{"id":"e6bce0e7-9284-49b6-b445-7e7ceac7c1a7","arxiv_id":"2411.14874","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Simulations of a triple-core gain-loss fiber show chiral and nonchiral mode conversion driven by encirclement of two connected second-order exceptional points, with the effect persisting for loops that pass near but do not encircle the exceptional points.","lead":"This paper uses simulations to design a triple-core optical fiber with gain and loss, and shows how light modes switch when the gain-loss pattern is varied along the fiber. It tests whether higher-order exceptional-point effects, previously mostly studied in chips, can work inside a fiber for all-fiber switching and mode conversion.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The EP3 claim rests only on two separate EP2s sharing one mode at distinct parameter points; no triple coalescence is demonstrated, so the 'third-order exceptional point' label is unsupported.","rationale":"The reader's weakest assumption is the unverified inference of an EP3 from two interconnected EP2s. My analysis confirms this is the single most load-bearing concern. The paper's claimed contribution — a third-order exceptional point in an optical fiber — depends entirely on this inference. Without direct evidence of triple eigenvalue and eigenvector coalescence, the central claim is not established. The observed 3-cycle permutation for Loop-3 is a necessary but not sufficient signature of an EP3; it also arises from encircling two distinct second-order branch points. The paper even hedges by calling it an 'analogous EP3' in Section E, which signals awareness of the gap. The concrete test I propose would settle the matter by explicitly searching for a triple-root point in the complex effective-index landscape. Since the reader already assigned a CONDITIONAL verdict conditioned on verification of third-order coalescence, my independent analysis does not change that verdict. The nonchiral dynamics and the persistence near but not enclosing EPs are interesting results that survive regardless, but they should be framed as multiple-EP2 phenomena unless the EP3 test passes.","tokens_in":15770,"tokens_out":4252,"duration_ms":42593,"concrete_test":"Re-run the FEM mode solver on a dense grid in the (γ,τ)-plane over the rectangle [0,5×10^-3]×[0.6,2.2] with spacing Δγ=2×10^-5 and Δτ=2×10^-3, and extract the three complex effective-index values neff_i(γ,τ). Compute the discriminant Δ=∏_{i<j}(neff_i-neff_j)^2. An EP3 exists if and only if there is a parameter point with Δ=0 and all three pairwise differences vanish simultaneously (a triple root). If the minimum of max(|neff_0-neff_1|, |neff_0-neff_2|, |neff_1-neff_2|) on the grid is nonzero, and the two EP2s remain isolated, then the EP3 label is falsified. Additionally, compute the determinant of the matrix whose columns are the three eigenvectors; only at a true EP3 does this determinant vanish while all three eigenvectors become linearly dependent at one point.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's title, abstract, and conclusions all rest on the claim that the designed fiber hosts an EP3 formed by two interconnected EP2s. The evidence in Section II B, however, consists solely of two independent EP2s: EP2(0,1) at (γ,τ)=(1.1×10^-3,1) between Ψ0 and Ψ1, and EP2(0,2) at (3.6×10^-3,2.008) between Ψ0 and Ψ2 (Fig. 2). Both involve Ψ0, but they occur at different parameter points. An EP3 is a single point in parameter space where three eigenvalues and their corresponding eigenvectors coalesce, giving a third-order branch point. The paper never shows such a point: no Fig. 2 panel, no grid scan, and no eigenvector analysis demonstrates triple coalescence. The 3-cycle permutation observed for Loop-3 (Fig. 3(d)) is exactly what one expects when a loop encloses two individual branch points; it does not imply a third-order branch point at a single location. In fact, for a generic two-parameter family of 3×3 non-Hermitian matrices, an EP3 is a codimension-4 phenomenon and does not follow from the presence of two EP2s, even if they share one eigenvector. The phrase 'analogous EP3' used later in Section E is more honest, but the abstract and conclusions state a real EP3. Thus the central novelty — 'dynamically encircled higher-order exceptional points in an optical fiber' — is not supported by the presented data. The results on nonchiral mode conversion near but not enclosing the EPs (Loop-5) may still be valid, but they concern multiple EP2s, not an EP3.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a triple-core optical fiber with a tailored gain-loss profile and uses FEM/BPM simulations to identify two exceptional points in the (γ, τ) plane: EP2(0,1) at (1.1×10^-3, 1) between Ψ0 and Ψ1, and EP2(0,2) at (3.6×10^-3, 2.008) between Ψ0 and Ψ2. It then studies adiabatic eigenvalue permutations and dynamical beam propagation along several loops: Loop-1 and Loop-4 around EP2(0,1), Loop-2 around EP2(0,2), Loop-3 around both, and Loop-5 near both but enclosing neither. The main claims are chiral asymmetric mode conversion for loops enclosing an EP2, nonchiral conversion to Ψ0 for Loop-2, Loop-3, and Loop-5, and persistence of EP-induced dynamics without enclosing the EPs. The interpretation of the dynamical results is built on the sign of a relative-gain factor ΔΓ through Eq. (4).","tokens_in":16163,"tokens_out":6885,"duration_ms":64851,"significance":"If the EP3 claim were directly supported, the paper would offer a simple all-fiber platform for higher-order exceptional-point dynamics and mode conversion with only two tunable parameters. The manuscript's strengths include explicit fiber parameters, systematic FEM/BPM simulations, several loop geometries with concrete parameter values, and falsifiable beam-propagation predictions. The Loop-5 result—nonchiral mode conversion without encirclement—is potentially interesting independently of the EP3 label. However, the central EP3 identification currently rests on two separate EP2s rather than on a demonstrated triple coalescence, and Eq. (4) is asserted rather than derived or independently validated. These two issues are load-bearing for the paper's main novelty, so the manuscript needs substantial rework or re-scoping.","major_comments":[{"comment":"The paper identifies EP2(0,1) at (γ, τ) = (1.1×10^-3, 1) and EP2(0,2) at (3.6×10^-3, 2.008), both involving Ψ0, and then concludes that 'such an interaction scheme indicates the emergence of an EP3.' This inference is not demonstrated. An EP3 is a single parameter point at which three eigenvalues and their eigenvectors coalesce, and the manuscript provides no eigenvalue scan, no grid search, and no eigenvector-overlap analysis showing such a point. In a generic 3×3 non-Hermitian Hamiltonian, an EP3 is a codimension-4 phenomenon in real parameter space and does not follow from two EP2s at disjoint parameter points. The 3-cycle permutation in Fig. 3(d) is exactly the monodromy expected when a loop encloses two separate branch points (a composition of two pairwise transpositions), so it cannot by itself certify a third-order branch point. The authors should either (i) provide direct evidence of triple coalescence at a single point—for example, show that the cubic eigenvalue equation reduces to (λ − λ_EP)^3 and that all three eigenvectors become parallel—or (ii) consistently re-label the object as an 'analogous EP3' or 'EP3-type topological behavior' throughout the title, abstract, and conclusions. As written, the abstract's claim of a 'third-order exceptional point (EP3), formed by two interconnected second-order exceptional points' is not supported by the presented data.","section":"Section II B, Fig. 2; also Abstract and Conclusions"},{"comment":"The nonadiabatic-correction relation N_{m→n} ∝ −exp[∫ ΔΓ_{m,n} dz] is stated without derivation and is then used in Sections II D and II E to explain every output channel (for example, 'ΔΓ_{0,1} < 0 ... allows the mode conversions {Ψ0, Ψ1} → Ψ0'). Because the coefficient is given only up to an unspecified proportionality and the sign convention for ΔΓ is defined verbally rather than derived, the explanations are post hoc. The principal dynamical predictions come from BPM simulations and are not independently checked against an equation-based model. Please derive Eq. (4) from the coupled-mode equations in the instantaneous eigenbasis, or alternatively treat it as a phenomenological summary and validate it quantitatively against at least one BPM case by predicting output mode fractions for, say, Loop-1 and comparing with Fig. 4(c). Without such support, the claimed mechanism is not load-bearing.","section":"Eq. (4), Section II C"},{"comment":"The persistence claim—that asymmetric mode conversion survives near, but not enclosing, the EPs—is supported only by the single Loop-5 case. The statement that below r = 0.2 'asymmetric conversions are no longer observed' is asserted without showing the r-dependence. Since the practical-feasibility conclusion (minimized accumulated gain and gain-loss contrast) depends on this threshold, please provide the output mode fractions or at least the ΔΓ sign pattern as a function of r (or of γ0 or τ0) between r = 0.2 and r = 0.9, and specify the numerical criterion used to define 'no asymmetric conversion.'","section":"Section II E, Loop-5 threshold"}],"minor_comments":[{"comment":"The sentence says the two outer cores are 'denoted as nL for the left core and nL for the right core'; the second symbol should be nR.","section":"Section II A, after Eq. (1)"},{"comment":"Section II B refers to 'neff-values associated with Ψ_j (j = 0, 2, 3)', which should be j = 0, 1, 2; Section II D similarly refers to 'Ψ3 remains unaffected' where Ψ2 is meant.","section":"Section II B and Section II D"},{"comment":"The caption lists '(c) along Loop-2' and then '(c) along Loop-3'; the second label should be (d), matching the in-text reference to Fig. 3(d).","section":"Fig. 3 caption"},{"comment":"Because the caption states that results are shown for only one of the three loops, please state explicitly in the caption which loop is displayed (Loop-2, Loop-3, or Loop-5) or show the three cases as separate panels.","section":"Fig. 5 caption"},{"comment":"The abstract says 'we report a triple-core specialty optical fiber,' but the work is a design and simulation study; consider wording such as 'we design and simulate' to avoid implying an experimental demonstration.","section":"Abstract and Introduction"},{"comment":"The phrase 'indices {m,n} signify the all possible transitions among Ψ_j' is grammatically incomplete, and the subscript in '∆Γ_..._n' should consistently be written as ∆Γ_{m,n}.","section":"Equation (4) notation"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is the overstatement of the EP3 claim. The BPM results themselves appear internally consistent and the Loop-5 non-encirclement result is the most interesting and citable contribution. If the authors can either demonstrate triple coalescence directly or re-scope the title, abstract, and conclusions to 'EP3-type behavior from two interconnected EP2s,' the manuscript could be publishable as a simulation study of mode conversion near multiple EP2s. The treatment of Eq. (4) also needs to be strengthened or explicitly demoted to a phenomenological interpretation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis is worth reading but not for the reason the authors advertise. The genuinely interesting result is that in this triple-core fiber, direction-independent mode conversion persists for parametric loops that pass near the two exceptional points without enclosing them, and with less gain-loss contrast. That observation, plus the simple two-parameter gain-loss scheme, is a real addition to the specialty-fiber photonics literature. The simulations are internally consistent, and the mode-conversion patterns in Figs. 4 and 5 match the sign-based nonadiabatic argument at a qualitative level.\n\nThe soft spot is the \"EP3\" label, and it is load-bearing. The paper shows two second-order EPs, EP2(0,1) at (1.1e-3,1) and EP2(0,2) at (3.6e-3,2.008), both involving mode 0 but at distinct points. It never shows three eigenvalues and eigenvectors coalescing at a single parameter point. An EP3 is a single third-order branch point; two EP2s sharing a common mode at separate locations do not constitute one. The authors seem half-aware of this: Section II E uses \"analogous EP3\" while the abstract and conclusions state an EP3. The 3-cycle permutation in Fig. 3(d) is what you'd expect for a loop that encloses two square-root branch points, so it doesn't rescue the claim. If the title and abstract say \"higher-order exceptional points,\" that is an overreach given the presented evidence.\n\nSecondary issues: Eq. (4) is asserted without derivation and is doing a lot of interpretive work. No code, data, or experiment is provided. These would matter less if the central claim were solid, but they compound the need for a rewrite.\n\nMy recommendation: send it to peer review, but with a clear message to the authors that the EP3 framing must be either proven (e.g., by showing an actual triple coalescence, which may require an extended parameter space) or removed. As a study of multiple interconnected EP2s and near-loop dynamics, it deserves serious attention. A good referee could turn this into a solid paper.\n\nWould I cite it? Probably not in the next year, but I'd bring it to the reading group. The overclaim is a useful case study.","headline":"A solid fiber-design simulation, but the EP3 claim is unsupported; the near-loop nonchiral result is the real news.","tokens_in":16731,"tokens_out":3539,"would_cite":false,"duration_ms":35226,"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":"This paper claims that a purpose-built triple-core optical fiber with spatially varying gain and loss can host two interconnected second-order exceptional points that together behave as a third-order exceptional point.","keywords":["exceptional point","third-order exceptional point","non-Hermitian photonics","mode conversion","optical fiber","gain-loss engineering","topological encirclement","beam propagation method"],"falsifier":"Compute the three complex effective indices of the fiber on a fine grid of $(\\gamma,\\tau)$ around the two claimed coalescence points and check whether all three eigenvalues converge to a single point, with the characteristic polynomial's discriminant vanishing to third order, at some parameter value; if the two EP2s remain distinct branch points that never meet, the EP3 label and the third-order topological interpretation are falsified even if the simulated mode-conversion outputs still match.","tokens_in":15541,"feed_emoji":"🔦","tokens_out":9412,"duration_ms":77519,"temperature":0.7,"pith_summary":"This paper claims that a purpose-built triple-core optical fiber with spatially varying gain and loss can host two interconnected second-order exceptional points (EP2s) that together behave as a third-order exceptional point (EP3). It argues that a single propagation pass through the fiber, during which the gain-loss parameters trace a loop around or near these exceptional points, produces asymmetric mode conversion: light entering in any of the three supported modes exits in a dominant mode that depends on propagation direction (chiral) or is the same in both directions (nonchiral). The authors further claim that the same conversion dynamics persist for parametric loops that pass close to the exceptional points without enclosing them, which would relax fabrication constraints. If the paper is right, this provides a fiber-based platform for higher-order mode converters, isolators, and circulators that exploit exceptional-point topology without complex parameterization.","feed_headline":"Triple-core fiber flips light modes in a single pass","feed_subtitle":"Gain-loss loops that only approach, not enclose, the exceptional points still convert all three guided modes.","key_machinery":"The load-bearing objects are the two second-order exceptional points EP2(0,1) and EP2(0,2) located at $(\\gamma,\\tau) = (1.1\\times10^{-3},1)$ and $(3.6\\times10^{-3},2.008)$ in the gain-loss parameter plane, both involving the fundamental mode $\\Psi_0$; their interconnection is the claimed genesis of an EP3. The dynamical mechanism is the mapping of a parametric loop in the $(\\gamma,\\tau)$ plane onto the propagation direction $z$ through $\\gamma(z) = \\gamma_0\\sin(\\pi z/L)$ and $\\tau(z) = \\tau_0 + r\\sin(2\\pi z/L)$, so that one full pass through the fiber equals one full encirclement. Nonadiabatic corrections are estimated from the relative gain factors $\\Delta\\Gamma_{m,n}$, defined by the difference between each mode's averaged imaginary effective index $\\Gamma_{\\mathrm{av}}$ over the loop; the signs of these factors decide which mode wins at the output and whether the conversion is chiral or nonchiral.","core_discovery":"The central discovery is that a three-core fiber segment, operating at 1.55 µm with core index 1.46 and cladding 1.45, supports three quasi-guided modes whose effective indices coalesce in pairs at two exceptional points, EP2(0,1) and EP2(0,2), when the gain-loss coefficient $\\gamma$ and loss-to-gain ratio $\\tau$ are tuned. Because both EP2s involve the fundamental mode $\\Psi_0$, the authors argue they are interconnected and together form an EP3 whose third-order branch-point topology is revealed by quasistatic encirclement along Loop-3, which permutes effective indices as $\\Psi_0 \\to \\Psi_2 \\to \\Psi_1 \\to \\Psi_0$ (clockwise) or $\\Psi_0 \\to \\Psi_1 \\to \\Psi_2 \\to \\Psi_0$ (counterclockwise). Dynamically, by mapping the loop onto the propagation axis, a single pass along the fiber realizes the encirclement; simulations show chiral conversion for loops enclosing only EP2(0,1), and nonchiral conversion for loops enclosing EP2(0,2), both EP2s, or passing near them without enclosing. The paper claims that whether the dynamics are chiral or nonchiral is governed by the relative gain factors $\\Delta\\Gamma_{m,n}$, which determine which nonadiabatic transitions dominate, and that the EP-induced conversion persists for a smaller loop (Loop-5) that does not enclose either EP2, as long as the relative-gain relations mirror those of the encircling loop.","pith_inferences":["The paper does not directly show three eigenvalues and eigenvectors coalescing at a single parameter point; a direct calculation of the full three-mode effective Hamiltonian's discriminant would settle whether the two EP2s genuinely form an EP3 or remain two independent branch points that the Loop-3 output merely traverses in sequence.","If the persistence claim generalizes, then 'encirclement' could be replaced by 'approach' in other non-Hermitian systems, letting engineers place the loop wherever the relative-gain sign pattern matches the encircling case, which would shorten devices and reduce absorbed power.","The relative-gain criteria $\\Delta\\Gamma_{m,n}$ are presented as an interpretive rule rather than derived from the mode-coupling equations, so a natural extension is to derive these sign rules from the fiber's coupling coefficients and test them against beam-propagation outputs across a wider family of loops.","Simulating the same fiber geometry with a different loop shape (for example a circle in the $(\\gamma,\\tau)$ plane) would test whether the nonchiral outcome is generic or an artifact of the particular sinusoidal mapping in Eq. (3)."],"forward_implications":["A single 35 mm propagation pass through the fiber realizes one full loop in parameter space, with forward propagation corresponding to clockwise encirclement and backward propagation to counterclockwise encirclement.","When the loop encloses only EP2(0,1), the two modes connected by that point are converted into one dominant mode whose identity depends on the propagation direction, while the third mode remains unchanged; enlarging the loop (Loop-4) draws the third mode into the conversion, making the dynamics fully chiral.","Loops that enclose EP2(0,2), both EP2s (Loop-3), or merely pass near both without enclosing them (Loop-5) all yield nonchiral conversion, with every input mode ending in $\\Psi_0$ regardless of direction.","The nonchiral conversion persists as the loop shrinks from $r=0.9$ down to $r=0.2$, demonstrating that EP-induced mode conversion does not require the loop to physically enclose the exceptional points.","Because only two tunable parameters ($\\gamma$ and $\\tau$) and a single fiber segment are needed, the scheme avoids the complex multi-parameter geometries previously used for higher-order exceptional-point devices."],"supporting_citations":[{"why":"Supplies the topological analogy that two interconnected EP2s form an EP3, which the paper relies on to label the third-order branch point.","marker":"[48]"},{"why":"Analyzes multiple exceptional points related to three interacting eigenmodes, supporting the interconnected-EP2 picture used for Loop-3.","marker":"[49]"},{"why":"Provides the time-asymmetric state-exchange mechanism that motivates the nonadiabatic correction formula in Eq. (4).","marker":"[34]"},{"why":"Claims chiral state conversion without encircling an exceptional point, the precedent that Loop-5 is designed to test.","marker":"[38]"},{"why":"Reports experimental observation of chiral state transfer without encircling an exceptional point, strengthening the basis for the persistence claim.","marker":"[39]"},{"why":"Prior demonstration of EP2-induced asymmetric mode conversion in a dual-core fiber, the geometry this triple-core design extends.","marker":"[36]"},{"why":"Experimental demonstration of dynamically encircling an exceptional point for asymmetric mode switching, the reference experiment for the chiral behavior.","marker":"[37]"}],"fun_headline_variants":["Triple-core fiber exploits EP3 for single-pass mode flips","Single fiber pass encircles EP3 and converts all three modes","Exceptional-point loops flip modes even without full encirclement","Loss-gain tuning in fiber drives higher-order exceptional dynamics","Specialty fiber demonstrates chiral mode conversion via EP3"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that two second-order exceptional points that share one mode necessarily merge into a single third-order exceptional point when both are encircled; the paper infers this interconnection but never shows all three modes collapsing at one parameter value.","fun_headline_variants_meta":{"raw":{"variants":["Triple-core fiber exploits EP3 for single-pass mode flips","Single fiber pass encircles EP3 and converts all three modes","Exceptional-point loops flip modes even without full encirclement","Loss-gain tuning in fiber drives higher-order exceptional dynamics","Specialty fiber demonstrates chiral mode conversion via EP3"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000678,"raw_usage":{"total_tokens":3162,"prompt_tokens":1103,"completion_tokens":2059,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":719,"completion_tokens_details":{"reasoning_tokens":1985}},"tokens_in":719,"tokens_out":2059,"duration_ms":18765,"temperature":1.0,"reasoning_tokens":1985,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:46:17.554683+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the three complex effective indices of the fiber on a fine grid of $(\\gamma,\\tau)$ around the two claimed coalescence points and check whether all three eigenvalues converge to a single point, with the characteristic polynomial's discriminant vanishing to third order, at some parameter value; if the two EP2s remain distinct branch points that never meet, the EP3 label and the third-order topological interpretation are falsified even if the simulated mode-conversion outputs still match.","supporting_citations":[{"cited_title":"Realizing exceptional points of any order in the presence of symmetry,","cited_arxiv_id":null,"evidence_quote":"Supplies the topological analogy that two interconnected EP2s form an EP3, which the paper relies on to label the third-order branch point."},{"cited_title":"Dynamically encircling exceptional points in a three-mode waveguide system,","cited_arxiv_id":null,"evidence_quote":"Analyzes multiple exceptional points related to three interacting eigenmodes, supporting the interconnected-EP2 picture used for Loop-3."},{"cited_title":"Time- asymmetric quantum-state-exchange mechanism,","cited_arxiv_id":null,"evidence_quote":"Provides the time-asymmetric state-exchange mechanism that motivates the nonadiabatic correction formula in Eq. (4)."},{"cited_title":"Enhancement of quantum heat engine by encircling a liouvillian excep- tional point,","cited_arxiv_id":null,"evidence_quote":"Claims chiral state conversion without encircling an exceptional point, the precedent that Loop-5 is designed to test."},{"cited_title":"Experimen- tal observation of the topological structure of exceptiona l points,","cited_arxiv_id":null,"evidence_quote":"Reports experimental observation of chiral state transfer without encircling an exceptional point, strengthening the basis for the persistence claim."},{"cited_title":"Furthermore, recent reports have ques- tioned whether it is essential to encircle an EP2 within a parametric loop to achieve asymmetric light dynamics [ 38, 39]","cited_arxiv_id":null,"evidence_quote":"Prior demonstration of EP2-induced asymmetric mode conversion in a dual-core fiber, the geometry this triple-core design extends."},{"cited_title":"Quantum state tomography across the exceptional point in a single dissipative qubit,","cited_arxiv_id":null,"evidence_quote":"Experimental demonstration of dynamically encircling an exceptional point for asymmetric mode switching, the reference experiment for the chiral behavior."}],"review_version":1}