{"id":"71e22ca4-f9b2-4bb8-8e20-19a86a58b34c","arxiv_id":"2506.05170","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The paper derives linear Fourier-Zernike constraints that make global DESC equilibria match a prescribed near-axis expansion through second order in the flux-surface radius.","lead":"A new method in the DESC stellarator equilibrium code directly imposes the mathematical near-axis expansion as constraints on the global flux-surface solution, instead of building an outer boundary from the approximate expansion. This preserves optimized core properties such as rotational transform and magnetic well at low aspect ratio, while leaving the outer plasma free to be determined by full equilibrium force balance.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central 'guarantee' depends on unenforced identification of DESC's θ with the Boozer angle; comparing DESC's λ1 to Eq. (3.10) would test whether the constraints actually enforce the intended NAE behavior.","rationale":"I read the paper in good faith as a numerical method: global DESC equilibria are constrained near the axis by linear constraints on Fourier-Zernike coefficients derived from a given NAE solution. The geometric derivation in Section 3 is careful, and the benchmarks are strong evidence that the method works for the tested cases, with Δι0 near 10^-5 and fB scaling as ρ^2/ρ^3. The load-bearing step is the identification of DESC's computational angle θ with the Boozer angle θ_B. This is not an imposed constraint; λ is a free variable in DESC (Section 2.1). The R/Z constraints (3.8, 3.13) are derived under this identification, so if the converged λ differs from -ι ν at first order, the constrained shapes are not the NAE shapes in Boozer coordinates. The paper checks only the on-axis λ0 (Eq. 3.9) and finds it close but not exact; the first-order λ1, which controls the angle mapping away from the axis, is not reported. The final relaxation step (Section 4) also means the presented equilibria are fixed-boundary solves, so the NAE match is an empirical outcome, not a guarantee. The proposed test directly measures λ1 and would determine whether the angle identification holds at the order needed. If it fails, the method requires the Appendix B generalization or a qualification of the central claim. Because the reader's CONDITIONAL verdict already rests on this same assumption and asks for a qualification of 'guarantee', my read leaves the verdict unchanged.","tokens_in":51,"tokens_out":13769,"duration_ms":282171,"concrete_test":"Extract λ_{1,±1,n} from the converged first-order constrained 'precise QA' equilibrium and compare to the NAE prediction λ^{(1)} = ι0 (x^φ_1/R0)(1 - λ0'/ι0) from Eq. (3.10), mapped to Fourier-Zernike coefficients via Eq. (2.13). If they differ at order ρ, θ deviates from θ_B at first order and the constraints do not enforce the intended Boozer-frame behavior. Repeat for the QI case (Δλ0 ≈ 5×10^-2) and for the second-order constrained 'precise QA+well' case, checking whether V'' and the fB scaling degrade.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim—that the construction 'guarantees the correct asymptotic behaviour' of a given NAE—depends on identifying DESC's computational poloidal angle θ with the near-axis Boozer angle θ_B via the stream-function relation λ = -ι ν (Eq. 2.9). This identification is not imposed: DESC solves for λ as part of the equilibrium (Section 2.1), and only R and Z are constrained (Section 4). The verification (Tables 1–2) checks only the on-axis value λ0 against Eq. (3.9), with small but nonzero deviations (e.g., Δλ0 = 5.35e-02 for the QI case). Because the geometric constraints (3.8) and (3.13) are written in θ, any first-order deviation of λ from -ιν means θ ≠ θ_B and the imposed R/Z constraints no longer correspond to the intended NAE behavior in Boozer coordinates. The authors acknowledge this restriction and defer a general-angle formulation to Appendix B and future work, so the abstract's unconditional 'guarantee' exceeds what the method enforces. The final fixed-boundary relaxation step (Section 4) then removes the constraints entirely, making the match to NAE purely empirical for the presented configurations.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a method for constructing global ideal-MHD equilibria in the DESC code that match a prescribed near-axis expansion (NAE) to zeroth, first, or second order in the distance from the magnetic axis. The authors derive linear constraints on DESC's Fourier-Zernike coefficients by (i) relating the NAE Taylor-Fourier basis to the Zernike basis, and (ii) transforming the NAE Frenet-Serret description into the cylindrical-coordinate description used by DESC. They implement the constraints using pyQSC and pyQIC inputs, solve constrained equilibria, and then relax the constraints and re-solve as fixed-boundary equilibria as a final check. Benchmarks for quasi-axisymmetric, quasi-helically symmetric, and quasi-isodynamic configurations show that on-axis quantities such as B0, iota0, and V'' match the NAE values much better than conventional surface-built equilibria, and that the quasisymmetry error scales as expected with the order of the imposed expansion.","tokens_in":24959,"tokens_out":5768,"duration_ms":74218,"significance":"If the results hold, this is a useful methodological advance: it gives a systematic, transparent route from near-axis optimized designs to global equilibria, avoiding the common practice of using the NAE at a finite radius where it is least valid. The theoretical derivations in Sections 3.2-3.3 and Appendices D-E are carefully presented, and the paper ships data and plotting scripts on Princeton Data Commons, which is commendable. The numerical benchmarks show the expected O(rho^2) and O(rho^3) scaling of the quasisymmetry error and large improvements in on-axis rotational transform and magnetic well compared with the fixed-boundary approach. The main caveats are that the abstract's 'guarantees' is stronger than what the implemented workflow actually enforces, and the identification of DESC's poloidal angle with the Boozer angle is not directly verified.","major_comments":[{"comment":"The abstract's claim that the construction 'guarantees the correct asymptotic behaviour' is not supported by the implemented workflow, because the near-axis constraints are relaxed after the constrained solve and the equilibrium is re-solved as a fixed-boundary problem. The verification numbers in Tables 1 and 2 therefore describe a solution that is no longer subject to the constraints; for example, Table 1 reports Delta_lambda0 = 5.35e-02 for the QI case and Delta_lambda0 = 2.54e-02 for the QH case. Please either present the constrained solution itself as the deliverable, quantify the drift introduced by the relaxation step, or replace 'guarantees' with language that accurately describes the constrained solve and the subsequent verification step.","section":"Section 4, paragraph before §4.1; Tables 1-2"},{"comment":"The identification of DESC's computational poloidal angle theta with the Boozer angle theta_B through lambda = -iota*nu is an assumption, not an enforced condition. DESC solves for lambda as part of the equilibrium, and at first order only R and Z are constrained; if the computed lambda deviates from -iota*nu at O(rho), then the imposed R/Z constraints no longer correspond to the intended NAE behavior in Boozer coordinates. The nonzero values of Delta_lambda0 in Table 1 show that this deviation is not negligible, especially for QH and QI. Please add a diagnostic that compares the first-order stream function lambda_1 with Eq. (3.10), or impose a constraint on lambda_1 as well, before claiming that the asymptotic behavior is guaranteed.","section":"§2.2, Eq. (2.9); §3.2.1, Eq. (3.10)"},{"comment":"The verification compares quantities such as B0, iota0, V'', and the fB scaling against the same NAE data that were used to construct the constraints. This is a legitimate consistency check of the implementation, but it cannot independently validate that the global equilibrium has the intended near-axis behavior in Boozer coordinates. A more independent test, such as a direct Boozer-coordinate Fourier analysis of the DESC solution or a comparison against a separately generated NAE solution, would substantially strengthen the paper's central claim.","section":"§4.1 and §5"}],"minor_comments":[{"comment":"The phrase 'Their benefits stride from the freedom' should be 'Their benefits stem from the freedom'.","section":"Section 1"},{"comment":"The word 'meaingful' should be 'meaningful'.","section":"Figure 4 caption"},{"comment":"The word 'equilibirum' should be 'equilibrium'.","section":"Section 5.2"},{"comment":"Equation (2.13) is numbered identically in the main text and in Appendix C.3; renumber one of them or reference the equation only once.","section":"Eq. (2.13) and Appendix C.3"},{"comment":"The symbols tau_tilde, kappa'_Z, and kappa_R are used in the equation but defined only in the surrounding prose; a brief definition immediately before the equation would improve readability.","section":"Eq. (3.16b)"},{"comment":"Several references contain LaTeX artifacts such as 'tex.ids=' and stray 'publisher:' fields (e.g., Boozer 1983, Anderson et al. 1995, Landreman 2022b); these should be cleaned up before publication.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for J. Plasma Phys. and represents a useful step toward coupling near-axis and global equilibrium descriptions. The main concern is the gap between the abstract's 'guarantees' and the actual procedure, which relaxes the constraints in the final step; this is fixable by rewording and by reporting the constrained solution or the relaxation drift explicitly. The angle-identification issue is acknowledged by the authors, but the main body would be substantially stronger with a direct lambda_1 comparison as suggested. The consistency-check nature of the verification is not fatal, but an independent check would increase confidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Key point: the paper delivers a real, useful bridge between near-axis expansions and global DESC equilibria, and the benchmarks show it works for QA, QH, and QI configurations at low aspect ratio. The soft spot is the word 'guarantees' in the abstract: the identification of DESC's poloidal angle with the Boozer angle is assumed, not enforced, and the final relaxation step makes the match empirical. That is a qualification, not a fatal flaw.\n\nWhat is new: the linear constraints on Fourier-Zernike coefficients derived from a geometric transformation between the Frenet-Serret/Boozer frame and the cylindrical lab frame. The derivation is careful, the implementation via feasible direction method is sound, and the verification is honest—they compare derived quantities (B0, iota0, lambda0, V'') rather than the constrained shapes themselves. The quasisymmetry error scaling and magnetic well matching are compelling evidence that the near-axis behavior is preserved in the global solution. Credit also for shipping data and scripts.\n\nSoft spots: (1) The central identification θ = θ_B (Eq 2.9) is not imposed on λ; only R and Z are constrained. The stress-test concern is fair: if λ deviates from -ι ν at first order, the imposed R/Z constraints no longer correspond to the intended Boozer-frame behavior. The verification checks λ0 only, and the QI case shows Δλ0 ≈ 5e-2, meaning a few percent deviation. This is acknowledged and deferred, but it undercuts the word 'guarantee'. (2) The final fixed-boundary relaxation step removes the constraints entirely, so the match is empirically demonstrated for the presented cases, not guaranteed in general. (3) The omitted second-order algebra in Appendix E is minor; the formulas are given and code is available.\n\nOverall, the method is a clear step forward for NAE-to-global coupling. It deserves serious peer review. Recommend accepting with revision: qualify the abstract, add a direct test of the θ/θ_B identification (e.g., comparing Δλ1 to Eq. 3.10), and clarify what is guaranteed vs. empirically verified.","headline":"Solid new bridge between near-axis expansions and global DESC equilibria; abstract's 'guarantee' overstates what the method actually enforces.","tokens_in":25426,"tokens_out":2122,"would_cite":true,"duration_ms":24753,"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":"Global stellarator equilibria can keep their near-axis design","keywords":["near-axis expansion","stellarator","global MHD equilibrium","Fourier-Zernike","DESC","quasisymmetry","magnetic well","linear constraints"],"falsifier":"Compute a second-order constrained equilibrium for a configuration with a strongly shaped, high-torsion axis (for example a quasi-isodynamic case), then Fourier-analyze the $R$ and $Z$ cross-sections at small fixed $\\rho$ and compare the second-order coefficients to the NAE-predicted lab-frame values; a systematic deviation as torsion grows would indicate that the leading-order angle identification is the cause.","tokens_in":24404,"feed_emoji":"🧲","tokens_out":6498,"duration_ms":69337,"temperature":0.7,"pith_summary":"The paper aims to make the near-axis expansion (NAE) of stellarator fields actionable for global equilibrium construction: instead of feeding a near-axis surface far from the axis into a fixed-boundary solver, it constrains the global solve of the DESC code directly at the axis, so the global equilibrium provably reproduces the prescribed near-axis asymptotic behavior. The central technical step is a set of linear constraints on DESC's Fourier-Zernike coefficients (Eqs. (3.8) and (3.13)) that encode the magnetic axis, the elliptical first-order cross-sections, and the second-order triangular shaping of a given NAE solution, while the solver remains free to find the equilibrium far from the axis. If correct, this gives a practical bridge between fast near-axis design and high-fidelity global equilibria, preserving curated near-axis properties at low aspect ratios where the standard fixed-boundary approach loses them; the benchmarks show that on-axis rotational transform, field strength, quasisymmetry-error scaling, and magnetic well all match the NAE far better than the conventional construction.","feed_headline":"Near-axis design can be imposed on global stellarator equilibria","feed_subtitle":"Linear constraints lock the near-axis field into DESC's global equilibrium solve while leaving the rest of the plasma free.","key_machinery":"The load-bearing object is the linear map between the NAE's Taylor-Fourier coefficients and DESC's Fourier-Zernike coefficients, Eq. (2.13). It converts the requirement 'this flux surface has shape f at order l' into a weighted sum of Zernike modes, with weights that grow like $k^l$ for high radial order $k$; convergence is controlled by a finite truncation. On top of that map, the paper derives a geometric transform (Eqs. (3.6)-(3.7) and (3.12)) that rewrites the axis-frame elliptical and triangular surface shapes in cylindrical coordinates, accounting for the inclination of the axis. These two pieces combine into the concrete linear constraints (3.8) and (3.13) that the solver enforces while minimizing the magnetohydrostatic force residual.","core_discovery":"The central claim is that a global ideal-MHD equilibrium in DESC can be forced to have exactly the near-axis behavior of a chosen NAE solution by imposing a handful of linear constraints on the solver's spectral coefficients. The near-axis Taylor-Fourier description of flux surfaces and the Zernike representation used by DESC are connected by an exact linear relation (2.13), and the near-axis surfaces, which are naturally written in the Frenet-Serret frame of the axis and in Boozer coordinates, are mapped to the cylindrical lab frame through the geometric relations (3.6)-(3.7) at first order and (3.12a)-(3.12b) at second order. The resulting constraints fix the axis shape, the elliptical cross-sections, and the second-order shaping (including the Shafranov shift and triangularity as seen in the lab frame) as linear combinations of Fourier-Zernike modes. With these constraints imposed through a feasible-direction method, DESC returns equilibria whose on-axis rotational transform, on-axis |B|, stream function, and magnetic well agree with the NAE to orders of magnitude better than equilibria built from a finite-radius near-axis boundary, and the quasisymmetry error scales as $O(\\rho^2)$ (first-order constraints) or $O(\\rho^3)$ (second-order constraints).","pith_inferences":["The same linear-constraint approach could be adapted to other spectral equilibrium codes or boundary-based solvers, as long as a radial-poloidal basis with the right near-axis regularity is available.","The angle-agnostic formulation sketched in Appendix B suggests a testable improvement: letting DESC use a generalized poloidal angle should reduce the toroidal-mode burden for quasi-isodynamic configurations with straight axis sections.","One could use the freedom left by the near-axis constraints to scan off-axis properties systematically, generating families of global equilibria that share the same core design but differ in boundary shaping, coils, or stability.","Soft or partial enforcement of the constraints, rather than exact imposition, may be useful during optimization to trade near-axis fidelity against other objectives."],"forward_implications":["Near-axis-optimized properties—rotational transform, on-axis field strength, quasisymmetry error, and magnetic well—survive in a global equilibrium at low aspect ratio, where finite-radius fixed-boundary construction degrades them.","The NAE-constrained equilibrium provides a much better initial condition for conventional stellarator optimization than a fixed-boundary solve from a large-radius near-axis surface.","Because the constraints are linear, they can be added to the same constrained-optimization machinery already used for fixed-boundary solves, with modest extra cost.","Second-order constraints transfer magnetohydrodynamic-stability-linked features such as the magnetic well from the near-axis design into the global solution.","The method generalizes beyond vacuum fields, since the constraint derivation itself does not require a vacuum or quasisymmetry assumption."],"supporting_citations":[{"why":"Supplies the near-axis construction in Boozer coordinates and the Taylor-Fourier coefficient form that the constraints are built from.","marker":"Landreman & Sengupta 2019"},{"why":"Foundational near-axis expansion theory establishing the asymptotic equilibrium description near the magnetic axis.","marker":"Garren & Boozer 1991c"},{"why":"Describes the DESC code's spectral equilibrium solve and force-residual minimization into which the constraints plug.","marker":"Panici et al. 2023"},{"why":"Provides the precise quasi-axisymmetric and quasi-helically symmetric configurations used as benchmark near-axis constructions.","marker":"Landreman & Paul 2022"},{"why":"Supplies the quasi-isodynamic near-axis construction used in benchmarks and the omnigenity framework it relies on.","marker":"Plunk et al. 2019"},{"why":"Relates the magnetic well $V''$ to near-axis quantities, used as a second-order verification measure.","marker":"Landreman & Jorge 2020"},{"why":"Provides the feasible-direction method used to impose the linear constraints during the equilibrium optimization.","marker":"Nocedal & Wright 1999"},{"why":"Defines the quasisymmetry error measure used to check the scaling of symmetry-breaking modes in the benchmarks.","marker":"Rodríguez et al. 2022a"}],"fun_headline_variants":["Near-axis design imposed exactly on global equilibria","Global stellarator equilibria forced to match near-axis theory","Linear constraints embed near-axis behavior into DESC","Exact near-axis constraints for global equilibria in DESC"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation assumes the poloidal angle used by DESC can be identified with the near-axis Boozer angle through the simple stream-function relation $\\lambda = -\\iota\\nu$, which is only guaranteed to leading order; if that angle identification fails at higher order, the geometric constraints will not enforce the intended near-axis behavior.","fun_headline_variants_meta":{"raw":{"variants":["Near-axis design imposed exactly on global equilibria","Global stellarator equilibria forced to match near-axis theory","Linear constraints embed near-axis behavior into DESC","Exact near-axis constraints for global equilibria in DESC"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000158,"raw_usage":{"total_tokens":1213,"prompt_tokens":922,"completion_tokens":291,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":538,"completion_tokens_details":{"reasoning_tokens":229}},"tokens_in":538,"tokens_out":291,"duration_ms":3682,"temperature":1.0,"reasoning_tokens":229,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:22:49.038606+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute a second-order constrained equilibrium for a configuration with a strongly shaped, high-torsion axis (for example a quasi-isodynamic case), then Fourier-analyze the $R$ and $Z$ cross-sections at small fixed $\\rho$ and compare the second-order coefficients to the NAE-predicted lab-frame values; a systematic deviation as torsion grows would indicate that the leading-order angle identification is the cause.","supporting_citations":[{"cited_title":"Journal of Plasma Physics 86 (5), 905860510","cited_arxiv_id":null,"evidence_quote":"Relates the magnetic well $V''$ to near-axis quantities, used as a second-order verification measure."},{"cited_title":"Springer","cited_arxiv_id":null,"evidence_quote":"Provides the feasible-direction method used to impose the linear constraints during the equilibrium optimization."}],"review_version":1}