{"id":"c425e4ca-42fd-4b8c-8d41-ecba5771e53a","arxiv_id":"2501.09458","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Naturally oxidized Cu generates damping-like orbital torques in Py/CuOx that rival Pt, and solid-state gating reversibly switches their sign.","lead":"This paper studies spin and orbital torques in magnetic heterostructures based on copper, a cheap and abundant metal. It shows that naturally oxidized copper can generate efficient torques, and that a voltage can reverse the torque direction, a step toward low-cost, voltage-controlled spintronic devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Gating mechanism unverified: the paper concedes Co interface oxidation 'is possible' (SI Note 8), and its own SI Note 5 control shows interfacial CoOx alone yields the same sign-opposite torque, yet the gated stack's Co/Cu redox states are never directly measured.","rationale":"The paper makes three connected claims: (i) naturally oxidized Cu produces a damping-like orbital torque of sign opposite to Cu and Pt; (ii) this torque in Py/CuOx reaches the efficiency of Py/Pt and exceeds Co/Pt for tPy near 5 nm; and (iii) solid-state gating reversibly tunes the torque via the Cu oxidation state. The most load-bearing unsecured link is (iii). I examined the alternative concern emphasized by the reader, bulk OHE versus interfacial orbital REE, and consider it less decisive: the thickness dependence (BDL rising with tFM up to 5-10 nm before saturating or decreasing in Py) discriminates against a purely interfacial mechanism, and both channels are orbital in nature, preserving the qualitative claims and the gating demonstration. The Co-oxidation confound is sharper because it attacks the specific redox assignment that makes the gating experiment interpretable. The authors, in good faith, provide partial counters: Device 3 shows no AHE/HS change across the cycle, self-torques are measured, and device-to-device variability is acknowledged. However, none of these is a direct chemical measurement of the Co or Cu oxidation state in the gated stack, and the SI Note 5 control undermines the assumption that the observed sign is uniquely attributable to CuOx. The paper's strengths should be credited: the harmonic-Hall methodology is standard, the thickness series and inverted-stack controls are well executed, the Py-based data reduce the FM-oxidation confound for the efficiency claim, and the authors clearly disclose the unresolved sign discrepancies and variability. For these reasons I would keep the CONDITIONAL verdict: the empirical claims are likely sound, but the mechanistic claim at the center of the gate-tunability demonstration should be secured by a direct oxidation-state measurement on the actual gated devices. My concern refines rather than replaces the reader's weakest assumption, so agreement is partial.","tokens_in":18626,"tokens_out":21977,"duration_ms":219616,"concrete_test":"Gate several identical Co(5)/CuOx(3)/GdOx(25)/Pt(5) devices to the pristine, +3 V, and -3 V states of Fig. 5, then acquire depth-resolved XPS (Co 2p, Cu 2p, O 1s) on each after removing the top Pt/GdOx by in-situ ion milling, or XMCD at the Co and Cu L-edges. If the torque reverses sign while Co 2p shows no oxide component in any state, the Cu-redox mechanism stands; if CoOx appears after negative biasing, the gating result is confounded. A complementary control is to repeat the gating cycle on a Py(5)/CuOx(3)/GdOx device, since SI Note 5 shows Py oxidation does not enhance torque; an identical sign reversal in the Py device would further exclude FM oxidation as the driver.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central novelty of the paper is the demonstration of reversible solid-state gating of the damping-like torque, attributed 'by precisely and reversibly controlling the oxidation state of the CuOx layer.' This attribution requires that oxygen ions driven toward the Co interface during negative bias oxidize the 3-nm Cu layer but stop short of the Co. The authors flag the fragility of this assumption in SI Note 8: 'minor oxidation of Co close to the Cu interface is possible and could occur in our experiments depending on the specific gating conditions and devices.' More telling is their own control experiment in SI Note 5: deliberately forming interfacial CoOx (Co exposed to air before Cu deposition, with no CuOx in the stack) produces a damping-like torque with the same sign as the claimed CuOx orbital torque, whereas Co/Cu is small and positive. Thus interfacial FM oxidation alone can reproduce the sign flip that the paper reads as CuOx-to-Cu conversion. The XPS characterization of SI Note 2, which showed no CoOx, was performed on the ungated Ti(1.5)/Co(5)/CuOx(3) stack; it was not performed on the actual gated device (Co(5)/CuOx(3)/GdOx(25)/Pt(5)) in the pristine, +3 V, or -3 V states. Moreover, the GdOx layer is deposited in an Ar:O2 plasma, so the initial interfacial state of the gated stacks is not necessarily the same as the air-exposed stack used for XPS. The HS/AHE monitoring in SI Note 8 is a poor probe for a sub-nm interfacial oxide and shows a small reproducible loop change in one device. The concern does not affect the Py/CuOx efficiency comparison or the thickness dependence, but it does threaten the mechanism assigned to the paper's most novel result, leaving the gate-tunability claim empirically sound but mechanistically unclosed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports harmonic Hall measurements of current-induced spin and orbital torques in Co and Ni80Fe20 (Py) films adjacent to Cu, naturally oxidized CuOx, and Pt layers. The main experimental findings are: (i) FM/CuOx bilayers exhibit a damping-like torque with sign opposite to that of FM/Cu and FM/Pt references; (ii) the damping-like torque efficiency grows with FM thickness in CuOx systems, with Py/CuOx reaching values comparable to Py/Pt and exceeding Co/Pt at tPy = 5 nm; and (iii) solid-state gating through a GdOx layer can reversibly change the sign and magnitude of the damping-like torque in Co(5)/CuOx(3), attributed to voltage-driven oxygen migration that switches Cu between oxidized and metallic states. The manuscript includes supporting thickness series, XPS characterization, control experiments for self-torques, and fits of orbital diffusion lengths.","tokens_in":19032,"tokens_out":4984,"duration_ms":54430,"significance":"If the central claims hold, this work would establish naturally oxidized Cu as a practical, low-cost, and voltage-switchable source of orbital torques, which is significant for the development of orbitronic devices using earth-abundant elements. The paper's strengths are its systematic thickness-dependent data, use of established harmonic Hall analysis, explicit control experiments for self-induced torques, and honest reporting of limitations (e.g., large device-to-device variability and possible Co oxidation). However, the most novel claim—full electrical tunability of the orbital torque via controlled oxidation of Cu—is not directly verified: the chemical state of the gated devices is inferred from resistance changes rather than measured, and the authors' own control experiments show that interfacial CoOx can produce the same sign of damping-like torque. The orbital-origin interpretation for the ungated CuOx bilayers is reasonable but not unique, as interface effects such as orbital Rashba-Edelstein or unintended FM oxidation could contribute. For these reasons, the paper is promising yet not definitive in its current form.","major_comments":[{"comment":"","section":"SI Note 8 and main-text 'Solid-state gated devices' (Fig. 4-5)"},{"comment":"","section":"Figure 5 and 'Electrical control of orbital torques'"},{"comment":"","section":"Figure 3c and accompanying text"}],"minor_comments":[{"comment":"","section":"SI Note 4"},{"comment":"","section":"SI Note 8"},{"comment":"","section":"Main text, 'Solid-state gated devices' section"},{"comment":"","section":"Main text, conclusion"},{"comment":"","section":"SI Note 8, last paragraph"},{"comment":"","section":"SI Note 6"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely and competitive topic, and the authors are well positioned in the field. The main concern is the mismatch between the strength of the central 'gate-tunable orbital torque' claim and the indirect evidence for the Cu oxidation mechanism, especially given their own control showing that CoOx produces the same sign. This is fixable with additional experiments or a revised, more cautious interpretation, so I do not recommend rejection. The paper may be suitable for this journal after the gating mechanism is clarified and the claims are scaled to the actual evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Good to talk. I've read the Damerio/Avci paper on Cu-based orbital torques. Short take: the ungated half of the paper is a solid, careful study that gives the field a useful Py-thickness series and a clean comparison between Cu, CuOx, and Pt. The gated half is an inventive proof-of-concept but the mechanism is not closed, and their own control data cuts against their interpretation.\n\nWhat's genuinely new: the Py(t)/CuOx thickness dependence, with xi_DL rising to Pt-like values at tPy ~ 5 nm, and the solid-state GdOx gating that reversibly flips the sign of the damping-like torque. The harmonic Hall analysis is standard but done carefully; they include self-torque controls, inverted stack checks, and a parallel-resistor Oersted subtraction, and they are refreshingly explicit about what they cannot explain (e.g., the BFL sign crossover at large tPy, the device-to-device variability in gating).\n\nThe soft spots are in the interpretation. The orbital-origin claim is inferred from the thickness dependence and sign reversal rather than directly evidenced; the authors acknowledge this. That would be acceptable for a first report, but the gating story has a bigger problem. The negative-bias state is assumed to oxidize Cu while leaving Co intact. The XPS showing no CoOx was done on the ungated stack, not on actual gated devices. Resistance changes and AHE loop shapes are indirect. More damaging, their own SI Note 5 control shows that intentionally forming interfacial CoOx produces a damping-like torque with the same sign as the claimed CuOx torque, yet in SI Note 8 they argue CoOx 'enhances but does not change sign.' That is inconsistent. If negative gating forms a sub-nm CoOx at the Co/Cu interface, it alone could reproduce the observed sign change, with no CuOx-to-Cu conversion needed.\n\nI don't think this kills the paper. The Py/CuOx efficiency result and the thickness dependence stand independently of the gating mechanism. But the paper's most novel claim - full electrical tunability via Cu oxidation - remains mechanistically open.\n\nMy recommendation: send it to peer review, and if I were the editor I'd ask for direct chemical state analysis (XPS or EELS) of the gated stack in pristine, +V, and -V states, or at least a recognition that CoOx formation is a viable competing explanation. As it stands, it's a conditional accept at best.\n\nI'd cite the Py/CuOx thickness data in my own work. Take it to reading group if you want to discuss what counts as evidence for OHE in oxides.","headline":"Solid orbital-torque study with a nice Py-thickness series, but the headline gating result lacks direct chemical-state evidence and is undercut by the authors' own CoOx control.","tokens_in":19541,"tokens_out":4379,"would_cite":true,"duration_ms":40799,"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":"Naturally oxidized Cu is claimed to generate damping-like orbital torque rivaling Pt, with sign and amplitude set by solid-state gating.","keywords":["spin-orbit torques","orbital Hall effect","orbital currents","copper oxide","solid-state gating","harmonic Hall measurements","magneto-ionics","spin-orbitronics"],"falsifier":"A decisive test: grow Py(5)/CuOx(t) with CuOx thickness t from about 0.5 to 10 nm under the same natural-oxidation protocol and measure xi_E_DL; if the negative damping-like torque does not increase with CuOx thickness and saturate, the bulk orbital Hall effect is not the dominant source.","tokens_in":18434,"feed_emoji":"🧲","tokens_out":8731,"duration_ms":112003,"temperature":0.7,"pith_summary":"The paper sets out to show that naturally oxidized copper, a light and abundant 3d metal, can generate current-induced torques on adjacent magnetic films as effectively as the heavy-metal standard platinum, and that the torque can be switched and tuned electrically. Its central experimental claim is that the damping-like torque in Co/CuOx and Py/CuOx bilayers has the opposite sign to that from metallic Cu and Pt, grows with ferromagnet thickness in a way that signals an orbital origin, and in Py/CuOx reaches an efficiency comparable to Py/Pt and greater than Co/Pt. The paper further claims reversible sign control of the damping-like torque in Co/CuOx/GdOx devices by solid-state gating, where voltage-driven oxygen-ion migration oxidizes or reduces the Cu layer. If these claims hold, cheap and sustainable Cu-based layers become viable torque sources for gate-tunable spintronic memory and logic.","feed_headline":"Cu oxide matches Pt torques, and gating flips their sign","feed_subtitle":"Voltage-driven oxygen migration flips the torque sign, opening gate-tunable spintronics with abundant metals.","key_machinery":"The load-bearing mechanism is the orbital Hall effect (OHE) in oxidized Cu: an electric current drives a transverse flow of orbital angular momentum in CuOx, and because orbital moments do not exchange-couple directly to the magnetization, the accumulated orbital angular momentum must be converted into a spin current in the adjacent ferromagnet via its orbital-to-spin conversion coefficient eta_FM. Harmonic Hall measurements separate the damping-like and field-like torque components through their different angular dependences in the second-harmonic signal, and the FM-thickness dependence distinguishes bulk orbital accumulation, which needs several nanometres of FM to convert, from interfacial effects. The gating demonstration uses an applied voltage to move O2- ions across a GdOx barrier, changing the Cu oxidation state and thereby turning the orbital-Hall contribution on, off, or reversed in sign.","core_discovery":"Using harmonic Hall measurements on Co(t)/NM and Py(t)/NM bilayers with NM = Pt, Cu, and naturally oxidized CuOx, the paper reports that the damping-like effective field (BDL) in FM/CuOx is negative, opposite to both metallic Cu and Pt, and grows in magnitude with FM thickness, contrary to the usual 1/tFM scaling of spin-Hall torques. The efficiency xi_E_DL in Py/CuOx rises sharply up to tPy approximately 5 nm and then saturates at a value comparable to Py/Pt and above Co/Pt; the authors attribute this thickness behavior to an orbital accumulation in CuOx that converts to a spin torque in the first nanometres of the ferromagnet, with Py's larger orbital-to-spin conversion coefficient accounting for the difference from Co. In solid-state gated Co(5)/CuOx(3)/GdOx/Pt devices, applying +3 V reduces CuOx toward metallic Cu and flips the damping-like torque from negative to positive, while subsequent -3 V reoxidizes the Cu and restores the negative sign; intermediate gating partially suppresses the torque. The authors conclude that the orbital Hall effect in naturally oxidized Cu is large enough to be useful and can be electrically controlled, while cautioning that the exact sign depends on the orbital texture and oxidation conditions.","pith_inferences":["A CuOx thickness series at fixed Py thickness would test the bulk-OHE picture: a bulk signal should grow and saturate with CuOx thickness, while an interface signal should saturate immediately.","The sign of the torque appears sensitive to oxidation conditions, so engineering the Cu oxidation state (for example Cu2O versus CuO) rather than merely having an oxide may allow deliberate sign selection.","The long gating times reported here, hundreds to thousands of seconds, suggest that immediate device use would target write-once or slowly refreshed configurations; faster ion conductors or thinner barriers could bring the effect closer to practical switching."],"forward_implications":["Py/CuOx bilayers offer torque efficiencies on par with Py/Pt, meaning a light-metal oxide can replace a heavy metal for current-induced torque generation.","The torque sign is determined by the Cu oxidation state, so a small gate voltage can reverse the torque direction without changing the magnetic layer.","Solid-state gating through GdOx works at room temperature with negligible gate leakage, offering a microelectronics-compatible route to tune orbital currents.","The FM thickness data identify Py as an efficient orbital-to-spin converter, making ferromagnet choice and interface quality central design levers.","The reduced-Cu torque state remains stable for at least the measurement time after gating, pointing toward non-volatile or semi-non-volatile operation."],"supporting_citations":[{"why":"Supplies the theoretical basis that the orbital Hall effect can occur without strong spin-orbit coupling, motivating Cu as an OHE candidate.","marker":"[6]"},{"why":"Predicts large orbital Hall conductivities in transition metals, supporting the expectation that Cu-related orbital currents can be sizable.","marker":"[7]"},{"why":"Provides the calculated spin and orbital conductivities for metallic Cu used to interpret the small positive torque in Co/Cu.","marker":"[8]"},{"why":"Establishes the conversion of an orbital Hall current into a spin current in a ferromagnet, the mechanism behind the observed damping-like torque.","marker":"[9]"},{"why":"Prior demonstration that natural oxidation of Cu creates a spin-torque generator; the direct experimental precedent for this study.","marker":"[16]"},{"why":"Reports the FM-thickness trend in CoFe/CuOx attributed to orbital currents, used here as a comparison for the Py and Co data.","marker":"[22]"},{"why":"Reports negative-sign orbital torque in Co/CuNx, providing a precedent for sign reversal in Cu-based orbital systems.","marker":"[24]"},{"why":"Explains why orbital angular momentum must be converted to spin angular momentum in the ferromagnet, a key step in the argument.","marker":"[30]"}],"fun_headline_variants":["Cu oxide torques rival Pt, gating flips sign","Gating flips torque sign in Cu-based heterostructures","Copper oxide matches Pt torques, gate-tunable","CuOx torques beat Pt, gating reverses them","Abundant Cu oxide torques surpass Pt, electrically switchable"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that the negative damping-like torque in FM/CuOx comes from a bulk orbital Hall effect in the oxidized copper, not from interfacial orbital Rashba-Eдельstein effects or from partial oxidation of the ferromagnet, and that voltage gating changes mainly the Cu oxidation state; if either assumption fails, the orbital-Hall interpretation and the gate-tuning conclusion would both be weakened.","fun_headline_variants_meta":{"raw":{"variants":["Cu oxide torques rival Pt, gating flips sign","Gating flips torque sign in Cu-based heterostructures","Copper oxide matches Pt torques, gate-tunable","CuOx torques beat Pt, gating reverses them","Abundant Cu oxide torques surpass Pt, electrically switchable"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000504,"raw_usage":{"total_tokens":2481,"prompt_tokens":985,"completion_tokens":1496,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":601,"completion_tokens_details":{"reasoning_tokens":1411}},"tokens_in":601,"tokens_out":1496,"duration_ms":12357,"temperature":1.0,"reasoning_tokens":1411,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:00:21.574151+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test: grow Py(5)/CuOx(t) with CuOx thickness t from about 0.5 to 10 nm under the same natural-oxidation protocol and measure xi_E_DL; if the negative damping-like torque does not increase with CuOx thickness and saturate, the bulk orbital Hall effect is not the dominant source.","supporting_citations":[{"cited_title":"Figure S6.1a shows a plot of the damping-like component of the torque as a function of Pt thickness ( tPt), which increases steeply for the first few nm and saturates above 3 nm","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical basis that the orbital Hall effect can occur without strong spin-orbit coupling, motivating Cu as an OHE candidate."},{"cited_title":"Therefore , we attempted to measure the orbital torques in Ni/CuOx bilayers","cited_arxiv_id":null,"evidence_quote":"Predicts large orbital Hall conductivities in transition metals, supporting the expectation that Cu-related orbital currents can be sizable."},{"cited_title":"Upon applying a negative VG, the Cu present in the stack gets oxidized and the resistance increases (light blue in Figure S8.1a)","cited_arxiv_id":null,"evidence_quote":"Provides the calculated spin and orbital conductivities for metallic Cu used to interpret the small positive torque in Co/Cu."}],"review_version":1}