{"id":"790df0f5-037b-453a-a37b-f653a7c43bda","arxiv_id":"2501.04235","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Current alone switches the perpendicular magnetization of Fe3GeTe2 on Ar-milled SrTiO3, and the switching direction can be set by a brief in-plane magnetic field initialization.","lead":"A magnetically ordered van der Waals crystal (Fe3GeTe2) placed on a treated strontium titanate surface can be flipped by an electric current with no magnetic field applied. The result points to a simpler route to magnetic memory devices that do not need external fields to write bits.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The mechanism claim rests on an unmeasured, undetected in-plane interface magnetization plus a schematic precession argument; the transport evidence alone cannot distinguish this from exchange-bias or canted-moment alternatives.","rationale":"The reader's conditional verdict already identifies the m_interface premise as the weak point, and this stress-test pass confirms that the concern is load-bearing. The paper's own supporting material admits that direct magnetic detection of m_interface was not possible, leaving only transport inference: polarity reversal, a small coercivity, and a weak exchange bias. These observations are consistent with the m_interface picture but are also consistent with an exchange-biased or canted interfacial moment, which would not require the proposed spin-precession mechanism. No quantitative model links the Rashba spin accumulation to a net out-of-plane polarization via precession around m_interface, so the mechanism chain is incomplete. The empirical observation of field-free switching is supported by multiple devices and controls (FGT/SiO2, no-Ar STO, thickness dependence, disappearance above Tc), so outright rejection is not warranted. Since the reader already recommended CONDITIONAL, the verdict should remain unchanged, with the condition being direct detection of m_interface or a quantitative spin-transport calculation demonstrating sufficient out-of-plane polarization.","tokens_in":14414,"tokens_out":6437,"duration_ms":75821,"concrete_test":"Fabricate an unpatterned FGT/Ar-STO stack with identical Ar-milling and exfoliation conditions, cool to 90 K, initialize with +8 kOe in-plane, then remove the field and measure the in-plane magnetization with SQUID/VSM at zero out-of-plane field; look for a remanent in-plane moment with a ~30 Oe coercivity matching the switching-polarity hysteresis in Fig. 4b. Also run the same measurement on FGT/SiO2 and non-Ar-milled STO controls. If no such remanent in-plane moment exists, or it does not track the polarity hysteresis, the m_interface premise collapses and the precession mechanism cannot be the cause.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mechanism requires a persistent in-plane interface magnetization m_interface at the FGT/Ar-STO interface that (1) survives after in-plane field initialization and (2) is strong enough to precess y-polarized Rashba spin accumulation into an out-of-plane polarization before it enters FGT. Supporting Note 3 explicitly concedes that XMCD detection of m_interface was not possible, and the offered evidence is purely inferential: switching polarity reversal after in-plane field initialization (Fig. 4a/b), a hysteresis with ~30 Oe coercivity (Fig. 4b), and a weak ~20 Oe exchange bias in AHE loops. These observations establish history-dependent symmetry breaking, but they do not establish that the symmetry-breaking object is a static in-plane magnetization rather than a canted interfacial FGT moment or an exchange-bias effect. Moreover, the mechanism of Fig. 3b is schematic: no spin-diffusion or precession calculation is given to show that a y-polarized spin current acquires a substantial z-component while traversing the interface region. If the precession efficiency is low, the field-free switching would need an alternative explanation (e.g., field-like torque from a canted moment), and the 'precession-facilitated SOT' central claim would be unsupported. The empirical field-free switching observation itself is well controlled, but the paper's headline interpretation is not.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports magnetic-field-free, current-induced perpendicular magnetization switching in exfoliated Fe3GeTe2 devices on an Ar-milled SrTiO3 surface. The empirical core is a set of transport measurements: writing-current sweeps produce hysteretic Rxy loops at zero field in multiple devices; control devices on SiO2, on non-milled STO, with thick FGT, and above Tc show no such switching; and AHE loops shift oppositely for opposite current directions. The authors attribute the effect to a precession-facilitated spin-orbit torque: Rashba spin accumulation at the FGT/STO interface precesses around an emergent in-plane interface magnetization m_interface, yielding an out-of-plane spin polarization that switches the perpendicular FGT magnetization. They further show that the switching polarity reverses after initialization with opposite in-plane fields and exhibits a ~30 Oe hysteresis, which they interpret as evidence for a persistent in-plane interface magnetization.","tokens_in":14771,"tokens_out":2458,"duration_ms":26536,"significance":"If the proposed mechanism is correct, the work would establish a new route to field-free SOT switching using a vdW magnet/oxide interface, with a reported critical current density of ~6e10 A/m2 that is competitive with prior art. The empirical switching observation itself is well supported: the inclusion of multiple devices, several control samples, and temperature/thickness dependence gives confidence that the zero-field switching is genuine and interfacial. The 'precession-facilitated SOT' interpretation, however, rests on an unmeasured interface magnetization and a schematic precession argument, with no quantitative model. The paper's headline claim is therefore only partially supported by the evidence presented.","major_comments":[{"comment":"The existence and persistence of the in-plane interface magnetization m_interface is the load-bearing premise of the mechanism, yet SI Note 3 concedes that XMCD detection was not possible and offers only transport inference: the switching-polarity reversal (Fig. 4a/b), the ~30 Oe coercivity of that reversal, and a ~20 Oe exchange-bias asymmetry. These observations establish history-dependent symmetry breaking, but they do not uniquely identify a static in-plane magnetization. A canted interfacial FGT moment or an exchange-bias-like coupling could produce the same transport signatures. Because m_interface is inferred from the same polarity-switching behavior it is invoked to explain, the mechanism claim is partly circular. The authors should either provide a direct probe of m_interface or a quantitative model whose predictions (polarity, coercivity, AHE loop shift) can be tested against the data without invoking m_interface ad hoc.","section":"SI Note 3"},{"comment":"The precession step from in-plane Rashba spin accumulation to out-of-plane spin polarization is asserted but not quantified. No spin-diffusion or spin-precession calculation is given to show that a y-polarized spin current obtains a substantial z-component while traversing the FGT/STO interface region. The torque tau_m = m_interface x sigma_Rashba merely describes the precession axis; the accumulated precession angle depends on the exchange field strength, the interface dwell time, and the spin diffusion length, none of which are estimated. If the precession efficiency is small, the proposed damping-like torque with out-of-plane polarization would be too weak to switch FGT, and the observed switching would need an alternative explanation. The authors should provide at least an order-of-magnitude estimate of the precession angle or a microscopic model supporting the 'ultimate out-of-plane spin polarization' claim.","section":"Figure 3b"},{"comment":"The switching-polarity hysteresis loop in Fig. 4b is central evidence for the magnetization-like behavior of m_interface, but it appears to be a single representative sweep without error bars or multiple reproducibility tests. The coercivity is only ~30 Oe, and the switching itself is stochastic (random peaks are acknowledged in SI Note 2). A single sweep cannot exclude that the polarity change is caused by a spurious field history effect or by partial magnetization reorientation in FGT. The authors should show multiple repeats of the polarity hysteresis and quantify the reproducibility, especially because the entire interpretation depends on this loop.","section":"Figure 4b"}],"minor_comments":[{"comment":"The captions for Fig. 2b and 2f contain contradictory device labels ('device 1 ... for device 3' and 'device 2 ... for device 1'), which obscures which device the data come from. Please correct these captions.","section":"Figure 2b/f captions"},{"comment":"SI Note 3 point (2) refers to 'Figures 3e and S5' for the exchange-bias effect, but there is no Figure 3e in the main text; the relevant figure appears to be Figure 2e. Please update the cross-reference.","section":"SI Note 3"},{"comment":"The sentence 'Our novel approach unitizes the large Rashba spin-orbit coupling' contains a typo; 'unitizes' should be 'utilizes'.","section":"Main text, Discussion"},{"comment":"The AHE loop shift of 14.6 Oe is reported without an uncertainty or repeated measurement; given that the loop width is of similar order, a quantitative statement of significance would strengthen the claim.","section":"Main text, Figure 2f"}],"recommendation":"major_revision","confidential_remarks":"The empirical observation of field-free switching is credible and well controlled, and the paper is likely of interest to the spintronics and vdW-magnet communities. However, the central mechanism is not sufficiently supported: the interface magnetization is unmeasured, the precession argument is schematic, and the polarity hysteresis lacks statistical backing. These are fixable with additional data and analysis, but they are load-bearing for the paper's headline interpretation. I also note that the paper's framing as the 'first' vdW-magnet/oxide SOT combination should be checked carefully against the literature before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know: this paper reports reproducible field-free switching of perpendicular FGT magnetization using an Ar-milled STO substrate, and it is backed by a proper set of controls. The switching observation is likely real. The mechanism attached to it—precession of Rashba-accumulated spins around an emergent interface magnetization—is not established by the data. Treat the device result as solid and the interpretation as a hypothesis.\n\nWhat is genuinely new: the FGT/oxide materials combination for SOT, specifically FGT on Ar-treated STO, and the polarity toggle via an in-plane initialization field with a ~30 Oe coercivity. Four devices show switching; FGT/SiO2, non-milled STO, above Tc, and 22-nm-thick FGT show none. That is a convincing empirical package. The reported critical current density around 6e10 A/m2 is respectably low, though not an order-of-magnitude improvement, and the authors acknowledge contributions from intrinsic FGT SOT and Joule heating.\n\nSoft spots: the mechanism is underdetermined. The interface magnetization m_interface is inferred from the same switching-polarity behavior it is invoked to explain; direct XMCD detection is explicitly conceded as not possible. The precession step in Fig. 3b is schematic, with no estimate of spin diffusion length, precession angle, or torque efficiency. A canted interfacial FGT moment or an exchange-bias field could plausibly produce the same symmetry breaking and polarity control; the weak ~20 Oe exchange bias they report does not rule that out. To their credit, SI Note 3 is honest about the missing detection, but the headline claim of precession-facilitated SOT remains one hypothesis among several.\n\nBottom line: the empirical contribution deserves publication and should be reproducible. The mechanism needs direct interface-sensitive measurement or a quantitative model before being accepted. A serious referee should engage; with revision separating the switching demonstration from the precession interpretation, this could be a useful addition to the vdW-magnet SOT literature.","headline":"A credible field-free SOT switching demonstration in FGT/Ar-STO with a mechanism claim that outruns the evidence.","tokens_in":15254,"tokens_out":2028,"would_cite":true,"duration_ms":20036,"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":"Placing the van der Waals ferromagnet Fe3GeTe2 on an argon-milled SrTiO3 surface makes a plain in-plane current capable of switching the perpendicular magnetization at zero applied field, with the switching direction set by a persistent…","keywords":["field-free switching","Fe3GeTe2","SrTiO3 heterostructure","Rashba spin-orbit coupling","interface magnetization","spin-orbit torque","van der Waals magnet","precession-facilitated SOT"],"falsifier":"Directly probe the Fe3GeTe2/SrTiO3 interface after removing an in-plane field, for example with X-ray magnetic circular dichroism on a stack thinned enough for the probe to reach the interface; if no in-plane moment remains at zero field, or if its reversal does not track the ~30 Oe polarity-coercivity, the precession mechanism is not the cause of the switching.","tokens_in":14254,"feed_emoji":"🧲","tokens_out":12065,"duration_ms":103081,"temperature":0.7,"pith_summary":"The paper claims that a current injected in the plane of a Fe3GeTe2/SrTiO3 heterostructure can switch the perpendicular magnetization of the Fe3GeTe2 layer while no external magnetic field is present. The reason, the authors argue, is that the Ar-milled SrTiO3 surface produces both a strong Rashba spin–orbit coupling and an in-plane magnetization at the interface with Fe3GeTe2; spins accumulated by the current precess around that interface magnetization and acquire an out-of-plane polarization that exerts a damping-like torque on the ferromagnet. The switching polarity is not fixed by the current direction alone but by the history of an in-plane initialization field, with a measured coercivity of about 30 Oe, which the authors take as evidence that a persistent interface moment is doing the symmetry breaking. If the mechanism is right, it makes field-free spin-orbit-torque switching accessible in a simple van der Waals magnet/oxide bilayer and at switching current densities near $6 \\times 10^{10}\\,\\text{A/m}^2$, which would interest anyone trying to build low-power magnetic memory.","feed_headline":"Current alone flips van der Waals magnet at zero field","feed_subtitle":"A spin torque from the Fe3GeTe2/SrTiO3 interface writes perpendicular magnetization with no magnetic field applied.","key_machinery":"The central mechanism is precession-facilitated spin-orbit torque built on an interface magnetization $\\mathbf{m}_{\\mathrm{interface}}$. A charge current creates a Rashba spin accumulation $\\boldsymbol{\\sigma}_{\\mathrm{Rashba}}$; the interface magnetization exerts a torque $\\boldsymbol{\\tau}_m = \\mathbf{m}_{\\mathrm{interface}} \\times \\boldsymbol{\\sigma}_{\\mathrm{Rashba}}$ on those spins, driving their precession so that the spins injected into Fe3GeTe2 acquire an out-of-plane component $\\sigma_z$. This out-of-plane spin polarization acts as a damping-like torque that deterministically switches the perpendicular magnetization, with the switching polarity controlled by the sign of $\\mathbf{m}_{\\mathrm{interface}}$.","core_discovery":"The authors claim that current-induced perpendicular magnetization switching at zero external field occurs in Fe3GeTe2/SrTiO3 heterostructures because the Ar-milled SrTiO3 surface provides a large Rashba spin–orbit coupling and, at the Fe3GeTe2/SrTiO3 interface, an emergent in-plane magnetization exists whose direction can be set (\"initialized\") by an in-plane magnetic field and persists after the field is removed. When a charge current flows, the Rashba-accumulated spins polarized in-plane precess around this interface magnetization, producing a net out-of-plane spin polarization; this polarization then exerts a damping-like spin-orbit torque on the Fe3GeTe2 layer and switches its perpendicular magnetization deterministically, with the polarity set by the direction of the interface magnetization. Evidence includes the current-driven hysteretic switching without a field, the opposite shifts of anomalous Hall loops for opposite current directions, the absence of switching in Fe3GeTe2/SiO2 and in Fe3GeTe2 on untreated SrTiO3, and the reversal of switching polarity upon in-plane field initialization with a ~30 Oe coercivity. The authors further show that the effect disappears in Fe3GeTe2 flakes outside roughly 9–15 nm thickness, pointing to the interfacial origin of the switching.","pith_inferences":["Extending the paper's mechanism to other van der Waals magnets, one would expect field-free switching to appear whenever a conducting oxide surface can host both a Rashba splitting and a persistent in-plane interface moment; a systematic control experiment changing the oxide while keeping Fe3GeTe2 would test this directly.","Because the authors could not directly probe the interface moment, a thinner-Fe3GeTe2 variant that lets X-rays reach the interface could look for the predicted remanent in-plane magnetization and its ~30 Oe switching field, testing the central premise.","The paper's transport data cannot fully distinguish a field-like torque from a damping-like torque in the anomalous Hall loop shift; a harmonic Hall-voltage measurement on the same devices would separate the two and place the precession picture on firmer footing.","If the interface moment is defect-driven, its magnitude and coercivity should depend on Ar-milling power and duration, so tuning the etching conditions offers a practical knob for lowering the switching current further."],"forward_implications":["Fe3GeTe2 flakes 9–15 nm thick on Ar-milled SrTiO3 can be switched by in-plane current alone at zero field, with a critical switching current density down to about $6 \\times 10^{10}\\,\\text{A/m}^2$.","The switching polarity is controlled by the direction of a previously applied in-plane magnetic field and reverses only when that field exceeds about 30 Oe, so the interface magnetization acts as a rewritable memory of the initialization direction.","No field-free switching appears in Fe3GeTe2/SiO2 or Fe3GeTe2 on untreated SrTiO3, and not in Fe3GeTe2 flakes outside the roughly 9–15 nm thickness window, indicating the effect is tied to the Ar-milled oxide interface rather than to the Fe3GeTe2 alone.","The anomalous Hall loop shift direction flips with the sign of the writing current, proving that the current produces a preferred out-of-plane direction rather than mere heating or domain noise.","The precession-torque scheme, previously requiring a ferromagnet/spacer/in-plane-ferromagnet stack, can be realized with a single interface that supplies its own in-plane moment, simplifying the device geometry."],"supporting_citations":[{"why":"It provides the theoretical model of spin-orbit precession torque in which in-plane spin accumulation acquires an out-of-plane component.","marker":"[13]"},{"why":"It shows large spin-orbit torque in Fe3GeTe2 devices that still requires an external magnetic field, defining the problem this paper attacks.","marker":"[21]"},{"why":"It establishes that magnetic-field-free switching requires an out-of-plane spin polarization, the ingredient the paper claims to generate.","marker":"[25]"},{"why":"It states the symmetry condition that field-free switching needs a broken mirror plane, which the interface magnetization supplies here.","marker":"[26]"},{"why":"It demonstrates perpendicular switching via precession torque in a ferromagnet/spacer/in-plane-ferromagnet stack that this work simplifies to one interface.","marker":"[27]"},{"why":"It supplies the observation that an Ar-milled SrTiO3 surface hosts sizable Rashba spin-charge conversion, the source of the in-plane spin accumulation.","marker":"[28]"},{"why":"It reports a spin Hall angle for the Ar-milled SrTiO3 surface about one order of magnitude larger than that of Pt, explaining the low switching current density.","marker":"[29]"},{"why":"It shows that Ar milling creates a conducting layer on SrTiO3 through oxygen vacancies, the fabrication step on which the device relies.","marker":"[30]"},{"why":"It documents magnetic phenomena at oxide heterostructure interfaces, supporting the plausibility of an emergent in-plane interface magnetization.","marker":"[31]"}],"fun_headline_variants":["Zero-field switching via current in vdW magnet/oxide stack","Zero-field magnetic switching via oxide interface spin precession","No-field toggle in Fe3GeTe2/SrTiO3 via Rashba precession","Current-driven switch in vdW magnet with no applied field"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a persistent in-plane magnetization exists at the Fe3GeTe2/Ar-milled SrTiO3 interface, can be aligned by an in-plane magnetic field, and remains aligned after the field is removed; the authors state that they could not measure this moment directly with X-ray magnetic circular dichroism and instead rely on transport evidence.","fun_headline_variants_meta":{"raw":{"variants":["Zero-field switching via current in vdW magnet/oxide stack","Zero-field magnetic switching via oxide interface spin precession","No-field toggle in Fe3GeTe2/SrTiO3 via Rashba precession","Current-driven switch in vdW magnet with no applied field"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000836,"raw_usage":{"total_tokens":3690,"prompt_tokens":1035,"completion_tokens":2655,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":651,"completion_tokens_details":{"reasoning_tokens":2579}},"tokens_in":651,"tokens_out":2655,"duration_ms":18761,"temperature":1.0,"reasoning_tokens":2579,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:37:41.353655+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Directly probe the Fe3GeTe2/SrTiO3 interface after removing an in-plane field, for example with X-ray magnetic circular dichroism on a stack thinned enough for the probe to reach the interface; if no in-plane moment remains at zero field, or if its reversal does not track the ~30 Oe polarity-coercivity, the precession mechanism is not the cause of the switching.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the theoretical model of spin-orbit precession torque in which in-plane spin accumulation acquires an out-of-plane component."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It establishes that magnetic-field-free switching requires an out-of-plane spin polarization, the ingredient the paper claims to generate."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It demonstrates perpendicular switching via precession torque in a ferromagnet/spacer/in-plane-ferromagnet stack that this work simplifies to one interface."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the observation that an Ar-milled SrTiO3 surface hosts sizable Rashba spin-charge conversion, the source of the in-plane spin accumulation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It reports a spin Hall angle for the Ar-milled SrTiO3 surface about one order of magnitude larger than that of Pt, explaining the low switching current density."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It shows that Ar milling creates a conducting layer on SrTiO3 through oxygen vacancies, the fabrication step on which the device relies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It documents magnetic phenomena at oxide heterostructure interfaces, supporting the plausibility of an emergent in-plane interface magnetization."}],"review_version":1}