{"id":"85da02f1-c5b1-40e4-8cad-67bbb10a66e7","arxiv_id":"2607.19930","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"In a reduced non-self-consistent Monte Carlo model of a planar magnetron, numerically integrating the finite permanent-magnet field yields a more sharply localised erosion racetrack, with FWHM close to a geometric estimate, while the dipole approximation broadens the profile.","lead":"An inexpensive Monte Carlo simulation of a planar magnetron compares two ways of representing its permanent magnets and reports that resolving the real magnet shapes predicts a sharper, more realistic erosion racetrack than a point-dipole approximation. The model is honestly semi-quantitative — its electron collision physics runs about 50% too fast — but positions magnetic-field representation as the dominant modelling choice for racetrack studies.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central racetrack claim (finite-magnet FWHM ≈ Eq. 5, dipole broader) is unquantified: no FWHM values, error bars, or particle-number convergence are reported, and §4 explicitly defers convergence studies; 2×10^4 electrons may be below the statistically converged ensemble.","rationale":"The reader's weakest assumption—statistical convergence of the 2×10^4-electron ensembles—is exactly the load-bearing concern. The paper's own §4 lists particle-number convergence studies as a future requirement, confirming that this check was not performed. Without convergence data or error bars, the central dipole/finite-magnet FWHM comparison could be Monte Carlo noise, and the paper provides no quantitative FWHM values to test Eq. (5). This does not require rejecting the paper, because the authors consistently describe their results as semi-quantitative and the relative comparison is plausible; it does require keeping the verdict at CONDITIONAL pending a convergence check. No other concern—the factor-1.5 drift overestimate, the inferred magnetisation, or the prescribed potential—is as directly tied to the headline claim as the statistical sufficiency of the ensemble. The recommended verdict is therefore UNCHANGED: the reader's conditional assessment correctly reflects the evidence, and a specific numerical test would settle whether the concern actually lands.","tokens_in":8698,"tokens_out":3688,"duration_ms":40141,"concrete_test":"Rerun the §3.7 racetrack calculation for both field representations with Ne = 2×10^4, 1×10^5, and 5×10^5, using at least five independently seeded ensembles per case; compute the radial erosion FWHM (in mm) for each field and the finite-magnet minus dipole FWHM difference with 95% confidence intervals (or bootstrap). Accept the claim only if the difference remains outside the CI and stabilises with Ne, and if the finite-magnet FWHM remains within, say, 10% of Eq. (5). Also report the actual FWHM values that Fig. 9 omits. If the code cannot produce these runs, the claim should be treated as unverified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline result is the difference in erosion-profile localisation between the finite-magnet field and the dipole approximation (§3.7). For this difference to be meaningful, the radial erosion profiles must be statistically converged and the FWHM difference must exceed Monte Carlo noise. The paper states only that the finite-magnet FWHM is 'close to' Eq. (5) and the dipole profile is 'broader'; no FWHM values, confidence intervals, or convergence study are reported. The seeds are 'at least 2×10^4' (§3.7), but this is a floor, not a convergence criterion. The conclusions explicitly list 'particle-number convergence studies' as a requirement for quantitative extension (§4), an admission that the sufficiency of 2×10^4 was not established. With no error bars, the observed broadening in the dipole case could be sampling noise, especially since the erosion statistic is an energy-weighted sum of rare high-energy impacts; variance in such heavy-tailed statistics can be large. A secondary but related gap is that the geometric benchmark ωRT (Eq. 5) is introduced without derivation and is compared only qualitatively. If the finite-magnet FWHM is not in fact close to Eq. (5), or if the dipole/finite-magnet difference collapses at higher Ne, the central claim weakens. The collision module's factor-1.5 drift overestimate is disclosed and affects both fields similarly, so it is not the bottleneck; the bottleneck is statistical evidence for the relative localisation claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"F. F. Locker and G. Strauß present a reduced-order, non-self-consistent Monte Carlo model of a circular planar magnetron discharge in argon. The model combines two magnetic-field representations (point-dipole superposition and numerically integrated finite-magnet fields) with a prescribed one-dimensional sheath–bulk potential, adaptive fourth-order Runge–Kutta orbit integration, and a null-collision treatment of electron–argon collisions. The collision module reproduces the reduced-field dependence of the electron drift velocity but overestimates its absolute value by about a factor of 1.5. Applied to a published magnetron geometry, the simulations yield a qualitatively two-temperature electron population and a racetrack erosion calculation in which the finite-magnet field produces a more sharply localised profile whose FWHM is 'close to' a geometric estimate, while the dipole approximation is broader. The paper explicitly frames results as semi-quantitative and lists convergence studies, measured field maps, and constrained cathode-interaction models as prerequisites for quantitative extension.","tokens_in":8953,"tokens_out":5553,"duration_ms":57962,"significance":"If substantiated, the paper would be a useful, computationally light tool for comparing magnetic-field representations in magnetron discharges. Its main potential contribution is identifying finite-magnet field resolution as the dominant modelling choice for racetrack localisation and showing that the dipole approximation broadens the predicted erosion profile. Strengths are the transparent disclosure of the failed drift test (§3.2), the clear discrimination-test framing, and the repeated, explicit caveats about the model's scope. The central racetrack claim, however, is not yet supported statistically: no FWHM values, confidence intervals, or particle-number convergence study are reported, and Eq. (5) is neither derived nor quantified. The paper is honest about these gaps, which makes the claims promising but not yet established.","major_comments":[{"comment":"The central claim—that the finite-magnet field yields an erosion FWHM close to Eq. (5) while the dipole approximation is broader—is not backed by statistical evidence. 'At least 2×10^4 seed electrons' is a floor, not a convergence criterion; no FWHM values, error bars, or multiple-seed variances are reported. Section 4 explicitly lists 'particle-number convergence studies' as a future requirement, an admission that the sufficiency of N_e was not established. Since the erosion profile is an energy-weighted sum of rare high-energy impacts, Monte Carlo noise could plausibly account for the reported broadening. Please add a convergence study (e.g., N_e from 2×10^4 upward, several independent seeds) and quantify the FWHM difference with confidence intervals.","section":"§3.7 and §4"},{"comment":"Equation (5), ω_RT ≈ 2√(2 r_L^e R_c), is introduced without derivation and without specifying how r_L^e and R_c are evaluated. No numerical value of ω_RT is given, and the comparison is only qualitative ('close to'). Because the geometric estimate is computed from the same nominal B_r ≈ 1.37 T field used in the simulation, it is not an independent check of the absolute width. Please state the definitions of all quantities, give the numerical value of ω_RT, and compare it quantitatively to the simulated FWHM for both field representations. Otherwise the 'close to' claim cannot be assessed.","section":"§3.7, Eq. (5)"}],"minor_comments":[{"comment":"The temperature calculation uses only 1200 seed and approximately 1400 ionisation-born electrons; no statistical uncertainty is reported. This is acceptable for the qualitative two-temperature claim, but a sentence quantifying the bin-to-bin variance would strengthen the comparison.","section":"§3.4"},{"comment":"The initial seed-electron energy interval (4.36–4.95 eV) is stated as a numerical assumption, and the drift test overestimates the absolute drift velocity by a factor of 1.5. A brief sentence relating the emission-spectrum assumption to the drift overestimate would help readers judge whether the two issues are coupled.","section":"§2.5 / §3.2"},{"comment":"The geometric racetrack-width estimate ω_RT is plotted or indicated, but the caption does not define the plotted quantity or specify units. Please add axis labels, units, and a short definition of the geometric estimate in the caption.","section":"Figure 9"},{"comment":"There are minor typographical inconsistencies: '2 × 10^4' appears without superscript formatting in the abstract, and R_C/RC is used inconsistently. These should be normalised.","section":"Abstract and §3.7"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is honest and well scoped, and the discrimination-test design is attractive. The load-bearing racetrack claim, however, is not yet supported statistically; I regard this as a major-revision issue rather than a rejection because the missing convergence and quantification are fixable within the manuscript's scope. If the authors supply FWHM values with error bars, a particle-number convergence study, and a derivation/quantification of Eq. (5), I would support acceptance. No concerns about authorship or scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Honest, clearly-scoped reduced-order simulation. The genuinely new thing here is the head-to-head comparison of two magnetic-field representations (dipole superposition vs numerical integration over finite magnet volumes) inside a single non-self-consistent electron Monte Carlo, with the finding that the field representation is the dominant modelling choice for predicted racetrack width: finite-magnet fields give a sharper trench, dipole fields broaden it. That is practically useful if true, and the authors are appropriately modest about the rest of the model.\n\nWhat earns credit: the paper is remarkably candid about its own limits. The factor-1.5 drift-velocity overestimate is disclosed in the abstract and Section 3.2, not hidden. The temperature results are explicitly qualitative. The racetrack result is framed as a discrimination test, not a prediction. No parameters are fitted to make the racetrack or temperature curves come out right. The authors also flag that no measured field map was available, so magnetisation is inferred from nominal remanence (Section 3.4).\n\nThe soft spots are concentrated in the headline claim. Section 3.7 reports that the finite-magnet FWHM is \"close to\" the geometric estimate from Eq. (5) and that the dipole profile is \"broader,\" but gives no FWHM numbers, no confidence intervals, and no particle-number convergence study. The paper's own Section 4 lists convergence studies as required for quantitative extension. That is an explicit admission that 2e4 seed electrons may not be enough, and with an energy-weighted erosion statistic the variance could be large. Eq. (5) is also introduced without derivation and compared only qualitatively. The concluding statement that resolving the finite magnet geometry is \"necessary to obtain the observed degree of racetrack localisation\" overreaches; \"necessary\" is too strong for a qualitative comparison to a published surface image. No code or data are shipped, only available on request.\n\nThese are real gaps, but they are not load-bearing flaws in the sense that the qualitative trend is almost certainly right. The model is computationally light, the collision physics is standard, and the comparison logic is sound. The central weak point is statistical evidence: the paper needs a convergence study and some estimate of uncertainty before the dipole-broadening claim can be taken as established. That is fixable.\n\nThe paper deserves peer review—a serious referee could push on the convergence issue. It is for magnetron modellers and experimentalists who want a cheap way to compare magnet configurations. I would not cite the central result in its current form, but I would follow the revision.\n\nRecommendation: send to peer review, with a request for convergence analysis and quantitative FWHM comparison.","headline":"A candid, clearly-scoped reduced-order magnetron simulation whose headline claim—finite-magnet field predicts a sharper racetrack than dipole—is plausible but not yet statistically demonstrated; the paper itself defers the needed convergence study.","tokens_in":9561,"tokens_out":2660,"would_cite":false,"duration_ms":24341,"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":"Resolving finite permanent-magnet geometry, not dipole approximations, is the dominant modeling choice for predicting racetrack erosion width in a reduced-order magnetron model.","keywords":["magnetron sputtering","Monte Carlo simulation","racetrack erosion","magnetic-field representation","electron heating","cathode recapture","reduced-order model","planar magnetron"],"falsifier":"Run the same racetrack calculation with an order of magnitude more seed electrons (e.g., 2×10^5 or 2×10^6) for both field representations; if the finite-magnet FWHM no longer stays close to the geometric estimate while the dipole profile broadens, the claim fails. Additionally, if a measured magnetic-field map for the experimental geometry becomes available and the finite-magnet simulation no longer matches the observed trench width, the inference about the magnetization would be overturned.","tokens_in":8411,"feed_emoji":"🧲","tokens_out":4332,"duration_ms":42023,"temperature":0.7,"pith_summary":"The paper tries to establish that in a reduced-order, non-self-consistent Monte Carlo model of a planar magnetron, the way the magnetic field is represented—point dipoles versus numerical integration over the finite magnet volumes—is the first-order determinant of where the erosion racetrack forms and how wide it is. If correct, this means that cheap test-particle simulations can rank magnet designs for racetrack localization without full particle-in-cell simulations, provided they resolve the actual magnet shape. The model also reproduces a qualitative two-temperature electron population (hot near the cathode, cold away from it) and shows that the cathode reflection probability controls electron multiplication. Because the collision module overestimates absolute drift velocities by roughly a factor of 1.5, the paper frames its predictions as semi-quantitative trend studies rather than absolute discharge predictions.","feed_headline":"Magnet geometry, not dipoles, sets computed racetrack width","feed_subtitle":"In a reduced-order Monte Carlo simulation, resolving real magnet shape yields sharper erosion than dipole approximation.","key_machinery":"The central mechanism is the comparison between two magnetic-field representations: a superposition of point dipoles versus a numerically integrated field from the finite permanent-magnet volumes via the magnetic vector potential. This is embedded in a reduced-order Monte Carlo framework with a prescribed one-dimensional Gaussian-plus-cosh sheath–bulk potential, adaptive fourth-order Runge–Kutta orbit integration, a null-collision Monte Carlo collision operator, and a generation-cycle bookkeeping scheme that treats cathode-return events through a reflection probability RC. The geometric racetrack-width estimate ωRT ≈ 2√(2 r_e L R_c), based on the electron Larmor radius and magnetic-field-lin","core_discovery":"For racetrack calculations initiated with at least 2×10^4 cathode-emitted electrons and a reflection probability RC = 0.5, the finite-magnet field produces a more sharply localized erosion profile whose full width at half maximum is close to the geometric estimate ωRT ≈ 2√(2 r_e L R_c). The dipole approximation yields a broader profile. This identifies magnetic-field geometry, not the prescribed sheath potential or collision details, as the dominant modeling requirement for the predicted racetrack localization. Additionally, the reduced model demonstrates that a prescribed one-dimensional sheath–bulk potential, combined with the finite-magnet field, is sufficient to produce a hot near-cathod","pith_inferences":["If the finite-magnet versus dipole difference persists under rigorous statistical convergence checks, then common dipole-based test-particle studies may systematically overestimate racetrack width, with direct consequences for target-utilization predictions in sputtering applications.","The same reduced-order setup could be extended to test unbalanced or moving magnet configurations as a design sweep, because the finite-magnet integration cost remains workstation-scale and the model already isolates magnetic-field effects from sheath nonlinearity.","The claim is conditional on the prescribed 1D sheath–bulk potential; a self-consistent or hybrid potential could alter radial confinement, so the identified dominance of magnetic-field geometry might be partly an artifact of freezing the potential.","A testable extension would be to replace the inferred magnetization (from nominal remanence) with a measured field map for the experimental source; this would separate the error from magnetization uncertainty from the error due to the magnetic-field representation itself."],"forward_implications":["If the central claim holds, reduced-order magnetron models should resolve the finite magnet geometry rather than rely on point-dipole approximations when predicting racetrack width and localization.","The finite-magnet calculation yields a racetrack FWHM close to the geometric estimate ωRT, suggesting that a simple geometric formula can serve as a rapid benchmark for erosion width once the field is resolved.","The cathode reflection probability RC is a sensitive control on electron availability for ionizing collisions; conclusions derived from a single RC value should be read as sensitivity results, not as material constants.","Because the drift velocity is overestimated by a factor of about 1.5, absolute transport predictions are not quantitative; the model's utility lies in comparative magnetic-configuration studies and mechanism identification.","The model is not a replacement for self-consistent PIC-MCC simulations, but its low computational cost makes it suitable for scanning magnet designs and exploring cathode-interaction parameters."],"fun_headline_variants":["Magnet shape sharpens erosion in Monte Carlo model","Finite-magnet field beats dipole for racetrack width","Racetrack width pinned by magnet geometry, not dipoles","Planar magnetron: real field localizes erosion better","Reduced model: finite magnets narrow racetrack erosion"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The racetrack comparison assumes that ensembles of at least 2×10^4 seed electrons yield statistically converged radial erosion profiles, but the paper contains no convergence study or variance analysis, so the claimed FWHM difference between dipole and finite-magnet fields could be Monte Carlo noise.","fun_headline_variants_meta":{"raw":{"variants":["Magnet shape sharpens erosion in Monte Carlo model","Finite-magnet field beats dipole for racetrack width","Racetrack width pinned by magnet geometry, not dipoles","Planar magnetron: real field localizes erosion better","Reduced model: finite magnets narrow racetrack erosion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00014,"raw_usage":{"total_tokens":1029,"prompt_tokens":809,"completion_tokens":220,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":553,"completion_tokens_details":{"reasoning_tokens":152}},"tokens_in":553,"tokens_out":220,"duration_ms":3452,"temperature":1.0,"reasoning_tokens":152,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T11:17:01.363857+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same racetrack calculation with an order of magnitude more seed electrons (e.g., 2×10^5 or 2×10^6) for both field representations; if the finite-magnet FWHM no longer stays close to the geometric estimate while the dipole profile broadens, the claim fails. Additionally, if a measured magnetic-field map for the experimental geometry becomes available and the finite-magnet simulation no longer matches the observed trench width, the inference about the magnetization would be overturned.","supporting_citations":[],"review_version":1}