{"id":"82b5c94e-231a-4182-9bb9-02b619278d46","arxiv_id":"2505.23482","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Nonlocal thermally driven magnon transport in few-layer NiPS3 exhibits sharp angular jumps near the b-axis and a 1/d^2 distance decay, interpreted as an anisotropic in-plane spin-flop transition plus intrinsic spin Seebeck effect.","lead":"In thin flakes of the antiferromagnet NiPS3, heat-generated magnon signals detected by platinum contacts show sharp jumps as the in-plane magnetic field direction is rotated, which the authors link to an anisotropic spin-flop transition. The result suggests nonlocal magnon transport can serve as a sensitive electrical probe of magnetic phase transitions in van der Waals magnets.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central interpretation hinges on an unobserved b-axis magnetic domain under the Pt electrodes; a local probe of the Néel vector under Pt would test the model.","rationale":"The reader's weakest assumption identifies exactly the b-axis magnetic domain under the Pt strips as the load-bearing element, and I agree. The paper's own SI text flags this as an assumption rather than a measured result, stating that the domain is 'reasonable to expect' from Pt-induced strain and thermal fluctuations, and explicitly notes the difficulty of detecting it with conventional optical probes. The proposed two-domain model is the sole mechanism that connects the observed sharp angular jumps to the anisotropic spin-flop transition; without the b-axis domain, the assignment of the B//b behavior to a spin-flop transition in a second domain is unsupported. The alternative explanations—such as coherent rotation of a single easy-axis domain or domain redistribution at low fields—are not excluded by the data presented. The concrete SMR test is feasible because the same Pt/NiPS3 devices are used for both transport and SMR measurements, and the first-harmonic SMR response is a direct probe of the interfacial Néel vector orientation. If SMR fails to show a b-axis anisotropy under Pt, the central interpretation would need substantial revision, possibly to a single-domain coherent-rotation model; if SMR confirms it, the conditional acceptance would be strongly justified. I therefore keep the reader's CONDITIONAL verdict unchanged rather than escalating it, because the concern is precisely the one already identified, and the proposed test is the appropriate way to resolve it.","tokens_in":17455,"tokens_out":5015,"duration_ms":48988,"concrete_test":"Measure first-harmonic spin Hall magnetoresistance (SMR) of the same Pt/NiPS3 device at 20 K as a function of in-plane field angle. SMR is a local probe of the Néel vector orientation under the Pt contact. If the b-axis domain in Fig. 4a exists, the SMR angular dependence should show an easy-axis-like response along the b-axis (or a symmetry rotated by 90° relative to a pristine a-axis flake), and the field sweep at B//b should show the same turnover field as the nonlocal signal. Compare SMR anisotropy on the Pt strip with that on a control flake without Pt. If SMR shows only an a-axis response, the two-domain model is falsified; if it shows a b-axis response, the interpretation is directly supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the angular jumps in the thermally driven nonlocal signal mark an anisotropic in-plane spin-flop transition—rests on the two-domain model in Fig. 4a. In that model, a strain-stabilized domain with spins along the NiPS3 b-axis must exist under the Pt strips. This domain is never directly observed. The SI section 'The origin of the proposed domain structure with spin along the b-axis' says only that such a domain is 'reasonable to expect' from Pt-induced strain and thermal fluctuations, citing Tan et al. and Lee et al., and admits that 'detecting such a small-scale magnetic domain structure at the Pt/NiPS3 interface is challenging using traditional polarized PL or LD techniques.' The Raman evidence (Eg(3)/Eg(4) disappearance) demonstrates structural disorder under Pt, not that the magnetic easy axis rotates by 90°. Without the b-axis domain, the spin-flop assignment for B//b is unsupported; alternative mechanisms such as field-driven coherent rotation or domain reorientation in a weakly anisotropic easy-plane system could produce similar jumps, and the model in Fig. 4a loses its basis. This is the load-bearing assumption: it is the link between the observed phenomenology (jumps near the b-axis, field-dependent upturn, device-orientation shift) and a specific magnetic phase transition. The other evidence—EBSD-calibrated axes, PL anisotropy, reproducibility across devices—does not constrain this domain's existence.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports nonlocal second-harmonic voltage measurements in few-layer NiPS3 devices with Pt injector/detector strips. The authors observe sharp jumps in the angular dependence of the thermally driven magnon signal when the in-plane magnetic field is near the b-axis at high field (9 T), and a weak upturn in the field dependence for B//b. They attribute these features to an in-plane anisotropic spin-flop transition occurring in a hypothesized b-axis-aligned magnetic domain under the Pt strips, while a coexisting a-axis domain undergoes coherent rotation. They also report that the nonlocal signal decays as 1/d^2 in a 12-nm-thick flake, which they ascribe to the intrinsic spin Seebeck effect. The manuscript includes EBSD and polarized photoluminescence characterization of the crystal axes, measurements on three devices with different Pt-strip orientations, and a YIG control sample.","tokens_in":17714,"tokens_out":3408,"duration_ms":36083,"significance":"If the interpretation is correct, the paper would demonstrate electrical detection of an anisotropic spin-flop transition in a van der Waals antiferromagnet via thermally driven magnon transport, which is a valuable addition to the growing effort on nonlocal magnon transport in 2D magnets. The reported reproducibility across multiple devices, the EBSD-PL axis calibration, the YIG comparison, and the field-dependent measurements are genuine strengths: they establish that the angular jumps are not a single-device artifact and that the phenomenon is tied to the b-axis direction. The 1/d^2 distance dependence, if robust, is also significant evidence for the intrinsic spin Seebeck mechanism in a thin antiferromagnet. However, the central claim is weakened by a load-bearing assumption that is not directly verified and by the lack of a quantitative model connecting the two-domain picture to the measured angular curves.","major_comments":[{"comment":"The central interpretation hinges on an unobserved magnetic domain with spins along the b-axis underneath the Pt strips. The SI explicitly states that such a domain is 'reasonable to expect' from Pt-induced strain and thermal fluctuations, but also acknowledges that 'detecting such a small-scale magnetic domain structure at the Pt/NiPS3 interface is challenging using traditional polarized PL or LD techniques.' The authors then use this postulated domain to explain the observed jumps (Figure 4a). This is circular: the domain is inferred from the same jumps it is invoked to explain, and no independent evidence is provided. The Raman data in Figure S7 demonstrate structural disorder under Pt, but do not establish a 90-degree rotation of the easy axis. Because the spin-flop assignment rests on this domain, the paper needs either direct experimental evidence for the b-axis domain (for example, spatially resolved magnetic imaging with sensitivity under the Pt, or an alternative geometry where the domain can be probed) or a substantial reframing of the conclusions as a hypothesis rather than a demonstration.","section":"Supporting Information, 'The origin of the proposed domain structure with spin along the b-axis of NiPS3'"},{"comment":"The two-domain model is presented only qualitatively. No equation or fitting procedure is given for the angular dependence V_NL(B, θ), nor for the jump positions across devices. The authors state that the jumps shift from 85° to 125° to 150° for devices 1-3 and that 'all these arrows are similarly aligned to the b-axis,' but the angular offset between the Pt strips and the b-axis is treated as a free parameter. Without a quantitative model, it is not demonstrated that a spin-flop transition in a b-axis domain, combined with coherent rotation in an a-axis domain, reproduces the sharp jumps, their angle-dependent amplitudes, or the near-zero signal at 90°. A quantitative fit—using the known anisotropy and exchange fields, or at least a minimal two-domain parametrization—is needed to make the interpretation convincing.","section":"§4, Figure 4a and Figure 3b"},{"comment":"The identification of the features as a spin-flop transition requires a comparison with the expected spin-flop field scale for NiPS3. The authors cite Basnet et al. for the bulk spin-flop transition but do not show that the observed threshold (between 7 and 9 T near the b-axis, and absent along the a-axis) is quantitatively consistent with the known H_E and H_A of NiPS3. The field-dependent upturn in Figure 4d is described as 'small' and is attributed to a 2-3 degree angular offset; this is a weak signature on which to base the central claim. Alternative mechanisms, such as field-driven domain-wall motion or coherent rotation in a multidomain easy-plane antiferromagnet, could produce abrupt changes in the net magnetization component and therefore in the nonlocal SSE signal. The authors should either identify a feature that is uniquely diagnostic of a spin-flop transition or explicitly discuss why these alternatives are excluded.","section":"§3, Figure 3a and Figure 4d"}],"minor_comments":[{"comment":"The abstract contains a typo: 'mag non transport' should read 'magnon transport'.","section":"Abstract"},{"comment":"The 1/d^2 fit in Figure 5c appears to be based on a small number of points without error bars. The authors should provide the number of data points, the fitting statistic, and ideally show the data points with uncertainties, to justify the claim that 1/d^2 is preferred over an exponential decay.","section":"Figure 5c"},{"comment":"The statement that for the 12-nm flake 'λm cannot be extracted by the typical exponential fitting' is vague. The authors should specify the fitting range and the criterion used (e.g., R^2, residual analysis) to support the conclusion that the decay is not exponential.","section":"§5, Figure 5 and Figure S6"},{"comment":"The manuscript would benefit from a table listing, for each device, the Pt-strip orientation with respect to the b-axis, the measured jump angle, and the inferred angular offset. This would make the device-to-device comparison in Figure 3b more transparent than the qualitative description.","section":"§3, Figure 3b"},{"comment":"The y-axis label uses an inconsistent notation for the normalized second-harmonic voltage; the authors should define V_NL^th/I^2 explicitly in the figure caption or at first use in the text.","section":"Figure 2c"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and reports a reproducible experimental phenomenon, but the central claim of a spin-flop transition relies on an unverified b-axis domain under the Pt contacts. The SI limitation passage should be weighed seriously; the authors could substantially strengthen the paper by adding a quantitative model or an independent test. If the domain cannot be verified, the framing should be moderated to 'consistent with' rather than 'demonstrates.' I recommend major revision to allow for these changes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a good experimental paper that deserves serious referee time, but the central conclusion should be softened. The data are genuinely new—nonlocal thermal magnon transport in few-layer easy-plane NiPS3 showing sharp angle jumps that track the b-axis across three devices, plus a field-dependent upturn near the b-axis and a 1/d^2 decay in a 12-nm flake. That is a useful addition to the 2D magnon-transport literature, which has so far focused on easy-axis MnPS3 and CrPS4 with out-of-plane fields.\n\nWhat the paper does well: the EBSD + polarized PL calibration gives a credible crystallographic reference frame; the jumps appear at angles that shift with device orientation and line up with the b-axis; the effect reproduces across devices and at different temperatures and currents. The 1/d^2 decay, though based on a few points, is consistent with the intrinsic spin Seebeck regime and is a reasonable claim given the short magnon relaxation length in thin NiPS3.\n\nThe soft spot is the interpretation. The model in Figure 4a requires a magnetic domain under the Pt strips with spins along the b-axis, and that domain is never directly observed. The SI says it is 'reasonable to expect' from Pt-induced strain and thermal fluctuations, citing Tan and Lee, but those works see domains at ±120° from the a-axis or 10–30° offsets, not a 90° b-axis domain. The Raman evidence shows structural disorder under Pt, not a rotated easy axis. So there is a circular element: the domain is inferred from the jumps and then used to explain them. The angular curves also lack error bars and the model is not fitted quantitatively. I don't think this is fatal—the field-dependent upturn near the b-axis is independent support for a spin-flop-like reorientation, and the reproducibility is a genuine internal check—but it leaves the paper as a strong experimental observation with a plausible-speculative assignment.\n\nBottom line: this is a paper for the magnon-transport and 2D-magnet communities. It should go to peer review, not be desk rejected. I would ask the authors to either image the domain under Pt, fit a quantitative two-domain model to the angular curves, or soften the spin-flop language to a field-induced magnetic reorientation. As is, I would cite it for the data and the method.","headline":"Sharp angular jumps in nonlocal magnon signal near the b-axis of NiPS3 are a real and reproducible effect, but the spin-flop explanation depends on a b-axis domain that is inferred, not observed.","tokens_in":18323,"tokens_out":2712,"would_cite":true,"duration_ms":24350,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"In few-layer NiPS3, the nonlocal voltage from thermally driven magnons jumps sharply at specific in-plane field angles near the b-axis, which the authors attribute to an anisotropic spin-flop transition; the $1/d^2$ decay marks the…","keywords":["magnon transport","spin-flop transition","NiPS3","van der Waals antiferromagnet","spin Seebeck effect","nonlocal magnon detection","magnetic anisotropy","two-dimensional magnets"],"falsifier":"A direct magnetic imaging measurement of the Pt-covered region of a similar NiPS3 device at 9 T and 20 K that sees no magnetic domain with spins along the b-axis would undercut the two-domain explanation for the angular jumps. Alternatively, if a field sweep with B exactly along the b-axis of a device whose b-axis is independently confirmed by EBSD shows no upturn or sharp feature, the spin-flop assignment would fail.","tokens_in":17234,"feed_emoji":"🧲","tokens_out":8237,"duration_ms":74942,"temperature":0.7,"pith_summary":"This paper tries to show that nonlocal magnon transport can detect a magnetic phase transition in the easy-plane van der Waals antiferromagnet NiPS3, where the spins lie in the plane of the flake. Heating a platinum strip injects thermally excited magnons into NiPS3, and a second platinum strip converts the arriving spin current into a voltage via the inverse spin Hall effect. The authors find that this nonlocal voltage, measured as the field angle is rotated at 9 T and 20 K, shows sharp jumps near the b-axis instead of the smooth $\\sin\\theta$ curve seen in YIG/Pt, and they attribute the jumps to an in-plane anisotropic spin-flop transition, a sudden reorientation of the antiparallel spin lattice at a threshold field whose value depends on field direction. In a 12 nm flake the signal decays as $1/d^2$ with injector-detector separation, which they take as evidence that the signal is dominated by the intrinsic spin Seebeck effect rather than magnon diffusion. The paper thereby positions nonlocal magnon transport as a sensitive probe of magnetic phase transitions in two-dimensional antiferromagnets, and it argues that a b-axis magnetic domain hidden under the platinum contacts is responsible for the jumps.","feed_headline":"Sharp jumps in magnon signal reveal spin-flop in NiPS3","feed_subtitle":"A 9 T in-plane field flips the spin texture only near the b-axis; signal falls as 1 over distance squared.","key_machinery":"The load-bearing object is the second-harmonic nonlocal voltage measured between two parallel platinum strips on an insulating NiPS3 flake. Joule heating in the injector creates a radial temperature gradient and a magnon chemical potential imbalance that drives thermally excited magnons toward the detector, where the inverse spin Hall effect converts the spin current into a transverse voltage. The argument rests on the angular dependence of this voltage as an in-plane magnetic field is rotated: at high field the response is modelled as the sum of contributions from an a-axis-aligned magnetic domain that rotates coherently and a b-axis-aligned domain that undergoes a spin-flop transition, with the jumps appearing at field directions where the spin-flop-induced magnetization component normal to the platinum strips changes sign or vanishes. The $1/d^2$ decay in the 12 nm flake is attributed to the intrinsic spin Seebeck effect from the bottom interface, which becomes dominant when the magnon diffusion length is short.","core_discovery":"The central claim is that the spin-flop transition in few-layer NiPS3 becomes visible in the angular dependence of thermally driven magnon transport. In devices with platinum strips oriented near the b-axis, the second-harmonic nonlocal voltage is not the smooth $\\sin\\theta$ response observed in YIG/Pt but shows sharp jumps when the in-plane magnetic field points near the b-axis; these jumps appear only above about 7 T and disappear toward lower fields, and the angle at which they occur rotates with the Pt-strip orientation and always maps back to the b-axis. Field sweeps with the field along the b-axis show a small upturn consistent with a spin-flop transition, while sweeps along the a-axis do not, showing that the spin-flop field is anisotropic. The authors explain the behaviour with two coexisting magnetic domains, one with spins along the a-axis, previously seen, and one with spins along the b-axis, induced by strain and thermal effects under the platinum strips, and they further report that in a 12 nm flake the signal amplitude obeys $1/d^2$ over 0.8–3.8 μm, which they attribute to the intrinsic spin Seebeck effect from the NiPS3/substrate interface.","pith_inferences":["Beyond the paper: the jump angle in the nonlocal signal could be used as a local, contact-level magnetometer for strain-induced magnetic anisotropy, since the jump position tracks the local easy direction rather than the global crystal axis.","Beyond the paper: if strain from the platinum is what stabilises the b-axis domain, then choosing different contact metals, strip widths, or annealing conditions could tune the domain population and thereby engineer when the spin-flop jump appears.","Beyond the paper: the $1/d^2$ regime at micrometre distances suggests that in other short-diffusion-length van der Waals antiferromagnets, nominally 'nonlocal' signals may contain a substantial intrinsic spin Seebeck component, and distance sweeps should be a routine control."],"forward_implications":["If the spin-flop assignment is right, angle-dependent nonlocal magnon measurements can map the in-plane anisotropy of the spin-flop field in easy-plane van der Waals antiferromagnets.","The $1/d^2$ decay means that micrometre-separated nonlocal signals in thin NiPS3 are dominated by the substrate-interface spin Seebeck effect, so distance-dependent measurements are necessary to separate diffusion from intrinsic spin Seebeck contributions.","The proposed b-axis domain implies that platinum deposition modifies the magnetic texture underneath the contacts, so contact engineering can change the observed magnon transport response.","The absence of a first-harmonic signal while second-harmonic thermal magnon signals are measurable indicates that thermally driven magnons are the accessible transport channel in this material system."],"supporting_citations":[{"why":"Reports the spin-flop transition in bulk NiPS3 by magnetometry, supplying the bulk reference the few-layer assignment is compared with.","marker":"[19]"},{"why":"Establishes nonlocal magnon transport in quasi-2D van der Waals antiferromagnets and gives the method for extracting the magnon diffusion length.","marker":"[24]"},{"why":"Detects a spin-flop transition in MnPS3 via thermally generated magnon transport, providing the direct precedent for this detection strategy.","marker":"[33]"},{"why":"Shows strain-dependent multidomain structures in thin NiPS3 flakes, used to justify the strained b-axis domain under the platinum.","marker":"[40]"},{"why":"Images small thermally fluctuating domains offset from the a-axis in NiPS3, supporting the possibility of a b-axis-oriented domain.","marker":"[41]"},{"why":"Links the polarization of NiPS3 photoluminescence to spin-induced in-plane magnetic anisotropy, grounding the b-axis identification in the devices.","marker":"[45]"},{"why":"Sets criteria for distinguishing magnon diffusion from the nonlocal spin Seebeck effect, used in interpreting the $1/d^2$ decay.","marker":"[53]"},{"why":"Demonstrates thermally driven long-range magnon currents in YIG from the intrinsic spin Seebeck effect, the reference $1/d^2$ behaviour.","marker":"[54]"},{"why":"Establishes the linear field dependence of the net magnetization in an easy-axis antiferromagnet, used to explain the field dependence of the nonlocal signal below the transition.","marker":"[58]"}],"fun_headline_variants":["Magnon jumps expose anisotropic spin-flop in NiPS3","Spin-flop transition in NiPS3 seen via sharp magnon voltage jumps","Unconventional magnon transport from anisotropic spin-flop in NiPS3","Sharp magnon jumps reveal spin-flop's anisotropy in NiPS3","Van der Waals antiferromagnet shows spin-flop in magnon transport"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model assumes that a small magnetic domain with spins aligned along the b-axis of NiPS3 exists underneath the platinum strips, even though no direct image of that domain is reported; remove that domain and the angular-jump mechanism has no support.","fun_headline_variants_meta":{"raw":{"variants":["Magnon jumps expose anisotropic spin-flop in NiPS3","Spin-flop transition in NiPS3 seen via sharp magnon voltage jumps","Unconventional magnon transport from anisotropic spin-flop in NiPS3","Sharp magnon jumps reveal spin-flop's anisotropy in NiPS3","Van der Waals antiferromagnet shows spin-flop in magnon transport"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000741,"raw_usage":{"total_tokens":3335,"prompt_tokens":998,"completion_tokens":2337,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":614,"completion_tokens_details":{"reasoning_tokens":2239}},"tokens_in":614,"tokens_out":2337,"duration_ms":16284,"temperature":1.0,"reasoning_tokens":2239,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:44:28.073613+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct magnetic imaging measurement of the Pt-covered region of a similar NiPS3 device at 9 T and 20 K that sees no magnetic domain with spins along the b-axis would undercut the two-domain explanation for the angular jumps. Alternatively, if a field sweep with B exactly along the b-axis of a device whose b-axis is independently confirmed by EBSD shows no upturn or sharp feature, the spin-flop assignment would fail.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the spin-flop transition in bulk NiPS3 by magnetometry, supplying the bulk reference the few-layer assignment is compared with."},{"cited_title":"C.; Han, W., Magnon Transport in Quasi-Two-Dimensional Van Der Waals Antiferromagnets","cited_arxiv_id":null,"evidence_quote":"Establishes nonlocal magnon transport in quasi-2D van der Waals antiferromagnets and gives the method for extracting the magnon diffusion length."},{"cited_title":"M.; van Wees, B","cited_arxiv_id":null,"evidence_quote":"Detects a spin-flop transition in MnPS3 via thermally generated magnon transport, providing the direct precedent for this detection strategy."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows strain-dependent multidomain structures in thin NiPS3 flakes, used to justify the strained b-axis domain under the platinum."},{"cited_title":"A.; Gao, H.; Li, J.; Kitadai, H.; Comin, R.; Ling, X., Observation of Three-State Nematicity and Domain Evolution in Atomically Thin Antiferromagnetic NiPS","cited_arxiv_id":null,"evidence_quote":"Images small thermally fluctuating domains offset from the a-axis in NiPS3, supporting the possibility of a b-axis-oriented domain."},{"cited_title":"R.; Simonet, V .; Ressouche, E.; Mcintyre, G","cited_arxiv_id":null,"evidence_quote":"Links the polarization of NiPS3 photoluminescence to spin-induced in-plane magnetic anisotropy, grounding the b-axis identification in the devices."},{"cited_title":"E.; Casanova, F., Large Room Temperature Spin-to-Charge Conversion Signals in a Few-Layer Graphene/Pt Lateral Heterostructure","cited_arxiv_id":null,"evidence_quote":"Demonstrates thermally driven long-range magnon currents in YIG from the intrinsic spin Seebeck effect, the reference $1/d^2$ behaviour."},{"cited_title":"M.; Rodríguez-Suárez, R","cited_arxiv_id":null,"evidence_quote":"Establishes the linear field dependence of the net magnetization in an easy-axis antiferromagnet, used to explain the field dependence of the nonlocal signal below the transition."}],"review_version":1}