{"id":"84b62780-4c04-479e-861c-562685eb62b4","arxiv_id":"2608.12505","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Antisite defects in MnSb2Te4 drive a strong antiferromagnetic exchange (J' ≈ 0.75 meV) that produces an optical magnon and dominates the magnetic excitations.","lead":"Neutron scattering on MnSb2Te4 crystals reveals that atomic-scale antisite defects create a magnetic bond about ten times stronger than the crystal's native magnetic coupling. This defect bond controls the spin waves, including a new high-energy optical magnon, and persists even above the magnetic ordering temperature.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quantitative J' value is not secured: the homogeneous tri-layer LSWT model is applied near the triangular-lattice percolation threshold, and the LLD validation uses J' as an input rather than as a fit-derived check.","rationale":"The central qualitative result—that antisite defects introduce a distinct magnetic subsystem and a gapped optical-like response—is well supported by the single-crystal INS maps and by the LLD simulations, and the dimer structure-factor comparison in Fig. A2 provides direct evidence for the Mn-Te-Mn interaction geometry. The soft spot is not the existence of the extra branch but the specific value J'≈0.75 meV and the claim that it is about ten times larger than J0. That number is obtained from a homogeneous, Jc=D=0 tri-layer LSWT fit to broad, disorder-dominated excitations. The strongest independent numerical check, LLD, was run with J' fixed at 0.75 meV, so it tests whether that value can reproduce the data, not whether the data require it. The percolation argument makes the homogeneous-layer assumption especially fragile, and the proposed inverse-test would quantify the bias. The reader's conditional verdict is therefore appropriate: the qualitative conclusion is plausible and well illustrated, but the quantitative headline coupling should not be accepted as fully established until the fitting bias is assessed. I also note that the data link in Ref. [49] is a placeholder ('[Link Here]'), which blocks independent reproduction of the fits. These issues do not justify rejection, but they do justify the conditional status already assigned.","tokens_in":13431,"tokens_out":15184,"duration_ms":156882,"concrete_test":"Run a closed-loop calibration with the same LLD code: generate synthetic S(Q,E) from the published random-disorder model (p=0.588, q=0.129; J0=-0.075 meV; Jc=0.002 meV; D=-0.002 meV) with a known J'=0.75 meV, then fit the synthetic peak centers with the homogeneous tri-layer LSWT model (Jc=D=0) exactly as done in Sec. IV.B/App. C. If the recovered J' deviates by more than about 20% (outside 0.60-0.90 meV), the homogeneous-layer bias is comparable to the claimed effect; repeat with J'=0.4 and 1.0 meV to map the bias. This directly determines whether the order-of-magnitude ratio J'/|J0| is an artifact of the fitting model rather than a robust property of the system.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative extraction of the headline coupling, J'≈0.75 meV, rests on the homogeneous tri-layer LSWT model of Sec. IV.B and App. C. That model replaces the heavily diluted Mn layers (p=0.588, q=0.129) by uniform layers with SA=pS and SB=qS and sets Jc=D=0. This approximation is particularly fragile here because the Mn layer forms a triangular lattice, whose site-percolation threshold is p_c=0.5, so p=0.588 is only about 0.09 above percolation; the percolating-cluster correlation length is of order 20 lattice spacings, and the observed crossover to broad, localized modes at H_c≈0.2 is a direct signature of strong-disorder physics rather than of a uniform dilute ferromagnet. Because the fitted optical gap is E2(q=0)=3J'(SA-2SB), any error in the effective local moments or in the coherence of the assumed ferrimagnetic block enters linearly in J'. The LLD simulations cited as support use J'=0.75 meV as an input and are not fitted to the data, so they cannot independently validate the numerical value. The data availability statement also points to a placeholder link [49], so an outside check of the fits is currently impossible.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"Jaishi et al. report inelastic neutron scattering measurements on single crystals of MnSb2Te4, an A-type antiferromagnet with substantial Mn/Sb antisite mixing. They observe a gapless acoustic spin-wave branch that broadens at short wavelengths and an additional gapped optical branch that they attribute to antiferromagnetic coupling J' between Mn_Sb antisite defects and Mn_Mn moments. Above T_N, the Q dependence of the scattering resembles the structure factor of next-nearest-neighbor Mn-Te-Mn dimers. The authors analyze the dispersions with a minimal tri-layer linear spin-wave model, obtaining J0 ≈ -0.075 meV and J' ≈ 0.75 meV, and reproduce the main features with Landau-Lifshitz dynamics simulations on quenched-disorder supercells. They conclude that antisite defects act as an intrinsic magnetic subsystem that dominates the spin dynamics of MnSb2Te4.","tokens_in":13650,"tokens_out":8016,"duration_ms":71570,"significance":"The central observation of a defect-induced optical magnon with a large gap is plausible, and the paper has several genuine strengths: the optical branch is directly visible in the constant-energy data, the above-T_N structure factor is compared to explicit dimer structure-factor calculations, the LLD simulations treat the quenched disorder explicitly and reproduce the qualitative broadening, and the extracted J' is consistent with prior high-field magnetization and dilute-Sb2Te3 results. If the quantitative extraction is confirmed, the work would establish antisite defects as a controllable second magnetic subsystem in the Mn(Bi,Sb)2Te4 family, with direct implications for defect-engineered topological magnetism. The main limitations are the reliance of J' on a homogeneous tri-layer LSWT model in a regime close to the triangular-lattice percolation threshold and the absence of an independent determination of J' from the disorder-resolved simulations.","major_comments":[{"comment":"The reported optical gap is inconsistent: the main text gives an onset Δ'≈1.5 meV (Section III.A), while Appendix C states that the fitted optic mode has a gap Δ'≈1.9 meV at q=0. Since the fitted J' scales linearly with this gap through E2(0)=3J'(SA-2SB), the discrepancy directly affects the headline value. Please reconcile the two numbers, and state whether the optical branch has a soft onset with a peak at higher energy or whether one of the values is a typo.","section":"III.A and Appendix C"},{"comment":"The homogeneous tri-layer LSWT model of Section IV.B and Appendix C replaces the diluted Mn layers (p=0.588, q=0.129) by uniform layers with S_A=pS and S_B=qS and sets J_c=D=0. For a triangular lattice the site-percolation threshold is p_c≈0.5, so the Mn layer sits only ≈0.09 above percolation, and the crossover to broad localized modes at H_c≈0.2 described in Section III.A is a direct sign of strong-disorder physics. As the optical gap depends linearly on (S_A-2S_B), any error in the effective moments enters linearly in J'. Please justify the homogeneous approximation quantitatively, for example by comparing analytic dispersions with explicit-disorder LLD simulations for a range of J', or by providing a percolation-aware uncertainty estimate for J'.","section":"IV.B / Appendix C"},{"comment":"Section IV.C uses J'=0.75 meV as an input to the LLD simulations, so the resulting agreement of the simulated optical mode is a consistency check rather than an independent validation of J'. To make the quantitative claim robust, the authors should show a sensitivity analysis over J' (and ideally over the effective moments) demonstrating how the simulated optical gap and linewidths change; the quoted ±0.03 meV from the LSWT fit does not capture the model uncertainty.","section":"IV.C"},{"comment":"The Data Availability statement cites reference [49] as 'Link Here', which is a placeholder with no usable URL. Since the central quantitative claim depends on fits whose results are not fully checkable from the figures, the data, fitted peak positions, and simulation outputs should be made accessible prior to publication.","section":"Data Availability"}],"minor_comments":[{"comment":"The relation p_eff≈(2H_c)^2=0.16 is asserted without derivation or reference; please provide the basis for this estimate.","section":"III.A"},{"comment":"The sign convention for the single-ion anisotropy D in Eq. (1) is not defined, and the text sets D=0 for the LSWT fit while using D=-0.002 meV in the LLD simulations; please clarify the convention and the rationale for the different values.","section":"Eq. (1) / IV.C"},{"comment":"The caption describes a dispersionless feature near 2.5 meV as spurious; since this energy is close to the reported optical branch, please explain how the spurious feature was identified and excluded from the peak fits.","section":"Fig. 1"},{"comment":"The model predicts a flat mode E3=3S_AJ'≈3.3 meV, but the paper does not discuss whether this mode is observed, falls outside the accessible energy window, or is overdamped; please address this.","section":"Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the central observation appears likely to be real. My main reservations are the internal inconsistency in the reported optical gap and the robustness of the homogeneous LSWT extraction of J', both of which are addressable in revision. I would not reject the manuscript on the current evidence, but the quantitative headline value needs additional support or explicit caveats."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's the short version: this is the first single-crystal INS work on MnSb2Te4 that resolves a defect-induced optical magnon and directly estimates the antisite exchange J' ~0.75 meV, about ten times the intrinsic FM J0. The qualitative story is convincing; the exact number is less nailed down than it looks.\n\nWhat's genuinely new: single-crystal INS data showing a gapped optic branch, a crossover to broad localized modes at Hc~0.2, and a dimer-like structure factor above TN that matches a nearly linear Mn-Te-Mn bond geometry. The LLD simulations with quenched disorder capture the dispersions and the paramagnetic correlations. That's real experimental progress. The extracted J' is consistent with high-field magnetization (Lai 2021) and with INS on dilute Mn in Sb2Te3 (Islam 2023), so the value has external support.\n\nSoft spots: the quantitative J' extraction rests on a homogeneous tri-layer LSWT model in which the heavily diluted Mn sublattice (p=0.588, q=0.129) is replaced by uniform sheets, with Jc=D=0. The stress-test point about proximity to the triangular-lattice percolation threshold (p_c=0.5) is fair; the crossover to broad modes at Hc~0.2 is direct evidence that disorder is not a small perturbation. Since the optical gap is proportional to 3J'(SA-2SB), errors in the effective moments feed linearly into J'. The LLD simulations use J'=0.75 meV as an input, so they validate the model consistency but not the value independently. The stiffness ratio prediction also reuses the same fitted parameters. And the data availability link is a placeholder, which blocks outside fitting checks. None of this sinks the central claim, but it means the reported J' should be treated as model-dependent until a more rigorous treatment of disorder, or a direct measurement, comes along. A sensitivity study varying Jc and the effective moments would help.\n\nWho benefits: anyone working on Mn(Bi,Sb)2Te4 magnetism or defect engineering in magnetic topological insulators. It's a serious experimental characterization paper that deserves a proper referee, with the conditions that the data become accessible and the model sensitivity be quantified.\n\nRecommendation: send to review. Ask for the data repository and a discussion of how J' shifts under nonzero Jc and under a more realistic disorder treatment.","headline":"First single-crystal INS study resolving a defect-induced optical magnon in MnSb2Te4; the qualitative picture is solid, but the quantitative J' is more model-dependent than it appears.","tokens_in":14295,"tokens_out":1660,"would_cite":true,"duration_ms":14705,"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":"In MnSb2Te4, antisite defects create a strongly antiferromagnetic Mn–Te–Mn bond about ten times stronger than the host ferromagnet, producing a large-gap optical magnon and controlling the measured spin dynamics.","keywords":["MnSb2Te4","antisite defects","inelastic neutron scattering","optical magnon","spin waves","magnetic topological insulator","ferrimagnetism","Mn-Te-Mn superexchange"],"falsifier":"Measure the spin excitations in MnSb2Te4 crystals with deliberately varied antisite concentrations: if $J'$ really is the defect-induced Mn–Te–Mn bond, the optical gap should track the defect fraction $q$, the stiffness enhancement should follow $A/A_0 = (3p^2 + 8pq|J'/J_0|)/(6(p - 2q))$, and a nearly stoichiometric sample should show no optical branch. Conversely, a high-resolution $(0,0,L)$ measurement that resolves the interlayer dispersion would test the $J_c = 0$ assumption, since a finite interlayer coupling would move the branches in a way the tri-layer model cannot reproduce.","tokens_in":13196,"feed_emoji":"🧲","tokens_out":20587,"duration_ms":160345,"temperature":0.7,"pith_summary":"MnSb2Te4, nominally a cousin of the antiferromagnetic topological insulator MnBi2Te4, is actually heavily mixed: neutron diffraction finds about 13% of Sb-layer sites occupied by Mn and about 41% of Mn-layer sites occupied by Sb. The paper argues that these antisite defects are the dominant magnetic ingredient: a defect Mn coupled to a layer Mn through a nearly linear Mn–Te–Mn bond feels an antiferromagnetic exchange $J' \\approx 0.75$ meV, about ten times larger than the intrinsic ferromagnetic coupling $J_0 \\approx -0.075$ meV. This strong bond makes each septuple block ferrimagnetic, producing a gapped optical magnon, closing the low-energy spin gap, and stiffening the acoustic spin waves. The same defect bond survives as dimer-like antiferromagnetic correlations above the Néel temperature. Establishing this matters because it identifies the microscopic interaction that defect engineering would tune to control magnetism, and with it the topological electronic phases, in the Mn(Bi,Sb)2Te4 family.","feed_headline":"Defect bond 10x stronger than the host sets MnSb2Te4 spin waves","feed_subtitle":"Single-crystal neutron data tie a large-gap optical magnon to antisite Mn-Te-Mn bonds.","key_machinery":"The load-bearing object is the effective tri-layer model of one septuple block: a central Mn layer with reduced moment $S_A = pS$, and two Sb-layer sheets with moments $S_B = qS$, where $p = 0.588$ and $q = 0.129$ come from diffraction. The central layer has ferromagnetic intralayer exchange $J_0 < 0$, and the defect sheets couple antiferromagnetically to it through $J' > 0$ along linear Mn–Te–Mn bonds; the analytical model neglects intralayer coupling of the dilute defect layers, the interlayer coupling $J_c$, and single-ion anisotropy $D$. Expressing the spins as bosons and diagonalizing the resulting quadratic form yields three branches—an acoustic mode $E_1$, a gapped optical mode $E_2$ whose gap sets $J'$, and a flat mode $E_3$—plus an analytic stiffness ratio $A/A_0 = (3p^2 + 8pq|J'/J_0|)/(6(p - 2q))$ that lets the fit explain the stiffened acoustic branch. Stochastic classical spin dynamics on a $40 \\times 40 \\times 4$ supercell with random vacancies and antisite spins is the disordered counterpart that captures the broad, localized short-wavelength excitations and the dimer-like correlations above $T_N$.","core_discovery":"Using inelastic neutron scattering on single crystals of A-type antiferromagnetic MnSb2Te4 ($T_N = 19$ K), the paper finds spin excitations qualitatively different from the nearly defect-free host MnBi2Te4: a gapless acoustic branch crosses over to broad, disorder-dominated excitations beyond $H_c \\approx 0.2$, and a second, gapped optical branch appears with onset $\\Delta' \\approx 1.5$ meV. The interlayer modulation of the scattering, with strong intensity at $L = 5.5$ and weak at $L = 11$, shows that moments in neighboring Mn and Sb layers are antiferromagnetically correlated within each septuple block. Fitting the acoustic and optical branches with an effective tri-layer linear spin-wave model yields $J_0 = -0.075(1)$ meV and $J' = 0.75(3)$ meV, and the model predicts a defect-driven enhancement of the acoustic stiffness by a factor $A/A_0 \\approx 3.6$, close to the observed $\\approx 3$. Classical spin-dynamics simulations with quenched disorder reproduce the ordered and paramagnetic scattering, and the measured $Q$-dependence above $T_N$ matches the structure factor of a magnetic dimer connected by the next-nearest-neighbor Mn–Te–Mn bond. The conclusion is that antisite defects form an intrinsic second magnetic subsystem whose exchange interactions dominate the intrinsic energy scales of the material.","pith_inferences":["A testable extension the paper does not pursue: at lower temperatures and higher resolution, the continuum observed above $T_N$ might resolve into discrete dimer spin-state excitations, since the dimer-like structure factor indicates bound Mn–Mn pairs with quantized levels.","If the paper's picture is right, the fitted $J'$ and $J_0$ may partly absorb a small interlayer coupling $J_c$, and a three-dimensional spin-wave fit on better-resolved $(0,0,L)$ data could separate $J_c$ and test whether the stiffness enhancement is entirely defect-driven.","The unresolved competition between the ferromagnetic defect–defect interlayer interaction and the antiferromagnetic Mn–Mn interlayer coupling could produce a defect-concentration-driven transition from A-type antiferromagnetism to net ferrimagnetism, which magnetization and transport measurements on tuned samples could probe.","High-field inelastic neutron scattering offers a further direct test: if the optical magnon gap is set by the ferrimagnetic exchange field, a field strong enough to flip the defect sublattice should collapse the optical branch."],"forward_implications":["Antisite concentration becomes a control knob for the magnetic Hamiltonian: crystals grown or annealed with different Mn/Sb mixing should show optical-gap and spin-stiffness variations that track $p$ and $q$.","Spin-wave measurements on Mn(Bi,Sb)2Te4 with intermediate Sb content should interpolate between the single well-defined branch of nearly defect-free MnBi2Te4 and the defect-dominated spectrum reported here, providing a direct test of the defect-subsystem picture.","Because the strong coupling runs through nearly linear Mn–Te–Mn bonds, the same defect-induced ferrimagnetism should appear in other quintuple- and septuple-layer topological insulators with Mn/Sb or Mn/Sn antisite mixing, as already suggested by Mn-doped Sb2Te3 and SnTe.","The broad, disorder-dominated short-wavelength excitations and the large optical gap are observable fingerprints that distinguish defect-rich from defect-poor crystals, so neutron and magnetization experiments can diagnose antisite content without crystallographic refinement."],"supporting_citations":[{"why":"Provides the neutron diffraction refinement giving p = 0.588 and q = 0.129, the antisite concentrations inserted into the spin model.","marker":"[27]"},{"why":"Supplies the high-field magnetization value Msat ≈ 1.5 µB and the earlier defect-ferrimagnetism picture used to estimate J'.","marker":"[26]"},{"why":"Supplies the MnBi2Te4 single-crystal inelastic neutron scattering used as the defect-free baseline for spin stiffness, gap, and dispersion.","marker":"[40]"},{"why":"Provides the competing-interaction parameters for MnBi2Te4 that anchor the intrinsic J0 and the anisotropy comparison.","marker":"[42]"},{"why":"Gives the dilute Mn-in-Sb2Te3 neutron result identifying antiferromagnetic Mn–Mn dimers, the direct precedent for J' and the linear-bond mechanism.","marker":"[25]"},{"why":"Supplies the magnetic dimer structure-factor formalism used to identify the next-nearest-neighbor Mn–Te–Mn bond from the Q-dependence above TN.","marker":"[43]"},{"why":"Provides the Sn0.95Mn0.05Te example where linear Mn–Te–Mn bonds form antiferromagnetic singlet dimers, supporting the bond-geometry conclusion.","marker":"[44]"},{"why":"Supplies the stochastic spin-dynamics integrator used to compute S(Q,E) on disordered supercells for both ordered and paramagnetic states.","marker":"[46]"}],"fun_headline_variants":["Antisite defects craft a gapped optical magnon in MnSb2Te4","Neutrons link Mn-Te-Mn bonds to magnetic defect modes","Disorder-driven magnon gap detected in MnSb2Te4 crystals","Optical magnon exposes hidden magnetic sublattice in MnSb2Te4"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fitted strengths of the two magnetic bonds come from a simplified model in which the messy, partly empty Mn and Sb layers are treated as two smooth uniform sheets with reduced magnetization, and two smaller energy terms—coupling between blocks and a direction-preference term—are set to zero; if the real disorder or those neglected terms bends the measured spin-wave curves, the reported bond strengths would shift.","fun_headline_variants_meta":{"raw":{"variants":["Antisite defects craft a gapped optical magnon in MnSb2Te4","Neutrons link Mn-Te-Mn bonds to magnetic defect modes","Disorder-driven magnon gap detected in MnSb2Te4 crystals","Optical magnon exposes hidden magnetic sublattice in MnSb2Te4"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000459,"raw_usage":{"total_tokens":2315,"prompt_tokens":976,"completion_tokens":1339,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":592,"completion_tokens_details":{"reasoning_tokens":1269}},"tokens_in":592,"tokens_out":1339,"duration_ms":9664,"temperature":1.0,"reasoning_tokens":1269,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:06:35.294236+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the spin excitations in MnSb2Te4 crystals with deliberately varied antisite concentrations: if $J'$ really is the defect-induced Mn–Te–Mn bond, the optical gap should track the defect fraction $q$, the stiffness enhancement should follow $A/A_0 = (3p^2 + 8pq|J'/J_0|)/(6(p - 2q))$, and a nearly stoichiometric sample should show no optical branch. Conversely, a high-resolution $(0,0,L)$ measurement that resolves the interlayer dispersion would test the $J_c = 0$ assumption, since a finite interlayer coupling would move the branches in a way the tri-layer model cannot reproduce.","supporting_citations":[{"cited_title":"Liu, L.-L","cited_arxiv_id":null,"evidence_quote":"Provides the neutron diffraction refinement giving p = 0.588 and q = 0.129, the antisite concentrations inserted into the spin model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the high-field magnetization value Msat ≈ 1.5 µB and the earlier defect-ferrimagnetism picture used to estimate J'."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the MnBi2Te4 single-crystal inelastic neutron scattering used as the defect-free baseline for spin stiffness, gap, and dispersion."},{"cited_title":"Li, J.-Q","cited_arxiv_id":null,"evidence_quote":"Provides the competing-interaction parameters for MnBi2Te4 that anchor the intrinsic J0 and the anisotropy comparison."},{"cited_title":"Islam, Y","cited_arxiv_id":null,"evidence_quote":"Gives the dilute Mn-in-Sb2Te3 neutron result identifying antiferromagnetic Mn–Mn dimers, the direct precedent for J' and the linear-bond mechanism."},{"cited_title":"Furrer and O","cited_arxiv_id":null,"evidence_quote":"Supplies the magnetic dimer structure-factor formalism used to identify the next-nearest-neighbor Mn–Te–Mn bond from the Q-dependence above TN."},{"cited_title":"Vaknin, S","cited_arxiv_id":null,"evidence_quote":"Provides the Sn0.95Mn0.05Te example where linear Mn–Te–Mn bonds form antiferromagnetic singlet dimers, supporting the bond-geometry conclusion."}],"review_version":1}