{"id":"be1b17e7-595f-44f2-a5f3-665d317260b7","arxiv_id":"1908.02811","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Diffusion Monte Carlo predicts an antiferromagnetic ground state for bulk LaCoO3 and shows that weak epitaxial lattice modulations can stabilize ferromagnetic phases.","lead":"Using diffusion Monte Carlo, the authors computed the magnetic ground states of bulk and epitaxially strained LaCoO3. They predict an antiferromagnetic ground state for the bulk material and show that weak lattice modulations under epitaxial strain can stabilize ferromagnetic order, addressing a long-running experimental controversy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed G-type HS-AFM ground state is only optimal among five collinear configurations; the paper itself notes a spin-canted structure may be lower (Sec. III A), and this untested possibility is the main load-bearing risk.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing concern: the ground state is only optimal within the studied set, and the paper itself flags Ref. [115]'s spin-canted structure as potentially more favorable. This is the single most consequential gap because the central claim is not just that an AFM state is low, but that a specific G-type HS-AFM state is the ground state. A lower-energy canted or noncollinear order would change the predicted magnetic structure and the interpretation of epitaxial ferromagnetism, even if the energy separation from LS remained large. The paper's other elements support the central claim: DMC convergence tests, formation enthalpy agreement within 0.07 eV, a consistent ordering across DFT functionals, and explicit care with finite-size effects. These do not, however, close the configuration-space gap. The concrete test is directly implementable with the existing methodology, since the competing state is already described in the literature and the authors have the computational workflow in place. Agreeing with the reader, the verdict should remain CONDITIONAL; no adjustment is needed because the condition was already identified.","tokens_in":20977,"tokens_out":2710,"duration_ms":34409,"concrete_test":"Construct the spin-canted magnetic configuration proposed in Ref. [115] (e.g., two Co sublattices with a noncollinear canting angle) and compute its DMC total energy using the same hard RRKJ pseudopotentials, LDA+U=6 eV trial wavefunctions, 0.01 Ha^-1 timestep, MPC interactions, and finite-size extrapolation to 90-atom cells used in Fig. 2. If the canted state is lower than HS-AFM by more than the 0.02 eV statistical error bar, the G-type HS-AFM ground-state claim is falsified. If noncollinear DMC is impractical, a fallback is to perform collinear DMC on the relaxed canted geometry with spin quantization axis chosen along the dominant moment direction; this still tests whether the geometric/electronic reorganization of the canted phase overturns the ordering.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that DMC predicts a G-type high-spin antiferromagnetic ground state for bulk LaCoO3, 0.40(2) eV/f.u. below the nonmagnetic state (Fig. 2, Table II). This conclusion is based on exactly five collinear spin configurations: HS-AFM, HS-FM, HS/LS-FM, IS-FM, and LS. In Section III A, the authors explicitly acknowledge that a spin-canted magnetic structure has been claimed to be energetically more favorable than the states they studied, citing Ref. [115]. If a canted or otherwise noncollinear magnetic order is indeed lower in energy, then the identification of the ground state as G-type HS-AFM is incorrect, and the quantitative energy ordering among the collinear states does not establish the paper's headline claim. The DMC methodology itself appears careful: timestep and finite-size checks are reported, the nodal surface is optimized with U as a variational parameter, and the formation enthalpy benchmark supports the pseudopotentials. The weakness is not the DMC machinery but the restricted magnetic search space. Because the competing canted state is known in the literature and is flagged by the authors themselves, the burden is on showing that it does not fall below HS-AFM within DMC error bars. Without this check, the ground-state claim remains conditional on the completeness of the five-state ansatz.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports diffusion Monte Carlo (DMC) calculations of the structural, electronic, and magnetic properties of bulk and epitaxial LaCoO3. For bulk LaCoO3, the authors find a G-type high-spin antiferromagnetic (HS-AFM) ground state, 0.40(2) eV per formula unit below the nonmagnetic low-spin state, in contrast to the long-held picture of bulk LaCoO3 as nonmagnetic. For epitaxial LaCoO3 on SrTiO3, they find that moderate uniaxial strain stabilizes ferromagnetic phases, that a lateral lattice modulation costs almost no energy for one mixed magnetic configuration, and that the experimentally observed large La-La modulation would require additional energy or defects. They also report DMC optical and quasiparticle gaps near 3.7 eV, larger than most experimental estimates, and argue that internal d-d transitions or defects may explain the discrepancy. The paper includes extensive DFT benchmarks, geometry relaxations with PBEsol+U, and DMC convergence tests for timestep and finite-size effects.","tokens_in":21249,"tokens_out":4140,"duration_ms":50849,"significance":"If the bulk ground-state prediction is correct, the paper overturns a textbook conclusion and reframes the epitaxial ferromagnetism debate, so the claim has high impact. The DMC methodology is handled carefully: the authors report timestep and finite-size convergence, a pseudopotential benchmark for the formation enthalpy (2.62(1) eV/f.u. versus the experimental 2.55(1) eV/f.u.), and a systematic check of the dependence of the phase ordering on the Hubbard U used to construct trial nodes (Fig. 9). The finding that the DMC phase ordering is stable for U from 2 to 10 eV is a genuine strength. The central weakness is that the ground state is identified only among five collinear spin configurations; the paper itself cites a claim that a spin-canted structure may be lower in energy, and that possibility is not tested in DMC. This limits the strength of the central claim until the competing magnetic order is assessed.","major_comments":[{"comment":"The central claim that bulk LaCoO3 has a G-type HS-AFM ground state is established only within the set of five collinear spin configurations studied. The text explicitly states that a spin-canted magnetic structure has been claimed to be energetically more favorable (Section III A, citing Ref. [115]), yet no DMC calculation or controlled estimate for such a noncollinear state is provided. Because the ground-state identification is the paper's headline result, this is a load-bearing gap. Please add DMC calculations of the canted state, or alternatively provide a quantitative estimate of the energy lowering it could produce and demonstrate that it cannot overcome the 0.40(2) eV gap to the LS state.","section":"Section III B, Fig. 3"},{"comment":"The epitaxial phase diagram is computed at geometries relaxed with PBEsol+U = 4 eV, and the manuscript itself notes that systematic contributions from DFT relaxations introduce scatter into the DMC energies. The PBEsol+U functional is benchmarked only for unstrained bulk volume and Co-O-Co bond angles, not for the strongly strained or modulated geometries used in Fig. 3 and Fig. 4. Because the strain-driven HS-AFM to HS-FM transition is central to the epitaxial conclusions, please provide a DMC-level check of representative relaxed geometries or an explicit sensitivity analysis using an alternative relaxation functional, especially near the crossing at a La-La distance of about 4.6 Å.","section":"Section II, Appendix"},{"comment":"The fixed-node approximation is the dominant systematic error in DMC energy differences among magnetic states. The manuscript optimizes the nodal surface by varying the Hubbard U in the trial wavefunction and shows that the phase ordering is stable, which is valuable, but it does not quantify the residual fixed-node bias for relative phase energies. A single-determinant nodal surface constructed from LDA+U orbitals could, in principle, favor collinear high-spin states relative to the true ground state. Please include a nodal-surface sensitivity test (for example, a multideterminant trial wavefunction for a smaller cell, or a comparison against a different nodal construction for a related cobalt oxide with known magnetic order) or state explicitly in the conclusions that the fixed-node bias is an unresolved uncertainty of this magnitude.","section":"Appendix"}],"minor_comments":[{"comment":"The text reports the DMC equilibrium volume as \"58.2(1) Å3 per formula unit (eV/f.u.)\"; the parenthetical unit should read \"Å3/f.u.\" rather than \"eV/f.u.\".","section":"Abstract"},{"comment":"The formation enthalpy is given as \"2.62(1) eV eV/f.u.\" in Section III A; the duplicate \"eV\" should be removed.","section":"Section IV"},{"comment":"The sentence \"This is in contrast to long-standing experiments; our experiments support this idea\" appears in the conclusions, but the manuscript does not present new experiments. Please rephrase to refer to recent experimental work (e.g., Refs. [49-54]) or to collaborations explicitly if new experimental data are intended.","section":"Section III C"},{"comment":"In the discussion of band gaps, the manuscript says the DMC gaps are roughly 3.7(2) eV for both LS and HS-AFM states, while Table I lists 3.65 ± 0.06 and 3.77 ± 0.12 eV; please ensure the text and table are consistent and clarify which entries correspond to the optical versus quasiparticle gap at the stated wavevector.","section":"Section III C"}],"recommendation":"major_revision","confidential_remarks":"The paper is methodologically serious and the DMC machinery appears well executed. The main risk to the central claim is the untested spin-canted magnetic order, which the authors themselves flag. I would be willing to consider a revision that adds a canted-state DMC test or a rigorous bound on its energy, plus a clearer statement of the fixed-node limitation. The epitaxial geometry sensitivity is a second concern but is likely addressable with targeted calculations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is the first DMC treatment of magnetism in LaCoO3, and it predicts a G-type HS-AFM ground state for bulk LCO, 0.40(2) eV/f.u. below the nonmagnetic LS state. That is a strong claim against the textbook picture, and it deserves a serious referee. But the ground-state identification is only as solid as the five collinear magnetic configurations they tested. The paper itself cites Ref. [115] (Seo et al.), which claimed a spin-canted structure is lower in energy; that configuration is not included in the DMC study. Until that or other noncollinear orders are ruled out within DMC error bars, the central conclusion should be labeled conditional.\n\nWhat the paper does well: the DMC calculations are careful. They report timestep and finite-size convergence, use U as a variational parameter for the nodal surface, and show the phase ordering is essentially independent of U across 2–10 eV. The formation enthalpy benchmark (2.62(1) vs 2.55(1) eV) is a meaningful check on the pseudopotentials. The epitaxial part is genuinely useful: a modulated superstructure with an HS-FM region and an HS-AFM region costs almost no energy, and the AFM-to-FM crossing under strain is tied to orbital ordering. The DMC gaps near 3.7 eV fit recent photoluminescence data, and the discussion of why older optical experiments see sub-eV features is sensible. Data are reported with statistical errors throughout. Citation pattern is clean: prior DFT and experimental work is covered, and the self-citations are to the group's method papers, which is appropriate.\n\nSoft spots, in proportion. The missing canted order is the load-bearing one, and the authors concede it in one sentence without follow-up. The fixed-node approximation is the usual DMC caveat, and for spin-state energy differences, which are notoriously sensitive, it deserves more than a passing note. Epitaxial geometries come from PBEsol+U relaxation, not DMC; defensible, but it adds a DFT layer to the conclusions. The abstract's phrase that a weak lateral modulation 'is sufficient to promote ferromagnetism' is a bit stronger than the text, which says the observed large modulation requires about 0.3 eV/f.u. of additional energy beyond magnetism. No input files or pseudopotential definitions are shipped, which makes reproduction harder. Minor: the conclusion says 'our experiments support this idea' without identifying the experiments; I assume neutron work from collaborators, but as written it is confusing.\n\nBottom line: this paper deserves peer review. I would send it out. The DMC machinery and benchmarks are credible, and the AFM ground state, while conditional, is a well-posed challenge to the field. Anyone working on cobaltites or strained oxide films should read it.","headline":"First DMC treatment of LaCoO3 predicts a G-type HS-AFM bulk ground state; careful and credible, but the untested canted-spin configuration leaves the headline claim conditional.","tokens_in":21806,"tokens_out":3949,"would_cite":true,"duration_ms":39158,"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":"Diffusion Monte Carlo calculations predict that bulk LaCoO3 is a G-type antiferromagnet, 0.40 eV per formula unit below the nonmagnetic state.","keywords":["lanthanum cobaltite","LaCoO3","diffusion Monte Carlo","spin-state transition","antiferromagnetism","epitaxial strain","orbital ordering","band gap"],"falsifier":"Cool a high-purity bulk LaCoO3 crystal and measure neutron diffraction: if no G-type antiferromagnetic Bragg peaks appear at low temperature, the predicted long-range AFM ground state is wrong. Alternatively, recompute the DMC energies with the spin-canted magnetic structure and find it below the G-type high-spin state, which would overturn the ordering within the paper's own method.","tokens_in":20779,"feed_emoji":"🧲","tokens_out":5312,"duration_ms":53315,"temperature":0.7,"pith_summary":"The paper uses diffusion Monte Carlo to settle a long-running dispute about lanthanum cobaltite, a material whose magnetism switches with strain. It claims that bulk LaCoO3 is not the textbook nonmagnetic low-spin insulator but a G-type high-spin antiferromagnet, 0.40(2) eV per formula unit lower in energy than the nonmagnetic state. If true, this overturns a half-century-old picture and reframes epitaxial ferromagnetism: a weak lateral modulation of the lattice, not oxygen vacancies alone, is enough to stabilize ferromagnetic order. The paper also argues that measured sub-electronvolt optical signals are likely internal d–d transitions or defects, because the true fundamental gap is near 3.7 eV.","feed_headline":"Bulk LaCoO3 is antiferromagnetic, diffusion Monte Carlo shows","feed_subtitle":"The calculated ground state sits 0.40 eV per formula unit below the nonmagnetic state, reshaping the strain story.","key_machinery":"The load-bearing tool is diffusion Monte Carlo with Slater–Jastrow trial wavefunctions, using LDA+U orbitals with the Hubbard U tuned as a variational parameter to improve the nodal surface; energies are compared at fixed experimental geometry and along fitted equation-of-state curves. The central objects are the five Co3+ spin configurations — high-spin (S=2), intermediate-spin (S=1), and low-spin (S=0) — with ferromagnetic, antiferromagnetic, and mixed orderings, and the central comparisons are the DMC energy differences between them. A secondary mechanism is orbital ordering in the minority-spin t2g orbitals of the high-spin ferromagnetic state, which the paper argues weakens antiferromagnetic superexchange and stabilizes ferromagnetism under strain.","core_discovery":"On the paper's own terms: using diffusion Monte Carlo on the experimental low-temperature structure, the authors compute total energies for five ordered spin states of Co3+ — high-spin antiferromagnetic (G-type, spins alternating along every direction), high-spin ferromagnetic, mixed high/low-spin ferromagnetic, intermediate-spin ferromagnetic, and low-spin nonmagnetic. The high-spin antiferromagnetic state is lowest by 0.40(2) eV per formula unit over the nonmagnetic state. Under uniaxial compression or expansion, ferromagnetic phases cross below it, and in the epitaxial superstructure a mixed high-spin-AFM/high-spin-FM pattern costs almost no energy, so ferromagnetism appears with small lattice modulation. The experimentally observed ~4.5 Å La–La stripe separation needs an extra ~0.3 eV per formula unit, which the authors attribute to defects. The same calculations put the optical and quasiparticle gap near 3.7 eV, matching photoluminescence rather than the sub-1 eV conductivity features.","pith_inferences":["If the antiferromagnetic ground state is confirmed, strain-engineering of LaCoO3 could select ferromagnetic order with far smaller lattice distortion than the observed stripes, changing the design rules for cobaltite films.","The paper's own caveat about spin-canted structures suggests a direct DMC test with noncollinear spin arrangements; such a calculation would settle whether the predicted ground state survives outside the five orders tested.","The predicted d–d transition near 0.7 eV could be tested directly by resonant inelastic x-ray scattering or optical spectroscopy under magnetic field, which would separate local spin excitations from charge excitations."],"forward_implications":["Bulk LaCoO3 should display short- or long-range G-type antiferromagnetic correlations at low temperature, and the nonmagnetic low-spin picture should be abandoned.","Epitaxial ferromagnetism in LaCoO3 thin films can arise from strain-induced crossing of ferromagnetic and antiferromagnetic energy curves; a uniform or weakly modulated in-plane strain is sufficient, so oxygen vacancies are not required for the magnetism itself.","The large lattice modulation observed in STEM images requires about 0.3 eV per formula unit beyond the magnetic energy landscape, pointing to defects as the driver of the superstructure.","Optical and transport measurements below 1 eV in nominally clean LaCoO3 should be reinterpreted as internal d–d excitations or defect states, since the DMC fundamental gap is about 3.7 eV."],"supporting_citations":[{"why":"Supplies the 4 K neutron-diffraction crystal structure used for the bulk DMC equations of state.","marker":"[66]"},{"why":"Provides the PBEsol+U=4 eV validation for structures and reports short-range AFM correlations that motivate the bulk magnetic study.","marker":"[49]"},{"why":"Claims a spin-canted structure can be more favorable, defining the main caveat to the predicted ground state.","marker":"[115]"},{"why":"Establishes DMC's superiority over DFT for CoO polymorphs, the methodological precedent for applying DMC to LCO.","marker":"[46]"},{"why":"Is the diffusion Monte Carlo formulation whose energy differences carry the central argument.","marker":"[26]"},{"why":"Supplies the orbital-ordering mechanism that explains why ferromagnetic order can beat antiferromagnetic superexchange.","marker":"[120]"},{"why":"Proposed orbital ordering in the intermediate-spin ferromagnetic state, the idea extended here to high-spin minority t2g orbitals.","marker":"[106]"}],"fun_headline_variants":["Strain flips LaCoO3 to ferromagnetic, bulk stays antiferromagnetic","Bulk LaCoO3 antiferromagnetic, epitaxial strain makes it ferromagnetic","Diffusion Monte Carlo: LaCoO3 ground state is antiferromagnetic","Lattice modulation enough to turn LaCoO3 ferromagnetic","LaCoO3: antiferromagnetic in bulk, ferromagnetic under strain"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predicted magnetic ground state is only the lowest of the five spin orders the paper tested; if a spin-canted or otherwise different magnetic order is actually lower in energy, the central conclusion collapses.","fun_headline_variants_meta":{"raw":{"variants":["Strain flips LaCoO3 to ferromagnetic, bulk stays antiferromagnetic","Bulk LaCoO3 antiferromagnetic, epitaxial strain makes it ferromagnetic","Diffusion Monte Carlo: LaCoO3 ground state is antiferromagnetic","Lattice modulation enough to turn LaCoO3 ferromagnetic","LaCoO3: antiferromagnetic in bulk, ferromagnetic under strain"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000385,"raw_usage":{"total_tokens":2032,"prompt_tokens":940,"completion_tokens":1092,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":556,"completion_tokens_details":{"reasoning_tokens":990}},"tokens_in":556,"tokens_out":1092,"duration_ms":10493,"temperature":1.0,"reasoning_tokens":990,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:33:44.561044+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cool a high-purity bulk LaCoO3 crystal and measure neutron diffraction: if no G-type antiferromagnetic Bragg peaks appear at low temperature, the predicted long-range AFM ground state is wrong. Alternatively, recompute the DMC energies with the spin-canted magnetic structure and find it below the G-type high-spin state, which would overturn the ordering within the paper's own method.","supporting_citations":[{"cited_title":"Razzaque Sarker, Int","cited_arxiv_id":null,"evidence_quote":"Claims a spin-canted structure can be more favorable, defining the main caveat to the predicted ground state."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the orbital-ordering mechanism that explains why ferromagnetic order can beat antiferromagnetic superexchange."},{"cited_title":"Yan, J.-S","cited_arxiv_id":null,"evidence_quote":"Proposed orbital ordering in the intermediate-spin ferromagnetic state, the idea extended here to high-spin minority t2g orbitals."}],"review_version":1}