{"id":"7cb78e04-5774-4189-96c6-3df2ca9711b3","arxiv_id":"2509.01522","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A Ruby-lattice artificial spin ice realizes Blume-Capel degrees of freedom via toroidal moments on triangular and hexagonal plaquettes, with a two-step ordering into a ferrotoroidic ground state.","lead":"This paper builds artificial spin ices on the Ruby lattice, where loops of nanomagnets form toroidal moments that can point clockwise, anticlockwise, or not form at all. The authors show that these three states realize the Blume-Capel model, a three-state spin model, and observe its crossover and phase transition in experiments and simulations.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The BC mapping is validated mainly against a coarse order parameter, and the modified BC model's high-temperature probabilities are fit; the effective two-parameter Hamiltonian itself is never directly tested.","rationale":"The reader's weakest assumption correctly identifies that the load-bearing premise is the coarse-grained three-state description with a two-parameter Blume-Capel Hamiltonian. My stress-test sharpens this: the validation loop is too weak to establish that premise. The comparison uses the order parameter Ψ, which is largely determined by single-site populations, and the modified BC model explicitly tunes those populations. The dipolar Monte Carlo results and experimental observations of toroidic phases are valuable and likely robust, but they support the existence of the phases rather than the specific claim that the interactions are faithfully represented by a temperature-independent J and Δ. A direct statistical test of the coarse-grained Hamiltonian would settle the question. This does not change the reader's conditional accept: the central claim remains plausible and experimentally motivated, but the quantitative BC mapping should be verified before the realization is asserted as established.","tokens_in":19000,"tokens_out":12308,"duration_ms":159282,"concrete_test":"Run the dipolar Monte Carlo for one a/b branch at several temperatures. From configurations, compute nearest-neighbor pair frequencies of t values and fit log-ratios to -βJ t_i t_j - (βΔ/2)(t_i²+t_j²). If the best-fit J or Δ varies by more than roughly 10% with temperature, or if pair correlations depend on a third plaquette, the two-parameter BC Hamiltonian is not quantitatively valid. Also rerun the modified BC simulation with an explicit Metropolis-Hastings proposal-ratio correction and check whether the high-temperature state probabilities remain 62/64 and 1/64; if not, the reported agreement is an artifact of the update rule.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that toroidal-moment plaquettes realize the two-parameter Blume-Capel Hamiltonian. The evidence in Figure 5b and Supplementary Figures S7/S8 is a comparison of color maps of the order parameter Ψ from full dipolar Monte Carlo and from Blume-Capel simulations, with J and Δ extracted from zero-temperature local energy differences. This is not a strong test of the mapping. Ψ depends only on single-site populations and the ±1 imbalance; it does not probe interaction structure, such as whether the probability of a t=0 plaquette depends on the t values of its neighbors through the -JΣtt' term. Moreover, the 'modified Blume-Capel model' used for the best agreement sets the high-temperature state probabilities to 62/64 and 1/64 from the ASI, so the high-temperature plateau Ψ≈2/64 is matched by construction rather than derived from the Hamiltonian. The Methods describe a non-symmetric proposal distribution (from 0: 62/64,1/64,1/64; from ±1: 1/3 each) without explicitly giving the Metropolis-Hastings proposal-ratio correction, so it is unclear what equilibrium measure this model actually samples. Thus the reported agreement does not establish that the coarse-grained configurations of the Ruby ASI are Boltzmann-distributed according to a single temperature-independent J and Δ. The mapping may be valid, but the paper's central evidence is insufficient as it stands.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the fabrication and characterization of a Ruby artificial spin ice (ASI) in which the lattice parameters a and b are independently tuned. It shows experimentally (MFM and PEEM) and via Monte Carlo simulations that the system can reach a ferrotoroidic ground state either directly or through an intermediate paratoroidic regime, depending on the ratio of interaction strengths within triangular and hexagonal plaquettes. The central claim is that fully-formed toroidal moments on closely-packed plaquettes realize the three states of the Blume-Capel (BC) spin variable t∈{-1,0,+1}, with effective parameters J and Δ extracted from micromagnetic energy differences, and that the observed two-step ordering and phase diagram are captured by the BC model. The paper reports a comparison of the order parameter Ψ between dipolar simulations of the Ruby ASI and BC simulations, including a 'modified' BC model with adjusted high-temperature state probabilities.","tokens_in":19333,"tokens_out":3237,"duration_ms":41028,"significance":"If the proposed mapping is valid, this would be a notable advance: a real-space, directly imageable realization of Blume-Capel degrees of freedom and a tunable toroidal-moment system with a Blume-Capel phase diagram. The experimental dataset is substantial, combining MFM and PEEM imaging over a wide range of lattice parameters, and the finite-size scaling analysis in Supplementary Figure 3 is a useful check of the critical behavior. The micromagnetic extraction of J and Δ from Mumax3 is clearly described and is a non-trivial step toward quantitative comparison. However, the validation of the central mapping is the weakest link, and the paper’s own modified BC construction partially fits the high-temperature behavior rather than predicting it.","major_comments":[{"comment":"The validation of the BC mapping rests on comparison of Ψ, which depends only on the single-site populations n_{+1}, n_{-1} and their imbalance. It does not test the interaction term -JΣt_i t_j or the Boltzmann structure of the coarse-grained variables. Agreement between the upper and lower panels of Fig. 5b therefore does not establish that neighboring toroidal moments interact through the two-parameter BC Hamiltonian. A direct test is needed, e.g. conditional probabilities P(t_i | t_j) or spin-spin correlators from the full dipolar Monte Carlo, compared with BC simulations using the extracted J and Δ without adjusting state probabilities.","section":"§'Blume-Capel degrees of freedom', Fig. 5b, Eq. (3)"},{"comment":"The modified BC model sets the high-temperature state probabilities to 62/64, 1/64 and 1/64 for hexagonal plaquettes (and 6/8, 1/8, 1/8 for triangular plaquettes). Therefore the high-temperature plateau Ψ≈2/64 or Ψ≈2/8 is reproduced by construction rather than derived from the Hamiltonian. This makes the agreement in the crossover region partly circular. The authors should either use the unmodified BC model and quantify the mismatch, or subtract the fitted baseline and show that deviations from the fit are captured by the model. As written, the best-agreement claim is weakened by this fitting.","section":"Methods, 'Monte Carlo simulations' (modified Blume-Capel model)"},{"comment":"The modified BC simulation uses a non-symmetric proposal distribution: from state 0 the proposed states are drawn with probabilities 62/64, 1/64, 1/64, while from ±1 they are drawn with 1/3 each. No Metropolis-Hastings proposal-ratio correction is given. If the acceptance probability is the standard min(1, exp(-ΔE/T)), this non-symmetric proposal changes the equilibrium measure unless compensated. The authors need to specify the detailed-balance condition or state explicitly that the proposal asymmetry is corrected. Without this, the reported BC results may not correspond to the intended Hamiltonian.","section":"Methods, 'Monte Carlo simulations' (proposal distribution)"},{"comment":"The effective temperature is obtained by matching the experimental vertex and flux-closed plaquette populations to the Monte Carlo simulation of the same model that is being validated. This fitting can mask systematic errors in the Hamiltonian or in the mapping. To support the claim that the experiments traverse a constant-T_eff line, an independent estimate of T_eff (e.g., from blocking temperature physics) or a consistency check across many lattice-parameter ratios should be provided. As it stands, the quantitative comparison along the dashed line in Fig. 3a is not fully predictive.","section":"Methods, 'Effective temperature'"}],"minor_comments":[{"comment":"Typo: 'having a positive fully-formed toroidal moment (state +1) is 1/64 and of having a positive fully-formed toroidal moment (state -1)' should read 'negative' for the -1 state.","section":"Methods, 'Monte Carlo simulations'"},{"comment":"Reference 28 is cited in the text ('effective thermal annealing ... 28,29') but is missing from the reference list; the numbering jumps from 27 to 29. Please add the missing reference or renumber.","section":"References"},{"comment":"Minor typos: 'and and the cell sizes' should be 'and the cell sizes'; also 'the probabilities of the proposed state are 1/3 for each of 0, +1 and -1' is clear but could be rephrased for readability.","section":"Methods, 'Micromagnetic simulations'"},{"comment":"The color maps of Ψ are useful but the color scale and contour levels are not defined in the caption. Adding a color bar and explicit contour lines for Ψ=2/3, 1, 2 would make the comparison quantitative.","section":"Fig. 5b and Supplementary Figs. S7/S8"}],"recommendation":"major_revision","confidential_remarks":"The experimental observations and the tuning of the two-step ordering are convincing and likely of interest to the artificial-spin-ice community. The gap is in the validation of the Blume-Capel mapping: the order parameter used is too coarse, and the modified model partially fits the high-temperature behavior. These issues are fixable with additional analysis (direct correlation tests, corrected proposal distribution, and an independent T_eff check), so I recommend major revision rather than rejection. The missing reference 28 and the proposal-distribution ambiguity should also be addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The Ruby ASI work is worth engaging with—the lattice design and the two-step ordering are real, and the Blume-Capel mapping is a good idea. But the paper's central evidence for the mapping is weaker than the abstract suggests, and a few small gaps should be fixed.\n\nWhat's new: the Ruby lattice ASI with two independently tunable lattice parameters controls whether hexagon or triangle toroidal moments form first. That gives a one-step or two-step ordering, with a crossover followed by a second-order transition into a ferrotoroidic ground state. The experiments—MFM and PEEM—are careful, and the Monte Carlo of the full dipole model reproduces the observed populations and the structure factors. That part is solid.\n\nThe mapping to Blume-Capel is plausible and the idea of using toroidal-moment plaquettes as three-state variables is a nice step beyond Ising and Potts ASIs. The J and Δ are extracted from micromagnetic energy differences, which is reasonable. But the validation in Fig. 5b relies on the order parameter Ψ, which only measures single-site populations and the ±1 imbalance. It doesn't test whether the interaction term -JΣtt' actually operates between neighboring plaquettes. The modified BC model sets the high-temperature state probabilities to the measured 62/64 and 1/64, so the high-temperature Ψ plateau is matched by construction. The Methods also describe a non-symmetric proposal distribution without spelling out the Metropolis-Hastings correction, so it's not fully clear what equilibrium measure that model samples. These are addressable, but they mean the mapping is not yet nailed down. I'd want to see a direct test, e.g., measuring the conditional probability of a plaquette's state given its neighbors and comparing to the BC Boltzmann weight.\n\nMinor: reference 28 appears in the text but is missing from the bibliography, and there's no data/code release.\n\nBottom line: this is a well-executed experimental study with an attractive conceptual claim. The claim is plausible but under-tested. Worth a serious referee, with requests for the interaction-structure test and a cleaned-up methods section.","headline":"A genuinely new ASI geometry that maps toroidal plaquettes to Blume-Capel states, with solid experiments but a too-weak test of the mapping itself.","tokens_in":19848,"tokens_out":1635,"would_cite":true,"duration_ms":19605,"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":"Magnetic loops in a Ruby lattice become Blume-Capel spins","keywords":["artificial spin ice","Blume-Capel model","toroidal moment","ferrotoroidic order","Ruby lattice","phase transitions","nanomagnet array","three-state spin model"],"falsifier":"Measure the heat capacity or order-parameter distribution of a large Ruby ASI through the paratoroidic-to-ferrotoroidic transition: a discontinuous jump in the toroidal order parameter, or a heat-capacity peak that grows faster than log L with system size, would falsify the second-order Blume-Capel branch.","tokens_in":18895,"feed_emoji":"🧲","tokens_out":9150,"duration_ms":97920,"temperature":0.7,"pith_summary":"The paper claims that the three possible magnetic states of a closely packed plaquette in a Ruby artificial spin ice—a clockwise head-to-tail loop, an anticlockwise loop, or no closed loop—are exactly the three states of the Blume-Capel spin variable t = +1, −1, 0. In this mapping, toroidal moments of hexagons or triangles become the spins, with the two lattice parameters a and b setting the effective interaction strength J and the anisotropy Δ. Monte Carlo simulations of the full dipolar model, of the Blume-Capel model, and real-space magnetic imaging together show the same sequence: a high-temperature paramagnetic regime, a paratoroidic crossover where plaquette loops form without long-range order, and a second-order transition into a ferrotoroidic ground state. If correct, this is the first real-space realization of Blume-Capel degrees of freedom and suggests that artificial spin ice superstructures can implement other multi-state spin Hamiltonians by design.","feed_headline":"Magnetic loops in a Ruby lattice become Blume-Capel spins","feed_subtitle":"Toroidal-moment plaquettes reproduce the three-state model's crossover and second-order transition in experiment and simulation.","key_machinery":"The load-bearing object is the plaquette toroidal moment: for each hexagon or triangle, the sense of circulation of its macrospins (clockwise, anticlockwise, or absent). These toroidal moments sit on a triangular lattice (hexagonal plaquettes) or a hexagonal lattice (triangular plaquettes) and act as the Blume-Capel variables. The mapping carries the argument by replacing the many-spin dipolar Hamiltonian with the two-parameter Blume-Capel Hamiltonian; J is set by a two-configuration energy difference (ground state versus one reversed toroidal moment), and Δ by the single-plaquette energy gap. The model then predicts the two-step ordering sequence, including the paratoroidic crossover and th","core_discovery":"For a plaquette of nanomagnets, the macroscopically distinct configurations are: all moments head-to-tail clockwise, all head-to-tail anticlockwise, or not all head-to-tail. These map onto t = +1, t = −1, and t = 0 of the Blume-Capel Hamiltonian H = −J Σ t_i t_j − Δ Σ t_i². J is fixed by the energy difference between two adjacent plaquettes with same-sign toroidal moments and one with a reversed moment; Δ is fixed by the gap between the average excited and ground plaquette energies. With J(a,b) and Δ(a,b) extracted from micromagnetic calculations, the Blume-Capel model reproduces the crossover and transition temperatures of the full dipolar Ruby ASI, and the experiments access a previously u","pith_inferences":["A direct test would be to measure the heat capacity of large Ruby ASI arrays under controlled thermal cycling: if the paratoroidic-to-ferrotoroidic peak sharpens with system size as log L, the second-order classification is confirmed; if it develops a jump, the Blume-Capel mapping needs a first-order branch or additional couplings.","The coarse-graining to J and Δ implicitly assumes intra-plaquette correlations can be integrated out once and for all; a natural extension is to compute temperature-dependent renormalized J(Δ) from the full dipolar model and compare with the fixed-parameter Blume-Capel predictions.","The manuscript cites a reference 28 for the deposition-annealing effective-temperature argument, but reference 28 does not appear in the bibliography; the effective-temperature comparison should be checked once the source is restored."],"forward_implications":["A Ruby ASI with tunable a/b is a real-space laboratory for Blume-Capel physics: imaging the signs of plaquette toroidal moments directly measures the Blume-Capel order parameter Ψ.","For large Δ/J the ferrotoroidic transition temperature is exactly known from the Ising model on triangular and hexagonal lattices (4J/ln 3 and 2J/ln(2+√3)), giving a quantitative check of the mapping.","The same superstructure principle—using the collective state of a group of nanomagnets as a single spin variable—can be applied to other multi-state Hamiltonians.","The positive-Δ region explored here excludes the tricritical point; reaching it would require superstructures with antiferromagnetic toroidal interactions."],"supporting_citations":[{"why":"Original 1966 model defining the three-state Hamiltonian and its first-order transitions, the target the paper aims to realize.","marker":"[9]"},{"why":"Companion 1966 model for an Ising system with triplet ions; together with [9] it defines the Blume-Capel degrees of freedom.","marker":"[10]"},{"why":"Exactly solved Ising critical temperatures used to predict the paratoroidic-to-ferrotoroidic transition for large Δ/J.","marker":"[31]"},{"why":"Micromagnetic solver used to compute pairwise interaction strengths from which effective J and Δ are extracted.","marker":"[42]"},{"why":"Supports the claim that as-grown nanomagnet arrays have undergone an effective thermal anneal, fixing a single effective temperature; in the bibliography only [29] is present.","marker":"[28,29]"},{"why":"Prior observation of toroidic phase transitions in an artificial spin ice, establishing toroidal order as a collective degree of freedom that the Ruby ASI extends.","marker":"[8]"}],"fun_headline_variants":["Ruby lattice turns magnets into 3-state Blume-Capel spins","Toroidal moments in artificial spin ice mimic Blume-Capel model","Three-state magnetic order from toroidal plaquettes in Ruby lattice","Blume-Capel realized in Ruby artificial spin ice via toroidal moments","Toroidal moments reproduce Blume-Capel phases in nanomagnet lattice"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that each plaquette's micromagnetic configuration is reducible to one three-state variable with two fixed parameters J and Δ taken from local energy differences, and that as-grown arrays all sit at one effective annealing temperature; if multiplet correlations renormalize J and Δ with temperature, the mapping fails.","fun_headline_variants_meta":{"raw":{"variants":["Ruby lattice turns magnets into 3-state Blume-Capel spins","Toroidal moments in artificial spin ice mimic Blume-Capel model","Three-state magnetic order from toroidal plaquettes in Ruby lattice","Blume-Capel realized in Ruby artificial spin ice via toroidal moments","Toroidal moments reproduce Blume-Capel phases in nanomagnet lattice"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000929,"raw_usage":{"total_tokens":3876,"prompt_tokens":867,"completion_tokens":3009,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":611,"completion_tokens_details":{"reasoning_tokens":2912}},"tokens_in":611,"tokens_out":3009,"duration_ms":21370,"temperature":1.0,"reasoning_tokens":2912,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:27:33.904153+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the heat capacity or order-parameter distribution of a large Ruby ASI through the paratoroidic-to-ferrotoroidic transition: a discontinuous jump in the toroidal order parameter, or a heat-capacity peak that grows faster than log L with system size, would falsify the second-order Blume-Capel branch.","supporting_citations":[],"review_version":1}