{"id":"2ae03e52-9cfe-470b-9998-93358e075e53","arxiv_id":"2602.00251","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A tight-binding Kubo calculation predicts gate-, Fermi-level-, and magnetization-tunable circular photogalvanic currents in Bi2Se3 slabs, with a mid-IR responsivity of about 0.17 μA/W.","lead":"This paper calculates how circularly polarized light generates electric current in a thin film of the topological insulator Bi2Se3, using a slab model that includes both surface and bulk states. It predicts the photocurrent can be tuned by gate voltage, Fermi level, and a magnetic layer, with a predicted mid-infrared responsivity near 0.17 μA/W.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section II.C's layer-resolved decomposition identity is unproved and likely false: G is full in layer space, so cross-layer projections do not vanish; this undermines the top/bottom CPGE separation and the R≈0.169 μA/W estimate.","rationale":"The reader's weakest assumption targets exactly Section II.C's layer-resolved cancellation, and I agree that this is the most load-bearing gap. The total Kubo formula in Eq. (14)/(22) may still be a valid route to injection currents, and the qualitative trends are plausible, but the paper's quantitative results—the top/bottom decomposition, the sign-flip/cancellation arguments, the magnetic-proximity surface-resolved maps, and the 0.169 μA/W responsivity—are all filtered through this unproved identity. The algebra is not merely missing a detail: the stated justification (G is block-tridiagonal) is incorrect, and the double-projection scheme is conceptually suspect. This is more specific than the abstract's unsupported σ_xyz mention or the absence of code, which are secondary reportability issues. I would not move the verdict to REJECT because the core Kubo derivation and the symmetry arguments may survive if the layer projection is corrected; however, the current conditional status is appropriate and should be retained until the identity is repaired or the affected results are recomputed under a defensible projection.","tokens_in":17984,"tokens_out":8127,"duration_ms":100976,"concrete_test":"For the 4-band Bi2Se3 slab with a small N_z (e.g., N_z=4) at a representative k-point and a Matsubara frequency with δ=0.005 meV, compute exactly the difference Δ = Tr[v_µ G v_α G v_β G] − Σ_l Tr[v_µ G (P_l v_α) G (P_l v_β) G] for one component, say µαβ=xxz, at ℏω=0.3 eV. If Δ≠0, the Section II.C cancellation is false. Then compute the alternative layer-resolved decomposition with only the current vertex projected, Σ_l Tr[(P_l v_µ) G v_α G v_β G], and verify whether it equals the total trace. If the double-projection sum deviates from the total by more than ~1%, the top/bottom CPGE and the R estimate based on the first two layers must be re-derived with the correct projection scheme.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claims—especially the top/bottom CPGE separation and the responsivity R≈0.169 μA/W attributed to the first two layers—rest on the layer-resolved decomposition in Section II.C. Equation (24) inserts projections P_l onto both v_α and v_β inside the triangle trace, and the text asserts that cross-layer terms vanish because \"the product of the last two terms ... would be zero if non-identical projection layers were used\" due to the \"block tri-diagonal form of G(k,iω) and v_α(k).\" This is not correct. Only H (and therefore v_α, the derivative of H) is block-tridiagonal; G(k,iω)=(iω−H)^{-1} is generally a full matrix in layer space even when H is tridiagonal. Consequently, terms like (P_l1 v_α) G (P_l2 v_β) with l1≠l2 do not vanish: v_α connects only neighboring layers, but G propagates the resulting amplitude to arbitrary layers. Thus Σ_l σ^Δ_{µαβ,l} ≠ σ^Δ_{µαβ} in general. Moreover, projecting both optical velocity vertices is physically questionable: a layer-resolved injection current should project the measured current vertex v_µ, not the optical transition vertices; the latter discards interlayer interference. The top/bottom results, the magnetic-proximity surface decomposition, and the responsivity estimate are therefore not cleanly derived from the total Kubo formula. The manuscript's own convergence caveat at ℏω→0 does not repair this algebraic step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a nonlinear Kubo-formula calculation of the second-order injection conductivity σ^inj_{µαβ}(ω) for a Bi2Se3 slab described by a symmetry-informed tight-binding model. The authors compute the circular and linear photogalvanic currents (Im σ_xxz and Re σ_xxy), study dependence on gate voltage, Fermi level, photon energy, and a magnetic proximity layer, and report a mid-IR responsivity R≈0.169 μA/W. They also introduce a layer-projected decomposition intended to isolate top/bottom surface contributions and use a phenomenological '8-band' model to study chirality-pairing effects.","tokens_in":18384,"tokens_out":9194,"duration_ms":103836,"significance":"This manuscript addresses an important gap: extending nonlinear optical-conductivity calculations for TIs beyond continuum models by using a slab tight-binding Hamiltonian fitted to Bi2Se3's band structure. The Kubo/injection-current framework (Eqs. (12)-(14)) is standard, and the paper is careful to avoid the numerically fragile low-photon-energy region. If the layer-resolved decomposition were correct, the predicted gate-tunable CPGE, the magnetization ridge, and the chirality dependence of the surface currents would be valuable and falsifiable. It is a strength that the tight-binding parameters come from band-structure fitting and that no target conductivity is used to set them. However, the central layer-projection sum rule is not established and is likely false as stated, so the quantitative surface-specific results and the device responsivity estimate are not yet supported.","major_comments":[{"comment":"The sum-rule claim Σ_l σ^Δ_{µαβ,l} = σ^Δ_{µαβ} is not justified. Only H(k) (and thus v_α = ∂H/∂k_α) is block-tridiagonal; the Matsubara Green's function G(k,iω)=(iω−H(k))^{-1} is generally full in layer space (e.g., any finite 1D chain). Hence terms P_{l1}v_α G P_{l2}v_β with l1≠l2 need not vanish. The layer-resolved quantities in Eq. (24) (with two projected optical vertices) are therefore not a proven decomposition of the total injection conductivity. Because Eq. (25), the top/bottom cancellation argument, and the responsivity R=0.169 μA/W attributed to the first two layers rest on this identity, the central quantitative surface-specific results require a proof or numerical verification of the sum rule, or a reformulation (e.g., projecting only the current vertex v_µ).","section":"Section II.C, Eq. (24)"},{"comment":"The claim that magnetization 'enables nonlinear conductivity tensor components (e.g., σ_xyz) that are normally forbidden by symmetry' is not backed by a computed σ_xyz or any other normally forbidden component. The magnetic-proximity results compute only Im σ_xxz. Either provide the forbidden-component calculation or revise the abstract to describe what is actually computed.","section":"Abstract and Section IV.C"}],"minor_comments":[{"comment":"The responsivity is quoted as R≈0.170 μA/W in the abstract and R=0.169 μA/W in the text. Reconcile and specify the photon energy and parameters used for the quoted value.","section":"Abstract vs Section IV.A"},{"comment":"The fermionic Matsubara frequency is written ω_l = (2n+l)πT; the standard notation is (2l+1)πT. Please correct.","section":"Section II.B, after Eq. (12)"},{"comment":"Typo: 'sutdies' should be 'studies'.","section":"Fig. 7 caption"},{"comment":"The main text does not define the constant σ0 used in the figure color scales beyond the figure captions; please state it explicitly in the text or in a common legend.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The key issue is the layer-decomposition sum rule in Section II.C. I believe the identity is false as stated, but if the authors can prove it or numerically verify it, or if they reframe the layer-resolved calculation as a well-defined projected quantity and also report the total σ, the manuscript could become publishable. The σ_xyz abstract claim also needs to be reconciled with the computed σ_xxz. Otherwise, the paper's quantitative surface-specific claims are not supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one. It's a tight-binding slab calculation of injection current contributions to CPGE/LPGE in Bi2Se3, with a layer-resolved twist and a magnetic-proximity knob. The unprojected Kubo results—gate voltage dependence, Fermi-level maps, the M_z-dependent resonance ridge—are interesting, broadly consistent with experiment, and do not look like fitting targets. The 8-band chirality-pairing analysis is speculative but clearly flagged as such.\n\nThe problem is Section II.C. The paper inserts projection operators onto both optical velocity vertices in the triangle diagram and claims the sum over layers equals the full conductivity because G and v_α are block-tridiagonal. That's wrong. Only H is block-tridiagonal; G is full in layer space. So cross-layer terms do not vanish. A proper layer-resolved injection current should project the measured current vertex, not both optical transition vertices. As written, the top/bottom separation and the R≈0.169 μA/W estimate attributed to the first two layers are not cleanly derived from the Kubo formula. The abstract's mention of σ_xyz is also not backed by any computed tensor element shown.\n\nIf you strip out the layer-resolved claims, the core framework still has value. The gate/Fermi/photon-energy trends are sensible, and the magnetic-proximity tuning is a nice suggestion. But the central quantitative promise—layer-resolved, surface-specific CPGE with a specific responsivity—needs a methodological fix before it can be taken at face value. No code or data shipped, so reproducibility is limited, though that's not unusual for this type of theory paper.\n\nThis deserves a serious referee: the topic is current, the approach is useful, and the flaw is addressable. I'd send it to review with a request to focus on Sec. II.C, and I would not cite it until that's resolved.","headline":"Promising TB-slab framework for CPGE, but the layer-resolved decomposition rests on a false identity, so the quantitative claims—including the responsivity—are not currently supported.","tokens_in":18889,"tokens_out":3212,"would_cite":false,"duration_ms":37317,"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":"A tight-binding slab model of Bi2Se3 computes the circular photogalvanic effect from first principles, including gate-voltage and magnetic-proximity tuning, and predicts a mid-IR responsivity near 0.17 μA/W.","keywords":["circular photogalvanic effect","topological insulator","Bi2Se3 slab","injection current","nonlinear optical conductivity","Kubo formalism","layer-resolved photocurrent","magnetic proximity"],"falsifier":"Compute σ^inj from the full slab Hamiltonian and compare it numerically with the layer-projected sum Σ_l σ^∆_{µαβ,l}; if they differ at the precision quoted, the central decomposition fails. Alternatively, measure bottom-surface CPGE in a magnetized-top-layer experiment: if the bottom contribution is not simply the sign-flipped top response when M_z is confined to the top surface, the symmetry argument breaks.","tokens_in":17859,"feed_emoji":"⚛️","tokens_out":4905,"duration_ms":53509,"temperature":0.7,"pith_summary":"The paper tries to show that the circular photogalvanic effect (CPGE) in a topological insulator can be computed directly from a slab tight-binding Hamiltonian using a nonlinear Kubo formula, without resorting to continuum models or DFT-fitting. It claims that this approach captures bulk-to-surface interband transitions a priori and yields layer-resolved contributions, so one can see how the top and bottom surfaces contribute opposite CPGE currents that cancel unless inversion or time-reversal symmetry is broken. The authors then use this tool to predict that gate voltage, Fermi level, photon energy, surface-state chirality, and an out-of-plane magnetic exchange field all tune CPGE, with a mid-wave infrared responsivity of roughly 0.17 μA/W that is comparable to reported topological photodetectors. If correct, the work gives a parameter-free route to designing chiral photodetectors and polarization-sensitive devices operating at room temperature in the THz-to-mid-IR range.","feed_headline":"Slab model captures gate-tunable chiral photocurrents in Bi2Se3","feed_subtitle":"Layer-resolved CPGE from a slab model; mid-IR responsivity near 0.17 μA/W without continuum fits.","key_machinery":"The central object is the injection-current contribution to the second-order optical conductivity tensor, σ^inj_{µαβ}, derived from the nonlinear Kubo formula with two photon-electron vertices on a single Green's function. Layer resolution is achieved by inserting projection operators P_l into the velocity vertices, so that the contribution from a given layer can be isolated; the block-tridiagonal structure of the slab Hamiltonian is the key assumption that makes this decomposition exact. The Bi2Se3 slab is described by a 4N_z × 4N_z tight-binding Hamiltonian built from four Dirac matrices, with an in-plane on-site part and interlayer hopping; an eight-band 'doubled' Hamiltonian is introduce","core_discovery":"The central claim is that the injection-current part of the second-order optical conductivity of a topological insulator can be evaluated in a slab geometry from a symmetry-informed tight-binding model, using the nonlinear Kubo formula with layer-resolved projection operators. In contrast to earlier continuum or DFT-based calculations, this slab approach inherently includes all interband transitions between bulk and surface states and allows the authors to separate the top and bottom surface contributions to the CPGE. They find that the top and bottom surface CPGEs have opposite signs and cancel in an inversion-symmetric slab; applying a gate voltage breaks inversion and resonantly enhances","pith_inferences":["The layer-decomposition identity (Σ_l σ^Δ_{µαβ,l} = σ^Δ_{µαβ}) is load-bearing; if interlayer velocity matrix elements or off-diagonal Green's function blocks make cross-layer terms nonzero, the top/bottom separation becomes approximate. A direct numerical comparison of the full tensor with the layer sum would settle this.","The formalism should transfer to other 3D topological insulators (Bi2Te3, Sb2Te3) by adjusting the tight-binding parameters; stronger hexagonal warping is expected to raise the linear photogalvanic (LPGE) response, which the present model predicts to be weak.","The magnetic-proximity prediction can be tested directly in a ferromagnet/TI heterostructure: varying the exchange strength should shift the CPGE resonance peak as ℏω ≈ 2|M_z| and suppress the low-energy response, a signature that would confirm the Berry-curvature redistribution mechanism.","The eight-band chirality-pairing argument suggests a design principle for selecting materials: ultraviolet or visible excitation between same-chirality surface states, such as in Bi2Te2Se, should produce enhanced CPGE with reduced Fermi-level sensitivity."],"forward_implications":["Gate voltage can act as a switch for the ratio of circular to linear photogalvanic current, because inversion breaking changes Im σ_xxz and Re σ_xxy differently.","Proximitizing a magnetic layer on one surface gives a practical, room-temperature knob for tuning or reversing CPGE, which had not been explored experimentally as a CPGE control.","Surface states with matched chirality yield additive CPGE contributions across the slab, suggesting that materials with same-chirality surface bands (e.g., certain Bi2Te2Se configurations) will have stronger and more robust circular photocurrents.","The slab approach captures bulk-surface transitions a priori, so it can predict photocurrents from terahertz to mid-infrared frequencies without manually enumerating transition channels.","The predicted responsivity near 0.17 μA/W indicates that measurable CPGE detection is possible without cryogenic cooling, aligning with several experimental TI photodetectors."],"fun_headline_variants":["Slab model reveals gate-tunable chiral photocurrents","Layer-resolved CPGE: gate voltage breaks inversion in Bi2Se3","Mid-IR responsivity 0.17 μA/W from slab model of Bi2Se3","Top and bottom surfaces cancel until gate voltage is applied","New slab approach computes circular photogalvanic effect"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper assumes that the layer-resolved contributions to the injection-current tensor add up exactly to the full tensor because the Green's function and velocity operators are block-tridiagonal, so all cross-layer terms in the trace vanish; if this cancellation is not exact, the top/bottom CPGE separation is not a clean decomposition of the total injection current.","fun_headline_variants_meta":{"raw":{"variants":["Slab model reveals gate-tunable chiral photocurrents","Layer-resolved CPGE: gate voltage breaks inversion in Bi2Se3","Mid-IR responsivity 0.17 μA/W from slab model of Bi2Se3","Top and bottom surfaces cancel until gate voltage is applied","New slab approach computes circular photogalvanic effect"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000281,"raw_usage":{"total_tokens":1547,"prompt_tokens":837,"completion_tokens":710,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":581,"completion_tokens_details":{"reasoning_tokens":616}},"tokens_in":581,"tokens_out":710,"duration_ms":7472,"temperature":1.0,"reasoning_tokens":616,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T06:06:25.906977+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute σ^inj from the full slab Hamiltonian and compare it numerically with the layer-projected sum Σ_l σ^∆_{µαβ,l}; if they differ at the precision quoted, the central decomposition fails. Alternatively, measure bottom-surface CPGE in a magnetized-top-layer experiment: if the bottom contribution is not simply the sign-flipped top response when M_z is confined to the top surface, the symmetry argument breaks.","supporting_citations":[],"review_version":1}