{"id":"29d77113-c586-42d5-ac0f-0af939b276f8","arxiv_id":"2505.13592","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Parity-invariant 3d CFTs gapped by a magnetic field satisfy c0 <= 0 and additional Wilson-coefficient bounds from dispersion relations, implying diamagnetism and positive background-monopole scaling dimensions.","lead":"Three-dimensional conformal field theories with a global U(1) symmetry are studied in a background magnetic field, under the assumption that the field opens a mass gap. The authors derive positivity bounds on the effective action coefficients, predicting diamagnetic behavior at large field and positive scaling dimensions for background monopole operators, and they compute the relevant coefficients for free scalars, free fermions, and a holographic model.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Universal bounds rest on the unproven gap premise; a nontrivial interacting gapped example is needed to test the assumed phase.","rationale":"The reader identified the mass-gap assumption as the weakest point; my reading agrees. I checked the apparent tension between the restriction ℓ≥4 in (5.28) and the c0≤0 bound: c0 enters the ω^4 coefficient through the k^2 part of the c0 current contact term after the iω^2 prefactor in the matrix, while c2,i enter at ω^6 and c4,i at ω^8. Thus the separation of derivative orders is consistent, and c0≤0 is not an artifact of an invalid subtraction order. The free-fermion and holographic counterexamples are explicitly traced to the absence of a gap, so they do not falsify the conditional theorem. What remains load-bearing is the applicability of the theorem: the assumed gapped phase is not demonstrated for any interacting CFT, and only a free, non-interacting scalar is shown to satisfy the bounds. Since the paper states the assumption clearly and the theorem is conditional, this concern does not invalidate the accepted result, but it does mark the natural place to seek a decisive check. An O(2N) large-N computation would test both the existence of the gap and the positivity bounds in an interacting, weakly coupled setting. Until such a check is done, the universality of the physical predictions retains a gap-shaped caveat.","tokens_in":41367,"tokens_out":25413,"duration_ms":271177,"concrete_test":"Compute the leading Wilson coefficients c0, c2,1, c2,2, c2,3 for the O(2N) Wilson-Fisher CFT in the 1/N expansion, using the free energy on S^1×S^2 with monopole flux and the current two-point function in a constant magnetic field, then test inequalities (5.30), (5.31) and the allowed region (5.55). This theory is weakly interacting and expected to develop the assumed gap. If its coefficients violate the bounds, either the gap premise or the dispersive derivation is wrong; if they obey the bounds, the universal claim gains nontrivial supporting evidence beyond the free scalar.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing premise is the mass-gap assumption stated in Section 2 as \"A critical assumption\": the magnetized CFT is gapped, so the effective action is local and the retarded Green's function has no upper-half-plane singularities. This premise is required for the dispersion-relation step (5.20)-(5.23) and for the claim in (5.25) that the δ(k+p_n) term does not affect the ω-expansion around ω=0. The paper's own examples show how fragile the premise is: the free Dirac fermion has a gapless lowest Landau level, and the correlator computation in Section 7.1 requires discarding an explicitly divergent zero-mode contribution by hand; the extremal holographic black hole in Section 8.2 has non-zero entropy and hence no gap. Both examples violate the derived bounds, exactly as the authors explain. The claimed universality of diamagnetic behavior and positive monopole dimensions therefore applies only to theories that actually enter the assumed gapped phase. For interacting CFTs this phase is asserted by citing [1], but no interacting gapped example is computed in this paper. The free complex scalar is the only fully gapped worked example. This is not an internal inconsistency, but it is the place where the central claim would fail if a plausible interacting theory turns out to be gapless in a magnetic field.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs a derivative expansion for the effective action of a parity-preserving 3d CFT with a global U(1) symmetry in a background magnetic field, assuming that the magnetic field drives the theory into a gapped phase. It builds the action to fourth order in derivatives, evaluates it in flat, monopole, squashed-sphere, and spinning-sphere backgrounds, computes current and stress-tensor two-point functions, and derives dispersive positivity constraints on the Wilson coefficients, including c0 ≤ 0 and an allowed region for the two-derivative coefficients shown in Fig. 1. It then computes these coefficients for the free complex scalar, the free Dirac fermion, and a holographic Einstein–Hilbert–Maxwell model. The free scalar obeys the bounds; the fermion and holographic examples violate them, and the authors attribute the violations to the absence of the assumed mass gap.","tokens_in":41587,"tokens_out":28475,"duration_ms":266021,"significance":"If the mass-gap assumption holds, the dispersive bounds provide a new universal input to the EFT of magnetized CFTs, with concrete falsifiable consequences: positivity of background monopole operator dimensions at large flux and diamagnetic response at large B. The derivation follows the standard analyticity-plus-unitarity route, and the free-scalar coefficients are cross-checked by several independent methods (flat-space derivative expansion, monopole background, S3 partition function, and current correlator). The paper is notably transparent about failure modes: the free fermion's gapless lowest Landau level and the extremal black hole's non-zero entropy are explicitly identified as violating the gap premise. The main limitation is that only one fully gapped example is computed, and the gap assumption for interacting theories rests on external arguments rather than on a worked interacting example.","major_comments":[{"comment":"The abstract and Section 1 present the results as 'universal predictions' for parity-preserving 3d CFTs, but the derivation of (5.30)–(5.39) relies on the mass-gap assumption stated in Section 2 as 'A critical assumption.' The paper's own examples show that the free Dirac fermion (gapless lowest Landau level, Section 7.1) and the extremal holographic model (non-zero entropy, Section 8.2) violate the bounds precisely because the gap is absent. The claims are therefore universal only within the class of CFTs that actually enter the assumed gapped phase, and I recommend that every occurrence of 'universal' in the abstract, introduction, and conclusion be accompanied by this conditionality.","section":"Abstract and Introduction"},{"comment":"The only fully gapped worked example in the paper is the free complex scalar; the fermion and holographic examples fail the gap premise, as the authors explain. The expectation that weakly interacting CFTs develop a gap in a magnetic field is attributed to reference [1], but no interacting gapped example is computed here. Since this premise is load-bearing for the central claim, I recommend either adding a nontrivial interacting test of the assumed phase (for example the O(2N) model in a 1/N expansion) or stating plainly in the conclusion that the applicability of the bounds to interacting CFTs remains an assumption.","section":"Section 2 and Section 9"}],"minor_comments":[{"comment":"The step in which the delta-function term in (5.25) is dropped when expanding around ω = 0 is cited to reference [5] as 'rigorously argued'; since this step is essential for the positivity bound (5.26), a one-sentence justification (support at |ω| ≥ m for a gapped spectrum) would make the paper self-contained.","section":"Section 5, around Eq. (5.25)"},{"comment":"The matching between the free-energy sum (6.2), the logarithm of the partition function (6.36), and the effective-action results (3.11) and (3.15) involves conventions for the sphere radius L, the thermal circle β, and the flux Q that are not stated explicitly; adding these conventions would help the reader verify the coefficient comparisons.","section":"Sections 3.2 and 6.2"},{"comment":"The matched asymptotic expansion used in the extremal background is only sketched; since the matching is delicate because the horizon becomes an essential singularity, a brief outline of the matching conditions or a more precise reference for the boundary-layer method would improve reproducibility.","section":"Section 8.1"},{"comment":"Reference [23] is listed as 'To appear' and is used to justify leaving the derivation of some holographic coefficients to future work; since it is an unpublished self-citation, the authors should either supply the missing computation or mark the citation more clearly as a forthcoming paper.","section":"References"},{"comment":"In matching the free-scalar correlator (6.15) to the EFT form factors (4.28), the sign and factor conventions between the Euclidean calculation and the Lorentzian EFT are not spelled out; a sentence clarifying this correspondence would prevent confusion.","section":"Section 6.1"}],"recommendation":"minor_revision","confidential_remarks":"The paper is technically sound, original, and unusually honest about its assumptions and their failure modes. The central derivation is credible, and the conditional nature of the claims is acknowledged in the body of the text. My only substantive request is that the scope of the 'universal' statements be made explicit at every point where they appear, and that the status of the mass-gap premise for interacting theories be stated as an assumption rather than as an established fact."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth your time. The genuinely new core is the complete four-derivative parity-invariant EFT for a 3d CFT in a magnetic field – 28 terms, constructed via hatted variables – and the dispersive positivity bounds on its Wilson coefficients. The derivation of c0 ≤ 0 and the kinked allowed region for c2,i from a spectral sum rule is coherent and does not fit any examples; the coefficients are then computed independently. The free complex scalar obeys the bounds, and the authors are admirably clear that the free fermion and the holographic model violate them because those theories lack a gap (gapless lowest Landau level, extremal black hole with entropy). That honesty is a real strength.\n\nThe soft spot is exactly where the stress-test points: the mass-gap premise is load-bearing and remains an assumption, not a result, for interacting theories. The paper calls it ‘a critical assumption’ and cites [1] for support, but no interacting gapped example is computed. So the universal predictions for diamagnetism and positive monopole dimensions are conditional: they hold for any parity-preserving 3d CFT that actually enters the gapped phase in a magnetic field. That is a legitimate, sharply formulated statement, but it is not the same as proving all (or even typical) interacting CFTs gap. A referee should probe this, ideally by asking for an interacting example in a 1/N or ε expansion that exhibits the gap and satisfies the bounds.\n\nMinor issues: the free-fermion c2,2 computation discards an explicitly divergent zero-mode contribution by hand; the paper explains why, but that coefficient is less secure than the scalar one. The 28-term expansion is long and was done in Mathematica; I cannot fully verify it from the text, but the ancillary file should make it checkable.\n\nBottom line: serious theory paper, central argument sound conditional on its stated premise. For readers working on magnetic-field response, monopole operators, or holographic transport, it will be a useful reference. It deserves a serious referee; send it out.","headline":"A careful, honest EFT-positivity paper where the central bounds are conditional on an explicit mass-gap assumption; the free scalar checks out, and the cleanest next step is an interacting gapped example.","tokens_in":42171,"tokens_out":2524,"would_cite":true,"duration_ms":24714,"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":"This paper shows that a parity-preserving three-dimensional CFT in a large background magnetic field, when gapped, must satisfy $c_0 \\le 0$, forcing diamagnetic large-field response and positive background monopole dimensions.","keywords":["conformal field theory","magnetic field","effective field theory","dispersive positivity bounds","diamagnetism","monopole operators","Wilson coefficients","three dimensions"],"falsifier":"Take any concrete parity-preserving 3D CFT, put it on a three-sphere with large magnetic flux $Q$, and compute the free energy, or equivalently the monopole scaling dimension, by exact diagonalisation or Monte Carlo: if a theory with a demonstrable gap gives $\\Delta<0$ or coefficients outside the region (5.55), the dispersive bound is wrong; conversely, violations in a gapless theory, such as the free Dirac fermion, do not test the bound because the premise fails.","tokens_in":41162,"feed_emoji":"🧲","tokens_out":8260,"duration_ms":79375,"temperature":0.7,"pith_summary":"The paper argues that any parity-preserving three-dimensional conformal field theory with a global $U(1)$ symmetry, once placed in a large background magnetic field, enters a gapped phase whose long-distance behaviour is a local effective action in the background fields. From analyticity and positivity of retarded two-point functions of the conserved current and stress tensor, the paper derives a set of inequalities on the Wilson coefficients, most importantly $c_0 \\le 0$. This single sign implies that the large-field free energy grows with the magnetic field, giving universal diamagnetic behaviour, and that background monopole operators have positive scaling dimension at large flux. The same inequalities carve out a finite allowed region for the two-derivative coefficients, verified explicitly for the free complex scalar, while the free Dirac fermion and a holographic model fall outside because their magnetic phases are not gapped.","feed_headline":"Magnetic fields make gapped 3D quantum field theories diamagnetic","feed_subtitle":"For any parity-preserving 3D conformal field theory with a mass gap, the leading Wilson coefficient must be negative—large fields repel.","key_machinery":"The load-bearing object is the generalised dispersion relation for retarded Green’s functions of the operator $O(x)=\\partial_0 J_\\mu(x)V^\\mu + T_{\\mu\\nu}(x)U^{\\mu\\nu}$, evaluated at momentum $k=\\omega(1,\\vec\\xi)$ with $|\\vec\\xi|<1$. In a gapped phase the retarded correlator is analytic in the upper-half $\\omega$-plane and grows like $\\omega^d$, so contour integration yields a positive spectral sum rule, $G_R^{(\\ell)}(0)\\ge 0$ for even $\\ell>d$, which translates into positive semi-definiteness of an $8\\times 8$ matrix built from current and stress-tensor two-point functions. Demanding that the $\\omega^4$, $\\omega^6$ and $\\omega^8$ coefficients of this matrix be positive semi-definite for all $|\\vec\\xi|<1$ produces the inequalities (5.30)–(5.39): $c_0\\le0$, $c_{2,3}\\le0$, and a sequence of quadratic and cubic bounds whose regions in coefficient space have piecewise boundaries.","core_discovery":"Weyl invariance forces the low-energy effective action $W[A,g]$ of a gapped magnetised CFT to be built from the hatted metric $\\hat g_{\\mu\\nu}=g_{\\mu\\nu}F$ and a rescaled field strength, giving one zero-derivative term, three two-derivative terms, and twenty-eight four-derivative terms. Computing the current and stress-tensor two-point functions from this action and imposing a generalised Kramers–Kronig positivity condition on the retarded Green’s functions yields $c_0\\le 0$ and the inequalities (5.30)–(5.39). In the $c_{2,1}/c_{2,3}$–$c_{2,2}/c_{2,3}$ plane the allowed region has a boundary with a kink at $(-3/2,0)$. The paper computes all second-order Wilson coefficients for the free complex scalar, the free four-component fermion, and a holographic model with a Maxwell field and negative cosmological constant; only the scalar satisfies the bounds, and the paper ties the fermion’s and holographic model’s violations to their gapless lowest Landau level and extremal horizon degeneracy, respectively.","pith_inferences":["The same positivity matrix can be used as a diagnostic: if a proposed gapped description of a magnetised CFT yields Wilson coefficients outside the allowed region, that is evidence the description has missed a light mode, even when no explicit zero mode has been found.","The kink at $(-3/2,0)$ in the allowed region is a natural place to look for extremal or solvable theories that saturate the bounds; computing three-point functions of $J_\\mu$ and $T_{\\mu\\nu}$ could tighten the allowed island and test whether any known theory sits exactly at the corner.","Adding a small chemical potential should preserve the EFT structure while shifting the coefficients; for $0<\\mu<\\sqrt{2|B|}$ the paper’s free-fermion analysis suggests only occupied Landau levels change, so the same positivity inequalities should hold with modified $c_{2,i}$, which is directly checkable in a proper-time computation."],"forward_implications":["The large-field free energy satisfies $E(B)\\sim -\\sqrt{2}\\pi c_0 Q^{3/2}/L$ with $c_0\\le0$, so $E(B)$ grows with $B$: parity-preserving gapped 3D CFTs are diamagnetic at large field.","Background monopole operators have positive scaling dimension at large flux, $\\Delta \\sim -\\sqrt{2}\\pi c_0 Q^{3/2}>0$, connecting analyticity of current and stress-tensor correlators to unitarity in the monopole sector.","Two-derivative Wilson coefficients must lie in the kinked allowed region of Fig. 5; the free complex scalar sits inside while the free fermion and holographic model sit outside, consistent with the gap assumption identifying when the EFT applies.","The EFT is universal to second order: only $c_0,c_{2,1},c_{2,2},c_{2,3}$ control long-distance response, and the paper fixes all four for three concrete theories."],"supporting_citations":[{"why":"Constructs the two-derivative effective action for a CFT in a magnetic field and argues that a gap opens, the framework the paper extends to four derivatives.","marker":"[1]"},{"why":"Gives the flat-space derivative expansion from which $c_0$ and $2c_{2,1}-c_{2,3}$ are read off for the free scalar.","marker":"[2]"},{"why":"Supplies the four-component fermion derivative expansion used to extract $2c_{2,1}-c_{2,3}$.","marker":"[3]"},{"why":"Establishes causality/analyticity positivity bounds on EFT coefficients that the dispersion argument adapts to magnetised backgrounds.","marker":"[4]"},{"why":"Provides the generalised Kramers–Kronig representation for spontaneously broken Lorentz invariance that yields the positive spectral sum rule.","marker":"[5]"},{"why":"Computes monopole scaling dimensions in free theories, giving the large-$Q$ expansion matched to the EFT.","marker":"[10]"},{"why":"Develops the proper-time method used to resum the magnetic field into the scalar and fermion propagators.","marker":"[17]"},{"why":"Shows how to match near-horizon and boundary expansions in extremal holographic backgrounds, used for the two-point function.","marker":"[22]"},{"why":"Provides the holographic dictionary relating boundary current response to bulk gauge fluctuations.","marker":"[24]"}],"fun_headline_variants":["3D CFTs become diamagnetic under strong magnetic fields","Magnetic fields force gapped CFTs to be diamagnetic","Diamagnetism is universal in gapped 3D CFTs","Positivity rules magnetised CFTs: only scalar survives","Weyl invariance and positivity bound magnetised CFTs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The magnetic field opens a real mass gap, so no massless excitations survive and the retarded Green’s function is analytic away from a mass threshold; without this, the positive spectral sum rule and all derived inequalities can fail.","fun_headline_variants_meta":{"raw":{"variants":["3D CFTs become diamagnetic under strong magnetic fields","Magnetic fields force gapped CFTs to be diamagnetic","Diamagnetism is universal in gapped 3D CFTs","Positivity rules magnetised CFTs: only scalar survives","Weyl invariance and positivity bound magnetised CFTs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000391,"raw_usage":{"total_tokens":2056,"prompt_tokens":945,"completion_tokens":1111,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":561,"completion_tokens_details":{"reasoning_tokens":1022}},"tokens_in":561,"tokens_out":1111,"duration_ms":8237,"temperature":1.0,"reasoning_tokens":1022,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:14:01.178568+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take any concrete parity-preserving 3D CFT, put it on a three-sphere with large magnetic flux $Q$, and compute the free energy, or equivalently the monopole scaling dimension, by exact diagonalisation or Monte Carlo: if a theory with a demonstrable gap gives $\\Delta<0$ or coefficients outside the region (5.55), the dispersive bound is wrong; conversely, violations in a gapless theory, such as the free Dirac fermion, do not test the bound because the premise fails.","supporting_citations":[{"cited_title":"Derivative Expansion of the Effective Action and Vacuum Instability for QED in 2+1 Dimensions","cited_arxiv_id":"hep-th/9409113","evidence_quote":"Gives the flat-space derivative expansion from which $c_0$ and $2c_{2,1}-c_{2,3}$ are read off for the free scalar."},{"cited_title":"On gauge invariance and vacuum polarization","cited_arxiv_id":null,"evidence_quote":"Develops the proper-time method used to resum the magnetic field into the scalar and fermion propagators."}],"review_version":1}