{"id":"e01ffe20-2f7e-426e-8a2c-8ad1f2d0409e","arxiv_id":"1908.05524","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Nonlocal spin transport in graphene/Cr2Ge2Te6 shows Hanle peak shifts and spin-lifetime anisotropy attributed to a perpendicular proximity exchange field, with supporting density-functional calculations that overestimate the field by orders of magnitude.","lead":"Experiments place graphene on the ferromagnetic insulator Cr2Ge2Te6 and observe spin-transport signatures attributed to a magnetic exchange field leaking into the graphene. The result points toward a path for imprinting magnetism on graphene for future spintronic and topological devices, but the inferred field strength is far smaller than the density-functional calculations predict.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The stray-field estimate in Supplementary note 2 uses a spatially averaged field to claim exchange dominates, but Hanle signals depend nonlinearly on local fields; edge stray fields in a 60%-overlap channel can mimic shifts and asymmetry, so the exchange attribution is not yet secured.","rationale":"I agree with the reader's conditional verdict and with their identification of the stray-field assumption as the weakest link, but I sharpen the objection: the spatial averaging in Supplementary note 2 is not a valid proxy for the Hanle signal's nonlinear, trajectory-dependent response. The additional DFT-versus-Hanle discrepancy (4.4-4.6 meV exchange splitting versus fitted B_ex of tens of mT) reinforces the need for caution, but it is secondary for this verdict because DFT may not represent the real 30 nm flake interface. The stray-field simulation is the more direct arbiter of the central claim. The verdict remains conditional pending that check.","tokens_in":17233,"tokens_out":12294,"duration_ms":131121,"concrete_test":"Numerically solve the spatially inhomogeneous Bloch equation used in Supplementary note 1 with B_ex = 0 but with a position-dependent B_s(x,y) computed from the actual CGT flake geometry and the measured magnetization (M = 164 kA/m or the SQUID M(T) of the same crystal), and compare the resulting forward/reverse Hanle curves with the 50 K data in Fig. 2e. If the stray-field-only simulation reproduces the observed peak separation and asymmetry to within noise, the proximity-exchange attribution is not established; if it cannot, the stray-field objection is answered.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the exclusion of stray fields in Supplementary note 2. The authors compute a spatially averaged perpendicular stray field B_s,eff from an idealized bar magnet and compare it with an averaged exchange field B_exch,eff = B_exch L/(L0+2λ) using B_exch = 30 mT. This comparison is not the right test. The nonlocal Hanle signal is a nonlinear functional of the local magnetic field along each spin trajectory, not a function of the channel-averaged field. The stray field from a perpendicularly magnetized 30 nm flake is concentrated near the flake edges, can reach values much larger than the 1-4 mT center-line values quoted, and changes sign across each edge; with only 60% channel overlap, spins are not uniformly exposed to these edge fields. Spins traversing one edge region can therefore contribute shifted or asymmetric Hanle components similar to what the exchange model produces. The comparison also assumes the very quantity at issue, B_exch = 30 mT, so it cannot establish that exchange dominates. Until a position-dependent stray-field calculation is fed through the same Bloch-equation model, the observed peak shifts and asymmetry do not uniquely indicate proximity exchange.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports nonlocal spin-transport measurements on graphene–Cr2Ge2Te6 van der Waals heterostructures. Below about 165 K, the Hanle spin-precession signals develop a hysteretic peak splitting and an asymmetry, which the authors attribute to an out-of-plane proximity-induced exchange field in the CGT-covered part of the graphene channel. The paper also reports a spin-lifetime anisotropy r = τ⊥/τ∥ ≈ 3.9 at B⊥ = 2.5 T, and supports the interpretation with DFT calculations yielding an exchange splitting of about 4.4–4.6 meV and with spin-orbit coupling parameters extracted from model fits. The manuscript includes temperature-dependent measurements, two devices with different CGT/graphene overlap, and extensive supplementary modeling.","tokens_in":17487,"tokens_out":5459,"duration_ms":51335,"significance":"If the exchange-field attribution survives closer scrutiny, this is a significant advance: it would demonstrate proximity-induced out-of-plane ferromagnetic exchange in graphene using a van der Waals magnetic insulator, with implications for spin filtering and topological states. The paper has real strengths: systematic temperature dependence of the Hanle features, a two-device overlap comparison, explicit discussion of stray fields, and DFT calculations with low-energy model fits. It is also candid about ambiguities, noting that the anisotropy could arise from SOC, exchange fluctuations, or both, and that the DFT exchange sign changes with supercell size. However, the quantitative foundation for the central claim is weak: the exchange field is a fitted parameter with a sixfold spread, the stray-field exclusion rests on a channel-averaged comparison, and the fitted mT-scale B_ex is not reconciled with the DFT meV-scale splitting.","major_comments":[{"comment":"The stray-field exclusion is not load-bearing in its current form. The comparison of a channel-averaged B_s,eff with B_exch,eff assumes that the Hanle signal depends on the mean field, but the nonlocal spin signal is a nonlinear functional of the local field along each spin trajectory; with only about 60% channel overlap, edge stray fields that change sign across each edge can generate peak shifts and asymmetries similar to those attributed to B_ex. The estimate also uses B_exch = 30 mT, a fitted value from Supplementary note 4, so it cannot independently establish that exchange dominates. A position-dependent stray-field calculation fed through the same Bloch-equation model is needed before the exchange attribution is secure.","section":"Supplementary note 2"},{"comment":"The quantitative content of the central claim is not settled. The four fits in Supplementary note 4 give B_ex = 33, 42, 58, and 209 mT depending on fitting constraints, a spread of more than a factor of six, and the fitting procedure uses B_ex as an input to generate the same Hanle curves from which it is then extracted. Moreover, these fitted mT-scale fields are three to four orders of magnitude smaller than the DFT exchange splitting of 4.4–4.6 meV, which corresponds to a spin-precession field of tens of tesla; the manuscript offers no reconciliation of this discrepancy.","section":"Supplementary note 4 / Fig. 2e"},{"comment":"The abstract and summary overstate the anisotropy result relative to the paper's own analysis. Supplementary note 6 derives τ_c = 4 ps from the measured r = 3.9 and then calculates r(B⊥ = 0) ≈ 1.002, explicitly concluding that the large anisotropy in Fig. 4d is driven by the external field and may not be intrinsic to the graphene/CGT interface; the main text similarly states that it is unclear whether SOC, exchange fluctuations, or a combination dominates. The claims of a 'proximity-induced anisotropic spin texture' and of an intrinsic larger perpendicular lifetime should be qualified or removed.","section":"Main text, anisotropic spin relaxation / Supplementary note 6"},{"comment":"The DFT support for the proximity exchange is weaker than presented: the exchange splitting changes sign between the 218-atom and 80-atom supercells, and the SOC parameters are obtained from a smaller cell with about 4% strain in CGT. This sign ambiguity and strain sensitivity are acknowledged in the supplement, but they mean the DFT calculations cannot independently confirm the direction or even the existence of a ferromagnetic exchange in the experimental system.","section":"Supplementary note 5"}],"minor_comments":[{"comment":"The Summary section contains a typo: 'existance' should be 'existence'.","section":"Summary"},{"comment":"The typesetting of the anisotropy ratio 'Δ𝑅NL⊥ Δ𝑅NL∥⁄ ∼ 10' and of the formula for r is difficult to parse; please clarify the notation and place the formula on a single line.","section":"Main text, anisotropic spin relaxation"},{"comment":"Equation (S1) is described as derived in reference 2 of the supplement, but the derivation is not reproduced; since this equation underlies all simulations, the authors should either include the derivation or ensure the companion paper (ref. 39 of the main text) is publicly available.","section":"Supplementary note 1"},{"comment":"The statement that B_exch is a constant depending only on magnetization and interface properties is not obviously consistent with the strong temperature dependence of the Hanle features in Fig. 3; a sentence explaining how B_exch is expected to vary with temperature would help.","section":"Supplementary note 2"},{"comment":"In the 218-atom supercell fit, the values λ_ex^A = 4.556 meV and λ_ex^B = 4.558 meV are almost identical, so the sublattice-resolved exchange term is essentially a uniform exchange; this near-degeneracy and its implications for the model should be commented on.","section":"Supplementary note 5"}],"recommendation":"major_revision","confidential_remarks":"The central experimental observation is interesting, but the paper's main quantitative claim needs a more rigorous stray-field treatment and a reconciliation of the fitted and DFT exchange scales before publication. The reliance on an unpublished companion paper for Eq. (S1) is also a reproducibility concern that the editor may wish to address."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is worth your time. It reports the first spin-transport evidence that putting graphene on the 2D ferromagnetic insulator Cr2Ge2Te6 creates an out-of-plane exchange field in the graphene, seen as hysteretic shifts and asymmetry in Hanle spin precession below roughly 165 K. That would be a real step toward proximity-induced quantum anomalous Hall physics. The two-region partial-overlap model that explains the asymmetric Hanle peak is a useful contribution, and the DFT work gives concrete microscopic expectations: exchange splitting near 4.5 meV plus Rashba and valley-Zeeman SOC. The authors also deserve credit for Supplementary Note 6, where they show that the strong spin-lifetime anisotropy could be driven by the external field quenching exchange fluctuations rather than an intrinsic spin texture. Not everyone would have included that.\n\nThe soft spot is the quantitative core. The exchange field extracted from Hanle fits ranges from 33 to 209 mT depending on fitting constraints, a wide spread the paper acknowledges but does not resolve. That range is two to three orders of magnitude below the 4.4–4.6 meV exchange splitting from DFT, which corresponds to tens of tesla. The paper never reconciles this discrepancy. It is possible the transport measurement sees only a small net component, or that the fits average over domains, but the gap is too large to leave open. The stray-field estimate in Supplementary Note 2 is also not decisive: it compares spatially averaged stray and exchange fields, while the Hanle signal depends nonlinearly on the local field along each spin trajectory. With a 60% overlap, edge-region stray fields could plausibly mimic the observed shifts and asymmetry. The authors need a position-dependent stray-field calculation fed through the same Bloch-equation model before the exchange attribution is secure.\n\nStill, the experimental observation is new and the paper engages honestly with alternative explanations. It deserves a serious referee, but the referee should demand a sensitivity analysis, a direct discussion of the DFT-versus-Hanle contradiction, and a more careful stray-field treatment. I would not cite it in its current form, but I would bring it to a reading group to argue over the interpretation.\n\nRecommendation: send it to peer review, but expect major revision.","headline":"First spin-transport evidence for out-of-plane proximity exchange in graphene/CGT, but the mT-scale fitted exchange field is not reconciled with DFT, and the stray-field control is too rough to fully rule out edge magnetism.","tokens_in":18096,"tokens_out":1759,"would_cite":false,"duration_ms":18243,"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":"The paper claims that graphene on the ferromagnetic insulator Cr2Ge2Te6 develops an out-of-plane proximity exchange field, seen as shifted and asymmetric Hanle spin-precession curves and as perpendicular spin lifetimes about 3.9 times…","keywords":["magnetic proximity effect","van der Waals heterostructure","graphene spintronics","Cr2Ge2Te6","Hanle spin precession","exchange interaction","spin lifetime anisotropy","two-dimensional ferromagnet"],"falsifier":"Fabricate the same device geometry but insert a thin hexagonal boron nitride spacer between graphene and CGT: the stray magnetic field of the flake stays nearly the same, while direct exchange proximity is cut off. If the Hanle peak shift and asymmetry disappear in the spacer device, the proximity-exchange reading is correct; if they persist unchanged, stray fields are the dominant cause.","tokens_in":17006,"feed_emoji":"🧲","tokens_out":6039,"duration_ms":59786,"temperature":0.7,"pith_summary":"This paper tries to show that graphene can inherit magnetic order simply by resting on a layered magnetic insulator, Cr2Ge2Te6, with no electrical contact or current. The evidence is nonlocal spin-transport data: below about 165 K the Hanle spin-precession curves shift and become asymmetric in a way that matches a proximity-induced out-of-plane exchange field acting on the graphene channel. A second signature is anisotropic spin relaxation, with out-of-plane spins living about 3.9 times longer than in-plane spins, which the authors attribute to a modified spin texture in the graphene. If true, this gives a practical route to making graphene magnetic by proximity, a step toward spin filters and topologically protected edge states that do not need an external magnetic field.","feed_headline":"A magnetic insulator imprints out-of-plane spin order on graphene","feed_subtitle":"Hanle peaks split below 165 K, and perpendicular spins live 3.9 times longer than in-plane spins.","key_machinery":"The load-bearing object is the nonlocal Hanle spin-precession signal, analyzed with a solution of the Bloch equation for a channel with two regions: bare graphene and graphene under CGT. In the covered region, the perpendicular field entering the Larmor precession frequency is $B_\\perp + B_\\mathrm{ex}$, where $B_\\mathrm{ex}$ is the proximity exchange field; hysteresis in $B_\\mathrm{ex}$ shifts the Hanle peaks between sweep directions, and the finite length of the covered region makes the central peak asymmetric. A second piece of machinery is the low-energy Hamiltonian for graphene near the Dirac points, with sublattice-resolved exchange coupling $\\lambda_\\mathrm{ex}^A$, $\\lambda_\\mathrm{ex}^B$ and spin-orbit terms; fitting this to DFT bands yields the ferromagnetic exchange splitting $(\\lambda_\\mathrm{ex}^A + \\lambda_\\mathrm{ex}^B)/2 \\approx 4.4$ to $4.6$ meV and the spin-orbit parameters $\\lambda_R$ and $\\lambda_{VZ}$.","core_discovery":"The central claim is that in a van der Waals heterostructure of graphene and the ferromagnetic insulator Cr2Ge2Te6, the graphene acquires a proximity-induced ferromagnetic exchange interaction with an out-of-plane easy axis. This is inferred from nonlocal Hanle spin precession: below the magnetic ordering temperature, the Hanle peaks for opposite field sweep directions are shifted relative to each other and the central peak is asymmetric, features that simulations reproduce only when the CGT-covered part of the channel experiences an exchange field $B_\\mathrm{ex}$ perpendicular to the graphene plane. The same heterostructure shows a spin-lifetime anisotropy $\\tau_\\perp/\\tau_\\parallel \\approx 3.9$, which the paper interprets as evidence of a proximity-induced anisotropic spin texture. Density functional theory calculations support the picture by giving a ferromagnetic exchange splitting of roughly 4.4 to 4.6 meV in the graphene layer, alongside a Rashba spin-orbit coupling of 0.253 meV and a valley-Zeeman coupling of 0.113 meV.","pith_inferences":["The tens-of-millitesla exchange field fitted from the Hanle data is about a thousand times smaller than the field corresponding to the DFT exchange splitting of roughly 4.5 meV, so the measured devices likely have a much weaker effective interface than the ideal DFT stack; thickness- and twist-angle-dependent studies could test this directly.","If stray fields were the real cause of the Hanle shift, the effect should scale with flake volume and edge configuration, so an hBN-spacer control device would cleanly separate stray-field effects from genuine exchange proximity.","The persistence of the signal above the bulk Curie temperature hints at interface- or surface-driven magnetic order, so local magnetometry on the same flakes could identify which region of the CGT is actually magnetically active.","The sign change of the DFT exchange splitting between the two supercells suggests that twist angle and strain control the proximity effect, making it a tunable parameter rather than a fixed material property."],"forward_implications":["Below about 165 K, graphene partially covered by CGT should act as a spin-transport channel with a built-in out-of-plane exchange field, modifying spin precession even at zero applied field.","The spin relaxation in such a channel becomes anisotropic, with perpendicular spins living roughly 3.9 times longer than in-plane spins; this anisotropy is a measurable fingerprint of a proximity-modified spin texture.","If the exchange and spin-orbit parameters can be tuned through the van der Waals gap, flake overlap, or double-sided coverage, the same platform could be pushed toward spin filtering and the quantum anomalous Hall state.","Hanle peak splitting and asymmetry provide a nonlocal electrical probe of interfacial magnetism that works down to small exchange fields."],"supporting_citations":[{"why":"Establishes that Cr2Ge2Te6 is an intrinsic two-dimensional ferromagnet with perpendicular anisotropy, motivating the choice of material.","marker":"[17]"},{"why":"Provides the model Hamiltonian and DFT framework for graphene on Cr2X2Te6 that the paper's calculations extend.","marker":"[31]"},{"why":"Supplies the Bloch-equation solution with spatially inhomogeneous spin transport used to simulate the Hanle curves.","marker":"[39]"},{"why":"Gives the pristine CVD graphene spin-transport baseline that the temperature dependence of the spin signal is compared against.","marker":"[38]"},{"why":"Documents static and fluctuating proximity exchange fields in graphene, the mechanism invoked for enhanced spin relaxation near the Curie temperature.","marker":"[15]"},{"why":"Introduces the giant spin-lifetime anisotropy mechanism from proximity-induced spin-orbit coupling used to interpret the measured anisotropy.","marker":"[44]"},{"why":"Provides the formula and method connecting nonlocal signal ratios to the spin-lifetime anisotropy parameter.","marker":"[43]"}],"fun_headline_variants":["Graphene gains out-of-plane spins from magnetic neighbor","Magnetic proximity gives graphene perpendicular spins","Graphene's spin order goes perpendicular via proximity","Proximity-induced out-of-plane spin order in graphene","Graphene gets longer-lived perpendicular spins from magnet"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument stands on the premise that the Hanle peak shifts and asymmetry are dominated by a proximity exchange field in the CGT-covered graphene, with stray fields from the flake contributing only a small correction, while the fitted exchange field of a few tens of millitesla is taken at face value even though the DFT exchange splitting corresponds to an exchange field tens of tesla, roughly a thousand times larger.","fun_headline_variants_meta":{"raw":{"variants":["Graphene gains out-of-plane spins from magnetic neighbor","Magnetic proximity gives graphene perpendicular spins","Graphene's spin order goes perpendicular via proximity","Proximity-induced out-of-plane spin order in graphene","Graphene gets longer-lived perpendicular spins from magnet"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000281,"raw_usage":{"total_tokens":1652,"prompt_tokens":919,"completion_tokens":733,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":535,"completion_tokens_details":{"reasoning_tokens":658}},"tokens_in":535,"tokens_out":733,"duration_ms":7471,"temperature":1.0,"reasoning_tokens":658,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:11:43.504268+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fabricate the same device geometry but insert a thin hexagonal boron nitride spacer between graphene and CGT: the stray magnetic field of the flake stays nearly the same, while direct exchange proximity is cut off. If the Hanle peak shift and asymmetry disappear in the spacer device, the proximity-exchange reading is correct; if they persist unchanged, stray fields are the dominant cause.","supporting_citations":[{"cited_title":"& Fabian, J","cited_arxiv_id":null,"evidence_quote":"Provides the model Hamiltonian and DFT framework for graphene on Cr2X2Te6 that the paper's calculations extend."}],"review_version":1}