{"id":"cde1361a-90fb-48e9-9bfd-da034b721800","arxiv_id":"2505.04932","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"The spectrum of a non-Hermitian monoclinic crystal is claimed to live on a real projective plane momentum space, supporting order-1 and order-4 exceptional manifolds, hybrid EMs, diabolic points, and bound states in the continuum.","lead":"This paper proposes that the momentum space of certain non-Hermitian low-symmetry crystals is shaped like the real projective plane, a non-orientable surface. The authors use this surface to catalogue how exceptional points merge into higher-order singularities as a material parameter changes.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"RP^2 identification of the k_z=-1 slice is unproven: the natural torus quotient by inversion is an orientable sphere-orbifold with Euler characteristic 2, not RP^2.","rationale":"The central claim is that the momentum space of low-symmetry media can be non-orientable, isomorphic to RP^2, and that this base space drives the exceptional-manifold phenomenology. The only support for the RP^2 identification is the visual Morin-surface argument in Fig. 1F. The mathematical facts cut the other way: for a compactified BZ slice, which is what the paper uses via box quantization, the inversion symmetry has four fixed points, and the quotient is an orientable sphere-orbifold, not RP^2. The paper does not prove that the physical momentum space is the quotient by inversion, nor that it should be compactified projectively; the symmetry Ω(k)=Ω(-k) only states that the spectrum is invariant, not that k and -k are the same point. Therefore the foundational claim is unsupported and, under the paper's own setup, likely false. The downstream predictions (order-1 EM, order-4 EM, charge conversion at a Möbius-like boundary, BIC strings, diabolic points) are computed from the Hamiltonian and may still describe the spectra on the torus, but their topological interpretation as phenomena on RP^2 is in question. The reader's weakest assumption identifies exactly this issue, and I agree. The proposed Euler-characteristic/orientability computation would settle whether the RP^2 claim is tenable.","tokens_in":11288,"tokens_out":12976,"duration_ms":135068,"concrete_test":"Construct the cell decomposition of the k_z=-1 momentum slice with the identifications used in Fig. 1F: start from the square [-1,1]^2, apply periodic boundary identifications, then impose inversion (kx,ky)~(-kx,-ky). Compute the Euler characteristic and orientability of the resulting quotient via the Riemann-Hurwitz formula for the four fixed points. Also repeat for the boundary-identification pattern in Fig. 1F without the torus identifications. If either computation yields χ=2 and an orientable surface, the identification with RP^2 fails. As a numerical cross-check, parallel transport an eigenvector frame around a noncontractible loop on the claimed RP^2 base and check whether the frame returns consistently.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Under the model in Eq. (2), the k_z=-1 slice with box quantization is a torus T^2 via periodic boundary identifications. The inversion symmetry Ω(kx,ky)=Ω(-kx,-ky) defines a Z2 action on this torus with four fixed points (kx,ky)=(0,0),(1,0),(0,1),(1,1) modulo the lattice. Riemann-Hurwitz gives χ(T^2/Z2)=2, and since the action is orientation-preserving, the quotient is an orientable sphere with four cone points, not RP^2, which has χ=1 and is non-orientable. The paper's Morin-surface construction (stretching, twisting, bending, gluing boundaries) is a visualization of RP^2 as an immersed surface, but it is not derived from the BZ topology or from the inversion symmetry of the eigenvalues; it silently replaces the periodic boundary identifications with different ones. The subsequent claims — order-1 EMs from merging identical charges, charge conversion across a Möbius-like boundary, and non-orientable band fluidity — are predicated on this identification. If the correct base space is instead a sphere-orbifold or simply R^2 without compactification, those conclusions do not transfer.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a non-Hermitian Maxwell-type model for low-symmetry (monoclinic) media with permittivity diag(a,b,c), permeability identity, and an off-diagonal imaginary magnetoelectric parameter d. The authors derive explicit eigenvalues in Eq. (2) and explore exceptional manifolds (EMs) as a single parameter b varies. They claim that on the k_z = -1 momentum slice, the inversion symmetry Omega(kx,ky) = Omega(-kx,-ky) makes the base space a non-orientable real projective plane RP^2, visualized through a Morin-surface gluing construction. On this base space, they report order-1 and order-4 exceptional manifolds, hybrid and anisotropic EMs, diabolic points, and bound states in the continuum, and they propose an 'expanded dihedral group' D4 to describe local band braiding. The paper combines analytic eigenvalue expressions with numerical trajectory tracking on the surface of a cube domain in momentum space.","tokens_in":11591,"tokens_out":7043,"duration_ms":72793,"significance":"If the central RP^2 base-space claim is correct, this would be a genuinely new contribution: non-Hermitian exceptional physics on a non-orientable momentum space, with concrete predictions for order-1 and order-4 EMs, charge conversion across non-orientable boundaries, and a two-band non-Abelian braiding group. The paper's explicit eigenvalue formula (Eq. 2) and the detailed parameter-dependent trajectory analysis are strengths, as is the connection to experimentally relevant concepts such as bulk Fermi arcs, BICs, C points, and V points. However, the topological identification of the momentum slice as RP^2 is not rigorously established and is in tension with the standard topology of the Brillouin zone; this issue is load-bearing for most of the paper's conceptual claims.","major_comments":[{"comment":"The central topological claim that the k_z = -1 momentum slice is isomorphic to RP^2 is not established and appears inconsistent with the standard Brillouin-zone topology. With periodic boundary conditions, the slice is a torus T^2. The inversion symmetry Omega(kx,ky) = Omega(-kx,-ky) is a symmetry of the eigenvalues, not a definition of an equivalence relation on the momentum space. If one does form the quotient of T^2 by this Z2 action, there are four fixed points (kx,ky) = (0,0), (1,0), (0,1), (1,1) modulo the lattice, and Riemann-Hurwitz gives an orientable sphere with four cone points (Euler characteristic 2), not RP^2, which is non-orientable with Euler characteristic 1. The Morin-surface construction in Fig. 1F (stretching, twisting, bending, gluing) describes an immersion of RP^2 in three dimensions, but it is not derived from the BZ topology or from Eq. (2); it replaces the periodic boundary identifications with a different gluing. Since the subsequent claims of order-1 EMs from merging identical charges, charge conversion across a Möbius-like boundary, and non-orientable band fluidity all depend on this identification, the authors must either provide a rigorous quotient-topology proof with an explicit equivalence relation on the BZ or substantially revise the topological claim.","section":"Results, Eq. (3) and the expanded dihedral group D4"},{"comment":"The algebraic description of local band fluidity needs a precise and consistent formulation. The set {c_m} in Eq. (3) is explicitly said to lack inverse elements and closure, so it is not a group; nonetheless, the text calls D4 an 'expanded dihedral group' and uses non-Abelian group language. In addition, the transfer matrix T is stated to be multivalued near point P and invalid near point O, so the statement that T 'belongs to D4' needs a well-defined domain of validity. Please define the quasi-group operations rigorously, state whether the c_m are matrix representations of an abstract algebra or merely a diagrammatic bookkeeping device, and specify the parameter and momentum ranges over which the non-Abelian property is actually proven.","section":"Results, Fig. 4E and the order-4 EM"},{"comment":"The order-4 EM at b = 1 with quarter dispersion is a headline result of the abstract ('higher-order exceptional manifolds'), but the main text only states the dispersion exponents and reports a polarization vorticity of 0.25. The derivation of the exponents epsilon_Re and epsilon_Im in the asymptotic expansion Omega_i(k_Z + delta k * k/|k|) - Omega_i(k_Z) proportional to delta k^{epsilon_Re} + i delta k^{epsilon_Im} is deferred to Supplementary section 3.3. To make the order-4 (and order-1) classification independently verifiable, the authors should include the explicit asymptotic expansion in the main text or ensure the supplementary derivation is self-contained and clearly cross-referenced.","section":"Results, box quantization and the cube domain (Fig. 2A)"},{"comment":"The reduction of the full three-dimensional evolution to trajectories on the surface of the cube k in [-1,1]^3 relies on the self-similarity H(eta k,b) = eta H(k,b) and on EMs forming straight lines through Gamma. This is plausible, but the text does not prove that every EM trajectory intersects the cube surface exactly once or that no interior degeneracies are missed. Moreover, the paper switches between the k_z = -1 slice (Fig. 1F) and the top surface k_z = 1 (Figs. 2A and 3) without explaining how these surfaces are related by the self-similarity. Please justify the box-quantization reduction and clarify the relation between the two surfaces, since the RP^2 claim refers to k_z = -1 while the trajectory analysis uses k_z = 1.","section":"Results, Fig. 2A"}],"minor_comments":[{"comment":"The text surrounding Eq. (2) contains typographical errors: 'alpha_± = ab ± ac − d^2, , beta_± = ab ± bc − d^2 ,and, gamma_±' has stray commas and spacing; please clean up the notation.","section":null},{"comment":"The caption states 'boundaries with the same color should be identified along the marked direction' but does not specify the gluing word; please provide the explicit identifications on the fundamental polygon so that the claimed RP^2 topology can be checked.","section":null},{"comment":"The term 'box quantization' is cited to Ref. [30], which concerns optical cross sections and appears unrelated to the mathematical box quantization or the periodic boundary conditions used here; please cite a standard reference for the procedure or rename it.","section":null},{"comment":"In Fig. 2A, nonzero EP trajectories are plotted outside the cube surface and zero EP trajectories inside it, which makes the figure difficult to interpret; consider separate panels or a transparent 3D view.","section":null},{"comment":"The sentence 'zero EMs forms an anisotropic or hybrid EM, while nonzero EMs forms a diabolic or Dirac point' conflates bound states in the continuum with exceptional manifolds; BICs are states, not degeneracies, and the sentence should be reworded for clarity.","section":null}],"recommendation":"major_revision","confidential_remarks":"The reader's conditional verdict is appropriate: the eigenvalue analysis and trajectory computations appear internally consistent, but the RP^2 identification is the paper's central novelty and is not supported by a rigorous quotient-topology argument. In fact, the natural quotient of the k_z = -1 torus by inversion has Euler characteristic 2 and is orientable, so as stated the central claim is likely wrong. This is a load-bearing issue, not a presentation issue, and it is not clear that a revised manuscript within the current scope can fix it without changing the main message. I would advise the editor that acceptance in the present form is not advisable, but a major revision that either supplies a rigorous construction of the RP^2 base space or reframes the paper around the mathematically valid base space (and adjusts the topological conclusions accordingly) could be considered."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know about arXiv:2505.04932.\n\nFirst, the headline claim — that a momentum slice in low-symmetry media forms a non-orientable RP^2 base space — is not established, and the natural topology of the Brillouin zone contradicts it. The picture with a Morin surface is a visualization, not a derivation. Second, underneath that claim is a concrete, explicit model of exceptional-manifold evolution that has real content and could be useful to people studying non-Hermitian photonics.\n\nWhat is actually new: the eigenvalue expression in Eq. (2) is written out cleanly, and the evolution of EMs as parameter b varies is tracked through a systematic series of collisions: inversion partners merging into an order-1 EM, zero and nonzero EPs colliding to form an order-4 EM, mirror partners producing hybrid/anisotropic EMs, BICs, and a diabolic point. The polarization-vorticity analysis adds an eigenvector-level check, which is more than most such papers do. The expanded dihedral group D4 is a reasonable way to encode the braid-like transitions of the four complex eigenvalues, though it is assembled after the fact, not predicted.\n\nThe soft spot is the load-bearing RP^2 identification. The stress-test note is correct: if you take the k_z=-1 slice as a torus (which the paper's own BZ discussion implies) and quotient by the inversion symmetry Ω(k)=Ω(-k), the quotient is an orientable sphere with four cone points — Euler characteristic 2 — not RP^2, which has χ=1 and is non-orientable. The Morin-surface construction in Fig. 1F replaces the periodic boundary identifications with different ones without justification. The paper's claims about identical charges merging, Möbius-like boundaries, and charge conversion all rely on that non-standard base space. If the correct base is a sphere-orbifold (or R^2 without compactification), those conclusions do not transfer.\n\nTwo further issues. First, no code or data is provided beyond 'available upon reasonable request'; for a paper with many detailed numerical trajectories, that is weak. Second, the 'quasi-group' structure behind D4 is not a group — it lacks closure and inverses — and the paper only restores a true dihedral group after imposing extra normalization. This isn't fatal to the braiding description, but it should be stated more carefully.\n\nWho gets value from this? Researchers working on exceptional-point dynamics may find the EM collision catalog useful, and the explicit model allows them to test the claims. But the RP^2 framing misleads more than it helps. I would not cite it in its current form. I would, however, send it to a serious referee: an expert can adjudicate the topology quickly, and the EM-evolution results are worth saving. If the authors can provide a rigorous quotient-topology derivation — or drop the RP^2 claim and present this as a model study — there is a publishable core here. As written, the central topological assertion fails the stress test.","headline":"The RP^2 claim is unproven and likely wrong, but the underlying EM-collision study has enough merit to warrant expert review if the topology is fixed.","tokens_in":12107,"tokens_out":8679,"would_cite":false,"duration_ms":82994,"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":"Low-symmetry non-Hermitian media can host a momentum space isomorphic to the real projective plane, and on that non-orientable base space exceptional manifolds of order 1 and 4, hybrid and anisotropic exceptional manifolds, diabolic…","keywords":["non-Hermitian optics","exceptional points","real projective plane","non-orientable momentum space","hyperbolic media","bound states in the continuum","dihedral group","topological charges"],"falsifier":"One concrete check is to compute the Euler characteristic and orientability of the $k_z=-1$ base space: the real projective plane has $\\chi=1$ and is non-orientable, whereas the quotient of a torus by the antipodal map has $\\chi=2$, four orbifold points, and is orientable. If a direct calculation of the identified boundaries in Fig. 1F reproduces the sphere's $\\chi=2$, the central $\\mathbb{R}\\mathbb{P}^2$ claim fails; if it reproduces $\\chi=1$, the order-1 and order-4 merging scenarios are topologically grounded.","tokens_in":2303,"feed_emoji":"🌀","tokens_out":4732,"duration_ms":107148,"temperature":0.7,"pith_summary":"Low-symmetry non-Hermitian crystals are usually described by a toroidal momentum space; this paper argues that in monoclinic hyperbolic media the relevant momentum slice is instead a non-orientable real projective plane. The argument follows an explicit Maxwell-equation model whose six eigenvalues are known in closed form, and it tracks the exceptional degeneracies as a single material parameter $b$ varies. On the $\\mathbb{R}\\mathbb{P}^2$-like surface, exceptional points with identical topological charges merge into an order-1 exceptional manifold; on mirror-symmetric faces, opposite charges can annihilate into bound states in the continuum, while identical charges produce an order-4 exceptional manifold, and nonzero charges can form diabolic or Dirac points. A sympathetic reader would care because the same simple medium then exhibits nearly the full catalogue of non-Hermitian degeneracies, and because the local band exchange is shown to obey a non-Abelian dihedral group with only two bands.","feed_headline":"Momentum space becomes a real projective plane","feed_subtitle":"In low-symmetry crystals, one parameter yields order-1, hybrid and bound-state exceptional degeneracies.","key_machinery":"The two load-bearing objects are the non-orientable base space and the dihedral transfer group. The base space is the real projective plane $\\mathbb{R}\\mathbb{P}^2$, realized as the $k_z=-1$ momentum slice of a monoclinic medium with inversion symmetry and represented as a Morin surface, an immersed $\\mathbb{R}\\mathbb{P}^2$ with self-intersections. Its non-orientability is what lets two exceptional points with identical topological charge merge instead of annihilating. The second object is the expanded dihedral group $\\mathbb{D}_4=\\{C_m\\}=\\{c_m\\}\\otimes\\{\\pm1,\\pm i\\}$, whose eight matrices $c_m$ together with global rotations describe how eigenvalue pairs swap orbits and rotate among four energy levels when the parameter $b$ runs from $+\\infty$ to $-\\infty$; the group is non-Abelian and is argued to give two-band non-Abelian band braiding. The explicit eigenvalues in Eq. (2) and the box-quantization reduction to a cube $[0,1]^3$ are the calculational scaffolding that turns these objects into observable trajectories.","core_discovery":"The paper's central claim is that, for low-symmetry media with inversion symmetry $\\Omega(k_x,k_y)=\\Omega(-k_x,-k_y)$, the momentum slice $k_z=-1$ is isomorphic to the real projective plane $\\mathbb{R}\\mathbb{P}^2$ rather than the usual torus; geometrically this is presented as the Morin surface obtained by stretching, twisting, bending, and gluing a spiral cylinder. On this non-orientable base space the paper discovers that exceptional manifolds evolve in a way controlled by the parameter $b$: equal-charge inversion partners collide to form an order-1 exceptional manifold with integer vorticity, a zero and a nonzero exceptional point with identical charge collide to form an order-4 exceptional manifold with quarter dispersion, and mirror partners on the torus-like faces yield a hybrid or anisotropic exceptional manifold, a diabolic or Dirac point, and a bound state in the continuum. The same evolution, viewed locally at fixed momentum, is described by an expanded dihedral group $\\mathbb{D}_4$ built from orbit interchanges and local rotations; the group is non-Abelian even though only two bands are involved. The paper also notes that conventional eigenvalue-braiding theory fails in this setting because imaginary parts jump across the bulk Fermi arc, but a squared-eigenvalue mapping restores continuity.","pith_inferences":["If the $\\mathbb{R}\\mathbb{P}^2$ identification is taken as a working hypothesis, a clean way to test it in any monoclinic platform is to measure the bulk Fermi arc on $k_z=-1$: on a non-orientable base space the arc should be one-sided, so crossing it flips the sign of $\\operatorname{Im}\\Omega$ without returning on the second side.","The paper's group $\\mathbb{D}_4$ is a quasi-group without inverse elements or closure; a future formal study could determine whether a genuine group structure exists after quotienting by the global rotation $\\{\\pm1,\\pm i\\}$, which would put two-band non-Abelian braiding on the same footing as multiband braiding groups.","One implication the authors leave implicit is that the same low-symmetry material should realize many of these degeneracies in the same sample at different frequencies or parameter biases, potentially turning a single crystal into a multifunctional device for sensing, lasing, and mode switching.","A rigorous quotient-topology treatment of the $k_z=-1$ slice would settle whether the base space is truly $\\mathbb{R}\\mathbb{P}^2$ or an orientable sphere with orbifold points; the physical distinction is whether identical-charge merging, rather than annihilation, is the generic outcome."],"forward_implications":["Sweeping $b$ in one monoclinic medium should produce, in order, creation of nonzero exceptional points, order-1 exceptional manifolds at $b=8$ and $b=0$, an order-4 exceptional manifold at $b=1$, a hybrid or anisotropic exceptional manifold and a bound state in the continuum at $b=0$, and a diabolic or Dirac point at $b=-4/7$.","Because the same eigenvalues undergo non-Abelian braiding under $\\mathbb{D}_4$ with only two bands, the paper's picture offers a route to non-Abelian processes and logic gates without requiring three or more bands.","On the $\\mathbb{R}\\mathbb{P}^2$-like surface, a closed loop around a merged exceptional manifold accumulates $\\pm\\pi$ phase, so a merged order-1 exceptional manifold carries integer topological charge inherited from two half-charge partners.","The bound states in the continuum found at the mirror-symmetric point $Y$ are not isolated points but, via the self-similarity $H(\\eta\\vec{k},b)=\\eta H(\\vec{k},b)$, extend along straight lines in momentum space.","The hybrid or anisotropic exceptional manifold at $Y$ has a $1/2$ slope in its imaginary dispersion except on $k_x=0$, where it is linear, distinguishing it from ordinary order-2 exceptional points."],"supporting_citations":[{"why":"Supplies the low-symmetry monoclinic crystal platform (beta-Ga2O3) and the constitutive form used for the non-Hermitian media.","marker":"[16]"},{"why":"Provides the classification of zero and nonzero exceptional coalescences that the paper uses to distinguish order-1, order-4, and hybrid degeneracies.","marker":"[17]"},{"why":"Defines the spectrum vorticity and winding number used to assign half-integer and integer topological charges to exceptional points.","marker":"[20]"},{"why":"Gives the Maxwell Hamiltonian formalism $H=M^{-1}R$ and the hyperbolic metamaterial band picture from which the eigenvalue problem is set up.","marker":"[14]"},{"why":"Supplies the Morin/Roman-surface immersion of the real projective plane that the paper identifies with the $k_z=-1$ momentum slice.","marker":"[19]"},{"why":"Underlies the bulk Fermi arc observation and the sign-flipping imaginary parts that mark one-sided band connectivity on the projective plane.","marker":"[18]"},{"why":"Establishes non-Abelian band topology in multiband systems, the baseline the paper contrasts with its two-band dihedral-group braiding.","marker":"[23]"},{"why":"Introduces the exceptional nexus with a hybrid topological invariant that the paper borrows for hybrid or anisotropic exceptional manifolds.","marker":"[33]"},{"why":"Provides the multiband exceptional non-Abelian braiding theory that the paper notes becomes invalid when imaginary parts jump across Fermi arcs.","marker":"[35]"}],"fun_headline_variants":["Non-Hermitian exceptional manifolds on RP^2","RP^2 momentum space hosts exceptional degeneracies","Exceptional physics on a projective momentum space","Momentum space turns real projective plane","Order-1 to bound-state exceptional manifolds on RP^2"],"cache_read_input_tokens":14208,"weakest_assumption_plain":"The argument assumes, without a full proof, that gluing antipodal points on the $k_z=-1$ momentum slice gives a real projective plane rather than a sphere-like surface with orbifold points; the paper supports this identification only with a geometric Morin-surface picture.","fun_headline_variants_meta":{"raw":{"variants":["Non-Hermitian exceptional manifolds on RP^2","RP^2 momentum space hosts exceptional degeneracies","Exceptional physics on a projective momentum space","Momentum space turns real projective plane","Order-1 to bound-state exceptional manifolds on RP^2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001228,"raw_usage":{"total_tokens":5038,"prompt_tokens":926,"completion_tokens":4112,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":542,"completion_tokens_details":{"reasoning_tokens":4037}},"tokens_in":542,"tokens_out":4112,"duration_ms":28490,"temperature":1.0,"reasoning_tokens":4037,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:18:51.917850+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"One concrete check is to compute the Euler characteristic and orientability of the $k_z=-1$ base space: the real projective plane has $\\chi=1$ and is non-orientable, whereas the quotient of a torus by the antipodal map has $\\chi=2$, four orbifold points, and is orientable. If a direct calculation of the identified boundaries in Fig. 1F reproduces the sphere's $\\chi=2$, the central $\\mathbb{R}\\mathbb{P}^2$ claim fails; if it reproduces $\\chi=1$, the order-1 and order-4 merging scenarios are topologically grounded.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the low-symmetry monoclinic crystal platform (beta-Ga2O3) and the constitutive form used for the non-Hermitian media."},{"cited_title":"C., Xiao, M., Zhang, Z","cited_arxiv_id":null,"evidence_quote":"Provides the classification of zero and nonzero exceptional coalescences that the paper uses to distinguish order-1, order-4, and hybrid degeneracies."},{"cited_title":"T., Zhen, B","cited_arxiv_id":null,"evidence_quote":"Defines the spectrum vorticity and winding number used to assign half-integer and integer topological charges to exceptional points."},{"cited_title":"P., Li, Z","cited_arxiv_id":null,"evidence_quote":"Gives the Maxwell Hamiltonian formalism $H=M^{-1}R$ and the hyperbolic metamaterial band picture from which the eigenvalue problem is set up."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Morin/Roman-surface immersion of the real projective plane that the paper identifies with the $k_z=-1$ momentum slice."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Underlies the bulk Fermi arc observation and the sign-flipping imaginary parts that mark one-sided band connectivity on the projective plane."},{"cited_title":"S., Soluyanov, A","cited_arxiv_id":null,"evidence_quote":"Establishes non-Abelian band topology in multiband systems, the baseline the paper contrasts with its two-band dihedral-group braiding."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the exceptional nexus with a hybrid topological invariant that the paper borrows for hybrid or anisotropic exceptional manifolds."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the multiband exceptional non-Abelian braiding theory that the paper notes becomes invalid when imaginary parts jump across Fermi arcs."}],"review_version":1}