{"id":"9667af3a-778b-44ee-aad9-a8aa12f28826","arxiv_id":"2411.14682","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For topological insulator surfaces with two or three Dirac cones, attractive interactions create conformal manifolds of quantum critical fixed points (a ring and two 2-spheres) in the one-loop renormalization group flow.","lead":"This paper analyzes the boundary between a gapless surface and a superconducting surface of a 3D topological insulator when electrons attract. It predicts that the quantum critical points on this boundary organize into continuous families, a ring or spheres, of scale-invariant theories rather than the usual isolated fixed points.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The one-loop conformal manifolds are not protected beyond O(epsilon); the paper's own partial two-loop calculation breaks them, and the complete two-loop beta function is not computed, so the physical-universality claim remains unestablished.","rationale":"The reader's weakest assumption correctly identifies the one-loop extrapolation as the main soft spot. My independent reading sharpens this into a concrete technical gap: the paper's own Eq. C19/IXB demonstrates that the leading field-renormalization correction breaks the ring manifold into isolated fixed points, and yet this calculation is not the full two-loop beta function because two-loop vertex diagrams are omitted. Therefore the one-loop conformal manifolds are neither protected by a symmetry of the microscopic theory nor shown to survive the complete leading beyond-one-loop correction. This is genuinely load-bearing because the central claim is about the universality of generic surface topological quantum criticality, not merely about a mathematical property of the one-loop RG. I do not recommend changing the reader's CONDITIONAL verdict: the one-loop analysis is internally consistent, the higher-loop fragility is explicitly acknowledged in the paper, and the results are framed within the epsilon expansion, so the work remains publishable as a one-loop classification. The condition is that the physical extrapolation to epsilon=1 and the complete two-loop fate of the manifolds must be clarified. The minor topological imprecision (the N=3 fixed-point set for V=epsilon e3 e3^T is actually RP^2, not S^2, because e3 and -e3 give the same matrix) does not affect the local counting of marginal operators or the stability analysis, so it is not the load-bearing issue.","tokens_in":45784,"tokens_out":9597,"duration_ms":109938,"concrete_test":"Compute the complete two-loop beta function for the N=2 and N=3 surface theory in 2+epsilon dimensions, including all two-loop four-fermion vertex diagrams in addition to the self-energy diagram of Fig. 6(c), using dimensional regularization and the MS scheme. Then solve for the fixed points of this full beta function and determine whether a continuous fixed-point manifold of dimension at least 1 (N=2) or 2 (N=3) survives at O(epsilon^2). If only isolated fixed points remain, the one-loop conformal-manifold claim is an artifact of the truncation; if a distorted manifold persists, the central claim gains support. A complementary nonperturbative check would be a lattice or functional-RG simulation of the effective Yukawa theory at epsilon=1, testing whether the phase boundary is controlled by a line/family of fixed points or by isolated points.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that generic surface topological quantum criticality is dictated by one-loop conformal manifolds (the ring for N=2 and the 2-spheres for N=3). For this to hold at physical dimension epsilon=1, the exactly marginal directions defining the manifolds must be protected, or the manifolds must survive the leading symmetry-breaking corrections. The paper itself shows they are not protected: Eq. C19 includes the two-loop field-renormalization term (eta_i+eta_j)V_ij with eta_i=(1/32) sum_k V_ik V_ki; for N=2 this reduces the emergent SO(2) to Z2 x Z2 and induces a flow around the ring, dtheta/dl = -2a V0^2 sin 2theta (Eq. 71), collapsing the ring into four isolated fixed points (Table V). At epsilon=1 the O(a epsilon^2) splittings are nonzero, and the erstwhile marginal operators acquire scaling dimensions D +/- a epsilon^2 (Appendix E), so no exactly marginal operators remain. The N=3 manifolds are subject to the same type of term. Moreover, Eq. C19 is not the complete two-loop beta function: it contains only the one-loop vertex renormalization plus the two-loop self-energy, while two-loop vertex diagrams with three interaction vertices (present in any four-fermion theory in 2+epsilon dimensions) are omitted. Thus the fate of the conformal manifolds at two loops is not actually settled by the paper. The central claim therefore rests on an unverified one-loop truncation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies attractive four-fermion interactions on the surface of a 3D topological insulator with N half-Dirac cones, focusing on N=2 and N=3. The authors derive a one-loop RG equation in 2+ε dimensions (Eq. 28) that possesses an emergent SO(N) symmetry, and they find that its interacting fixed points form conformal manifolds: a ring for N=2 and two rank-one projector manifolds (labeled \"2-spheres\") for N=3. They argue that these manifolds, together with their irrelevant directions, form the D_p−1 dimensional phase boundary separating the gapless surface from superconducting states, and that the infrared-stable manifold dictates the universality of generic surface topological quantum critical points. The paper also analyzes higher-loop symmetry-breaking effects, finding that a certain three-loop one-particle-irreducible term only distorts the ring, while a two-loop field-renormalization term breaks the ring into four isolated fixed points.","tokens_in":46053,"tokens_out":7796,"duration_ms":76302,"significance":"If the one-loop conformal-manifold picture survived at physical dimension, this would be a substantial contribution: it would provide a geometric, coset-space classification of surface topological quantum criticality, with explicit marginal operators, scaling dimensions, and an emergent single-boson Yukawa description. The derivation of the one-loop beta function in Appendix C and the fixed-point stability analysis are internally consistent, and the coset-space intuition (Section VI) is attractive and partially explanatory. However, the paper's own partial two-loop calculation destroys the central conformal manifolds, and the complete two-loop beta function is not computed. The physical-universality claim is therefore not established beyond the one-loop ϵ-expansion, and the manuscript's conclusions overstate what is proven.","major_comments":[{"comment":"This section shows that including the two-loop field-renormalization term (Eq. C19) induces a flow dθ/dl = −2a V_0^2 sin 2θ around the N=2 ring, collapsing it into four isolated fixed points (Table V), with the formerly marginal operators acquiring scaling dimensions D ± aε^2 (Appendix E). At physical dimension ε=1 these splittings are of order one, so no exactly marginal operators remain. This directly contradicts the conclusion in Section X(B) that the conformal manifolds \"dictate all the universal properties of generic surface topological critical points\"; at face value, the paper's own calculation shows the one-loop manifolds are not stable beyond O(ε). The central claim therefore needs either a complete two-loop calculation or a substantial qualification that the conformal-manifold universality is only a leading-order ϵ-expansion statement.","section":"Section IXB, Eqs. (69)–(71), Table V"},{"comment":"Eq. (C19) is not the complete two-loop beta function. It combines the one-loop vertex renormalization with the two-loop self-energy term (η_i+η_j)V_ij, but it omits two-loop vertex diagrams with three interaction vertices, which are present in any four-fermion theory in 2+ε dimensions. Consequently, the paper does not actually settle the fate of the conformal manifolds at O(ε^2): the breakdown reported in Section IXB is established only within a partial two-loop truncation, and the alternative that the complete two-loop beta function leaves a deformed manifold or restores some marginal directions remains open. Since the physical-dimension universality claim is the paper's main result, this omission is load-bearing.","section":"Appendix C, Eq. (C19)"},{"comment":"The fixed-point manifolds V_c = ε e_3⊗e_3 and V_c = ε I − ε e_3⊗e_3 are sets of rank-one projectors in R^3. Since n⊗n = (−n)⊗(−n), the manifold is the real projective plane RP^2, not the 2-sphere S^2. Correspondingly, the stabilizer of a rank-one projector in SO(3) is O(2) (embedded with determinant +1), not SO(2), so the coset is SO(3)/O(2) = RP^2 rather than SO(3)/SO(2) = S^2. The dimension and the local stability counts (2 marginal, 3 irrelevant, 1 relevant) are unchanged, but the geometry stated in Table III, Table IV, and the text following Eq. (57) is incorrect and should be corrected.","section":"Table IV, Eqs. (50)–(57), Section VIIIB"},{"comment":"The statement that a distorted co-dimension-two manifold persists for all higher-loop one-particle-irreducible effects of Type A (N_L = 4,5,6,...) is not proven; only a single three-loop term is solved explicitly. The phrase \"one can show\" is not backed by an explicit argument or by a general theorem, so the robustness conclusion for this class is overstated and should be either proven or clearly labeled as a conjecture.","section":"Section IXA, after Eq. (66)"},{"comment":"The two-loop field-renormalization analysis is carried out explicitly only for N=2. The conclusion that the N=3 two-sphere/RP^2 manifolds also break into isolated fixed points under Type B effects is asserted without an analogous calculation. Given that the N=3 manifolds are a central result, this missing analysis is a gap; at minimum the paper should state that the N=3 fate is inferred by analogy, or provide the calculation.","section":"Section X(D) and Section IXB"}],"minor_comments":[{"comment":"The text says \"Let n = √ε (cos θ, sin θ) ... be the unit vector,\" but n has norm √ε, not 1; this is a typo.","section":"Section VIID, after Eq. (43)"},{"comment":"The sentence \"the phase boundary is a 2-dimensional surface in the 3-dimensional parameter space, meaning a D_p = 1 manifold\" should read \"a D_p−1 = 2 dimensional manifold.\"","section":"Section VIIIE, first paragraph"},{"comment":"The N=2 marginal operator in Eq. (37) is written without an explicit h.c. term, unlike the analogous N=3 expressions in Eq. (58); please harmonize the notation.","section":"Eq. (37) and Eq. (58)"},{"comment":"The entries labeled \"2-sphere\" should be labeled \"RP^2\" (see major comment); the parenthetical \"S2\" in Section VIIIB should be updated accordingly.","section":"Table III and Table IV"},{"comment":"The axes in the flow diagrams should be labeled directly in the figure panels (V0, Vx, Vz), and the blue/green lines indicating irrelevant and relevant directions should be identified in the figure rather than only in the caption.","section":"Figs. 7 and 9"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's one-loop content is sound and the coset-space framework is appealing, but the self-reported two-loop breakdown of the central conformal manifolds, together with the incomplete two-loop beta function, makes the physical-universality claim premature. If the authors can either complete the two-loop calculation or explicitly recast the paper as a one-loop ϵ-expansion result with the higher-loop caveats moved to the abstract and conclusions, I would be willing to revisit. There is also a nontrivial geometry correction (S^2 vs RP^2) that must be fixed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: the one-loop RG analysis is careful, internally consistent, and the fixed-point-manifold classification for N=2,3 is a real addition to the surface-TI literature. But the central selling point, that generic surface topological quantum criticality is dictated by conformal manifolds, is not established beyond one loop. The paper's own two-loop field-renormalization calculation drives a flow around the ring and splits it into four isolated fixed points; the complete two-loop beta function is not computed, so the physical story at epsilon=1 is open.\n\nWhat is new and good: the explicit construction of the ring for N=2 and the two rank-one fixed-point manifolds for N=3 from the one-loop RGE, the conical phase boundary, the Yukawa description, and the coset-space stability bookkeeping. The derivation from the lattice model in the appendices is real work, and the paper is honest that the SO(N) symmetry is emergent and that higher loops break it. For the one-loop equation, the fixed-point analysis checks out and the stability counts match the coset-space dimensions. The paper also correctly identifies that the emergent symmetry is sufficient but not necessary for a manifold, and the three-loop distortion example is a useful counterpoint.\n\nSoft spots, in proportion: (1) The main universality claim rests on a one-loop truncation. The two-loop eta term in Eq. C19 is computed and shown to induce dtheta/dl proportional to sin 2theta on the ring, leaving four isolated fixed points. At epsilon=1 this is not a small correction, so the ring is not a conformal manifold in the physical dimension unless a yet-uncomputed full two-loop beta function restores marginality. That is the crux, and it is unresolved. (2) Eq. C19 is not the complete two-loop beta function; it includes the two-loop self-energy but omits two-loop vertex diagrams. The stress-test note is right that the fate of the manifolds at two loops is not settled by this paper. The authors should either compute the full two-loop beta function or explicitly present the one-loop classification as provisional. (3) Calling the N=3 manifolds 2-spheres is topologically sloppy: the rank-one projectors form RP^2, not S^2. This is a minor fix but worth making.\n\nThis is a paper for people working on RG approaches to interacting topological surfaces and on emergent symmetries in epsilon expansions. It deserves a serious referee, and the referee should ask for a clearer separation between the one-loop mathematics and the physical universality claims. If the authors can either compute the missing two-loop terms or recalibrate their claims, the one-loop core is publishable.","headline":"A careful one-loop RG classification of N=2,3 surface topological criticality, with real substance, but the conformal-manifold universality claim is not protected beyond one loop and the paper's own two-loop field renormalization already breaks the ring.","tokens_in":46629,"tokens_out":3595,"would_cite":true,"duration_ms":41533,"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":"The quantum critical surface of a topological insulator is governed by a ring and two spheres of fixed points.","keywords":["surface topological quantum criticality","conformal manifold","renormalization group","emergent symmetry","topological insulator surface","fixed point manifold","epsilon expansion","Yukawa field theory"],"falsifier":"Evaluate the two-loop $\\beta$ function in Eq. (69) on the one-loop ring: if any continuous arc of points remains exactly fixed rather than flowing to the four discrete fixed points, the claimed splitting of the conformal manifold is incorrect. Equivalently, compute the scaling dimension of the candidate marginal operator in Eq. (37) at order $\\epsilon^2$; any nonzero correction away from $D$ removes exact marginality and would mean the manifold is only approximately conformal.","tokens_in":45496,"feed_emoji":"⚛️","tokens_out":7300,"duration_ms":72012,"temperature":0.7,"pith_summary":"This paper asks what universality classes control the quantum phase transition from a gapless surface to a superconducting surface of a three-dimensional topological insulator with attractive interactions. For surfaces carrying two or three half-Dirac cones, it argues that generic points on the phase boundary flow to a continuous family of strongly interacting fixed points, a ring for two cones and two 2-spheres for three cones, rather than to an isolated Wilson-Fisher-type fixed point. These conformal manifolds, whose dimension equals the number of exactly marginal operators, set the infrared dynamics on the whole co-dimension-one phase boundary. Isolated strong-coupling fixed points also exist, but the paper identifies them as multi-critical rather than generic critical points. If correct, the result changes what to look for experimentally: power-law correlations of marginal operators, and universality classes labelled by manifolds rather than single critical points.","feed_headline":"Topological surface criticality lives on rings and spheres","feed_subtitle":"In one-loop RG flow, N=2 and N=3 topological-insulator surfaces flow to continuous fixed-point manifolds, not isolated points.","key_machinery":"The central object is the one-loop renormalization-group equation for the interaction matrix, $d\\hat{V}/dl = -\\epsilon \\hat{V} + \\hat{V}^2$, written for an $N\\times N$ real symmetric coupling matrix $\\hat{V}$ in $D=2+\\epsilon$ spacetime dimensions. Its emergent $SO(N)$ flavor-rotation symmetry, combined with the invariant subgroup $H$ of a fixed-point matrix, produces continuous fixed-point manifolds as coset spaces $SO(N)/H$; the number of exactly marginal operators at a point equals the dimension of the tangent space and fixes the manifold's co-dimension in the $D_p=N(N+1)/2$-dimensional interaction parameter space.","core_discovery":"The central claim is that, in the one-loop approximation, surface topological quantum criticality of an attractively interacting topological-insulator surface is captured by conformal manifolds of fixed points rather than by isolated fixed points. The renormalization-group equation $d\\hat{V}/dl = -\\epsilon \\hat{V} + \\hat{V}^2$ for the $N\\times N$ interaction matrix has an emergent $SO(N)$ symmetry, and any fixed point invariant under a subgroup $H$ generates a manifold of fixed points shaped as the coset space $SO(N)/H$. For $N=2$ this gives a ring of fixed points sitting on the conical phase boundary in the three-dimensional parameter space; for $N=3$ it gives two 2-spheres in the six-dimensional parameter space. The number of exactly marginal operators equals the dimension of the manifold, and the scaling dimensions of the operators near the manifold are position-independent, so the whole manifold carries a single universality class. The paper further shows that certain higher-loop one-particle-irreducible effects only distort these manifolds, while fermion field renormalization can break them down into discrete fixed points.","pith_inferences":["If the conformal-manifold picture survives beyond the one-loop approximation, analogous fixed-point manifolds should appear for other $N$ and for other symmetry-protected boundary transitions, with the coset rule $SO(N)/H$ predicting their shape and co-dimension from the emergent symmetry alone.","The two-loop splitting of the ring into discrete fixed points suggests a general mechanism: when exact marginality is broken, a continuous universality class softens into a slow flow along the former manifold, so intermediate-energy experiments might see approximate power laws before a crossover to the discrete fixed points.","Because the marginal operators have scaling dimension exactly equal to the spacetime dimension, they are natural candidates to appear as non-decaying soft modes in surface transport or noise spectroscopy, which could test the distinction between conformal manifolds and isolated critical points.","The paper's separation of generic criticality (continuous manifolds) from multi-criticality (isolated strong-coupling fixed points) may apply more broadly to interacting gapless boundaries, not just topological-insulator surfaces with attractive interactions."],"forward_implications":["For $N=2$, the phase boundary is a two-dimensional cone in a three-dimensional parameter space, and its universality is set by the ring manifold; for $N=3$, a five-dimensional boundary in a six-dimensional space is governed by the infrared-stable 2-sphere manifold with one relevant, three irrelevant, and two marginal operators.","The infrared-stable conformal manifolds support exactly marginal operators with scaling dimension $D$, so their correlation functions obey $\\langle O_M(x)O_M(0)\\rangle \\sim |x|^{-2D}$; observing such power laws would be direct evidence for a conformal manifold.","Isolated strong-coupling fixed points such as $\\hat{V}_c = \\epsilon \\hat{I}$ have only relevant directions, so they describe multi-critical surface topological quantum critical points rather than generic ones.","Near the conformal manifolds the four-fermion interaction is factorizable, so the infrared physics of the phase boundary is captured by an emergent single-complex-boson Yukawa field theory with two or three flavors of surface fermions, with the boson mass playing the role of the single relevant direction.","The number of surface Dirac cones is stable under local symmetric perturbations, because connecting surfaces with different $N$ for the same bulk topological invariant requires non-local unitaries that break crystal symmetries and introduce infinite-range hopping."],"supporting_citations":[{"why":"Supplies the renormalization-group method and $2+\\epsilon$ expansion from which the one-loop beta function for the interaction matrix is obtained.","marker":"[68–70]"},{"why":"Provides the dimension-reduction derivation of the effective two-dimensional interaction matrix $V_{ij}$ from a three-dimensional bulk interaction.","marker":"[64]"},{"why":"Defines the four $\\mathbb{Z}_2$ invariants and the surface Dirac-cone counting used to identify the $N=1,2,3$ surfaces of the topological insulator.","marker":"[7]"},{"why":"Supplies the minimal cubic-lattice model whose surface spectra are solved to demonstrate distinct surfaces for the same bulk topological invariant.","marker":"[58]"},{"why":"Gives the parity-eigenvalue method used to compute the $\\mathbb{Z}_2$ invariants at the time-reversal-invariant momenta.","marker":"[59]"},{"why":"Establishes the interacting ultraviolet fixed point for the $N=1$ surface against which the $N=2,3$ conformal-manifold results are contrasted.","marker":"[15–18, 65–67]"}],"fun_headline_variants":["Topological surface criticality flows to rings and spheres","Conformal manifolds replace isolated fixed points on surfaces","Emergent symmetry shapes surface criticality into manifolds","One-loop RG: surface criticality lives on rings and spheres","For N=2, ring; for N=3, sphere: surface criticality"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything important follows from the one-loop $\\beta$ function $d\\hat{V}/dl = -\\epsilon \\hat{V} + \\hat{V}^2$ and its emergent $SO(N)$ symmetry; the paper assumes this one-loop fixed-point structure, solved at order $\\epsilon$ and then extrapolated to $\\epsilon=1$, identifies the generic surface criticality, even though its own two-loop field-renormalization calculation breaks $SO(N)$ and splits the ring into isolated fixed points.","fun_headline_variants_meta":{"raw":{"variants":["Topological surface criticality flows to rings and spheres","Conformal manifolds replace isolated fixed points on surfaces","Emergent symmetry shapes surface criticality into manifolds","One-loop RG: surface criticality lives on rings and spheres","For N=2, ring; for N=3, sphere: surface criticality"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000741,"raw_usage":{"total_tokens":3389,"prompt_tokens":1109,"completion_tokens":2280,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":725,"completion_tokens_details":{"reasoning_tokens":2195}},"tokens_in":725,"tokens_out":2280,"duration_ms":16872,"temperature":1.0,"reasoning_tokens":2195,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:01:23.770658+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the two-loop $\\beta$ function in Eq. (69) on the one-loop ring: if any continuous arc of points remains exactly fixed rather than flowing to the four discrete fixed points, the claimed splitting of the conformal manifold is incorrect. Equivalently, compute the scaling dimension of the candidate marginal operator in Eq. (37) at order $\\epsilon^2$; any nonzero correction away from $D$ removes exact marginality and would mean the manifold is only approximately conformal.","supporting_citations":[{"cited_title":"Zhou, Emergent u(1) symmetries in gapless fermionic superfluids and superconductors, Phys","cited_arxiv_id":null,"evidence_quote":"Provides the dimension-reduction derivation of the effective two-dimensional interaction matrix $V_{ij}$ from a three-dimensional bulk interaction."},{"cited_title":"We follow closely the path outlined in Ref.[64]","cited_arxiv_id":null,"evidence_quote":"Defines the four $\\mathbb{Z}_2$ invariants and the surface Dirac-cone counting used to identify the $N=1,2,3$ surfaces of the topological insulator."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the minimal cubic-lattice model whose surface spectra are solved to demonstrate distinct surfaces for the same bulk topological invariant."},{"cited_title":"Jian, C.-H","cited_arxiv_id":null,"evidence_quote":"Gives the parity-eigenvalue method used to compute the $\\mathbb{Z}_2$ invariants at the time-reversal-invariant momenta."}],"review_version":1}