{"id":"b50c33e5-6d65-42cf-a7d4-95b081f032af","arxiv_id":"1908.10626","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In field-driven pinwheel artificial spin ice, island end-states produce a field-induced Heisenberg-like pseudo-exchange coupling that sets the reversal path and explains emergent anisotropies, with strength tunable by island size.","lead":"This paper shows that in a pinwheel pattern of tiny magnetic islands, the stray fields from the curved magnetisation at each island's ends create an effective nearest-neighbour coupling that controls how the whole array flips under an external magnetic field. It explains a previously puzzling tilt of the array's magnetic easy axes and shows that shrinking the islands tunes this behaviour, a useful lever for future magnetic devices.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The pseudo-exchange claim is inferred from switching-field astroids, not from an energy functional; dynamic thresholds with metastable noise may not establish a Heisenberg-like coupling.","rationale":"The paper has genuine supporting evidence: time-resolved reversal shows end-state-mediated synchronized switching, the size-reduction sequence in Sec. VIII shows the strong-coupling window narrowing as islands approach point dipoles, and the reversal patterns match experiment qualitatively. These make the end-state mechanism credible and justify a conditional rather than a reject verdict. The load-bearing weakness is not merely the exact plateau values, but that the central interaction is inferred from switching fields rather than from an energy functional. That matters because the claim is specifically that a Heisenberg pseudo-exchange term exists and governs the reversal; a dynamic threshold proxy cannot by itself establish an energy term, and the acknowledged metastable noise in footnote 34 shows the proxy is noisy. The proposed constrained-energy simulations and averaged sweeps would settle whether the coupling is real and whether the quoted anisotropy angles are reliable. If the fitted J is absent or small, the paper's main interpretation would need to be weakened to a field-induced synchronization effect. This is addressable with additional computation, so the reader's conditional verdict should stand unchanged.","tokens_in":17973,"tokens_out":10335,"duration_ms":123737,"concrete_test":"Using the same MuMax3 parameters, constrain the average magnetization directions of the two islands in the T-shaped dimer (and, separately, the four islands in the pinwheel vertex) to a grid of orientations around the easy axes, relax all other degrees of freedom (end states) at H = 0 and at an applied field just below Hs, and record the total micromagnetic energy. Fit E(theta1,theta2) to a bilinear Heisenberg term plus single-ion anisotropy terms, e.g. E = -J cos(theta1-theta2) + K1 sin^2(theta1) + K2 sin^2(theta2), with theta_i measured from each island long axis. Check that J is nonzero with the sign and angular range implied by the strongly coupled regime, and that the stationary points of E reproduce the reported anisotropy axes within 0.25°. Independently, repeat the Hs(theta) sweeps of Figs.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Central claim: end-state dipolar fields produce a Heisenberg-like inter-island coupling that drives the strongly coupled regime, the misaligned anisotropy axes, and the avalanche reversal. The paper states in Sec. II that it 'mainly use[s] the magnetisation switching field values as proxies to the net interactions', and all quantitative outputs—the strongly coupled regime boundaries, the anisotropy axes (38.21° for the T-dimer, 49.50° for the vertex, 43.25°/48.00° plateau angles), and the cubic/uniaxial decomposition—are read from maxima and spacings of Hs(θ) astroids. A switching field is a dynamic reversal threshold, not an equilibrium energy derivative; it depends on the field-step protocol, damping, and metastable states. Footnote 34 acknowledges that the Hs curves contain 'noise' from metastable states and that the data were obtained from a single set of simulations. Nothing in the paper extracts an interaction energy from the micromagnetic configurations or fits the angular dependence of the energy to a Heisenberg-like term. The named 'pseudo-exchange' is therefore an interpretation of switching-field phenomenology rather than a demonstrated energy contribution. If the energy landscape lacks the inferred bilinear coupling, or if metastable/dynamic artifacts shift the Hs maxima, the central mechanistic claim is not supported at the level asserted.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a micromagnetic simulation study of field-driven reversal in pinwheel artificial spin ice, using 470 nm × 170 nm × 10 nm Permalloy-like islands with 420 nm center-to-center spacing. It examines T-shaped dimers, four-island pinwheel vertices, finite arrays with symmetric and asymmetric edges, and periodic-boundary arrays, and compares selected reversal patterns with Lorentz transmission electron microscopy experiments. The central claims are that magnetization end-states on extended islands produce a 'Heisenberg pseudo-exchange interaction' between nearest neighbors, that this interaction creates a strongly coupled angular regime near 45°, that it misaligns emergent anisotropy axes from the geometrical axes (38.21° for the T-dimer, 49.50° for the vertex, and plateau values 43.25°/48.00° for large arrays), and that it drives corner-mediated avalanche reversal. The paper also shows that reducing island size narrows the strongly coupled regime, moving the behavior toward the point-dipole limit.","tokens_in":18214,"tokens_out":6045,"duration_ms":69841,"significance":"If substantiated, the proposed mechanism would explain previously puzzling superferromagnetic reversal and anisotropy-axis misalignment in pinwheel ASI, and it would establish island size as a tuning knob for collective reversal behavior. The paper has several clear strengths: the simulation campaign is systematic across geometries and sizes; the use of periodic boundary conditions to restore the 45° symmetry axis is a well-designed control; the comparison between micromagnetic, uniformly magnetized, and point-dipole models in Fig. 11 directly isolates the role of end-states; and the qualitative comparison to LTEM experiments supports the relevance of the predicted avalanche reversal. The main weakness is that the term 'Heisenberg pseudo-exchange' is asserted from switching-field phenomenology rather than demonstrated from an energy functional, and the quantitative anisotropy angles and regime widths rest on single-run astroids with acknowledged metastable 'noise'.","major_comments":[{"comment":"The central 'Heisenberg pseudo-exchange' claim is not established as an energy contribution. Section II states that the paper 'mainly use[s] the magnetisation switching field values as proxies to the net interactions,' and the evidence for strong coupling—the snapping together of Hs(θ) curves and synchronized reversal in Figs. 4–6—comes from dynamic reversal thresholds. Such thresholds depend on the field-step protocol, damping (set to 0.02), and the particular metastable states encountered; footnote 34 explicitly attributes 'noise' in Hs(θ) to metastable states from a single set of simulations. A synchronized reversal threshold does not uniquely demonstrate a bilinear Heisenberg-like interaction. Please compute the interaction energy directly from the micromagnetic configurations (for example, total energy as a function of the relative orientation of the two island moments at fixed applied field for the T-dimer) and fit it to a J m_i·m_j form, or revise the title and abstract to describe an emergent synchronization effect rather than an interaction. The paper's own phrasing in Sec. VII that the effect 'can in some ways be regarded as analogous' is more cautious than the headline claim.","section":"Secs. II, VII"},{"comment":"The decomposition into cubic and uniaxial anisotropy contributions is inferred from angular spacings between Hs maxima rather than from an energy landscape. The 90° spacing between anisotropy axes in the pinwheel vertex (Fig. 6) and the alternating 103.5°/76.5° spacings in the T-shaped array (Fig. 4) are consistent with cubic-plus-uniaxial symmetry breaking, but axis positions alone do not determine the functional form or amplitudes of the anisotropy terms; many combinations of anisotropy invariants can reproduce the same axes. In addition, the plateau angles 43.25° and 48.00° in Fig. 8 are read from single-run astroids at 0.25° angular resolution, under conditions that footnote 34 acknowledges contain metastable-state 'noise.' Please quantify the repeatability and uncertainty of these angles, ideally by repeating simulations with different initializations or by fitting an energy model, and, if possible, extract anisotropy constants from energy-vs-angle calculations.","section":"Secs. V–VI"},{"comment":"The quantitative boundaries of the strong-coupling regime are protocol-dependent observables, but the paper does not test this sensitivity. The quoted regime widths (14.00°, 15.0°, <0.25°, and 1.5°) are obtained at a single damping value (0.02), a single field-step size (25 µT), and without averaging over initial micromagnetic states. Switching fields can shift with damping, field ramp rate, and the presence of metastable states, especially in near-degenerate pinwheel geometries. Please demonstrate robustness by varying damping and field-step size, and by sampling multiple initial states, and state explicitly whether the reported regime widths are intrinsic properties of the geometry and material or are protocol-dependent thresholds.","section":"Secs. III, VI, VIII"}],"minor_comments":[{"comment":"The sentence 'this system aught to be particularly sensitive' contains a typo: 'aught' should be 'ought.'","section":"Sec. I"},{"comment":"The section heading 'TOW ARDS POINT DIPOLES' has a typo and should read 'TOWARDS POINT DIPOLES'; the same section contains the misspelling 'microoromagnetic' instead of 'micromagnetic.'","section":"Sec. VIII"},{"comment":"The main text gives the T-dimer anisotropy axis as 38.25° while the higher-resolution simulation and Fig. 5 caption give 38.21°; please clarify whether these are different runs/resolutions or one number is a typographical error.","section":"Sec. III / Fig. 4"},{"comment":"The data availability statement says 'DOI TBA'; a working data DOI should be provided before publication.","section":"Conclusions"},{"comment":"The sentence 'the symmetry of the astroids is increased for the asymmetric arrays' is confusing because it comes immediately after stating that the reversal process is non-reciprocal; please reword to explain why the switching-field symmetry is higher than the reversal-process symmetry.","section":"Sec. II"},{"comment":"The conclusion states that the exact anisotropy angles are 'relatively sensitive to imperfections,' but this is not quantified in the main text; please refer explicitly to the supporting Supplemental Material results and, if available, to the estimated uncertainty.","section":"Sec. IX"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper has a real result: the field-driven reversal behavior of pinwheel artificial spin ice that point-dipole models could not explain is traced to end-state stray fields. The micromagnetic evidence is layered and mostly convincing: the T-shaped dimer shows synchronized reversal at 38.21°, the periodic-boundary array restores the anisotropy axis to exactly 45°, the strongly coupled regime narrows as islands are shrunk toward point dipoles, and the B-field maps directly show end-state lobes that differ dramatically from uniformly magnetized or point-dipole islands. The comparison to the earlier Lorentz TEM experiments is fair and reproduces the corner-nucleated avalanche reversal. That is a genuine advance for the ASI field.\n\nThe soft spot is real but not fatal: the 'Heisenberg pseudo-exchange' is an analogy drawn from Hs(θ) astroids, not a fitted or derived energy functional. A switching field is a dynamic threshold that depends on the field-step protocol, damping, and metastable states; footnote 34 concedes the curves contain noise from a single set of simulations. So the specific plateau angles (43.25° vs 48.00°) and the angular width of the strong-coupling regime should be read as approximate, not precise. That said, the symmetry restoration under PBC and the monotonic narrowing with island size are strong evidence that the mechanism is genuinely end-state-mediated coupling, whatever one labels it. The missing data DOI is a minor but easily fixable omission.\n\nThe citation pattern is honest: references to their own prior work are appropriate because those earlier results are what this paper explains, and the new mechanism is not fitted to those data. There is no circularity problem.\n\nWho gets value from this: anyone working on artificial spin ice, especially pinwheel geometry, and anyone interested in the limits of point-dipole approximations for extended magnetic islands. It deserves a serious referee. I would accept it with a request for the data release and some tempering of the pseudo-exchange language, but the central claim about end-state-controlled anisotropic reversal holds up.\n\nBring it to reading group; I'd cite it if I worked on ASI reversal.","headline":"A solid micromagnetic study that traces pinwheel ASI reversal to end-state dipolar fields; the 'pseudo-exchange' label is an interpretation rather than a derived energy term, but the core phenomenology is convincingly established.","tokens_in":18785,"tokens_out":1911,"would_cite":true,"duration_ms":22085,"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":"End-states, not dipoles, drive pinwheel spin-ice reversal","keywords":["artificial spin ice","pinwheel","micromagnetism","superferromagnetism","anisotropy","end-states","Heisenberg pseudo-exchange","dipolar coupling"],"falsifier":"If the island ends are sharpened so that end-state curvature is suppressed but net moment is unchanged, the strongly coupled regime and the anisotropy-axis offset should vanish; should they persist, the Heisenberg pseudo-exchange explanation fails.","tokens_in":17789,"feed_emoji":"🧲","tokens_out":7580,"duration_ms":64648,"temperature":0.7,"pith_summary":"This paper proposes that the surprising ferromagnetic reversal in pinwheel artificial spin ice, at island rotation around 45 degrees, is driven not by the net magnetic dipoles of the islands but by the curved magnetisation 'end-states' at the island tips. Using micromagnetic simulations compared with electron-microscopy experiments, it shows that under an applied field these end-states create stray fields that effectively couple nearest-neighbour islands like an exchange interaction. This 'Heisenberg pseudo-exchange' produces a superferromagnet whose reversal nucleates at array corners and spreads in avalanches, and it shifts the array's anisotropy axes away from the geometric axes. The effect is absent in point-dipole models and disappears as island size shrinks, so the paper also shows how to tune the array's behaviour.","feed_headline":"End-states, not dipoles, drive pinwheel spin-ice reversal","feed_subtitle":"Explains superferromagnetic reversal and axis misalignment in rotated square ice via end-state dipolar fields.","key_machinery":"The central object is the magnetisation end-state: the curvature of the magnetisation away from the island long-axis at the tips of each physically extended island. In field-driven reversal, these end-states (s-type, leaf, and c-type) grow and mediate the reversal of each island, and the stray field they emit is strong and spatially lobed, unlike the field of a uniformly magnetised dipole. From this, the paper defines a Heisenberg pseudo-exchange: an effective nearest-neighbour coupling induced by end-state dipolar fields that acts like an exchange interaction in a continuous ferromagnet, but which exists only under an applied field. The switching-field astroid, the curve of reversal field versus applied-field angle, is the probe used to map the anisotropies that this pseudo-exchange produces.","core_discovery":"The central claim is that magnetisation end-states on extended islands induce a Heisenberg pseudo-exchange interaction that governs inter-island coupling and the field-driven reversal of pinwheel artificial spin ice. Under an applied field, the s-type and leaf end-states at island ends produce stray fields that add asymmetrically at each nearest-neighbour site, so that the two collinear subarrays reverse together only when the field is tilted away from the geometric 45-degree axis. This correlated coupling reduces the switching field below the single-island value, creates a strongly coupled angular regime, and lowers the local reversal barrier at array corners, where avalanche reversal begins. The emergent anisotropy is a combination of cubic and uniaxial contributions whose symmetry mirrors the array's edge symmetry but whose axes are misaligned with the geometric ones. The paper argues that previous experimental observations of superferromagnetism and of anisotropy misalignment, unexplained by point-dipole Monte Carlo, are direct consequences of these end-state fields.","pith_inferences":["If end-state stray fields dominate the field-driven coupling, the same mechanism should appear in other low-coupling artificial spin ice geometries, such as kagome arrays at larger island spacings, where the point-dipole model also fails to explain reversal.","Varying island shape (for example, rounding or sharpening the ends) should continuously tune the anisotropy-axis offset, since the end-state strength is set by the tip geometry rather than by the net moment.","At remanence the end-state-induced fields are weaker than under applied field, but they will still bias thermalisation experiments; measuring the correlation between end-state orientation and subsequent thermal flips could test the role of pseudo-exchange in zero-field dynamics.","The avalanche propagation direction being tied to array corners suggests pinwheel arrays could act as field-steerable conduits for magnetic domain walls, a property that might be harnessed in reconfigurable logic or neuromorphic devices if corner nucleation can be controlled lithographically."],"forward_implications":["Reducing the in-plane island size weakens the pseudo-exchange: the width of the strongly coupled angular regime shrinks from 14 degrees to below the 0.25-degree resolution, and the array-edge anisotropy is effectively removed.","Array reversal proceeds by corner-nucleated avalanches through nearest-neighbour islands, forming mesoscopic domain walls that run perpendicular to the applied field, with wall speeds close to individual island magnetisation-component speeds.","The anisotropy-axis plateau angles differ by edge symmetry: 43.25 degrees for asymmetric arrays and 48.00 degrees for symmetric arrays, both offset from the 45-degree geometric axis.","Periodic (edge-free) arrays restore the anisotropy axis to exactly 45 degrees, so the misalignment is a finite-size edge effect associated with corner islands.","The paper's mechanism explains the previously reported superferromagnetic reversal and the failure of point-dipole Monte Carlo to reproduce it."],"supporting_citations":[{"why":"Establishes the pinwheel geometry and its point-dipole-level prediction of near-degenerate ferromagnetic ordering, the baseline this work departs from.","marker":"[10]"},{"why":"The experiment that reported superferromagnetic reversal and anisotropy misalignment in pinwheel arrays, which this paper explains via end-state pseudo-exchange.","marker":"[14]"},{"why":"Supplies the experimental Fresnel-imaging hysteresis data used for direct comparison with the simulations.","marker":"[23]"},{"why":"Demonstrates experimentally that island end-states exist and are visible in the magnetic induction, supporting the premise that extended-island effects matter.","marker":"[12]"},{"why":"Supplies the micromagnetic simulation software used to generate all numerical reversal and astroid results.","marker":"[30]"},{"why":"Provides the astroid formalism used to extract anisotropy axes from the angle-dependence of switching fields.","marker":"[31]"}],"fun_headline_variants":["End-state fields set pinwheel spin-ice reversal path","Pinwheel spin ice: end-states drive Heisenberg pseudo-exchange","Emergent anisotropy misaligned in pinwheel artificial spin ice","Corner avalanches from end-state coupling in pinwheel ice","Island size tunes pinwheel spin-ice coupling regime"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The switching field Hs, measured from a single micromagnetic run per field angle without averaging over metastable states, is assumed to be a faithful proxy for the inter-island interaction energy, even though the paper notes that metastable states add noise to the Hs(θ) curves.","fun_headline_variants_meta":{"raw":{"variants":["End-state fields set pinwheel spin-ice reversal path","Pinwheel spin ice: end-states drive Heisenberg pseudo-exchange","Emergent anisotropy misaligned in pinwheel artificial spin ice","Corner avalanches from end-state coupling in pinwheel ice","Island size tunes pinwheel spin-ice coupling regime"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000311,"raw_usage":{"total_tokens":1800,"prompt_tokens":1000,"completion_tokens":800,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":616,"completion_tokens_details":{"reasoning_tokens":716}},"tokens_in":616,"tokens_out":800,"duration_ms":6930,"temperature":1.0,"reasoning_tokens":716,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:38:46.572034+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If the island ends are sharpened so that end-state curvature is suppressed but net moment is unchanged, the strongly coupled regime and the anisotropy-axis offset should vanish; should they persist, the Heisenberg pseudo-exchange explanation fails.","supporting_citations":[{"cited_title":"Mac\\^ e do , author G","cited_arxiv_id":null,"evidence_quote":"Establishes the pinwheel geometry and its point-dipole-level prediction of near-degenerate ferromagnetic ordering, the baseline this work departs from."},{"cited_title":"pinwheel","cited_arxiv_id":null,"evidence_quote":"The experiment that reported superferromagnetic reversal and anisotropy misalignment in pinwheel arrays, which this paper explains via end-state pseudo-exchange."},{"cited_title":"pinwheel","cited_arxiv_id":null,"evidence_quote":"Supplies the experimental Fresnel-imaging hysteresis data used for direct comparison with the simulations."},{"cited_title":"Phatak , author A","cited_arxiv_id":null,"evidence_quote":"Demonstrates experimentally that island end-states exist and are visible in the magnetic induction, supporting the premise that extended-island effects matter."},{"cited_title":"Thiaville ,\\ title title Coherent rotation of magnetization in three dimensions: A geometrical approach , \\ 10.1103/PhysRevB.61.12221 journal journal Phys","cited_arxiv_id":null,"evidence_quote":"Provides the astroid formalism used to extract anisotropy axes from the angle-dependence of switching fields."}],"review_version":1}