 'in preparation' (Ref. [25]) should be flagged as such.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is well-executed in its methodology and the NiPS3 results are likely solid. The CrPS4 central claim needs a self-consistent handling of the lattice in the spiral state and an honest treatment of the U dependence. The authors have disclosed the limitation, which is commendable, but the issue is load-bearing; without additional evidence the most striking conclusion remains conditional. I recommend major revision rather than rejection because the requested checks are within the scope of the study and the methodology itself is sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The CrPS4 spin-spiral claim is the paper's headline, and it has a load-bearing weak spot: the exchange tensors come from a lattice relaxed in a ferromagnetic cell that cannot accommodate the spiral. The NiPS3 result—zigzag order appearing only when J5 is included—is more solid. The paper deserves a serious referee, but the CrPS4 conclusion should not be taken as established until the spiral is checked on a spiral-consistent lattice.\n\nWhat's new: they extract exchange tensors directly via LKAG rather than fitting total energies, and they show shell-by-shell convergence of the Monte Carlo results. That is the right way to test the three-shell truncation most of the field uses. The NiPS3 J5 mechanism is clean: antiferromagnetic J5 ≈ 2J1 supplies the frustration that a three-shell fitted model would wrongly attribute to J1. The disclosure is also unusually honest: they state the FM-cell limitation and the fact that U=0 for CrPS4 was chosen to match the experimental Tc.\n\nSoft spots, in order of size.\n\nFirst, the spin spiral in CrPS4 rests on exchange interactions computed at the FM-relaxed geometry. The four-Cr cell cannot host a q=0.124 Å^-1 spiral, so the dimerization d12=3.75 Å vs d34=3.60 Å is the FM value. Their own interpolation (Eq. 3) shows the spiral is stabilized by the J1a/J1b splitting; reduce that splitting and the spiral unwinds into FM domains. If the true spiral equilibrium has different dimerization, J1a, J1b, J4 and the rest shift, and q and Tc could move substantially. The authors acknowledge this, but acknowledging a limitation does not remove it. This needs to be addressed: relax a large supercell with the spiral allowed, or at least vary the dimerization and show the spiral survives.\n\nSecond, the quantitative agreement with experiment is partly produced by the U choice. U=0 was selected because it gave Tc close to 23 K, so reporting 21 K as agreement is weaker than it looks. No U=1 or 2 eV robustness check is given. This is a moderate concern, not a fatal one—the spiral may well survive, but the claim of excellent agreement should be softened.\n\nNiPS3 is better. U=4 eV is calibrated to the magnon bandwidth, but the J5 mechanism is about sign pattern and shell structure, and the five-shell convergence is clean. The tiny domain degeneracy numbers are inherited from the Hamiltonian, not fit.\n\nWho is this for? Anyone building spin models for 2D magnets, and anyone using fitted short-range Heisenberg models for thiophosphates. I would send it to peer review and ask the authors to fix the CrPS4 lattice problem and add U robustness before accepting the spiral.","headline":"CrPS4 spin-spiral rests on a ferromagnetically relaxed lattice; the NiPS3 J5 result is solid — worth a serious referee, but the CrPS4 conclusion needs more work.","tokens_in":15722,"tokens_out":2690,"would_cite":true,"duration_ms":23860,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that the magnetic ground states of CrPS4 and NiPS3 monolayers are set by long-range exchange interactions, not by the few nearest-neighbor couplings used in earlier models: CrPS4 becomes a spin spiral and NiPS3's zigzag ord","keywords":["CrPS4 monolayer","NiPS3 monolayer","spin spiral","zigzag antiferromagnetism","long-range exchange interactions","exchange frustration","two-dimensional van der Waals magnets","Monte Carlo spin model"],"falsifier":"Relax the CrPS4 monolayer lattice self-consistently in the spin-spiral magnetic state and recompute the exchange tensors and Monte Carlo transition temperature; if the spiral and the ~21 K Tc do not survive this relaxation, the paper's central claim for CrPS4 fails.","tokens_in":14691,"feed_emoji":"🧲","tokens_out":9610,"duration_ms":71834,"temperature":0.7,"pith_summary":"The paper sets out to show that the standard practice of truncating magnetic exchange interactions after the first few neighbor shells fails for two thiophosphate monolayers. By extracting exchange tensors directly from density functional theory and pushing the interaction range until the results converge, the authors find that CrPS4's true ground state is a long-wavelength spin spiral rather than the ferromagnet predicted by short-range models, and that NiPS3's experimentally observed zigzag order appears only when fifth-neighbor couplings are included. If right, these results mean that quantitatively predictive spin models for two-dimensional van der Waals magnets must include long-range exchange, and that earlier short-range predictions for CrPS4's ordering and transition temperature were qualitatively wrong.","feed_headline":"CrPS4 is a spin spiral, not a ferromagnet","feed_subtitle":"Long-range exchange beyond three shells also sets NiPS3's zigzag order and matches measured transition temperatures","key_machinery":"The central object is the set of tensorial exchange interactions extracted directly from the electronic structure rather than fitted to total energies; the isotropic part of each pair's exchange tensor is included shell by shell until ground state and critical temperature converge. In CrPS4, the decisive coupling is the fourth-shell antiferromagnetic interaction J4, which opposes the ferromagnetic first-neighbor bonds along the b-direction; combined with a structural dimerization that splits the first-neighbor couplings into two inequivalent values, it drives the spin spiral. In NiPS3, the decisive coupling is the fifth-shell antiferromagnetic interaction J5 ≈ 2 J1, which supplies the frustr","core_discovery":"The paper claims that in monolayer CrPS4, a fourth-neighbor antiferromagnetic exchange couples second neighbors along the b-direction and competes directly with the ferromagnetic first-neighbor couplings, frustrating the collinear ferromagnet and stabilizing a spin spiral with a wavelength of about 6.9 lattice constants. When exchange interactions are extended to numerical convergence (about the tenth shell), the Monte Carlo critical temperature settles near 21 K, matching the experimental value around 23 K. In monolayer NiPS3, the paper finds that a model truncated at the third shell produces a staggered antiferromagnetic state; the experimentally observed zigzag order appears only when the","pith_inferences":["If the CrPS4 spiral is real, zero-field experiments that reported out-of-plane ferromagnetism may have been probing the field-polarized state; a zero-field local probe (spin-polarized scanning tunneling microscopy, for instance) could directly image the spiral and its ~6.9-unit-cell wavelength.","The paper's finding that CrPS4's spiral unwinds when the dimerization is artificially removed suggests uniaxial strain along the b-direction could tune the spiral pitch or suppress it entirely, a testable extension.","The systematic failure of short-range fitted models in these two compounds implies that apparent discrepancies in other 2D magnets may also trace back to omitted long-range exchange, not to missing anisotropy or interlayer coupling; applying the same convergence protocol to those materials could resolve similar puzzles."],"forward_implications":["Monolayer CrPS4 is predicted to be a spin spiral at zero field, with a critical temperature of about 21 K; the experimentally reported ~23 K transition is reproduced only when exchange interactions are included out to the tenth shell.","The (B,T) phase diagram of monolayer CrPS4 contains a spin-spiral phase, a canted spin-spiral phase with a net out-of-plane moment, and a field-induced ferromagnetic phase, so magnetic fields can be used to switch between non-collinear orders.","Monolayer NiPS3 requires exchange interactions up to the fifth shell to obtain the experimentally observed zigzag antiferromagnetic order; a three-shell model gives the wrong (staggered) ground state, while the five-shell model is sufficient and yields Tc ≈ 61 K.","The spin-wave spectra computed from the full long-range models reproduce the main measured magnon features for both materials, whereas truncating the exchange range or averaging out the structural dimerization in CrPS4 degrades the agreement."],"fun_headline_variants":["Long-range exchange flips CrPS4 to spin spiral","CrPS4's ferromagnetism undone by distant neighbors","Spin spiral emerges when exchange goes far in CrPS4","NiPS3 zigzag order needs fifth-shell exchange","Thiophosphate magnets: long-range exchange decides"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"For CrPS4, the entire spin-spiral picture rests on a lattice that was relaxed in a ferromagnetic four-atom unit cell, which cannot represent the spiral state; if the true spiral equilibrium has different bond lengths, the exchange tensors, wavevector, and transition temperature could shift.","fun_headline_variants_meta":{"raw":{"variants":["Long-range exchange flips CrPS4 to spin spiral","CrPS4's ferromagnetism undone by distant neighbors","Spin spiral emerges when exchange goes far in CrPS4","NiPS3 zigzag order needs fifth-shell exchange","Thiophosphate magnets: long-range exchange decides"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000106,"raw_usage":{"total_tokens":868,"prompt_tokens":728,"completion_tokens":140,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":472,"completion_tokens_details":{"reasoning_tokens":61}},"tokens_in":472,"tokens_out":140,"duration_ms":2254,"temperature":1.0,"reasoning_tokens":61,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T09:46:49.577817+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Relax the CrPS4 monolayer lattice self-consistently in the spin-spiral magnetic state and recompute the exchange tensors and Monte Carlo transition temperature; if the spiral and the ~21 K Tc do not survive this relaxation, the paper's central claim for CrPS4 fails.","supporting_citations":[],"review_version":1}