{"id":"3ce07ea2-6622-4745-b261-ae69920bfe9d","arxiv_id":"1908.01916","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"At zero temperature, non-collinear magnetic orderings are predicted to be more stable than the observed type-I collinear order in six 4d/5d double perovskite oxides, with strain switching between two such states in Ba2YOsO6.","lead":"Using density functional theory, the authors predict that at zero temperature six frustrated double perovskite oxides should adopt non-collinear magnetic arrangements instead of the collinear order seen in experiments at higher temperatures. In Ba2YOsO6 they predict a small uniaxial strain can switch between two of the non-collinear states.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Zero-temperature ordering and strain-switch predictions rely on a single 4-magnetic-ion supercell in which J2 is a constant, so untested longer-period or incommensurate orders could still be lower in energy.","rationale":"The reader's weakest assumption correctly identifies the restricted configuration space as the key vulnerability. The paper's own text admits that only a 4-magnetic-ion supercell is used and that J2 is a constant, which means the DFT energies cannot discriminate states that differ in second-neighbor correlations or propagation vectors. This is directly load-bearing: if an untested longer-period or incommensurate order is lower in energy, then the claimed zero-temperature non-collinear ground states are not the true ground states, and the strain tuning between E2a and E3a would be controlling metastable states rather than the equilibrium ordering. The concern does not require rejecting the paper. The paper has real independent support: the same energy hierarchy appears with different functionals and U values, the spin-model explanation is transparent and archived data are provided. The finite-temperature transition is not computed, but the zero-temperature energy hierarchy is the foundation, so the appropriate verdict remains conditional pending a wider configuration search. No ad hominem is involved; this is a technical limitation explicitly acknowledged in the manuscript.","tokens_in":12485,"tokens_out":15816,"duration_ms":210980,"concrete_test":"Recompute total energies of Ba2YOsO6 with a 2x2x2 (or larger) supercell containing at least 8 Os ions, using the same PBEsol+SOC settings, and add candidate orderings that are sensitive to second-neighbor and longer-range couplings: type-II (q=(1/2,1/2,0)), type-III (q=(1/2,0,0)), a spiral with q along (1,0,0), and a non-collinear state with a period larger than one tetrahedron. If any added ordering is more stable than E3a, the zero-temperature ground state and the 0.1% strain-switch prediction fail. If E3a/E2a remain the two lowest states and the E2a-E3a splitting retains the same sign and magnitude, the central claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central predictions are built on DFT calculations in a 40-atom supercell containing exactly four magnetic ions that form one fcc tetrahedron (Section II and Fig. 1). All magnetic states compared—E1, E2, E3 and their SOC-split variants—are commensurate with this single tetrahedron. As the authors state in Section III.A, 'the second nearest-neighbor interaction is a constant in our DFT calculations' because second-neighbor spins are periodic images of the same ion; hence the DFT energies cannot constrain J2 or any longer-range coupling. The spin model in Eq. (2) likewise contains only a nearest-neighbor J1 and a combined biquadratic/ring term α1, and it reproduces the DFT ordering E3 < E2 < E1 < EFM for positive J1 and α1. But that only proves these four ansatz states are ordered as computed; it does not establish that the global fcc ground state lies within this manifold. Frustrated fcc antiferromagnets are known to host type-II, type-III, and incommensurate spiral orders with larger magnetic unit cells, none of which is sampled in a one-tetrahedron cell. If any such state has lower energy than E3a, then the predicted zero-temperature non-collinear ground state and the E2a/E3a strain switching would not describe the physical ground-state manifold. The two-parameter ring parameterization of Eq. (1) is also not the most general zero-total-spin configuration of four spins, so even within the 4-site cell the search is not exhaustive.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports DFT+U spin-polarized calculations, with and without spin-orbit coupling, for six ordered double perovskite oxides (Sr2ScRuO6, Sr2YRuO6, Ba2YRuO6, Sr2ScOsO6, Sr2YOsO6, Ba2YOsO6). Using a 40-atom supercell containing four magnetic ions on an fcc tetrahedron, it finds that at zero temperature non-collinear coplanar and non-coplanar antiferromagnetic orderings (denoted E2 and E3) are lower in energy than the type-I collinear ordering observed experimentally at finite temperatures. A classical Heisenberg plus biquadratic and four-spin ring-exchange model is shown to reproduce the ordering E3 < E2 < E1 < EFM for positive nearest-neighbor J1 and positive combined coefficient α1. After including SOC, nine spin-orbit-split states are identified, with E3a and E2a almost degenerate in Ba2YOsO6; the paper predicts that roughly 0.1% uniaxial strain can switch between these two states. The central predictions are that a collinear-to-noncollinear magnetic transition occurs at low temperatures and that strain offers a mechanical route to control non-collinear magnetic orderings.","tokens_in":12817,"tokens_out":8752,"duration_ms":101977,"significance":"If the zero-temperature ground-state claim and the strain-switching prediction survive scrutiny, this would be a valuable contribution: it identifies concrete candidate non-collinear magnetic structures in frustrated double perovskites that could be tested by low-temperature neutron scattering, and it proposes a practical strain-based control mechanism. The DFT evidence is internally consistent: the ordering E3 < E2 < E1 < EFM is stable across different functionals, Hubbard U values, and the presence or absence of SOC, at least for Ba2YOsO6. The analytic spin-model derivation is transparent and requires no fitted parameter values, only signs. The authors also make their data openly available in Zenodo. The significance is tempered, however, by two gaps: all tested magnetic states are commensurate with a single four-magnetic-ion tetrahedron, and the finite-temperature transition is inferred without any free-energy calculation.","major_comments":[{"comment":"The search over magnetic orderings is restricted to a 40-atom supercell containing four magnetic ions, and the manuscript itself states that 'the second nearest-neighbor interaction is a constant in our DFT calculations'. Consequently, the energies compared in Figs. 3, 5, 7, and 8 only establish the relative stability of states commensurate with a single fcc tetrahedron. Frustrated fcc antiferromagnets can host type-II, type-III, and incommensurate spiral orderings with larger magnetic unit cells, and none of these is sampled here. This limitation is load-bearing because the abstract and conclusions claim that at zero temperature non-collinear orderings are 'more stable' than the type-I ordering, which implies a statement about the global ground state. The authors should either perform calculations in larger supercells containing candidate longer-period orders, or explicitly restrict the claim to the four-sublattice manifold. As written, the global ground-state claim is unproven.","section":"Section III.A, Fig. 1, and Eq. (2)"},{"comment":"The finite-temperature conclusion is inferred without any free-energy calculation. The paper states that an 'entropy-driven collinear-to-noncollinear magnetic transition' could occur, but no partition function, Monte Carlo sampling, or high-temperature expansion is presented for either the spin model or the DFT-derived states. Since the experimentally observed state at finite temperature is the collinear type-I order, the key physical puzzle is why entropy would favor a state that is higher in energy at T=0. The paper's assertion is plausible but is not supported by a calculation. To make the central claim about a collinear-to-noncollinear transition load-bearing, the authors should provide at least a classical Monte Carlo or analytical free-energy estimate showing that the collinear state becomes entropically selected at the relevant temperatures.","section":"Section III.B and Section IV"},{"comment":"The strain-switching prediction rests on a very small energy difference, reported as about 0.10 meV/f.u. between E2a and E3a, and on a linear fit of DFT+SOC energies as a function of uniaxial strain. The critical strain is about 0.1%, which is comparable to the scale on which the two states are nearly degenerate. The manuscript does not report convergence of this energy difference with respect to k-point density, plane-wave cutoff, smearing, or structural relaxation tolerances, nor does it provide error bars or a measure of the quality of the linear fit. Because the sign of E2a - E3a under strain is the entire basis for the proposed mechanical switching, the authors should demonstrate that this small energy difference is numerically robust and not an artifact of computational settings. Without such tests, the strain-switching prediction is fragile even though the underlying near-degeneracy may be real.","section":"Section III.B and Fig. 9"}],"minor_comments":[{"comment":"The abstract says 'a wide range of magnetically frustrated oxides' and 'universal' language appears in the text, but only six compounds are studied and functional/SOC robustness is demonstrated in detail only for Ba2YOsO6; the wording should be matched to the evidence.","section":"Abstract and Section I"},{"comment":"The space groups are listed with different measurement temperatures in the parent references; a short note on whether the calculations use the low-temperature or room-temperature structures for each compound would improve reproducibility.","section":"Table I"},{"comment":"The strain is defined as a percentage in Eq. (6), but the x-axis label in Fig. 9 is 'εxx(%)'; please make the definition consistent and state whether the fitted line uses the DFT points at all plotted strains.","section":"Fig. 9 and Eq. (6)"},{"comment":"The energy zero is set to E3a in the main text but to E2a in Fig. S1; this is not an error, but the convention should be stated in every caption to avoid confusion.","section":"Supplementary Fig. S1"},{"comment":"Some reference titles and journal names, especially Ref. [16], contain garbled or non-standard characters ('Zeitschrift fr anorganische und allgemeine Chemie'); these should be corrected.","section":"References"},{"comment":"The parameterization of the four-spin configuration is clear, but the text should state explicitly that this parameterization covers the zero-total-spin head-to-tail ring configurations considered in this work and that other closed configurations are not intended.","section":"Eq. (1) and Fig. 2"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a worthwhile DFT paper with a clean result and one honest limitation. The material-specific predictions—non-collinear ground states in six double perovskites and a strain switch in Ba2YOsO6—are new, and the internal evidence is solid. The soft spot is that the search space is limited to the four-sublattice tetrahedron of the 40-atom cell, so J2 is a constant and longer-period or incommensurate orders are never tested. That does not kill the paper, but it does mean the ground-state claim is conditional.\n\nWhat the paper does well: the energy ordering E3 < E2 < E1 < EFM is robust across PBEsol, PBE, LDA, across Hubbard U values, and with and without SOC. The spin model with Heisenberg plus biquadratic and ring exchange is simple and transparent; the fact that only the signs of J1 and α1 are needed to reproduce the ordering is a nice piece of physics. The archived data (Zenodo) makes the numbers checkable, and the experimental citations for the type-I order in these six oxides are relevant and current.\n\nThe soft spots: (1) the four-site cell misses wavevectors that are not commensurate with one tetrahedron; type-II, type-III, and spiral orders with larger cells are known to be competitive in fcc antiferromagnets, and the paper acknowledges J2 is a constant but does not attempt to bound its effect. If one of those states is lower, the predicted ground state changes. This is a genuine limitation, though not an unusual one for DFT. (2) The collinear-to-noncollinear transition at low temperature is inferred from an entropic argument, not computed; no free-energy or Monte Carlo is presented. Reasonable, but unproven. (3) The strain-switching prediction rests on an energy difference of ~0.1 meV/f.u. That is smaller than typical DFT error bars, though the near-degeneracy is independent of functional and the linear response to strain is physically plausible. I would flag this as promising but needing experimental confirmation, not as a logical flaw.\n\nWho this is for: condensed-matter theorists and experimentalists working on frustrated double perovskites and 4d/5d magnetism. It deserves a serious referee—the predictions are concrete and the internal consistency is strong. I would want the authors to discuss the restricted search space more explicitly and, ideally, run one or two larger-cell checks on Ba2YOsO6 to see if the E3a/E2a states hold against type-II or spiral order. Even if those checks fail, the paper would still be useful because it maps the energy landscape within the 4-sublattice manifold. Recommending: send to peer review, with a request for those checks or a clear caveat.","headline":"Solid DFT prediction of non-collinear order in six double perovskites, but the claimed ground states are only proven within a four-sublattice ansatz and the strain switch sits on a very small energy gap.","tokens_in":13338,"tokens_out":2652,"would_cite":true,"duration_ms":27048,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"At zero temperature, non-collinear spin order beats collinear order in six oxides","keywords":["frustrated magnetism","double perovskite oxides","non-collinear magnetic ordering","biquadratic exchange","ring exchange","spin-orbit coupling","uniaxial strain","density functional theory"],"falsifier":"Neutron scattering at millikelvin temperatures on Ba2YOsO6 or a similar compound: if the magnetic order stays type-I collinear down to the lowest temperature and no transition to the predicted coplanar or non-coplanar states appears, the central prediction is contradicted. Strain-dependent diffraction would further test the predicted switch at about 0.1% uniaxial compression.","tokens_in":12291,"feed_emoji":"🧲","tokens_out":11007,"duration_ms":87502,"temperature":0.7,"pith_summary":"Using density-functional calculations, this paper argues that in a family of six frustrated double perovskite oxides the magnetic state seen in experiments—a type-I collinear antiferromagnet—is not the lowest-energy state at zero temperature. Instead, non-collinear orderings, both coplanar and fully three-dimensional, are predicted to be more stable, and their stability is driven by biquadratic and four-spin ring exchange that become important for the extended 4d and 5d orbitals of Ru and Os. The paper also shows that in Ba2YOsO6, a uniaxial strain of about 0.1 percent can shift the balance between two nearly degenerate non-collinear states. If correct, this predicts a collinear-to-non-collinear spin transition at low temperatures in these oxides and offers a mechanical way to control magnetic order.","feed_headline":"Non-collinear spins win at zero temperature in six oxides","feed_subtitle":"Calculations say strain can switch Ba2YOsO6 between two nearly equal magnetic states.","key_machinery":"The argument is carried by a classical vector-spin Hamiltonian on the four spins of an fcc tetrahedron: Heisenberg exchange plus nearest-neighbor biquadratic and four-spin ring exchange, with the biquadratic and ring coefficients combined into a single parameter $\\alpha_1 = a_1 + a_2$. On the fcc tetrahedron the Heisenberg term is a constant independent of spin direction, so the energy landscape is set entirely by the higher-order terms. Inserting a general two-angle parametrization of a four-spin ring yields an analytic energy $E/N = -2J_1 + \\alpha_1\\left(\\frac{13}{4} - \\cos\\theta + \\frac{7}{4}\\cos 2\\theta + 2\\cos^4(\\theta/2)\\cos 2\\phi\\right) + E_0$, whose three extrema are exactly the collinear state $E_1$, the coplanar state $E_2$, and the non-coplanar state $E_3$. Positive $J_1$ and $\\alpha_1$ give the ordering $E_3 < E_2 < E_1 < E_{FM}$ independent of their magnitudes. Adding spin-orbit coupling splits the states into nine distinct orderings; the near-degeneracy of two of them is what strain acts on.","core_discovery":"The central claim is that at zero temperature, non-collinear antiferromagnetic orderings are thermodynamically more stable than the type-I collinear ordering experimentally observed at finite temperatures in six frustrated double perovskite oxides: Sr2ScRuO6, Sr2YRuO6, Ba2YRuO6, Sr2ScOsO6, Sr2YOsO6, and Ba2YOsO6. The paper identifies two distinct non-collinear states of a tetrahedron of four $S=3/2$ spins—a coplanar state $E_2$ and a non-coplanar state $E_3$ with all spins at an angle $\\arccos(1/3) \\approx 71^\\circ$—and shows that both sit lower in energy than the collinear state, independent of Hubbard $U$, exchange-correlation functional, and the presence or absence of spin-orbit coupling. With spin-orbit coupling included, the degeneracies split and nine stable orderings appear; in Ba2YOsO6 the two lowest, coplanar $E_{2a}$ and non-coplanar $E_{3a}$, are almost degenerate at about 0.10 meV per formula unit. Uniaxial strain moves the system across this near-degeneracy: tensile strain favors $E_{3a}$, while roughly 0.1% compressive strain makes $E_{2a}$ more stable.","pith_inferences":["The four-sublattice model confines the search to states compatible with a 40-atom supercell; the paper notes that second-neighbor exchange is a constant in such a cell, so other ordering wavevectors, longer-range interactions, or incommensurate states could be even lower in energy and would change the predicted ground state.","The strain-switching result suggests a general design rule: look for frustrated oxides with accidental near-degeneracies between non-collinear orders, since even small lattice distortions can select the ground state; Ba2YOsO6 is the first concrete candidate.","If the predicted low-temperature phases are confirmed by neutron scattering, they would give a direct measurement of biquadratic and four-spin ring exchange in 5d oxides, whose magnitudes are currently inferred mainly indirectly."],"forward_implications":["At sufficiently low temperatures, each of the six oxides should undergo a collinear-to-non-collinear magnetic transition, with the collinear state stabilized by entropy at finite temperature.","The non-collinear states $E_{3a}$ and $E_{2a}$ are predicted to be the zero-temperature ground states in all six compounds, not just Ba2YOsO6.","In Ba2YOsO6, the energy difference between the two lowest non-collinear states is about 0.10 meV per formula unit; uniaxial strain of about 0.1% compressive selects $E_{2a}$ and tensile strain selects $E_{3a}$.","The energy ordering is robust against the choice of exchange-correlation functional (PBEsol, PBE, LDA), against Hubbard $U$ from 0 to 5 eV, and against whether spin-orbit coupling is included."],"supporting_citations":[{"why":"Supplies the experimental neutron-scattering evidence that Sr2ScOsO6 orders in the type-I collinear structure at finite temperature, the baseline the zero-temperature prediction must beat.","marker":"[11]"},{"why":"Reports type-I collinear ordering and Néel temperature for Sr2YOsO6, another of the six oxides the calculations target.","marker":"[15]"},{"why":"Gives the experimental magnetic structure, Néel temperature, and Os moment for Ba2YOsO6, used as the strain-tuning test case.","marker":"[20]"},{"why":"Shows that a nearest-neighbor Heisenberg model on the fcc lattice leaves the four-sublattice antiferromagnetic states continuously degenerate, motivating the higher-order exchange terms.","marker":"[34]"},{"why":"Provides the derivation that biquadratic and four-spin ring-exchange interactions arise from the t/U expansion for extended 4d and 5d orbitals, grounding the model Hamiltonian.","marker":"[38]"},{"why":"Demonstrates that continuous uniaxial strain up to about 1% is experimentally achievable with piezoelectric actuators, establishing the feasibility of the predicted strain switching.","marker":"[46]"}],"fun_headline_variants":["Zero-temperature non-collinear order wins in six oxides","Strain switches Ba2YOsO6 between two non-collinear states","Non-collinear magnetic order stable at zero K in six oxides","Strain tunes near-degenerate non-collinear states in Ba2YOsO6","Higher-order exchange drives non-collinear order in frustrated oxides"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predictions assume that the magnetic energy of each oxide is fully captured by the relative angles of four spins on one tetrahedron, with interactions only between nearest neighbors; longer-range or non-repeating magnetic patterns were not tested and could be lower in energy.","fun_headline_variants_meta":{"raw":{"variants":["Zero-temperature non-collinear order wins in six oxides","Strain switches Ba2YOsO6 between two non-collinear states","Non-collinear magnetic order stable at zero K in six oxides","Strain tunes near-degenerate non-collinear states in Ba2YOsO6","Higher-order exchange drives non-collinear order in frustrated oxides"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00085,"raw_usage":{"total_tokens":3740,"prompt_tokens":1035,"completion_tokens":2705,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":651,"completion_tokens_details":{"reasoning_tokens":2614}},"tokens_in":651,"tokens_out":2705,"duration_ms":18523,"temperature":1.0,"reasoning_tokens":2614,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:00:20.548050+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Neutron scattering at millikelvin temperatures on Ba2YOsO6 or a similar compound: if the magnetic order stays type-I collinear down to the lowest temperature and no transition to the predicted coplanar or non-coplanar states appears, the central prediction is contradicted. Strain-dependent diffraction would further test the predicted switch at about 0.1% uniaxial compression.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the experimental neutron-scattering evidence that Sr2ScOsO6 orders in the type-I collinear structure at finite temperature, the baseline the zero-temperature prediction must beat."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports type-I collinear ordering and Néel temperature for Sr2YOsO6, another of the six oxides the calculations target."},{"cited_title":"Kermarrec, C","cited_arxiv_id":null,"evidence_quote":"Gives the experimental magnetic structure, Néel temperature, and Os moment for Ba2YOsO6, used as the strain-tuning test case."},{"cited_title":"Oguchi, H","cited_arxiv_id":null,"evidence_quote":"Shows that a nearest-neighbor Heisenberg model on the fcc lattice leaves the four-sublattice antiferromagnetic states continuously degenerate, motivating the higher-order exchange terms."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates that continuous uniaxial strain up to about 1% is experimentally achievable with piezoelectric actuators, establishing the feasibility of the predicted strain switching."}],"review_version":1}