{"id":"9b738e81-b45b-43e7-b8c0-bc4236412cb8","arxiv_id":"2506.11878","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Orbital currents are generated directly from magnetization dynamics of the ferrimagnetic insulator BiYIG and detected in naturally oxidized Cu, with spin currents shown to dominate in Cr.","lead":"This paper reports a new way to generate pure orbital currents: by pumping them from the magnetic insulator BiYIG using magnons, with no extra conversion layer. This matters because it separates the orbital degree of freedom in angular momentum transport, which could lead to more efficient orbitronic and spintronic devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The direct-orbital-pumping assignment rests on the assumption that the Cu/CuOx detector does not convert the majority spin current injected from BiYIG into orbital current; the YIG/Cu* control rules this out only if YIG injects a comparable spin current through a comparable interface.","rationale":"The paper is a careful experimental study with strong internal controls: the Cr detector shows that spin pumping dominates the total angular-momentum current, the pure-Cu control shows that the Cu/CuOx interface is essential for detection, the Ar+ etching enhancement is consistent with improved interfacial transmission, and the YIG/BiYIG comparison controls for a material with much smaller orbital character. The g-factor difference between YIG and BiYIG and the known enhancement of magneto-optical properties in BiYIG provide independent motivation for expecting a larger orbital component in BiYIG. However, the decisive attribution of the BiYIG/Cu* signal to direct orbital pumping—rather than to spin-to-orbital conversion inside the metallic detector—depends on the purity of Cu* as an orbital detector and on the comparability of spin injection between YIG and BiYIG through the same Ar+-etched interface. The paper's statement that spin-to-orbital conversion is inactive in insulators addresses conversion in the insulator only; the relevant conversion would occur in the metal, where conduction electrons exist. The YIG/Cu* control is the right empirical test, but the main text does not fully establish that the Cu(5)/Pt(5) transparency control was performed on YIG, nor does it quantify the expected signal from spin-to-orbital conversion against the observed YIG/Cu* upper bound. These are missing quantitative steps rather than demonstrated errors. The central claim is plausible and well motivated, but the evidence as presented supports conditional acceptance with a specific additional calibration, which is exactly the reader's verdict. No change to the verdict is needed; the proposed test would settle the residual ambiguity.","tokens_in":14622,"tokens_out":8525,"duration_ms":144556,"concrete_test":"Fabricate, on the same chip with identical Ar+ etching, YIG/Cu(5)/Pt(5), YIG/Cu*, and BiYIG/Cu* detectors. Measure the YIG/Cu(5)/Pt(5) spin-pumping signal and the damping-normalized BiYIG/Cu(5)/Pt(5) signal to confirm that YIG injects a spin current into Cu at least as large as BiYIG's after the same interface treatment. Then, using the literature value of spin-to-orbital conversion efficiency at Cu/CuOx interfaces, compute the expected YIG/Cu* signal if spin-to-orbital conversion were active in the detector. If the measured YIG/Cu* signal is at least an order of magnitude below this expectation, the control rules out the conversion channel and the direct-orbital-pumping assignment is supported; if the YIG/Cu* signal is comparable to or not far below this expectation, the origin claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central step is the sentence in the introduction: 'The latter process is not active in insulators due to the absence of conduction electrons.' This statement excludes spin-to-orbital conversion inside the insulator, but the pumped spin current—shown by the Cr result to be the majority angular-momentum channel—enters the metallic Cu/CuOx detector, where conduction electrons and interfacial spin-orbit coupling are present. A spin-to-orbital conversion at the Cu/CuOx interface would produce an orbital current detected via the inverse orbital Rashba-Edelstein effect, giving the same signal as direct orbital pumping. The YIG/Cu* control is the only empirical guard against this alternative, and its validity requires YIG to inject a spin current into Cu* comparable to BiYIG's after identical Ar+ etching, with comparable interface transparency. The main text states that a Cu(5)/Pt(5) detector with an identically etched interface shows a large signal, but it does not establish that this control was performed on the YIG interface; if it was performed only on BiYIG, it does not constrain YIG spin injection. Moreover, the upper bound on the YIG/Cu* signal is not quantified against the signal expected if Cu* converted spin into orbital current with a realistic efficiency. Without this calibration, the observed BiYIG/YIG difference could reflect different spin injection or interface transparency rather than direct orbital pumping from precessing orbital magnetization.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports experiments on nonlocal angular-momentum pumping from the ferrimagnetic insulator BiYIG into nanoscale detectors made of Pt, Cr, Cu, and naturally oxidized Cu (Cu*). The authors claim to detect pure orbital currents generated by both coherent and thermal magnons in BiYIG, and attribute the signal to direct orbital pumping from the orbital magnetization dynamics of BiYIG rather than to spin-to-orbital conversion. Evidence includes the contrast between Cu* (large signal) and pure Cu (negligible signal), the comparison between BiYIG and unsubstituted YIG (negligible in YIG), the sign analysis for Cr detectors, and the enhancement of signals upon Ar+ etching of the garnet/metal interface.","tokens_in":14931,"tokens_out":6430,"duration_ms":76323,"significance":"If the interpretation is correct, the paper provides the first demonstration of direct orbital pumping from a magnetic insulator without a converter layer, establishing the orbital magnetization dynamics of BiYIG as a source of pure orbital currents. The experimental methodology, based on broad-wavevector nonlocal pumping, enables detection of small signals and is supported by multiple controls: Cu vs Cu*, YIG vs BiYIG, angle and power dependence, and Cr sign analysis. The data are made openly available. The central claim is important for orbitronics and magnon transport, and the experimental approach is a useful contribution even independently of the interpretation.","major_comments":[{"comment":"The argument that spin-to-orbital conversion cannot occur because \"the latter process is not active in insulators due to the absence of conduction electrons\" addresses conversion only inside the insulator. It does not rule out conversion of the pumped spin current into an orbital current at the metallic Cu/CuOx detector interface or within the Cu layer, where conduction electrons and interfacial spin-orbit coupling are present. The Cr result shows that spin currents are the majority angular-momentum channel pumped from BiYIG, so a spin-to-orbital conversion in the detector would produce the same charge signal as direct orbital pumping. The YIG/Cu* control is the only empirical guard against this alternative, but its validity requires that YIG injects a spin current comparable in magnitude to BiYIG through a comparable interface. The paper does not provide a direct measurement of the spin mixing conductance for YIG/Cu* versus BiYIG/Cu*, nor does it quantify the upper bound on the YIG/Cu* signal against the signal expected if Cu* converted spin to orbital with a realistic efficiency. Please provide either a YIG/Cu(5)/Pt(5) control with the same etched interface to establish comparable spin injection, or a quantitative estimate of the expected YIG/Cu* signal under the spin-to-orbital-conversion hypothesis and show that the observed null lies well below that expectation.","section":"Introduction and section 3 (paragraph beginning \"These results lead to two important deductions\")"},{"comment":"The normalization of coherent and thermal pumping signals by α_eff^2 and α, respectively, accounts for differences in damping, but it implicitly assumes that the spin mixing conductance (and thus the proportionality constant between precession angle and pumped spin current) is the same for YIG and BiYIG. The paper states that unequal magnetic dissipation is considered, but the spin mixing conductance can differ between YIG and BiYIG due to different electronic structure, lattice matching, and interface properties. If the spin current injected by YIG into Cu* is much smaller than that from BiYIG after the stated normalizations, the null YIG/Cu* signal would be consistent with spin-to-orbital conversion in the detector rather than with the absence of direct orbital pumping. Please provide evidence, for instance from a Cu(5)/Pt(5) detector on YIG with the same etched interface, that the spin injection efficiency from YIG is comparable to that from BiYIG after the α normalizations, or explicitly discuss the uncertainty introduced by unknown mixing-conductance ratios.","section":"Section 3, Fig. 3 and accompanying text"},{"comment":"The text states that a Cu(5)/Pt(5) detector with an identically etched interface exhibits a large signal, thereby ensuring that interfacial transparency in the YIG or BiYIG/Cu or Cu* devices is substantial. However, the sentence does not specify whether this Cu(5)/Pt(5) control was fabricated on YIG, on BiYIG, or on both. If the control was performed only on BiYIG, it does not constrain the YIG/Cu* interface transparency, and the vanishing YIG/Cu* signal could be due to a poor YIG/Cu* interface rather than to the absence of spin-to-orbital conversion in Cu*. Please clarify which interfaces were used for this control and, if not already done, provide the equivalent YIG/Cu(5)/Pt(5) measurement.","section":"Section 3, sentence beginning \"Furthermore, a Cu(5)/Pt(5) detector with an identically etched interface\""}],"minor_comments":[{"comment":"There is a typo: \"unsubtituted YIG\" should be \"unsubstituted YIG\".","section":"Section 3"},{"comment":"Reference [94] (Lyalin and Kawakami, \"Interface transparency to orbital current\") is cited as Phys. Rev. B 110, 104418 (2014); the volume number 110 corresponds to 2024, so the year should be corrected to 2024.","section":"Reference list"},{"comment":"The axis labels in Fig. 3 (e.g., \"α2effκp (10-6 × nA/mW)\") are difficult to read because of the mix of superscripts and Greek letters; please format them more clearly, for example with explicit multiplication symbols and parentheses.","section":"Figure 3"},{"comment":"The ×10 annotations in Fig. 4 are unclear; please specify whether they indicate scaling factors for the plotted values or for the axes, and ensure all panel labels are consistent with the units stated in the text.","section":"Figure 4"},{"comment":"Reference [96] (Huang et al., \"Orbital Current Pumping From Ultrafast Light-driven Antiferromagnetic Insulator\") lacks volume and page information; please complete the citation.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is from a well-known group and the experiments appear carefully executed, but the central claim of direct orbital pumping hinges on the YIG/Cu* control ruling out spin-to-orbital conversion in the Cu* detector. The current text does not clearly establish that YIG injects a comparable spin current through a comparable interface, nor does it quantify the expected signal under the alternative hypothesis. This is fixable by adding the missing control or a quantitative calibration, but without that the interpretation is not fully supported. I would not recommend rejection, as the experimental data and methodology are valuable, but the claim as stated is stronger than the presented evidence allows."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: The paper likely reports the first direct observation of orbital pumping from a ferrimagnetic insulator without a converter layer, and the experimental controls are strong enough that the claim deserves serious referee time. The main caveat is whether the YIG control fully rules out spin-to-orbital conversion in the CuOx detector.\n\nWhat is new: nonlocal magnon pumping in BiYIG with naturally oxidized Cu, pure Cu, Pt, and Cr detectors. The key results are that BiYIG/Cu* gives a large signal while YIG/Cu* is negligible; Cr shows a negative signal indicating the spin current dominates; and Ar+ etching enhances both spin and orbital transparency. The separation of coherent and thermal magnon contributions is done carefully with power and angle dependences. The g-factor comparison between YIG and BiYIG is a nice supporting argument.\n\nThe main soft spot is the inference that Cu* is a pure orbital detector. The argument that spin-to-orbital conversion is inactive in insulators only applies inside the insulator, not at the metal interface where conduction electrons exist. The YIG/Cu* control is meant to catch that, but it only works if YIG injects a comparable spin current. The paper normalizes by damping, which is standard, but doesn't directly measure spin injection in YIG vs BiYIG. A quantitative estimate of the expected YIG/Cu* signal under a realistic conversion efficiency would tighten the argument. The Cu(5)/Pt(5) control is described as having an 'identically etched interface,' but it is not clear which garnet it was on.\n\nMinor issues: comparison plots in Figs. 3 and 4 lack error bars, which matters for claims about relative magnitudes. The Cr sign analysis is clean but doesn't quantify the relative spin and orbital contributions.\n\nNone of these are fatal. The evidence is credible and the interpretation is plausible. The paper is in good shape for a strong journal, provided the authors add error bars, clarify the interface control, and discuss spin-to-orbital conversion in the detector explicitly.\n\nWho this is for: researchers in orbitronics, spin-charge conversion, and magnonics. I would cite it if published.\n\nRecommendation: send to peer review with these requests. It deserves a proper referee.","headline":"First direct orbital pumping from a ferrimagnetic insulator, credible but with a control that needs sharper quantification.","tokens_in":15443,"tokens_out":5110,"would_cite":true,"duration_ms":56454,"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 shows that the ferrimagnetic insulator BiYIG pumps pure orbital currents, generated by coherent and thermal magnons, directly into adjacent metals without any spin-to-orbital converter layer.","keywords":["orbital pumping","ferrimagnetic insulator","magnons","orbital Rashba-Edelstein effect","spin pumping","BiYIG","orbital Hall effect","orbitronics"],"falsifier":"Insert a 1–2 nm spacer between BiYIG and Cu* that is known to transmit spin currents but block orbital currents, and measure the nonlocal converted voltage; if the pure-orbital signal survives, the claim that it comes from direct orbital pumping is falsified.","tokens_in":14474,"feed_emoji":"🧲","tokens_out":8693,"duration_ms":90265,"temperature":0.7,"pith_summary":"This paper reports the detection of pure orbital currents generated by both coherent and thermal magnons in the ferrimagnetic insulator Bi-doped yttrium iron garnet (BiYIG). The authors argue that these orbital currents are pumped directly by the orbital component of the magnetization dynamics in the insulator, rather than produced by spin-to-orbital conversion, a process they state is inactive in insulators because no conduction electrons are available. The experimental signature is a sizeable converted charge current in naturally oxidized Cu (Cu*) detectors, which are sensitive to orbital currents but not spin currents, while pure Cu and unsubstituted YIG give negligible signals. If the claim holds, it shows that a magnetic insulator can inject orbital angular momentum into an adjacent metal without any converter layer, and that orbital and spin transport can be separated in non-equilibrium angular-momentum devices.","feed_headline":"Magnetic insulator pumps pure orbital currents from magnons","feed_subtitle":"BiYIG sends orbital angular momentum into adjacent copper without a converter layer, separating orbital from spin transport.","key_machinery":"The load-bearing mechanism is the precessing orbital magnetization of the ferrimagnetic insulator BiYIG, which acts as a source of orbital angular-momentum current; its finite orbital character is indicated by the g-factor exceeding 2. The detection is carried out by the inverse orbital Rashba-Edelstein effect at a Cu/CuOx interface, which converts an orbital current into a transverse charge voltage without requiring spin-to-charge conversion. The experimental platform is a nonlocal nano-stripe device in which a microwave antenna excites both coherent propagating magnons and thermal magnons in BiYIG, and separate detector stripes of Pt, Cr, pure Cu, and naturally oxidized Cu convert the pumped angular-momentum currents; comparing signal sizes and signs across detectors separates the orbital and spin channels.","core_discovery":"The core claim is that the orbital component of magnetization dynamics in BiYIG directly pumps a pure orbital current into an adjacent metal, detected as a charge voltage via the inverse orbital Rashba-Edelstein effect at a Cu/CuOx interface. The evidence is comparative: BiYIG/Cu* shows a large coherent and thermal signal; YIG/Cu* is small; pure Cu detectors are negligible; and Cr detectors, which have opposite signs for spin and orbital Hall conductivities, give a negative signal, showing that spin pumping still dominates the total angular-momentum current. Because BiYIG is an insulator and because the g-factor of BiYIG (about 2.03) is larger than that of YIG (about 2.01), the authors assign the orbital signal to the finite orbital magnetization of BiYIG itself. They further show that Ar+ etching of the interface enhances both spin and orbital pumping efficiencies by roughly an order of magnitude.","pith_inferences":["The same nonlocal pumping geometry could be used to search for orbital pumping in other insulators with finite orbital moments, including antiferromagnetic insulators; the paper's note added points to recent terahertz evidence from alpha-Fe2O3 that supports this direction.","The paper does not derive a quantitative scaling of the orbital pumping amplitude with the g-factor shift; measuring a Bi-substitution series would test whether the signal tracks the orbital moment and would harden the direct-pumping assignment.","Because spin currents dominate in BiYIG, devices that want pure orbital currents may need to suppress spin transmission through interface engineering rather than simply maximize total transparency.","If orbital and spin mixing conductances respond differently to surface treatments, then Ar+ etching could be used to tune the orbital-to-spin ratio of the pumped current, a possibility the paper leaves open."],"forward_implications":["Orbital currents can be generated and detected from insulating magnon systems without converter layers, extending orbitronics beyond metallic magnets.","Naturally oxidized Cu acts as a near-pure orbital detector, giving a simple experimental way to separate orbital from spin transport.","In BiYIG the orbital current is a minority component of the pumped angular-momentum current, since Cr detectors show the negative spin Hall effect dominating the positive orbital Hall effect.","Interface treatments such as Ar+ etching improve transparency for both spin and orbital currents, increasing pumping efficiencies by about an order of magnitude.","Doping that increases orbital magnetization, such as Bi substitution in YIG, is a practical handle for turning on orbital pumping from an insulator."],"supporting_citations":[{"why":"Provides the theoretical concept of orbital pumping by magnetization dynamics that this paper tests in an insulating source.","marker":"[47]"},{"why":"Supplies the companion theory of orbital pumping used to interpret the BiYIG signals.","marker":"[48]"},{"why":"Prior experimental observation of orbital pumping in a metallic system, establishing the detection approach extended here to insulators.","marker":"[49]"},{"why":"Demonstrates the inverse orbital Rashba-Edelstein effect, the orbital-to-charge conversion mechanism behind the Cu* signal.","marker":"[50]"},{"why":"Shows that CuOx interfaces mediate orbital-charge conversion, supporting the use of naturally oxidized Cu as an orbital detector.","marker":"[51]"},{"why":"Identifies the orbital Rashba effect in surface-oxidized Cu, the microscopic origin of the Cu* detection signal.","marker":"[57]"},{"why":"Provides the broad-wavevector nonlocal pumping technique used to excite and detect coherent and thermal magnons.","marker":"[67]"},{"why":"Documents the orbital character of magnetization in garnet films, supporting the g-factor argument for BiYIG.","marker":"[76]"},{"why":"Links Bi substitution in garnets to enhanced magneto-optical response and finite orbital magnetization, grounding the YIG/BiYIG comparison.","marker":"[77]"}],"fun_headline_variants":["Magnons pump pure orbital currents from a ferrimagnet","Thermal and coherent magnons pump orbital currents in BiYIG","Orbital currents from magnons, separated from spin in BiYIG","Improved interface boosts both orbital and spin pumping in BiYIG"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument stands on the premise that spin-to-orbital conversion requires conduction electrons, so a spin current leaving an insulator cannot become an orbital current at the copper interface; if the metal interface could perform that conversion, the Cu* signal would not by itself prove that the orbital current originated from the insulator's magnetization dynamics.","fun_headline_variants_meta":{"raw":{"variants":["Magnons pump pure orbital currents from a ferrimagnet","Thermal and coherent magnons pump orbital currents in BiYIG","Orbital currents from magnons, separated from spin in BiYIG","Improved interface boosts both orbital and spin pumping in BiYIG"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000975,"raw_usage":{"total_tokens":4118,"prompt_tokens":895,"completion_tokens":3223,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":3149}},"tokens_in":511,"tokens_out":3223,"duration_ms":30625,"temperature":1.0,"reasoning_tokens":3149,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T01:01:41.341835+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Insert a 1–2 nm spacer between BiYIG and Cu* that is known to transmit spin currents but block orbital currents, and measure the nonlocal converted voltage; if the pure-orbital signal survives, the claim that it comes from direct orbital pumping is falsified.","supporting_citations":[{"cited_title":"Orbital Pumping Incorporating Both Orbital Angular Momentum and Position","cited_arxiv_id":"2311.00362","evidence_quote":"Supplies the companion theory of orbital pumping used to interpret the BiYIG signals."},{"cited_title":"Hayashi, D","cited_arxiv_id":null,"evidence_quote":"Prior experimental observation of orbital pumping in a metallic system, establishing the detection approach extended here to insulators."},{"cited_title":"El Hamdi, J.-Y","cited_arxiv_id":null,"evidence_quote":"Demonstrates the inverse orbital Rashba-Edelstein effect, the orbital-to-charge conversion mechanism behind the Cu* signal."},{"cited_title":"Santos, J","cited_arxiv_id":null,"evidence_quote":"Shows that CuOx interfaces mediate orbital-charge conversion, supporting the use of naturally oxidized Cu as an orbital detector."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Identifies the orbital Rashba effect in surface-oxidized Cu, the microscopic origin of the Cu* detection signal."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the broad-wavevector nonlocal pumping technique used to excite and detect coherent and thermal magnons."},{"cited_title":"Rogalev, J","cited_arxiv_id":null,"evidence_quote":"Documents the orbital character of magnetization in garnet films, supporting the g-factor argument for BiYIG."},{"cited_title":"Li and G.-Y","cited_arxiv_id":null,"evidence_quote":"Links Bi substitution in garnets to enhanced magneto-optical response and finite orbital magnetization, grounding the YIG/BiYIG comparison."}],"review_version":1}