{"id":"2ce617a5-595e-4b7c-b245-1160862a7134","arxiv_id":"2607.28310","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"Ag spacers raise Bi charge-to-spin conversion to bulk-theory levels by preserving Bi integrity rather than via Bi/Ag Rashba coupling, with bulk spin Hall angle ≈1.","lead":"Ag spacers between epitaxial Bi and ferromagnets boost damping-like spin-orbit torque by over 10× to ~2×10⁵ (ℏ/2e) S/m, matching bulk-Bi theory. The gain comes from protecting Bi’s structure, not from Bi/Ag Rashba coupling, clarifying a long-standing controversy and guiding Bi-based SOT devices.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The manuscript supplies a coherent multi-technique structure–property map that directly ties the order-of-magnitude SOT enhancement to the presence/absence of solid-state dewetting, alloying and oxidation. The inverted Ni/Bi(012) control (robust SOT without any Ag) and the Cu/Al spacer failures are especially decisive against a Bi/Ag-interface-origin story. The interpretive softness around residual interface transparency after oxidation and the use of a single-layer conductivity for the effective SHA are acknowledged limitations, not hidden contradictions. They do not falsify or critically weaken the strongest claim. Reader’s ACCEPT / high-confidence verdict is therefore left unchanged.","tokens_in":28734,"tokens_out":578,"duration_ms":12243,"concrete_test":"Compare ξ_E_DL of Bi(001)/Ag/FeCo against an otherwise identical stack in which a thin (~1–2 nm) amorphous or polycrystalline Ag spacer is replaced by a lattice-matched, spin-transparent spacer known not to form Rashba alloys with Bi (if available) or, more simply, re-measure the oxidized Bi(001)+O₂/Ag/FeCo series at two additional O₂ doses bracketing 4.5 L and 360 L while confirming surface-state quenching by in-situ ARPES on sister samples; if ξ_E_DL remains within ~30% of the pristine value once surface states are gone, the bulk-dominance reading is reinforced.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption (that residual SOT after 360 L O₂ plus ARPES absence of ordered BiAg₂ cleanly establishes bulk dominance while treating Ag as a passive stabilizer) is real but already flagged by the authors and is not load-bearing for the central claim. The claim that Ag enhances ξ_E_DL by >10× primarily by preserving Bi integrity (rather than by generating a dominant Bi/Ag REE) is independently supported by: (i) STEM/EDX discontinuity, roughening, oxidation and AlOx intrusion without Ag; (ii) Bi–Ni alloying and (001)→(012) conversion without Ag; (iii) failure of Cu/Al spacers; (iv) ~60% drop for polycrystalline Bi; (v) sizable positive SOT in inverted Ni/Bi(012) that is unchanged by adding Ag on top; and (vi) ARPES showing no BiAg₂ alloy or enhanced Rashba splitting. The oxidation experiment is only one supporting strand and is not required for the integrity-not-Rashba conclusion. SHA≈1 via single-layer σ_Bi is an effective lower-bound estimate, not a precision intrinsic value, and is presented as such.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript investigates charge-to-spin conversion in epitaxial Bi(001), polycrystalline Bi, and Bi(012)-textured magnetic heterostructures, with and without Ag (and Cu/Al) spacers, using polar MOKE magnetometry of the damping-like SOT together with HAADF-STEM/EDX, XRD, RHEED, AFM, ARPES, SQUID, and harmonic Hall comparisons. The central claim is that inserting an Ag spacer between Bi(001) and FeCo or Ni raises the effective spin Hall conductivity by more than an order of magnitude to ≈(1.7–1.9)×10⁵ S/m, matching bulk Bi theory, because Ag preserves the structural and chemical integrity of Bi rather than because of a dominant Rashba–Edelstein contribution at Bi/Ag. Supporting strands include solid-state dewetting/oxidation/alloying without Ag, failure of Cu/Al spacers, a ~60% drop for polycrystalline Bi, ARPES showing no ordered BiAg₂ or enhanced Rashba splitting, residual torque after a 360 L O₂ dose that destroys Bi(001) surface states, and a sizable positive SOT in inverted Ni/Bi(012) that is unchanged by adding Ag on top. An effective Bi spin Hall angle of order 1 is extracted using the conductivity of a Bi single layer.","tokens_in":28998,"tokens_out":1283,"duration_ms":36005,"significance":"If the result holds, the work supplies a coherent structural explanation for the long-standing scatter in Bi and Bi/Ag spin–charge interconversion efficiencies and supplies concrete growth/design rules (Ag as a chemically inert, spin-transparent diffusion barrier; preference for epitaxial Bi integrity) for Bi-based SOT devices. Strengths include orthogonal probes on the same device set, E-field-normalized ξᴱ_DL that avoids dubious current partitioning, explicit comparison of MOKE versus harmonic Hall (including magnetothermal/magnonic artifacts), and ARPES that directly tests the BiAg₂ scenario. The correlation of atomic-scale integrity with torque magnitude is a clear advance over prior spin-pumping and THz studies that often lacked depth-sensitive structural characterization.","major_comments":[{"comment":"Sec. III.D.2 and IV.B–C: The 360 L O₂ experiment is presented as evidence for bulk-dominated generation (28% drop, residual ξᴱ_DL ≈ 1.22×10⁵ S/m). The manuscript correctly notes that oxidation can also reduce interface spin transparency, so the residual cannot cleanly partition surface versus bulk. For the SHA estimate ξ^{j}_Bi_DL = ξᴱ_DL/σ_Bi ≈ 0.78–1.2, please state more explicitly in the main text (not only SM/discussion) that σ_Bi is taken from a separate AlOₓ-capped Bi(001) single layer and that the quoted SHA is therefore an effective lower-bound figure of merit, not an intrinsic bulk SHA after interface transparency correction. A short sensitivity check (range of σ_Bi or parallel-resistor bounds) would make the claim robust without changing the integrity-not-Rashba conclusion.","section":"Sec. III.D.2, IV.C"},{"comment":"Body text vs. figure numbering: In Sec. III.D the MOKE line scans, B_DL(j) plots, and ξᴱ_DL bar chart are repeatedly cited as “Figures 6(b–h)” and “Fig. 6”, yet the corresponding caption is FIG. 4 and the summary table is FIG. 6. This mis-citation makes the central SOT data hard to locate and must be corrected before publication; it is load-bearing for readability of the main result.","section":"Sec. III.D, Figs. 4 and 6"}],"minor_comments":[{"comment":"Abstract and Sec. V state an effective Bi spin Hall angle of “approximately 1”; Sec. IV.C gives both 1.2±0.2 (average of FeCo/Ni) and 0.78±0.03 (oxidized). Align the abstract wording with the more conservative oxidized lower bound or quote the range.","section":"Abstract, Sec. IV.C, Sec. V"},{"comment":"Fig. 6 summary table: the optical micrographs are useful but low-resolution in the manuscript rendering; ensure final production quality so that pinholes versus continuous films remain distinguishable.","section":"Fig. 6"},{"comment":"Sec. III.E.2: The statement that a spin diffusion length ≪ 3 nm in Bi(012) cannot be excluded is appropriate; a brief citation or estimate of λ_sf from related Bi literature would help the reader judge whether the null effect of the top Ag layer is expected.","section":"Sec. III.E.2"},{"comment":"Typographical consistency: “Bi 0.9Sb0.1” vs “Bi0.9Sb0.1”, “¯h/2e” formatting, and occasional missing spaces before units appear in several places; a copy-edit pass is warranted.","section":null},{"comment":"Note 4 of the SM is cited for the MOKE–HHR discrepancy; a one-sentence quantitative summary (e.g., factor of misestimation) in the main text would help readers who do not open the SM.","section":"Sec. III.D"}],"recommendation":"minor_revision","confidential_remarks":"The work is thorough and suitable for a high-quality condensed-matter/materials journal. The figure-numbering slip in Sec. III.D is the only issue that currently impedes review; once fixed and the SHA language tightened, accept is appropriate. No concerns about novelty disclosure or scope."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The punchline is simple: the huge scatter in Bi and Bi/Ag spin–charge numbers is mostly sample integrity, not mysterious interface physics. Put Ag between epitaxial Bi(001) and FeCo or Ni and you get ξ_E_DL ≈ 1.7–1.9 × 10^5 S/m, matching bulk Bi theory; drop the Ag and the Bi dewets, alloys, oxidizes, or reorients, and the torque collapses or flips sign.\n\nWhat is actually new is the controlled matrix—epitaxial vs poly, oxidized surface, inverted Ni/Bi(012), Cu/Al spacers—tied to the same MOKE line-scan protocol plus STEM/EDX, XRD, RHEED, AFM, and ARPES on the same stacks. ARPES shows Ag-on-Bi(001) does not form ordered BiAg2 or boost the Rashba splitting. Inverted Ni/Bi(012) already gives a solid positive torque that does not change when Ag is added on top. Cu and Al fail as spacers. Polycrystalline Bi drops ~60%. That package is stronger than another spin-pumping Bi/Ag paper.\n\nThey chose MOKE for the right reason: Bi’s Nernst and Ni’s magnons make HHR messy; the SM comparison shows it. Normalizing to E rather than a guessed j_Bi is the honest move given the microstructure mess. Data and citations look clean; self-cites are to their own BiSb and MOKE methods and are used as benchmarks, not crutches.\n\nSoft spots are real but secondary. The 360 L O2 dose only cuts torque ~28%; they themselves note oxidation can also hurt spin transparency, so it is not a clean bulk/surface partition. SHA ≈ 1 via single-layer σ_Bi is an effective number, not an intrinsic precision value. Neither undercuts the central integrity-not-Rashba claim, which stands on the STEM discontinuity, alloying, failed alternate spacers, poly drop, and inverted stack.\n\nThis is for people who grow or measure SOT stacks with Bi, BiSb, or topological-insulator/FM interfaces. Worth a serious referee. I would engage.","headline":"Solid structure–torque map that largely settles the Bi/Ag controversy: Ag works as a diffusion barrier, not a Rashba engine.","tokens_in":29701,"tokens_out":547,"would_cite":true,"duration_ms":11713,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Silver multiplies bismuth’s spin-orbit torque by more than ten by keeping the bismuth film intact, not by Rashba interface physics.","keywords":["spin-orbit torque","bismuth","spin Hall effect","Bi/Ag interface","charge-to-spin conversion","magneto-optic Kerr effect","epitaxial thin films","spin Hall angle"],"falsifier":"Grow a Bi/Ag/FM stack in which Bi is demonstrably continuous and unalloyed, then deliberately destroy only the Bi bulk spin-Hall channel (for example by amorphization or extreme thinning while keeping the Bi/Ag interface) and check whether the large positive damping-like torque disappears; if it survives, the bulk-origin claim fails.","tokens_in":29599,"feed_emoji":"🧲","tokens_out":1052,"duration_ms":29651,"temperature":0.7,"pith_summary":"Bismuth is predicted to convert charge current into spin current about as efficiently as platinum, yet experiments disagree wildly and the large signals seen in Bi/Ag stacks are usually blamed on a Rashba interface. This work shows that the real bottleneck is structural: when a ferromagnet is deposited directly on Bi(001), the bismuth dewets, alloys, oxidizes or reorients, collapsing the torque. A thin silver spacer stops that damage, restores an effective spin Hall conductivity near 2×10⁵ (ℏ/2e) S/m—matching bulk theory—and leaves the Bi surface states essentially unchanged. Oxygen doses known to wipe out those surface states cut the torque by only ~28 %, so the spin current is mostly bulk-generated. The practical message is that reliable Bi-based spin-orbit-torque devices require an atomic-scale barrier that preserves Bi integrity rather than an engineered Rashba interface.","feed_headline":"Silver rescues bismuth’s spin torque by keeping the film intact","feed_subtitle":"Ag spacer multiplies damping-like SOT tenfold; bulk Bi, not Rashba interface, supplies the spin current","key_machinery":"Scanning polar MOKE magnetometry of the damping-like effective field, calibrated against the Oersted field and free of magnetothermal artifacts, correlated sample-by-sample with HAADF-STEM, EDX, XRD and ARPES that track whether Bi remains continuous, unalloyed and chemically intact.","core_discovery":"Inserting an Ag spacer between epitaxial Bi(001) and FeCo or Ni raises the damping-like spin-orbit-torque efficiency by more than an order of magnitude to ~2×10⁵ (ℏ/2e) S/m. The gain comes from preserving the structural and chemical integrity of Bi, which is otherwise destroyed by direct contact with the ferromagnet; it is not produced by enhanced Rashba spin–orbit coupling at the Bi/Ag interface. Bulk spin-Hall generation in intact Bi therefore dominates, giving an effective Bi spin Hall angle of order 1.","pith_inferences":["Any future claim of giant interfacial REE in Bi/Ag must first prove, with cross-sectional chemistry, that Bi has not simply been rescued from degradation.","The same integrity-first logic likely applies to other low-melting, surfactant-prone spin-orbit metals (e.g., pure Sb or Bi-rich alloys) whose reported torques are equally scattered.","If Ag’s only essential role is diffusion blocking, thinner or discontinuous Ag, or alternative inert barriers with still longer spin diffusion length, should recover the same bulk Bi torque.","Harmonic Hall methods on Bi stacks will continue to mis-estimate damping-like torque unless magnetothermal and magnon backgrounds are independently subtracted or MOKE-style checks are added."],"forward_implications":["Bi-based SOT devices need a chemically inert, spin-transparent spacer (Ag works; Cu and Al do not) rather than direct FM contact.","Literature scatter in Bi and Bi/Ag spin-to-charge efficiencies is largely explained by uncontrolled dewetting, alloying and oxidation, not by intrinsic material differences.","Polycrystalline or rough Bi films deliver substantially lower torque than epitaxial Bi(001), so crystalline quality is a first-order design variable.","An effective Bi spin Hall angle near 1 becomes a realistic target once structural integrity is secured, comparable to the best topological-insulator reports.","Stacking order matters: depositing Bi on Ni can preserve integrity where depositing Ni on Bi does not."],"fun_headline_variants":["Ag spacer lifts Bi SOT tenfold by keeping the film intact","Intact Bi yields bulk spin Hall angle ~1 with Ag shield","Ag multiplies Bi damping-like torque by blocking FM damage","Bi spin current jumps when Ag preserves epitaxial order","Ag-protected Bi hits 2×10⁵ spin Hall conductivity from bulk"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The claim that the large torque is bulk-dominated rests on the idea that a heavy oxygen dose mainly removes surface states while leaving bulk generation and spin transparency essentially intact—yet the same oxidation can also degrade interface transparency, so the residual signal is not a clean bulk-only number.","fun_headline_variants_meta":{"raw":{"variants":["Ag spacer lifts Bi SOT tenfold by keeping the film intact","Intact Bi yields bulk spin Hall angle ~1 with Ag shield","Ag multiplies Bi damping-like torque by blocking FM damage","Bi spin current jumps when Ag preserves epitaxial order","Ag-protected Bi hits 2×10⁵ spin Hall conductivity from bulk"]},"model":"grok-4.5","effort":"low","cost_usd":0.003825,"raw_usage":{"total_tokens":1336,"prompt_tokens":939,"num_sources_used":0,"completion_tokens":70,"cost_in_usd_ticks":38248000,"prompt_tokens_details":{"text_tokens":939,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":327,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":939,"tokens_out":70,"duration_ms":5406,"temperature":1.0,"reasoning_tokens":327,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T11:37:52.512532+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Grow a Bi/Ag/FM stack in which Bi is demonstrably continuous and unalloyed, then deliberately destroy only the Bi bulk spin-Hall channel (for example by amorphization or extreme thinning while keeping the Bi/Ag interface) and check whether the large positive damping-like torque disappears; if it survives, the bulk-origin claim fails.","supporting_citations":[],"review_version":1}