{"id":"e6ddd520-271c-49d6-a57d-b7a442c97072","arxiv_id":"2505.20475","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"The pyAQSC code builds near-axis expansions of anisotropic-pressure quasisymmetric equilibria to arbitrary order and demonstrates a 6th-order quasi-axisymmetric equilibrium.","lead":"This paper presents pyAQSC, the first code that solves near-axis expansions of quasisymmetric stellarator equilibria with anisotropic pressure to arbitrary order, demonstrated on a 6th-order quasi-axisymmetric case. If correct, it gives fusion researchers a fast tool to construct and optimize exact quasisymmetric configurations that isotropic-pressure near-axis codes cannot reach.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Manual Fourier truncation makes the residual-based convergence check non-independent, and the n=4 DESC outlier is the one external point left unexplained; the validation of the sixth-order claim therefore rests on a single unverified order.","rationale":"The reader's weakest assumption about filter cutoffs is real, but the sharper formulation is that the residual diagnostics are non-independent because they are evaluated after the same manual low-pass filter that defines the numerical solution. The one external validation point that disagrees is the n = 4 DESC comparison, and the paper's dismissal of it as a DESC numerical error is not backed by a resolution study. That makes the central claim, as stated, rest on an unexplained outlier. I agree with the CONDITIONAL verdict: pyAQSC is a genuine contribution with open data, a detailed derivation, and expected scaling for five of six orders, but the n = 4 anomaly should be resolved before the validation is treated as complete. The proposed test is targeted and would either clear the code or localize the defect to order 4.","tokens_in":29584,"tokens_out":5761,"duration_ms":69679,"concrete_test":"Recompute the n = 4 series of Fig. 3 with (i) doubled Fourier-Zernike and radial resolution in DESC plus a force-balance tolerance tightened by an order of magnitude, and (ii) pyAQSC reruns at Mmax,4 = 30, 40, and 50 with fixed nφ. If Bdiff at the matched ψcrit,4 points changes appreciably or the slope moves toward O(A^-5), the manual filter or resolution choices explain the n = 4 outlier; if it remains off while n = 5, 6 stay aligned, a pyAQSC recursion or filtering defect is indicated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing weakness is that the in-code residual and convergence diagnostics are not independent of the manual Fourier truncation. In §6 and App. B.3, Mmax,n is hand-picked from plots of iteration error versus retained FFT modes, at the \"elbow\" where error stops decreasing. Because every unknown is low-pass filtered before its residual is evaluated, Fig. 1's small residuals and exponential coefficient growth cannot by themselves distinguish \"the recursion is correct\" from \"the filter removed the offending modes.\" The external check is Fig. 3: DESC truncation-error scaling behaves as O(A^-(n+1)) for n = 1, 2, 3, 5, 6, but n = 4 visibly departs. The paper attributes this to DESC numerical error, yet gives no resolution study or error bars to support that claim. Since n = 5 and n = 6 consume n = 4 data, their good scaling does not fully exonerate n = 4. A second independent check—recomputing n = 4 with higher DESC resolution and varied Mmax,4—would settle whether the outlier is a DESC artifact or a pyAQSC/filter issue. This does not invalidate the code's substantial contribution; it means the \"6th-order equilibrium validated against DESC\" claim is currently supported by five orders, with order 4 unverified.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents pyAQSC, a numerical implementation of the Rodriguez-Bhattacharjee near-axis expansion (NAE) for quasisymmetric magnetohydrostatic equilibria with anisotropic pressure. The authors claim that pyAQSC is the first code to solve this expansion to arbitrary order, and they demonstrate a sixth-order quasi-axisymmetric (QA) equilibrium with anisotropic pressure. The paper describes the recursion relations, the numerical methods (pseudo-spectral in the toroidal angle, finite Fourier series in the poloidal angle, JAX-based automatic differentiation), and the empirical low-pass filtering used to control high-frequency noise. Validation is attempted in two ways: internal residuals of the ordered governing equations (Fig. 1) and comparison against global DESC equilibria whose boundaries are constructed from the near-axis solution (Fig. 3). The paper also reports a measured radius of convergence, a study of flux-surface self-intersection, and a proof-of-principle DESC fit showing good quasisymmetry quality. The central claims are that pyAQSC correctly evaluates the recursion relations to sixth order and that it can serve as an initial-state tool for anisotropic-pressure quasisymmetric stellarator design.","tokens_in":29875,"tokens_out":4115,"duration_ms":44805,"significance":"If the central claims hold, this is a valuable contribution. The near-axis expansion with anisotropic pressure is a promising route around the Garren-Boozer overdetermination problem, and a robust, efficient, arbitrary-order numerical implementation would open new possibilities for stellarator optimization and for studying high-order properties of quasisymmetric equilibria. The paper's strengths include: (i) a detailed, self-contained presentation of the recursion relations and numerical implementation; (ii) internal residual checks at every order; (iii) an external comparison with the independent global solver DESC for five of six orders; (iv) the use of auto-differentiation and GPU vectorization, which are practically important for optimization; and (v) an open validation dataset. However, the validation is not yet fully convincing: the residual diagnostics are not independent of the manual Fourier truncation, and the n=4 DESC comparison visibly departs from the expected scaling. Because orders n=5 and n=6 consume n=4 data, the sixth-order validation currently rests on an unverified order.","major_comments":[{"comment":"The n=4 data in Fig. 3 do not follow the expected O(A^-(n+1)) scaling, while n=1,2,3,5,6 do. The authors attribute this to 'numerical errors in the global equilibrium solver' but provide no resolution study, error bars, or convergence test to support that attribution. Since the n=5 and n=6 coefficients depend on n=4 data, the claim that pyAQSC is validated to sixth order cannot be fully sustained until the n=4 outlier is explained. I recommend recomputing the n=4 case with varied DESC resolution and varied Mmax,4, and either showing that the outlier disappears or identifying the source of the discrepancy. Without this, the 'sufficiently validated' conclusion in §7.3 is premature.","section":"§7.3, Fig. 3"},{"comment":"The manual selection of the low-pass filter cutoff Mmax,n at the 'elbow' of the iteration-error curve introduces a circularity in the validation logic. As described in App. B.1, every unknown is low-pass filtered before its residual is evaluated; therefore, the small residuals in Fig. 1 and the apparent exponential coefficient growth cannot, by themselves, distinguish a correct recursion from a truncation that removes the offending high-frequency modes. The authors should quantify the sensitivity of the results to the cutoff choice, for example by reporting coefficient magnitudes, residual errors, and the radius-of-convergence estimate over a range of Mmax,n values, or by adopting a regularization criterion that is independent of the residual being measured. This would address the concern that the convergence properties in Fig. 1 are artifacts of the filter.","section":"§6 and App. B.3"},{"comment":"The DESC fitting procedure used in §7.4 is explicitly noted by the authors as designed for isotropic-pressure NAE ([31]), while the paper's novel capability is anisotropic pressure. The reported normalized force-balance error (6.058×10^-6) is computed with this isotropic-fitting method, so the demonstration that pyAQSC can serve as an initial-state tool for anisotropic-pressure equilibria is not yet fully supported. The authors should either implement and use a proper anisotropic-pressure fitting procedure or, failing that, temper the abstract and conclusion claims of 'comparing the RB method with DESC equilibria with anisotropic pressure' and present the §7.4 result as a proof-of-concept with a clear statement of this limitation.","section":"§7.4 and Table 3"}],"minor_comments":[{"comment":"There is a typo: 'piAQSC' should be 'pyAQSC'.","section":"§2"},{"comment":"The manuscript refers to '3.2' in several places where a section number is meant (e.g., 'as discussed in 3.2' in §5.2 and §A.6, and 'see App. A.6' in §5.2). These should be updated to consistent cross-references.","section":"§5.2, §A.5, §A.6"},{"comment":"The notation in the sentence 'the aspect ratio An ∝ αn' is confusing: An is used both for the asymptotic coefficient magnitude and for the aspect ratio A in the DESC comparison. Please use distinct symbols and define them.","section":"§7.3"},{"comment":"The captions are difficult to parse; for example, Fig. 2(b) and Fig. 3 contain 'crit, n' and 'crit, 1 = ...' without clear units or a definition of the plotted quantity. Please improve the captions so that the reader can understand the curves without referring to the text.","section":"Fig. 2 and Fig. 3 captions"},{"comment":"The data availability statement cites a Zenodo dataset, but no repository or availability statement is given for the pyAQSC source code itself. Since the paper's central contribution is a code, please provide a code availability link or state clearly where the code can be obtained.","section":"§9 and Introduction"},{"comment":"The table contains entries like '10 3×' and '10 2 ∼ 103×' that appear to be missing superscript formatting; these should be rendered as 10^3× and 10^2–10^3× for readability.","section":"§7.2, Table 2"}],"recommendation":"major_revision","confidential_remarks":"The two issues most needing attention are the unexplained n=4 outlier in the DESC comparison and the circularity introduced by the manual filter-cutoff selection. Both are addressable with additional numerical experiments and would substantially strengthen the paper. A secondary point is the mismatch between the paper's abstract claim about DESC comparisons with anisotropic pressure and the isotropic-pressure fitting procedure used in §7.4. If the authors can resolve the n=4 discrepancy and demonstrate robustness of the results to filter-cutoff choices, the paper would likely be suitable for acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"PyAQSC is the real thing: a working arbitrary-order anisotropic-pressure NAE code, demonstrated to sixth order, with explicit recursion relations and a Zenodo dataset. The paper earns credit for implementing the RB scheme, for the careful account of the looped equations, and for validating against DESC rather than only self-residuals. The 6th-order QA equilibrium with measured convergence radius is new, and the code is fast, auto-differentiable, and interfaced with DESC and Simsopt. The central claim—that the recursion can be evaluated to any order—is largely supported.\n\nThe soft spots are real but not fatal. The internal residual checks are not independent of the low-pass filter: Mmax,n is hand-picked from the elbow of iteration-error curves, and filtering happens before residuals are computed, so small residuals cannot fully distinguish a correct recursion from a filter that removes offending modes. The external DESC comparison is the better check, and it passes for n=1,2,3,5,6. The n=4 outlier is the one unexplained point, and because n=5 and 6 consume order-4 data, the sixth-order claim currently rests on five good orders plus an assumption. The paper attributes the deviation to DESC numerical error, but gives no resolution study or error bars to back that up. A focused rerun at higher DESC resolution and varied Mmax,4 would settle it. The section 7.4 demonstration also uses an isotropic-pressure fitting procedure, which the authors explicitly flag; that limits what the 'initial state tool' claim can mean for anisotropic equilibria until a proper anisotropic fit exists.\n\nNone of this undercuts the contribution. The recursion relations are detailed enough to reimplement, the code is out, and the validation is mostly convincing. This is a paper that deserves a serious referee, not a desk reject. I would want a referee to push for the n=4 resolution study and a sensitivity analysis of the filter cutoffs; if those come back clean, the 'any order' claim will be on solid ground.","headline":"First working arbitrary-order anisotropic-pressure NAE code, with a solid sixth-order demonstration; the n=4 DESC outlier and hand-picked filter cutoffs are real concerns but fixable.","tokens_in":30411,"tokens_out":2420,"would_cite":true,"duration_ms":25047,"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":"Pressure anisotropy removes the third-order overdetermination that blocks near-axis quasisymmetric equilibria, and this paper's new code, pyAQSC, carries the anisotropic expansion to arbitrary order, demonstrated at sixth order.","keywords":["quasisymmetry","near-axis expansion","anisotropic pressure","stellarator design","magnetohydrostatic equilibrium","overdetermination problem","high-order asymptotic expansion","automatic differentiation"],"falsifier":"Recompute the sixth-order equilibrium with the low-pass filter cutoffs increased or removed entirely, and check whether the ordered-equation residuals remain small and whether the apparent growth rate $\\bar{\\alpha} \\approx 320 \\pm 110$ and radius of convergence $\\psi_{\\mathrm{conv}} \\approx (9.76 \\pm 6.65)\\times 10^{-6}\\,\\mathrm{T\\,m^2}$ persist; if the residuals inflate or the measured divergence rate changes sharply with the cutoff, the convergence claim is filter-induced.","tokens_in":29359,"feed_emoji":"🧲","tokens_out":13943,"duration_ms":115369,"temperature":0.7,"pith_summary":"Quasisymmetry is a property of a magnetic field in which the field strength, though not the full field vector, has a direction of symmetry; quasisymmetric plasma configurations lose much less energy through particle collisions. This paper argues that the long-standing obstruction to constructing globally quasisymmetric magnetohydrostatic equilibria—the overdetermination of the near-axis expansion at third order—disappears when the plasma pressure is allowed to be anisotropic. On that basis it presents pyAQSC, the first code that solves the near-axis expansion of anisotropic-pressure quasisymmetric equilibria to arbitrary order. The paper demonstrates the claim with a sixth-order quasi-axisymmetric equilibrium with anisotropic pressure, showing that residuals of the ordered governing equations are small and that the error between the near-axis solution and a global equilibrium solver scales as expected. If correct, the code opens a route to exact global quasisymmetry and supplies a fast, auto-differentiable tool for stellarator design and optimization.","feed_headline":"Anisotropic pressure unlocks arbitrary-order quasisymmetric equilibria","feed_subtitle":"Pressure anisotropy sidesteps a three-decade overdetermination barrier, yielding a validated sixth-order near-axis stellarator.","key_machinery":"The machinery is the ordered set of recursion relations obtained by substituting power-Fourier series in $\\epsilon$ into inverse-coordinate forms of the governing equations. At each order the iteration solves a small linear system: an ODE for the covariant component $B_{\\psi n-2}$, algebraic expressions for $Z_n$ and $X_n$, a linear first-order ODE for $Y_n$, and—for the force-balanced problem—a coupled linear system called the looped equations that determines $\\{B_{\\theta n}, B_{\\psi n-2,0}, Y_n^{\\mathrm{free}}\\}$ together with the average anisotropy $\\bar{\\Delta}_{n,0}$ two orders at a time. The recursion is evaluated pseudo-spectrally, storing the toroidal-angle dependence on a grid, and repeated derivatives are controlled by an empirical Fourier low-pass filter applied after each order.","core_discovery":"The central claim is that pressure anisotropy removes the third-order overdetermination that has been interpreted as forbidding non-axisymmetric globally quasisymmetric magnetohydrostatic equilibria. The near-axis expansion, written in a quasisymmetry-adapted magnetic coordinate system with effective minor radius $\\epsilon = \\sqrt{\\psi}$, is no longer forced to stop; at each order the magnetic equations $\\{J, C_b, C_\\kappa, C_\\tau\\}$ and the force-balance projections $\\{I, II, III\\}$ yield linear recursion relations that determine the flux-surface shape coefficients $X_n, Y_n, Z_n$, the covariant field components $B_{\\theta n}, B_{\\psi n-2}$, and the pressure coefficients $p_{\\perp n}, \\Delta_n$ from lower-order data. The paper validates this with a sixth-order quasi-axisymmetric equilibrium: the residuals of the ordered equations stay at the iteration-error level at every order, and the difference from a global anisotropic equilibrium solver follows the expected $O(A^{-(n+1)})$ truncation scaling for orders $n=1,2,3,5,6$. The measured radius of convergence is $\\psi_{\\mathrm{conv}} \\approx (9.76 \\pm 6.65)\\times 10^{-6}\\,\\mathrm{T\\,m^2}$, and the critical flux-surface radius where surfaces self-intersect appears to approach this value as the order increases.","pith_inferences":["If the observed approach of $\\psi_{\\mathrm{crit},n}$ to $\\psi_{\\mathrm{conv}}$ is real, flux-surface self-intersection is the mechanism that ends the series; a test is whether one more order narrows the gap further, and whether suppressing self-intersection by axis shaping enlarges the convergent volume.","Optimizing in the anisotropic solution space for low pressure anisotropy could locate equilibria that are nearly isotropic yet globally quasisymmetric, effectively sidestepping the isotropic overdetermination obstruction in the limit relevant to fusion plasmas.","The same recursion machinery is likely portable to other near-axis objectives—quasi-isodynamic or other omnigenous designs—where overdetermination also truncates isotropic expansions.","The code's auto-differentiability makes the radius of convergence itself a differentiable design objective, so enlarging the usable volume could be formulated as a gradient-based optimization problem."],"forward_implications":["Quasisymmetric equilibria with exact global QS and finite anisotropy can be constructed to arbitrarily high order, making high-order properties such as magnetic shear and detailed flux-surface geometry accessible for the first time in the anisotropic setting.","A sixth-order near-axis solution is convergent only within a small volume around the axis, but the radius of convergence can in principle be enlarged by optimizing the axis shape and profiles.","Truncated near-axis solutions can be extended to global equilibria by fitting boundary shapes and profiles, reaching quasisymmetry quality comparable to several existing optimized configurations without direct QS optimization.","Because the code is auto-differentiable and can be vectorized on GPUs, it can scan thousands of configurations per second and supply initial states that make global stellarator optimization markedly faster."],"supporting_citations":[{"why":"States the near-axis overdetermination of isotropic quasisymmetric equilibria, the problem this paper sets out to resolve.","marker":"[1]"},{"why":"Reports numerical isotropic-pressure equilibria with nearly exact quasisymmetry, the apparent contradiction the anisotropic formulation reconciles.","marker":"[2]"},{"why":"Introduces the quasisymmetry-adapted magnetic coordinate system in which the expansion is performed.","marker":"[5]"},{"why":"Establishes the anisotropic-pressure force-balance formulation whose recursion relations this code implements.","marker":"[6]"},{"why":"Supplies the global equilibrium solver used to compute the reference equilibria for validation.","marker":"[7]"},{"why":"Provides the high-order near-axis expansion methodology and the O(A^{-n}) surface-fitting error estimate used to interpret validation residuals.","marker":"[13]"},{"why":"Derives the all-orders weakly quasisymmetric magnetic recursion relations that form the backbone of the code.","marker":"[28]"},{"why":"Shows the overdetermination is resolved for circular-axis anisotropic equilibria, the precedent for the general implementation.","marker":"[29]"},{"why":"Describes the procedure for extending near-axis equilibria into global fits with matching profiles, used in the initial-state demonstration.","marker":"[31]"}],"fun_headline_variants":["Anisotropic pressure breaks the quasisymmetry barrier","PyAQSC: first code for arbitrary-order anisotropic quasisymmetric equilibria","Sixth-order quasisymmetric equilibria with pressure anisotropy","Anisotropy resolves the Garren-Boozer conundrum","Near-axis expansion with anisotropy unlocks higher-order equilibria"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the manually selected low-pass filter cutoffs—chosen by scanning candidate frequencies for low iteration error—do not hide genuine divergence or suppress physical modes; if the filters are too aggressive, the claimed convergence and validation residuals would be artifacts of the filtering rather than properties of the expansion.","fun_headline_variants_meta":{"raw":{"variants":["Anisotropic pressure breaks the quasisymmetry barrier","PyAQSC: first code for arbitrary-order anisotropic quasisymmetric equilibria","Sixth-order quasisymmetric equilibria with pressure anisotropy","Anisotropy resolves the Garren-Boozer conundrum","Near-axis expansion with anisotropy unlocks higher-order equilibria"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000851,"raw_usage":{"total_tokens":3809,"prompt_tokens":1164,"completion_tokens":2645,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":780,"completion_tokens_details":{"reasoning_tokens":2558}},"tokens_in":780,"tokens_out":2645,"duration_ms":20757,"temperature":1.0,"reasoning_tokens":2558,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:53:38.341083+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the sixth-order equilibrium with the low-pass filter cutoffs increased or removed entirely, and check whether the ordered-equation residuals remain small and whether the apparent growth rate $\\bar{\\alpha} \\approx 320 \\pm 110$ and radius of convergence $\\psi_{\\mathrm{conv}} \\approx (9.76 \\pm 6.65)\\times 10^{-6}\\,\\mathrm{T\\,m^2}$ persist; if the residuals inflate or the measured divergence rate changes sharply with the cutoff, the convergence claim is filter-induced.","supporting_citations":[{"cited_title":"Generalized Boozer coordinates: A natural coordinate system for quasisymmetry","cited_arxiv_id":null,"evidence_quote":"Introduces the quasisymmetry-adapted magnetic coordinate system in which the expansion is performed."},{"cited_title":"Constructing stellarators with quasisymmetry to high order","cited_arxiv_id":null,"evidence_quote":"Provides the high-order near-axis expansion methodology and the O(A^{-n}) surface-fitting error estimate used to interpret validation residuals."},{"cited_title":"Weakly quasisymmet- ric near-axis solutions to all orders","cited_arxiv_id":null,"evidence_quote":"Derives the all-orders weakly quasisymmetric magnetic recursion relations that form the backbone of the code."}],"review_version":1}