{"id":"dbe35fcb-7bfe-49db-b640-86c904efcfb1","arxiv_id":"2412.04920","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A DFT-based model fitted on 32 known magnets predicts Curie-temperature trends in seven alloy families, including a new prediction for Fe1-xTcx.","lead":"The authors test a previously fitted quantum-mechanics-based model for predicting Curie temperatures on seven magnetic alloy families, including a prediction for a radioactive iron-technetium alloy that is hard to measure. The model captures composition trends in most systems, but requires manual exceptions and fails quantitatively for two of the seven.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"FeCr/FeV agreement may be an artifact of manually relaxing the DLM moment constraint; need a constrained-moment control calculation.","rationale":"The reader's weakest assumption already identifies the 0.75 μB threshold and manual exemptions for Cr/V as the core risk to the central claim. My stress-test narrows this to a concrete mechanistic concern: unconstrained Cr/V moments in the DLM state change the physical meaning of ΔE and Smag for two of the seven systems, so the successful FeCr and FeV predictions may not test the model's transferability at all. This is a genuine soft spot because the paper itself documents the collapse (1.3→0.2 μB) and the convergence failure that forced the exception. The proposed control calculation—a constrained-moment DLM with a convergence-tolerant protocol—would directly determine whether the exemption is load-bearing. If the control calculation reproduces the current TC, the model is more robust than the workflow suggests; if not, the claim of 'generally capable' predictions would need to be qualified to exclude or re-parameterize these itinerant-moment systems. The reader's CONDITIONAL verdict remains appropriate: the paper is honest and valuable, but this unaddressed sensitivity is exactly the kind of condition that should be resolved before full acceptance. No ad hominem is implied; the critique targets the procedure, not the authors.","tokens_in":14158,"tokens_out":2401,"duration_ms":29084,"concrete_test":"For Fe0.875Cr0.125 and Fe0.75V0.25, repeat the DLM calculation with Cr/V moments constrained near their ground-state magnitudes, using a more robust convergence strategy (e.g., start from the converged unconstrained DLM state and ramp the Ma-Dudarev λ in small increments, or use fixed-spin-moment DLM). Compare the resulting TC from Eq. (1) with the values in Table I. If TC shifts by more than ~100 K or the composition trend changes, the manual relaxation is a load-bearing adjustment; if it stays within the model's stated 126 K error, the concern is mitigated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of broad transferability depends on the DLM state faithfully representing the paramagnetic state. For Fe1-xCrx and Fe1-xVx, the workflow is explicitly altered: Cr and V moments with ground-state magnitudes above the 0.75 μB threshold are left unconstrained because constrained DLM runs fail to converge (Sec. V). In FeCr with x=0.125, the Cr moment collapses from ~1.3 μB in the ground state to ~0.2 μB in the DLM state. This changes both ingredients of Eq. (1): ΔE now includes the energy cost of moment collapse rather than purely magnetic disorder, and Smag (Eq. 2) excludes Cr/V because only constrained moments count toward Nmag. The model's good agreement for these two systems could therefore be a fortuitous cancellation rather than evidence that Eq. (1) transfers to diverse chemistries. Since the abstract claims 'less hands-on adjustments compared to other theoretical approaches,' the manual exemption for two of seven tested systems is a load-bearing exception that undermines the clean test of the model's generality.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper applies a previously developed four-parameter model, Eq. (1), to predict the Curie temperature of seven substitutionally disordered alloy systems (Fe-Co, Fe-Cr, Fe-V, Ni-Cu-MnSb, Ni-MnSb, Ti-Cr-N, and Co-Al), comparing the predictions with experimental values and also presenting a prediction for the experimentally unexplored Fe-Tc system. The model, with parameters A=0.85, B=0.69, C=0.14, and D=124 K fitted to 32 known magnets in an earlier work (Ref. 20), uses the DLM-ground-state energy difference, a magnetic-entropy term, and the number of nearest magnetic neighbors. The authors report qualitative agreement with the composition dependence of TC in most systems, some quantitative agreement, and explicitly discuss cases where the model fails (Co-Al) or where the standard workflow had to be altered (unconstrained Cr/V moments in Fe-Cr and Fe-V, modified λ schedule for Co-Al). The central claim is that the method is broadly applicable with less hands-on adjustment than competing approaches.","tokens_in":14348,"tokens_out":4648,"duration_ms":49798,"significance":"If the central claim is supported, the model provides a computationally efficient, out-of-sample screening tool for alloy Curie-temperature trends, complementing more expensive Heisenberg-based methods and potentially guiding experimental alloy design. The work has several strengths: none of the tested alloys are in the 32-material fitting set, so the test is genuinely out-of-sample; the authors are unusually honest in reporting the Co-Al failure and the workflow modifications; the Fe-Tc prediction is a falsifiable experimental target; and the Python scripts for data extraction are openly available. The main reservation is that the evidence for 'broad applicability requiring less hands-on adjustments' is weakened by the very exceptions the paper acknowledges, and at least one of the claimed qualitative successes (Fe-V) does not actually reproduce the experimental composition trend.","major_comments":[{"comment":"The central claim that the model reproduces 'the qualitative effect on TC of altering the alloy composition' is contradicted by the Fe1-xVx results. From x=0.125 to x=0.25, the experimental TC decreases (1111 K to 1053 K), while the predicted TC increases (1053 K to 1091 K). The text in Sec. V states that 'our method correctly predicts this observation' and attributes the discrepancy to the changes being 'minimal,' but the predicted slope has the opposite sign to the experimental slope. This is a qualitative failure, not a quantitative offset. The discussion should explicitly acknowledge this and explain why the model nevertheless is considered to capture the trend.","section":"Sec. V and Table I"},{"comment":"The unconstrained Cr and V moments in Fe1-xCrx and Fe1-xVx break the prescribed workflow of Sec. II C, and the good agreement for these systems is therefore not a clean out-of-sample test of Eq. (1). Because Cr and V are not constrained in the DLM runs, they are excluded from the magnetic entropy in Eq. (2), while their moment collapse (e.g., Cr from 1.3 to 0.2 μB at x=0.125) contributes to the energy difference ΔE in a way that a constrained run would not. A control calculation with constrained Cr/V moments, or a sensitivity analysis showing how TC depends on the choice of constraint, is needed to establish that the agreement is not fortuitous.","section":"Sec. IV A and Sec. V"},{"comment":"The abstract's claim that the method requires 'less hands-on adjustments compared to other theoretical approaches' is overstated given the manual interventions described in the paper: the 0.75 μB moment threshold is relaxed for Cr and V in Fe-Cr and Fe-V, and the λ schedule is modified for Co-Al. These are exactly the type of system-specific adjustments the method aims to avoid. Please either soften the claim to accurately reflect the observed level of intervention, or provide a quantitative measure of the manual tuning needed per system.","section":"Abstract and Sec. V"}],"minor_comments":[{"comment":"The word 'cobolt' in the caption of Fig. 3 should be 'cobalt'.","section":"Fig. 3 caption"},{"comment":"The description of the parameter fit states that 'A linear fit to the experimental TC subsequently gives the parameters A and D,' but it does not specify which experimental data are used; a cross-reference to Ref. 20 would clarify the procedure.","section":"Sec. II A"},{"comment":"The legend in Fig. 1 lists 'This work' three times, which is redundant and could be simplified to a single entry or a note explaining that all filled symbols are from this work.","section":"Sec. IV A"},{"comment":"The details of the SQS supercells (size, number of configurations, and generated k-points) are not fully specified, which limits reproducibility; providing the SQS parameters or a reference to the generation code would be helpful.","section":"Sec. III"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about its limitations, which is commendable, but the central claim of broad transferability with minimal manual adjustment is not fully supported by the evidence. The Fe-V trend mismatch and the unconstrained Cr/V moments are load-bearing issues. If the authors can provide a control calculation or a clear sensitivity analysis for the constraint choice, and revise the overstatement in the abstract, the paper could become a valuable contribution. I do not see any circularity problem, since the parameters are fixed from prior work and the tested alloys are out-of-sample."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this if you care about screening Curie temperatures in alloys. The paper is a validation and extension of the authors' own fixed four-parameter model (Eq. 1, fit to 32 materials in Ref. 20) on seven new alloy systems, none in the fitting set. That makes it a genuine out-of-sample test, and it mostly passes: qualitative composition trends are reproduced for FeCo, FeV, FeCr, NiMnSb, NiCuMnSb, and TiCrN, with errors within the model's previously quoted 126 K MAE in the Fe binaries. The new Fe1-xTcx prediction is a concrete, testable target with a plausible rationale (similarity to FeMo and FeRu). The paper also ships Python scripts for data extraction, which is a real plus.\n\nWhat it does well: it reports its failures honestly. The CoAl case is explicitly diagnosed as an itinerant-magnetism breakdown, and the TiCrN overestimation is correctly tied to the model's 124 K floor. The authors do not oversell the quantitative agreement.\n\nThe soft spot the stress-test flags is real but not fatal. For FeCr and FeV, constrained DLM runs would not converge, so the Cr and V moments were left unconstrained. In FeCr with 12.5% Cr, the Cr moment drops from roughly 1.3 to 0.2 μB in the DLM state, which changes both ΔE and Smag. Those two systems are therefore not clean tests of the model as written; the good agreement there could include a fortuitous cancellation. The abstract's claim of 'less hands-on adjustments' is undercut by this manual exception plus the adjusted λ schedule for CoAl. The authors disclose it, but they do not quantify the effect. A control calculation with a constrained-moment DLM, even at one composition, would materially strengthen the transferability claim.\n\nThe bigger weakness is the absence of error bars or sampling analysis: one SQS cell and one DLM configuration per composition. With that setup, the difference between a 60 K and a 170 K error is not robustly interpretable. That said, the paper frames itself as a trend-prediction tool, and the trends look credible.\n\nMy take: this deserves peer review. It is a legitimate, honest extension with a new experimental prediction. A referee should push for a discussion (or better, a calculation) addressing the Cr/V moment collapse and ask the authors to soften the 'less hands-on' claim to match what they actually did. The central physics claim—that a simple ΔE/Smag model captures composition trends across diverse chemistries—is plausible and mostly supported.","headline":"Honest out-of-sample test of a fixed four-parameter TC model; trends hold in most systems, but the Cr/V manual exceptions and lack of error bars keep the 'broad applicability' claim from being fully proven.","tokens_in":14873,"tokens_out":2234,"would_cite":true,"duration_ms":23754,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82D40"],"pacs":["75.30.Kz","75.50.Bb","71.15.Mb"],"model":"deepseek-v4-flash","headline":"A four-parameter model fit to 32 known magnets reproduces Curie-temperature trends in seven untested alloy families and predicts a decrease in Tc for iron-technetium.","keywords":["Curie temperature","substitutional disorder","disordered local moments","density functional theory","magnetic entropy","special quasirandom structures","Heusler alloys","iron-technetium"],"falsifier":"Measure the Curie temperature of a bcc Fe$_{0.9}$Tc$_{0.1}$ sample; a value above rather than below the pure-Fe prediction would contradict the claimed monotonic decrease, and a cheaper check is to recompute one Fe-V or Fe-Cr composition with several independent SQS cells and DLM configurations to see whether the scatter exceeds the trend the model attributes to composition.","tokens_in":13934,"feed_emoji":"🧲","tokens_out":10470,"duration_ms":100856,"temperature":0.7,"pith_summary":"This paper argues that a four-parameter physics model, fit once to 32 known magnets in Ref. 20, transfers to substitutionally disordered alloys it never trained on. The model uses the density-functional energy difference between the disordered-local-moments (DLM) paramagnetic state and the magnetic ground state, the magnetic entropy of the DLM state, and the number of nearest magnetic neighbors. Applied to Fe$_{1-x}$Co$_x$, Fe$_{1-x}$Cr$_x$, Fe$_{1-x}$V$_x$, Ni-based Heusler alloys, Ti$_{1-x}$Cr$_x$N, and Co$_{1-x}$Al$_x$, it reproduces the qualitative composition trends of the measured Curie temperature and often lands within the model's stated mean absolute error of about 126 K. The same fixed-parameter model predicts a monotonic decrease of $T_\\text{C}$ with solute content in the experimentally unexplored bcc alloy Fe$_{1-x}$Tc$_x$, matching the pattern seen in Fe-Mo and Fe-Re analogues.","feed_headline":"Four fitted numbers transfer Curie-temperature trends to new alloys","feed_subtitle":"A fixed physics model reproduces composition trends across seven alloys and flags Fe-Tc for experiment.","key_machinery":"The load-bearing machinery is the DLM supercell representation of the paramagnetic state together with Eq. (1). In one DLM run, atomic spin directions are fixed in near-random orientations with short-range order close to zero using the constraint method of Ref. 29, while magnitudes relax; the magnetic ground state comes from a non-collinear ground-state search (Ref. 23), and substitutional disorder is encoded in special quasirandom structures (Ref. 28). From these calculations the paper extracts $\\Delta E = E_\\text{DLM} - E_\\text{GS}$, the magnetic entropy $S_\\text{mag} = k_B \\sum_i \\ln(m_i+1)/N_\\text{mag}$ over constrained moment magnitudes $m_i$ in Bohr magnetons, and $N_N$, the number of nearest magnetic neighbors defined by a 0.75 $\\mu_B$ ground-state moment threshold. The empirical factor $(1 - B/N_N^C)$ is the model's correction for short-range-order effects that matter when the magnetic energy is distributed over few neighbors, and $D$ sets a 124 K floor.","core_discovery":"The central claim is that Eq. (1), with the fixed constants $A=0.85$, $B=0.69$, $C=0.14$, and $D=124$ K, is generally capable of reproducing the qualitative dependence of the Curie temperature on alloy composition across diverse chemistries and crystal structures, and in several systems the absolute values are close to experiment. None of the seven alloys studied here belongs to the 32 materials used to fit those constants, so the agreement is presented as a transferability test. The formula $T_\\text{C} = A(\\Delta E/S_\\text{mag})(1 - B/N_N^C) + D$ K ties the ordering temperature to the energy penalty for magnetic disorder, divided by the magnetic entropy, corrected by the number of nearest magnetic neighbors, and offset by a constant floor. The paper's own agreement ranges from roughly 20-80 K for the Fe binaries to within about 170 K for the Heusler alloys, while fcc Co$_{1-x}$Al$_x$ clearly fails and Ti$_{1-x}$Cr$_x$N is overestimated because its low $T_\\text{C}$ approaches the model's 124 K lower limit.","pith_inferences":["A natural next stress test is an ordered intermetallic or a strongly localized-moment oxide, where the single-DLM-configuration approximation is less demanding; if Eq. (1) holds there, the entropy and neighbor-count terms are doing genuinely physical work rather than absorbing the fit.","The predicted Fe$_{1-x}$Tc$_x$ curve could be sharpened by repeating one composition with several independent SQS cells and DLM configurations; if the scatter exceeds the roughly 50 K drop the model predicts, the trend is not yet a robust target.","The manual exemptions for Cr and V moments suggest a principled extension: an itinerancy-aware constraint threshold, or a separately parameterized DLM moment magnitude, could remove the convergence-related exceptions and let the model speak for systems like Fe-V and Co-Al."],"forward_implications":["For the six families with experimental data, the model returns the correct sign of $\\partial T_\\text{C}/\\partial x$ in every case, so it can rank compositions by magnetic ordering temperature before synthesis.","The fixed parameters mean each new alloy family is an independent test; the observed agreement within about 170 K across Heusler alloys and Fe binaries indicates the model is not being re-fit per system.","The failure on fcc Co$_{1-x}$Al$_x$ and the too-high values for Ti$_{1-x}$Cr$_x$N delimit the method's domain: robust local moments and $T_\\text{C}$ well above the 124 K floor.","For Fe$_{1-x}$Tc$_x$, the predicted drop in $T_\\text{C}$ with technetium content is a concrete, testable experimental target, despite the radioactivity of Tc."],"supporting_citations":[{"why":"Supplies the four-parameter model of Eq. (1), the fitted constants A=0.85, B=0.69, C=0.14, D=124 K, and the 32-material fitting set that the seven alloys here are designed to test.","marker":"[20]"},{"why":"Defines the supercell disordered-local-moments implementation used to simulate the paramagnetic state and obtain E_DLM and S_mag.","marker":"[21]"},{"why":"Supplies the ground-state search method: non-collinear relaxation from random initial moment directions, used to identify the magnetic ground state and the 0.75 μB threshold moments.","marker":"[23]"},{"why":"Provides special quasirandom structures that model the substitutional chemical disorder of each alloy at the supercell sizes used here.","marker":"[28]"},{"why":"Gives the Ma-Dudarev constrained-moment method that pins spin directions while allowing magnitudes to relax in the DLM runs.","marker":"[29]"},{"why":"Supplies the compiled experimental Fe-Co Curie-temperature data that anchor the comparison for the bcc and fcc Fe1-xCox trends.","marker":"[38]"},{"why":"Supplies the collected experimental Fe-V Curie-temperature data used to judge the predicted Fe1-xVx trend.","marker":"[41]"},{"why":"Supplies the experimental ferromagnetism and Curie-temperature data for Ti1-xCrxN thin films against which the model's overestimation is identified.","marker":"[53]"}],"fun_headline_variants":["Fixed constants forecast Curie temperatures across seven alloys","One formula, seven alloys: Curie temperature trends hold","Transferable Curie-temperature formula tested on seven alloys","Four fitted numbers, no refitting: Curie trends transfer"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that one DLM supercell with near-zero spin short-range order, a fixed 0.75 $\\mu_B$ threshold for constraining moments, and the four fitted constants faithfully represent the paramagnetic state of every tested chemistry; the paper's own hand-adjusted exemptions for chromium and vanadium moments show where that premise is strained.","fun_headline_variants_meta":{"raw":{"variants":["Fixed constants forecast Curie temperatures across seven alloys","One formula, seven alloys: Curie temperature trends hold","Transferable Curie-temperature formula tested on seven alloys","Four fitted numbers, no refitting: Curie trends transfer"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000785,"raw_usage":{"total_tokens":3514,"prompt_tokens":1044,"completion_tokens":2470,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":660,"completion_tokens_details":{"reasoning_tokens":2405}},"tokens_in":660,"tokens_out":2470,"duration_ms":20160,"temperature":1.0,"reasoning_tokens":2405,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:08:02.077591+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the Curie temperature of a bcc Fe$_{0.9}$Tc$_{0.1}$ sample; a value above rather than below the pure-Fe prediction would contradict the claimed monotonic decrease, and a cheaper check is to recompute one Fe-V or Fe-Cr composition with several independent SQS cells and DLM configurations to see whether the scatter exceeds the trend the model attributes to composition.","supporting_citations":[{"cited_title":"Arale Br\\\" a nnvall, G","cited_arxiv_id":null,"evidence_quote":"Supplies the four-parameter model of Eq. (1), the fitted constants A=0.85, B=0.69, C=0.14, D=124 K, and the 32-material fitting set that the seven alloys here are designed to test."},{"cited_title":"Alling, T","cited_arxiv_id":null,"evidence_quote":"Defines the supercell disordered-local-moments implementation used to simulate the paramagnetic state and obtain E_DLM and S_mag."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the ground-state search method: non-collinear relaxation from random initial moment directions, used to identify the magnetic ground state and the 0.75 μB threshold moments."},{"cited_title":"Nishizawa, K","cited_arxiv_id":null,"evidence_quote":"Supplies the compiled experimental Fe-Co Curie-temperature data that anchor the comparison for the bcc and fcc Fe1-xCox trends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the collected experimental Fe-V Curie-temperature data used to judge the predicted Fe1-xVx trend."},{"cited_title":"Inumaru, K","cited_arxiv_id":null,"evidence_quote":"Supplies the experimental ferromagnetism and Curie-temperature data for Ti1-xCrxN thin films against which the model's overestimation is identified."}],"review_version":1}