{"id":"44183c4f-c9ad-46c0-8215-28c1db12cef7","arxiv_id":"1908.09255","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Spin waves in a nearly flat Haldane-Hubbard band become topological when the band develops dispersion, giving acoustic magnons with Chern number ±1 and chiral in-gap modes.","lead":"The paper predicts that magnons, the collective spin waves of a magnet, can be topological in an itinerant ferromagnet described by the quarter-filled Haldane-Hubbard model with a nearly flat electron band. This matters because it extends topological band theory to metallic magnets, where conventional spin-wave theory fails, and suggests new routes to chiral spin transport.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Projection onto the lower band is used at UA=UB=1.2t while the single-particle gap is Δ≈1.15t, violating the paper's own U<Δ condition for the central topological-magnon result.","rationale":"The reader and I converge on the same weak assumption. The central novelty claim is strong (\"first theoretical realization of 2D itinerant topological magnons\"), so the numerical machinery should be examined where it is least protected. The projected single-magnon ED (Eq. 2) is an exact diagonalization in a restricted Hilbert space; it is not an uncontrolled approximation for that restricted model, but the restriction itself is the issue. The paper's flatness-maximizing parameters give Δ≈1.15t and the main runs use U=1.2t, so the projection is applied outside the paper's own stated regime. This matters because the interaction term in the full model couples lower and upper electron bands at order U, and at U≈Δ the perturbative justification of projection fails. The domain-wall calculation does not rescue the bulk result: it is explicitly illustrative (footnote 54), uses U≈2-2.7t, and therefore cannot corroborate the exact parameters. I would keep the verdict CONDITIONAL because the concern is addressable: rerunning the projected ED at U<Δ, or benchmarking against full ED on a small cluster, would show whether the topological magnon band survives outside the invalid regime. I do not see a stronger internal inconsistency; the flatband exact bases and the 2x2 effective model are a genuine analytical contribution, and the Chern numbers within the projected model are internally consistent. The single decisive check is the unprojected comparison described above.","tokens_in":12358,"tokens_out":9927,"duration_ms":107928,"concrete_test":"On a 12-site cluster with t'=0.3155, φ=0.656, quarter filling, and UA=UB, perform exact diagonalization of the full (unprojected) Haldane-Hubbard Hamiltonian for U=1.2t and for U=0.8t (the latter satisfying U<Δ≈1.15t). Compare the low-energy S=1 spectrum and, via twisted boundary conditions, the Chern number of the acoustic magnon branch with the projected Eq. (2) result. If the unprojected spectrum at U=1.2t differs substantially from the projected one, or if no gapped C=-1 branch appears, the central claim relies on an invalid projection; if the two agree, or a C=-1 branch appears at U=0.8t, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the projection of H onto the lower electron band (Eq. 2), which underlies every bulk spectrum and Chern number in Figs. 2 and 4. The paper itself states the validity condition: \"the parameter space is restricted to the region where Δ is larger than both UA and UB, so that ... the whole Hamiltonian H can be projected onto it.\" The topological result of Fig. 2(a2) is obtained at t'=0.3155, φ=0.656, UA=UB=1.2t. At the Dirac points Δ=2(3t' sinφ)≈1.15t, so UA is about 4% larger than Δ. The projected Hamiltonian omits all processes that pass through the upper band; with U/Δ≈1.04, second-order corrections enter at order U^2/Δ≈1.25t, which is an order of magnitude larger than the electron bandwidth W≈Δ/7≈0.16t and is not a controlled small expansion. The only bulk-edge check (Fig. 3) is run at artificially large U in a regime the authors concede is outside the projection (footnote 54: \"the projection onto the lower electron band does not apply in this case\"). Thus, if projection error is significant at U=1.2t, neither the nonflatness-induced gap nor the C=-1 acoustic magnon band is established for the full Haldane-Hubbard model; the mass-inversion explanation rests on the same projected effective model. This is an internal inconsistency with the paper's own criterion, not a disagreement with external consensus.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies the quarter-filled Haldane-Hubbard model on the honeycomb lattice with a nearly flat lower electron band, and claims the first theoretical realization of two-dimensional itinerant topological magnons. The authors project the full Hamiltonian onto the lower electron band, obtain spin-wave excitations over the ferromagnetic ground state, and compute magnon dispersions and Chern numbers. In the flatband limit they find Dirac magnons; including the electron-band dispersion opens a topological gap, yielding an acoustic magnon band with Chern number -1. An effective 2x2 model, built from 'sublattice particle-hole vectors,' is used to attribute the nontrivial topology to a mass-inversion mechanism at K and K'. They also present domain-wall calculations showing in-gap chiral magnon modes.","tokens_in":12655,"tokens_out":3284,"duration_ms":36645,"significance":"If the central claims hold, this would be a valuable and conceptually interesting result: a concrete microscopic model for itinerant topological magnons in two dimensions, with an analytic effective description rather than only a numerical one. The paper's strengths include the exact reduction of the flatband spin-wave problem to a pair of sublattice particle-hole vectors, the resulting closed-form 2x2 effective Hamiltonian, and the explicit connection between the signs of the Dirac masses at K and K' and the magnon Chern number. The phase diagram in Fig. 2(g) is also a useful organizing summary of the different ferromagnetic and nonferromagnetic regions. However, the projection-based derivation is used outside the regime the authors themselves state is required, and the domain-wall verification is performed in a parameter regime the authors concede is outside that projection. These issues directly affect the paper's main claim, so the result is not yet established to the standard the manuscript claims.","major_comments":[{"comment":"The paper states that the parameter space is restricted to Δ larger than both U_A and U_B so that the Hamiltonian can be projected onto the lower band, but the central topological-magnon result of Fig. 2(a2) uses U_A = U_B = 1.2t with t' = 0.3155, φ = 0.656. For these parameters the single-particle gap at the Dirac points is Δ ≈ 2(3t' sinφ) ≈ 1.15t, so U_A exceeds Δ by about 4%. The projection omits all processes through the upper band, and at U/Δ ≈ 1.04 the omitted second-order corrections are of order U^2/Δ ≈ 1.25t, which is larger than the lower-band width W ≈ Δ/7 ≈ 0.16t. The perturbation expansion underlying the projected Hamiltonian is therefore not controlled at the parameters used for the main claim that nonflatness opens a topological magnon gap with Chern number -1. Please either recompute the central results for parameters satisfying Δ > U_A, U_B (for example, smaller U or larger t' with correspondingly larger Δ), or provide explicit evidence that the omitted interband processes are negligible, such as a comparison with full exact diagonalization on small clusters or a systematic second-order calculation showing small corrections.","section":"Section on the model and Eq. (2); Fig. 2(a2)"},{"comment":"The domain-wall calculation is presented as the bulk-edge correspondence check for the topological magnon bands, but it is carried out at U_A, U_B between 2.0t and 2.7t, and footnote [54] explicitly states that 'the projection onto the lower electron band does not apply in this case.' This means the in-gap magnon modes in Fig. 3 are computed in a regime the paper itself excludes from the projection validity, so they cannot serve as evidence for the existence of topological magnons at the parameters where the Chern number -1 is claimed. Please either provide domain-wall spectra within the stated projection regime (using a geometry that avoids the electronic edge-state difficulty in another way), or clearly characterize Fig. 3 as a heuristic illustration and remove it from the support for the abstract's central claim.","section":"Fig. 3 and footnote [54]"},{"comment":"The effective 2x2 Hamiltonian is derived by treating the electron-dispersion term M1 and the sublattice-imbalance term M2 as first-order perturbations to the flatband effective Hamiltonian. While this is a useful diagnostic, the perturbation parameter is not small in the regime of interest: the lower-band width W ≈ 0.16t is not negligible compared with the flatband magnon energy scale U/2 = 0.6t, and the omitted upper-band corrections of order U^2/Δ are larger still. The paper should state the precise small parameter controlling this perturbation expansion and quantify the resulting error in the mass term signs, since the mass-inversion conclusion is drawn from this first-order effective model.","section":"Supplemental Eq. (S7) and the effective model"}],"minor_comments":[{"comment":"The symbol h2(k) appears in the eigenvectors but is never defined; it should be h_y(k).","section":"Supplemental Eqs. (S2)-(S3)"},{"comment":"The derivation of M2 = U/2 in the flatband limit assumes U_A = U_B = U; the text should state this explicitly before presenting the exact bases |v^A(q)> and |v^B(q)> and the 2x2 reduction of Eq. (6), to avoid giving the impression that the exact-basis construction holds also for U_A ≠ U_B.","section":"Main text after Eq. (2)"},{"comment":"The NFM phase boundary is obtained from the projected model; since the projection itself is questionable when U exceeds Δ, the NFM boundary in the upper part of Fig. 2(g) may be an artifact of the projection. A brief comment on this limitation would be helpful.","section":"Phase diagram in Fig. 2(g)"},{"comment":"Minor typographical and wording issues: 'which is attribute to' should be 'which is attributed to'; also the phase labels TFM+ and TFM− are introduced in the phase diagram but their precise parameter ranges are not given in the text.","section":"Summary paragraph"}],"recommendation":"major_revision","confidential_remarks":"The paper's strongest contribution is the exact flatband spin-wave basis and the effective 2x2 mass-inversion description. However, the central numerical claim is computed at parameters that violate the paper's own projection condition, and the only boundary check is run in a regime the authors admit is outside the projection. These are internal consistency problems rather than disputes with external consensus. The authors should be able to address them either by shifting to parameters satisfying Δ > U or by providing a controlled unprojected check; without one of these, the claimed 'first realization of itinerant topological magnons' is not established. The novelty statement also leans heavily on the authors' own prior work (Refs. [32,44]); the editor may wish to verify that the increase over those papers is sufficiently substantial."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is a serious paper, not a crank. It claims the first 2D itinerant topological magnons, in the quarter-filled Haldane-Hubbard model with a nearly flat band, and supports that with an exact flatband basis plus an effective 2×2 model. The flatband construction is genuinely new and useful; the decomposition of the projected Hamiltonian into sublattice particle-hole vectors is clean and allows a mass analysis at K/K' without heavy numerics. The ED spectra on the projected Hamiltonian (60×60) are consistent, and the phase diagram with Chern numbers ±1/0 makes sense.\n\nThe soft spot is the reader's flagged one, and it is real. The paper states that projection onto the lower electron band requires Δ > U_A, U_B. The main topological result in Fig. 2(a2) uses t'=0.3155, φ=0.656, U_A=U_B=1.2t, while the single-particle gap at the Dirac points is Δ=2(3t' sinφ)≈1.15t. So U exceeds Δ by about 4%, and second-order processes through the upper band scale as U²/Δ≈1.25t, which is larger than the electron bandwidth W≈0.16t. That means the projected Hamiltonian may miss important corrections, and the C=-1 acoustic magnon is not established for the full Haldane-Hubbard model. The domain-wall check is explicitly outside the projection (footnote 54), so it does not resolve the issue. This is not fatal to the internal logic of the projected model, but the abstract and title claim more than what is shown.\n\nThe fix is straightforward: redo the key figures at U significantly below Δ, say U≈0.4t, and show that the gap and Chern numbers survive, or justify the projection beyond leading order. The domain-wall part should be labeled qualitative only.\n\nI see no circularity. The Chern numbers are computed from the projected magnon Hamiltonian, and the effective model is derived from the same object; the mass-inversion explanation is a diagnosis, not an input. The citation pattern is fine, including the self-citations to Ref. [32] and [44], which are directly relevant prior work.\n\nWho should read it: people working on correlated topological bands and magnon topology. It deserves a serious referee. The referee should ask for the parameter fix, but the core idea and the flatband basis are worth engaging.","headline":"Serious and potentially important paper, but the main topological result sits just outside the paper's own projection validity, so it needs a parameter fix before I'd trust it for the full model.","tokens_in":13212,"tokens_out":2028,"would_cite":true,"duration_ms":19913,"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 claims that the slight curvature of a nearly flat electron band opens a topological gap at the Dirac points and turns the acoustic magnons of a quarter-filled Haldane-Hubbard model into the first two-dimensional itinerant…","keywords":["itinerant topological magnons","Haldane-Hubbard model","nearly-flat electron band","Chern number","mass inversion mechanism","Dirac magnons","projected exact diagonalization","domain-wall magnon modes"],"falsifier":"A numerical spectrum of the full unprojected Haldane-Hubbard model on a finite cluster with $t'=0.3155$, $\\phi=0.656$, and $U_A=U_B=1.2t$ would settle the point: if the acoustic magnon band no longer shows a gap at the K and K′ points with Chern number $-1$, the projection itself produced the topology.","tokens_in":12140,"feed_emoji":"🧲","tokens_out":9506,"duration_ms":88831,"temperature":0.7,"pith_summary":"The paper sets out to show that the collective spin excitations of an itinerant ferromagnet can carry a nontrivial band topology. Its setting is the quarter-filled Haldane-Hubbard model on a honeycomb lattice, tuned so the lower electron band is nearly flat. In the flatband limit the magnon bands touch at the Dirac points, and the paper's central claim is that the electron band's small curvature opens a topological gap there, giving the acoustic magnon band Chern number $-1$. If correct, this is the first theoretical realization of two-dimensional itinerant topological magnons, a phenomenon previously studied only in local-spin magnets.","feed_headline":"Band curvature creates topological magnons in itinerant magnets","feed_subtitle":"A slightly curved lower electron band opens a Chern −1 magnon gap at the Dirac points.","key_machinery":"The load-bearing object is the sublattice particle-hole vector $|v^a_i(q)\\rangle=\\sqrt{U_a/N}\\,\\mu^*_{a k_i-q\\downarrow}\\mu_{a k_i\\uparrow}$, which measures the amplitude to create a spin-1 particle-hole pair in the lower electron band on sublattice $a$. The projected interaction matrix $M^3_{ji}(q)$ splits into a sum of projectors onto these vectors, so in the flatband limit the spin-wave problem reduces to a $2\\times2$ matrix spanned by them. The nonflatness of the electron band enters through the kinetic piece $M^1_i(q)$ and supplies opposite-signed mass terms at the K and K′ points; this mass inversion, familiar from the electron Haldane model, is the mechanism that makes the magnon band topological.","core_discovery":"Working in a projected basis where each spin-1 excitation is created by $d^\\dagger_{k_i-q,\\downarrow}d_{k_i,\\uparrow}|FM\\rangle$, the paper finds an exact description of the magnons in the flatband limit as a $2\\times2$ Dirac-like Hamiltonian, with massless nodes at K and K′. Including the dispersion of the lower electron band adds terms that act as Dirac masses with opposite signs at the two nodes; by the mass inversion mechanism, the acoustic magnon band acquires Chern number $-1$. The paper further shows that a sublattice Hubbard imbalance closes and reopens the magnon gap, changing the Chern number from $-1$ to $0$, while tuning the next-nearest-neighbor hopping flips it between $+1$ and $-1$ with the electron band topology unchanged. Domain walls between regions with different magnon Chern numbers host chiral in-gap modes, with one mode per unit of Chern difference.","pith_inferences":["A natural test of the mechanism is to repeat the projected spin-wave calculation in other nearly-flat Chern bands, such as kagome or checkerboard lattice models, and ask whether the same band-curvature mass inversion appears.","Because the topological gap is controlled by band curvature rather than by spin-orbit-type local interactions, fine-tuning the flatness in an optical-lattice or moiré realization could switch the magnon Chern number and the sign of a thermal Hall response without changing the electron band topology.","If the projection is robust slightly beyond its formal regime, the same effective mass-term picture may survive in moderately correlated metals, connecting itinerant topological magnons to Fermi-surface instabilities."],"forward_implications":["If the central claim is right, two-dimensional itinerant magnets can host topological magnons without Dzyaloshinskii-Moriya or other local spin-spin interactions; the curvature of the electron band is sufficient.","The predicted phase diagram contains a trivial ferromagnet and two topological ferromagnetic phases with Chern numbers $\\pm1$, separated by lines where the magnon gap closes and reopens as $\\Delta U=U_A-U_B$ or $t'$ is tuned.","Magnetic domain walls between regions with different magnon Chern numbers should carry chiral in-gap magnon states, and their count equals the Chern-number difference between the two sides.","The magnon Chern number can change while the electron band topology stays fixed, so the spin-wave topology is not a simple copy of the underlying electron band topology."],"supporting_citations":[{"why":"Supplies the single-particle Haldane model whose complex next-nearest-neighbor hopping realizes the mass inversion mechanism that the paper transfers to magnons.","marker":"[11]"},{"why":"Provides the nearly-flat-band tuning condition and the projection onto the lower band that the calculation relies on.","marker":"[23]"},{"why":"Supplies the projected spin-wave formalism and the itinerant ferromagnetic ground-state stability analysis used to obtain the magnon spectra.","marker":"[32]"},{"why":"Documents the lack of an effective linear-spin-wave description for itinerant magnets, the gap the paper's 2x2 effective model fills.","marker":"[44]"},{"why":"Establishes the ferromagnetic ground-state criterion for the quarter-filled band on which the excitation calculation is built.","marker":"[53]"}],"fun_headline_variants":["Band curvature flips magnon Chern number in Hubbard","Tuning electron band curvature drives topological magnons","Mass inversion triggers Chern -1 magnons in near-flat band","Curved band yields topological magnons via mass inversion"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation projects the full Hamiltonian onto the lower electron band, which is valid only when the Hubbard interactions are smaller than the electron band gap; the headline parameters use $U_A=1.2t$ with a gap of roughly $1.15t$, placing the main results at the edge of that assumption.","fun_headline_variants_meta":{"raw":{"variants":["Band curvature flips magnon Chern number in Hubbard","Tuning electron band curvature drives topological magnons","Mass inversion triggers Chern -1 magnons in near-flat band","Curved band yields topological magnons via mass inversion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001221,"raw_usage":{"total_tokens":5007,"prompt_tokens":916,"completion_tokens":4091,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":532,"completion_tokens_details":{"reasoning_tokens":4026}},"tokens_in":532,"tokens_out":4091,"duration_ms":27540,"temperature":1.0,"reasoning_tokens":4026,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:17:13.880344+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A numerical spectrum of the full unprojected Haldane-Hubbard model on a finite cluster with $t'=0.3155$, $\\phi=0.656$, and $U_A=U_B=1.2t$ would settle the point: if the acoustic magnon band no longer shows a gap at the K and K′ points with Chern number $-1$, the projection itself produced the topology.","supporting_citations":[{"cited_title":"Su, Z.-L","cited_arxiv_id":null,"evidence_quote":"Supplies the projected spin-wave formalism and the itinerant ferromagnetic ground-state stability analysis used to obtain the magnon spectra."},{"cited_title":"Su, Z.-L","cited_arxiv_id":null,"evidence_quote":"Documents the lack of an effective linear-spin-wave description for itinerant magnets, the gap the paper's 2x2 effective model fills."},{"cited_title":"Tasaki, Phys","cited_arxiv_id":null,"evidence_quote":"Establishes the ferromagnetic ground-state criterion for the quarter-filled band on which the excitation calculation is built."}],"review_version":1}