betaFeedback

WorksPhysics and Astronomy

The Free Scalar as a Thermal Envelope for Three-Dimensional CFTs

Abstract

We conjecture that every unitary three-dimensional conformal field theory obeys cS/CT≤16πζ(3)c_S/C_T\leq16πζ(3), where s=cST2s=c_ST^2 is the thermal entropy density and CTC_T the stress-tensor two-point coefficient, with equality for free scalars. We embed this in a stronger conjecture: on S1×H2S^1\times H^2, the CTC_T-normalized hyperbolic heat capacity is bounded above by that of the free scalar for all T≥(2πR)−1T\geq(2πR)^{-1}. Near the conformal temperature, for theories smooth from the high-temperature side and without boundary terms in the modular Hamiltonian, the heat-capacity bound follows from conformal-collider positivity. We test the conjectures in free, interacting large-NN, and holographic theories. Monte Carlo data for the three-dimensional Ising and O(N)O(N) models also obey the flat-space bound.

Introduction

In two spacetime dimensions the leading thermal free energy is fixed by the central charge. In three dimensions this simplification disappears. The sphere free energy FF, the stress-tensor two-point coefficient CTC_T, and the flat-space thermal entropy coefficient cSc_S are independent pieces of CFT data [1], all of which can be regarded as measures of the number of degrees of freedom. It is therefore natural to ask whether unitarity and conformal invariance nevertheless impose inequalities among them.

A striking proposal in this direction is the conjecture of Ref. [2] that CT/FC_T/F is maximized by the free scalar. Here we propose a complementary inequality involving the thermal coefficient,

cSCT≤16πζ(3),s(T)=cST2.(1)\frac{c_S}{C_T} \le16\pi\zeta(3), \qquad s(T)=c_ST^2 . \tag*{(1)}

Although our detailed tests are in three dimensions, the free-scalar upper bound has a natural analogue in d≥3d \ge3. A perturbative test near four dimensions, including possible infrared and thermal-ordering loopholes, is described in Supplemental Material.

We use the normalization

⟨Tμν(x)Tρσ(0)⟩=CTx6Iμν,ρσ(x),(2)\langle T_{\mu\nu}(x)T_{\rho\sigma}(0)\rangle= \frac{C_T}{x^6}I_{\mu\nu,\rho\sigma}(x), \tag*{(2)}

for which one real conformal scalar has CT=3/(32π2)C_T=3/(32\pi^2). Its thermal entropy coefficient is cS=3ζ(3)/(2π)c_S=3\zeta(3)/(2\pi) and hence

cSCT∣s=16πζ(3)=60.4219….(3)\left.\frac{c_S}{C_T}\right|_s = 16\pi\zeta(3)=60.4219\ldots. \tag*{(3)}

The free-scalar ratio is unchanged by adding decoupled free scalars. Further, a decoupled topological sector contributes neither to cSc_S nor to CTC_T.

Several familiar theories lie below this value. A free Dirac fermion has

cSCT=3416πζ(3)=45.3165…,(4)\frac{c_S}{C_T}=\frac{3}{4}16\pi\zeta(3)=45.3165\ldots, \tag*{(4)}

Einstein holography gives (see, e.g., Refs. [1, 3, 4, 5])

cSCT=4π527=45.3363…,(5)\frac{c_S}{C_T}=\frac{4\pi^5}{27}=45.3363\ldots, \tag*{(5)}

and the critical O(N)O(N) model at N=∞N=\infty gives [6, 7]

cSCT=4516πζ(3)=48.3376….(6)\frac{c_S}{C_T}=\frac{4}{5}16\pi\zeta(3)=48.3376\ldots. \tag*{(6)}

Although these examples support the conjecture, a comparison of finitely many CFTs provides limited evidence for a universal bound. In this Letter we propose a stronger inequality that constrains the thermal response on hyperbolic space at every temperature above the conformal temperature. We test this inequality both in individual CFTs and across continuous families of interacting theories.

Before turning to hyperbolic space, where FF, CTC_T, and cSc_S can all be embedded in a single thermal function, we note that the thermal coefficient also has a vacuum interpretation. As emphasized by Shaghoulian [8] (see also Ref. [9]), exchanging the thermal and spatial circles relates the leading high-temperature free energy to the vacuum energy on a geometry with one compact spatial circle. In three dimensions,

Evac(SL1×R)=−εvacVL2,εvac=cS3.(7)E_{\mathrm{vac}}(S^1_L \times\mathbb{R})=-\frac{\varepsilon_{\mathrm{vac}}V}{L^2}, \qquad\varepsilon_{\mathrm{vac}}=\frac{c_S}{3}. \tag*{(7)}

Here VV is the length of the non-compact direction, and fermions are antiperiodic around SL1S^1_L, as inherited from the thermal circle. A periodic spatial spin structure defines a different Casimir problem, so the vacuum form of the conjecture refers specifically to this Neveu–Schwarz sector. Thus, the proposed inequality can equally well be written as

εvacCT≤16πζ(3)3.(8)\frac{\varepsilon_{\mathrm{vac}}}{C_T}\le\frac{16\pi\zeta(3)}{3}. \tag*{(8)}

Both quantities in Eq. (8) are vacuum data. The conjecture bounds the Casimir energy on a circle in units of CTC_T.

Further thermal-OPE and operator-counting formulations of the bound are given in Supplemental Material.

A single hyperbolic thermal function. Place the theory on Sβ1×HR2S^1_\beta\times H^2_R and define

T0=12πR,x=TT0,Φ(x)=R2sH(T),(9)T_0=\frac{1}{2\pi R}, \qquad x=\frac{T}{T_0}, \qquad\Phi(x)=R^2s_H(T), \tag*{(9)}

where sH(T)s_H(T) is the entropy density on HR2H_R^2. The vacuum density matrix reduced to a disk of radius RR is conformally equivalent to a thermal density matrix on Rt×HR2\mathbb{R}_t \times H_R^2 at T0=(2πR)−1T_0 = (2\pi R)^{-1} [10]. In Euclidean signature this state is represented by S2πR1×HR2S_{2\pi R}^1 \times H_R^2. The same conformal map takes the qq-sheeted replica geometry, which computes the Rényi entropy SqS_q, to S2πqR1×HR2S_{2\pi qR}^1 \times H_R^2. The Rényi index is therefore q=T0/T=1/xq = T_0/T = 1/x. We normalize

Ψ(x)=12π3CTΦ(x).(10)\Psi(x) = \frac{12}{\pi^3 C_T}\Phi(x). \tag*{(10)}

Three apparently independent CFT quantities are then encoded in elementary properties of one function:

Ψ(1)=6Fπ4CT,Ψ′(1)=1,Ψ(x)x2→x→∞3π5cSCT.(11)\Psi(1) = \frac{6F}{\pi^4 C_T}, \qquad\Psi'(1) = 1, \qquad\frac{\Psi(x)}{x^2} \xrightarrow{x\to\infty} \frac{3}{\pi^5}\frac{c_S}{C_T}. \tag*{(11)}

The first relation follows from the equality between the universal disk entanglement entropy and the sphere free energy [10]. The second is the universal first derivative of spherical Rényi entropy and follows from the stress-tensor two-point function [11]. The third is the high-temperature, or small-Rényi-index, limit [12, 13], where the curvature radius becomes irrelevant:

R2sH(T)∼cSR2T2=cS4π2x2.(12)R^2s_H(T) \sim c_SR^2T^2 = \frac{c_S}{4\pi^2}x^2. \tag*{(12)}

Equation (11) gives a useful geometric interpretation of the two conjectured CFT bounds. The bound of Ref. [2] on CT/FC_T/F says that the free scalar minimizes Ψ(1)\Psi(1). Our conjecture says that the free scalar maximizes the coefficient of the large-xx growth of Ψ(x)\Psi(x). Since every CFT also obeys Ψ′(1)=1\Psi'(1)=1, the natural quantity to compare is the normalized hyperbolic heat capacity itself.

We therefore conjecture

Ψ′(x)≤Ψs′(x),x≥1,Ψ′(x)=12R2π3CTcV(T)x,(13)\Psi'(x) \leq\Psi'_s(x), \qquad x \geq1, \qquad\Psi'(x) = \frac{12R^2}{\pi^3 C_T}\frac{c_V(T)}{x}, \tag*{(13)}

where cV=T dsH/dTc_V = T\,ds_H/dT and the subscript ss in Ψs′(x)\Psi'_s(x) denotes the free scalar. Thus, at fixed CTC_T the scalar bounds the hyperbolic heat capacity, with the universal factor 1/x1/x dictated by our normalization.

Integrating from x=1x=1 gives the weaker entropy envelope

Ψ(x)−Ψ(1)≤Ψs(x)−Ψs(1),x≥1.(14)\Psi(x)-\Psi(1) \leq\Psi_s(x)-\Psi_s(1), \qquad x \geq1. \tag*{(14)}

Taking the high-temperature limit recovers Eq. (1). This statement is logically independent of the conjectured CT/FC_T/F bound because the value at x=1x=1 has been subtracted.

The two conjectures nevertheless have a useful combined consequence. For theories to which the CT/FC_T/F conjecture applies and for which Ψ(1)>0\Psi(1)>0, positivity of the heat capacity gives Ψ(x)−Ψ(1)≥0\Psi(x)-\Psi(1)\geq0. Together with Ψ(1)≥Ψs(1)\Psi(1)\geq\Psi_s(1) and Eq. (14), this yields

Ψ(x)Ψ(1)≤Ψs(x)Ψs(1)⟹CSF≤(CSF)s=8.9949…,(15)\frac{\Psi(x)}{\Psi(1)} \leq\frac{\Psi_s(x)}{\Psi_s(1)} \Longrightarrow\frac{C_S}{F} \leq\left(\frac{C_S}{F}\right)_s = 8.9949\ldots, \tag*{(15)}

where the second inequality is the x→∞x\to\infty limit. Thus, the CT/FC_T/F bound controls the starting value at T0T_0, while Eq. (13) controls the subsequent response.

The heat-capacity conjecture is stronger than ordinary thermodynamic stability. Positivity gives Ψ′(x)≥0\Psi'(x)\geq0, but even together with the universal normalization Ψ′(1)=1\Psi'(1)=1 this does not place an upper bound on the large-xx slope. A proof of Eq. (13) must therefore use constraints specific to conformal field theory. The first such constraint appears already in the second derivative at x=1x=1, which is fixed by stress-tensor three-point data.

A local theorem from ⟨TTT⟩\langle TTT\rangle. We now show that Eq. (13) holds near x=1x=1 for every CFT whose thermodynamics is smooth above x=1x=1 and whose modular Hamiltonian has no boundary term on the entangling surface. Let hqh_q denote the scaling dimension of the codimension-two twist defect that implements the qq-fold replica construction across the spherical entangling surface. Its first derivative at q=1q=1 is fixed by CTC_T, while its second derivative is determined by an integrated three-point function of the stress tensor [14]. In three dimensions, writing the parity-even ⟨TTT⟩\langle TTT\rangle structure in terms of the standard conformal-collider parameter t4t_4 [15, 16] (using Eqs. (2.21), (2.45), (2.46), and (3.14) of Ref. [14]), one has

hqCT=π324(q−1)−π311520(420+t4)(q−1)2+⋯ .(16)\frac{h_q}{C_T} = \frac{\pi^3}{24}(q-1) - \frac{\pi^3}{11520}(420+t_4)(q-1)^2 + \cdots. \tag*{(16)}

The thermodynamic identity hq=q2∫1/q1dx xΦ′(x)h_q = \frac{q}{2}\int_{1/q}^{1} dx\,x\Phi'(x) gives

h1′=12Φ′(1),h1′′=−12[Φ′(1)+Φ′′(1)],(17)h'_1 = \frac{1}{2}\Phi'(1), \qquad h''_1 = -\frac{1}{2}\left[\Phi'(1)+\Phi''(1)\right], \tag*{(17)}

and hence

Ψ′′(1)=34+t4240.(18)\Psi''(1) = \frac{3}{4}+\frac{t_4}{240}. \tag*{(18)}

For parity-even theories conformal-collider positivity gives −4≤t4≤4-4\leq t_4\leq4. For parity-violating theories the pair (t4,αT)(t_4,\alpha_T) lies inside the collider disk [16], so in particular t4≤4t_4\leq4. Therefore, every theory without such a boundary term obeys

1115≤Ψ′′(1)≤2330.(19)\frac{11}{15}\leq\Psi''(1)\leq\frac{23}{30}. \tag*{(19)}

The upper endpoint is the bosonic parity-even TTTTTT structure, while Einstein gravity has t4=0t_4=0 and hence Ψ′′(1)=3/4\Psi''(1)=3/4.

The conformal scalar is exceptional in precisely the way analyzed in Refs. [17, 18]. The improved scalar stress tensor is associated with a boundary contribution to the modular Hamiltonian,

H=Hcov+2πξ∫∂H2ϕ2.(20)H = H_{\mathrm{cov}} + 2\pi\xi\int_{\partial H^2} \phi^2 . \tag*{(20)}

This term does not change the first derivative at q=1q = 1, because the relevant separated two-point function ⟨Tμνϕ2⟩\langle T_{\mu\nu}\phi^2\rangle vanishes, but it contributes at the next order. In the conventions of Ref. [17], the difference between the full scalar result and the result obtained from the bulk stress tensor alone is

δh1′′=−π512.(21)\delta h_1^{\prime\prime} = -\frac{\pi}{512}. \tag*{(21)}

Using CT=3/(32π2)C_T = 3/(32\pi^2) gives exactly

δΨ′′(1)=12.(22)\delta\Psi^{\prime\prime}(1) = \frac{1}{2}. \tag*{(22)}

Thus,

Ψs′′(1)=1915=2330+12.(23)\Psi_s^{\prime\prime}(1) = \frac{19}{15} = \frac{23}{30} + \frac{1}{2}. \tag*{(23)}

The excess of Ψs′′(1)\Psi_s^{\prime\prime}(1) over the conformal-collider upper bound is thus accounted for by the scalar modular-Hamiltonian boundary term. What is new here is its role in separating the free-scalar thermal envelope from the ordinary conformal-collider bound.

Combining Eqs. (19) and (23) gives

Ψs′(x)−Ψ′(x)≥12(x−1)+O((x−1)2),(24)\Psi_s^{\prime}(x) - \Psi^{\prime}(x) \geq\frac{1}{2}(x - 1) + O((x - 1)^2), \tag*{(24)}

provided the hyperbolic thermodynamics admits a right-sided Taylor expansion through this order at x=1x = 1. Thus, the differential conjecture holds sufficiently close to x=1x = 1, with a finite margin; integration gives an entropy gap 14(x−1)2+O((x−1)3)\frac{1}{4}(x - 1)^2 + O((x - 1)^3). The remaining qualification is a boundary term built from a symmetry-allowed dimension-one scalar, as ϕ2\phi^2 is for the free scalar. Such an operator is marginal on the entangling surface. Theories with such a term require separate analysis and are not covered by Eq. (24).

Equation (24) also shows why the restriction to x≥1x \geq1 in Eq. (13) is essential. For a theory whose thermodynamics is smooth across x=1x = 1, the difference Ψs′(x)−Ψ′(x)\Psi_s^{\prime}(x) - \Psi^{\prime}(x) has the same expansion for x<1x < 1. Its linear term is then negative, so Ψ′(x)>Ψs′(x)\Psi^{\prime}(x) > \Psi_s^{\prime}(x) just below x=1x = 1. In the free Dirac and Einstein examples below, the ordering stays reversed down to x=0x = 0. At x=0x = 0 the scalar has Ψs′(0)=0\Psi_s^{\prime}(0) = 0, because its density of states on H2H^2 vanishes like E2E^2. The fermion’s density of states tends to a constant, which gives ΨF′(0)=16/(3π2)\Psi_F^{\prime}(0) = 16/(3\pi^2). Einstein holography gives ΨEin′(0)=2/(33)\Psi_{\mathrm{Ein}}^{\prime}(0) = 2/(3\sqrt{3}). Some large-NN saddles, such as the conventional critical O(N)O(N) saddle below, have no continuation to x<1x < 1.

Global tests. We first test Eq. (13) in theories whose hyperbolic response can be computed. We then test the flat-space bound, Eq. (1), in theories whose hyperbolic response is unknown.

For free fields, the calculation follows by differentiating the thermal free energies and using the H2H^2 spectral densities of Ref. [19]. For a one-particle mode of energy EE, let z=βEz = \beta E be the dimensionless thermal energy. The bosonic and fermionic single-mode entropies are

sB(z)=zez−1−log⁡(1−e−z),(25)s_B(z) = \frac{z}{e^z - 1} - \log(1 - e^{-z}), \tag*{(25)}
sF(z)=log⁡(1+e−z)+zez+1.(26)s_F(z) = \log(1 + e^{-z}) + \frac{z}{e^z + 1}. \tag*{(26)}

For later use, define the dimensionless effective mass mm by Ep=R−1p2+m2E_p = R^{-1}\sqrt{p^2 + m^2}, and write the bosonic entropy as

ΦB(x;m)=∫0∞dp2πptanh⁡(πp)sB(2πp2+m2x),(27)\Phi_B(x;m) = \int_0^\infty\frac{dp}{2\pi}p\tanh(\pi p)s_B\left(\frac{2\pi\sqrt{p^2 + m^2}}{x}\right), \tag*{(27)}

where pp is the dimensionless principal-series momentum on H2H^2. For the free conformal scalar m=0m = 0, so Φs(x)=ΦB(x;0)\Phi_s(x) = \Phi_B(x;0); for the interacting models below, mm is fixed by a gap equation. The free-Dirac entropy is instead

ΦF(x)=∫0∞dpπpcoth⁡(πp)sF(2πp/x).(28)\Phi_F(x) = \int_0^\infty\frac{dp}{\pi}p\coth(\pi p)s_F(2\pi p/x). \tag*{(28)}

At high temperature, the free scalar exhibits a logarithmic correction,

Ψs(x)=48ζ(3)π4x2−83π2log⁡x+O(1).(29)\Psi_s(x) = \frac{48\zeta(3)}{\pi^4}x^2 - \frac{8}{3\pi^2}\log x + O(1). \tag*{(29)}

This logarithm reflects the absence of a spectral gap. The conformal coupling ξR=−1/(4R2)\xi R = -1/(4R^2) cancels the gap 1/(4R2)1/(4R^2) of the H2H^2 Laplacian, so the scalar spectrum extends down to E=0E = 0. Without interactions, no thermal mass screens these low-energy modes.

Differentiating the corresponding Bose and Fermi entropy integrals gives the exact heat-capacity curves shown in Figure 1. The scalar lies strictly above the fermion for all x>1x > 1; this ordering also follows analytically from the spectral integrals (Supplemental Material).

Figure 1. Normalized hyperbolic heat-capacity response Ψ′(x)−1\Psi'(x)-1. All curves start at zero because Ψ′(1)=1\Psi'(1)=1. The free scalar is an upper envelope for the critical O(N)O(N) model, the BMB endpoint of the stable sextic line, a free Dirac fermion, and Einstein holography. The dashed Einstein curve nearly coincides with the Dirac curve.

Einstein holography supplies a very different example. The thermal state on H2H^2 is represented by the hyperbolic Schwarzschild-AdS black hole [10, 20]. Let rhr_h denote the horizon radius and L~\widetilde{L} the AdS curvature radius. Defining y=rh/L~y = r_h/\widetilde{L}, its temperature satisfies

x=3y2−12y,y≥1,(30)x = \frac{3y^2 - 1}{2y}, \qquad y \geq1, \tag*{(30)}

and, after the normalization in Eq. (10),

ΨEin(x)=y(x)2.(31)\Psi_{\mathrm{Ein}}(x) = y(x)^2 . \tag*{(31)}

Here y(x)=(x+x2+3)/3y(x) = (x + \sqrt{x^2 + 3})/3, and the solution exists for all x≥0x \ge0. The Einstein curve obeys ΨEin′′≤8/9\Psi''_{\mathrm{Ein}} \le8/9, whereas Ψ′′>96ζ(3)/π4≈1.18\Psi'' > 96\zeta(3)/\pi^4 \approx1.18 for x≥1x \ge1. Integrating from x=1x = 1 proves Eq. (13) for all $x > 1 (Supplemental Material).

A considerably sharper test is provided by the critical O(N)O(N) model at N=∞N = \infty. At leading order its sphere free energy and CTC_T are those of NN free scalars, and its parity-even stress-tensor three-point function is at the bosonic collider endpoint t4=4t_4 = 4. Thus, on the high-temperature side, its entropy function Ψ(x)\Psi(x) shares the free scalar’s value and first derivative at x=1x = 1, while its curvature Ψ′′(1+)=23/30\Psi''(1+) = 23/30 saturates the collider bound (19) without the free scalar’s additional ϕ2\phi^2 boundary contribution. The boundary contribution is absent because the interaction replaces ϕ2\phi^2 by the Hubbard–Stratonovich field σ\sigma, whose dimension is two.

On S1×H2S^1 \times H^2 the conventional saddle generates a dimensionless thermal mass m(x)m(x), determined by the large-NN gap equation (Eq. (3.9) of Ref. [21])

0=∫0∞dp[ptanh⁡(πp)p2+m2coth⁡(πp2+m2x)−1].(32)0 = \int_0^\infty dp \left[\frac{p\tanh(\pi p)}{\sqrt{p^2 + m^2}}\coth\left(\frac{\pi\sqrt{p^2 + m^2}}{x}\right) - 1\right]. \tag*{(32)}

For x≥1x \ge1 this saddle is well-behaved, with m(1)=0m(1) = 0. At leading order, the entropy is Φ(x)=NΦB(x;m(x))\Phi(x) = N\Phi_B(x;m(x)), using Eq. (27). The saddle reproduces this curvature from the x>1x > 1 side, even though the thermodynamics is not analytic across x=1x = 1. On this conventional saddle m(x)m(x) increases with xx, and the single-mode heat capacity decreases with mass. These facts establish Eq. (13) analytically for all x>1x > 1 at leading NN (Supplemental Material). At high temperature, 2πm(x)/x→2log⁡φ2\pi m(x)/x \to2\log\varphi, where φ=(1+5)/2\varphi= (1 + \sqrt{5})/2, reproducing the flat-space thermal mass [6], and the entropy coefficient tends to Eq. (6). The conventional saddle has no real solution for x<1x < 1 [21]; we use the x≥1x \ge1 branch.

The large-NN sextic line gives a particularly direct test because it contains the free scalar itself. For the interaction g6(ϕ⋅ϕ)3/(6N2)g_6(\phi\cdot\phi)^3/(6N^2), define α=g6/(4π)\alpha= \sqrt{g_6}/(4\pi). The stable large-NN conformal line is 0≤α≤10 \le\alpha\le1, ending at the Bardeen–Moshe–Bander point [22, 23, 24]. On the conformal vacuum saddle, FF and CTC_T retain their free-field leading values. Write I(m,x)I(m,x) for the integral in Eq. (32). The sextic saddle then obeys [21, 23]

m=αI(m,x).(33)m = \alpha I(m,x). \tag*{(33)}

The same monotone-mass argument establishes Eq. (13) on the conventional saddle throughout this line for x>1x > 1 (Supplemental Material). At fixed x>1x > 1 the response decreases with the sextic coupling; the BMB endpoint is shown in Figure 1.

At the BMB endpoint, α=1\alpha= 1, the flat-space limit gives m=x2πlog⁡2m = \frac{x}{2\pi}\log2, and hence

cSCT∣BMB=14πζ(3)=52.8692…(34)\left.\frac{c_S}{C_T}\right|_{\mathrm{BMB}} = 14\pi\zeta(3) = 52.8692\ldots \tag*{(34)}

Thus, the stable sextic line runs from free-scalar saturation down to 7/87/8 of its thermal coefficient, rather than to the critical-O(N)O(N) value. Its right-sided curvature at the conformal temperature is ΨBMB′′(1+)=19/15−π2/[2(2+π)2]=1.0799967…\Psi''_{\mathrm{BMB}}(1+) = 19/15 - \pi^2/[2(2+\pi)^2] = 1.0799967\ldots (see Supplemental Material), below the free scalar but above the ordinary collider upper bound 23/3023/30. This illustrates the separate behavior of theories with a dimension-one ϕ2\phi^2 operator, not covered by Eq. (24). At finite NN the fixed-line interpretation is more subtle, so we use this example only as a strict large-NN test.

In every example with a known hyperbolic response we also find the stronger ordering Ψ′′(x)≤Ψf′′(x)\Psi''(x) \le\Psi''_{\mathrm{f}}(x) for x>1x > 1, which integrates to Eq. (13). Unlike positivity of the heat capacity, no general CFT principle is known to imply it.

The remaining tests concern only the flat-space bound, because the hyperbolic response is not known. In holography, homogeneous planar Einstein-scalar black branes obey the stronger bound cS/CT≤4π5/27c_S/C_T \le4\pi^5/27, even with source-free scalar hair. This follows from warp-factor monotonicity (Supplemental Material).

Large-NN Chern–Simons (CS) vector models provide further tests. Write λ=N/k\lambda= N/k for the ’t Hooft coupling, where kk is the renormalized CS level, and restrict by parity to 0≤λ<10 \le\lambda< 1. At planar order the flat-space bound holds analytically for CS theories with a single fundamental scalar or fermion, at either the regular or the critical fixed point (Supplemental Material). The thermal free energies follow from the planar saddle of Ref. [25], and the stress-tensor normalization is

CT(λ)=CT(0)sin⁡(πλ)πλ.(35)C_T(\lambda) = C_T(0)\frac{\sin(\pi\lambda)}{\pi\lambda}. \tag*{(35)}

The regular-boson (RB) theory also has a marginal sextic coupling λ6(ϕ†ϕ)3/(6N2)\lambda_{6}(\phi^{\dagger}\phi)^{3}/(6N^{2}) at N=∞N=\infty. Varying λ6\lambda_{6} leaves the planar ⟨TT⟩\langle TT\rangle and separated-point ⟨TTT⟩\langle TTT\rangle unchanged [26], but changes cSc_{S}. For example, two stable couplings at λ=1/2\lambda=1/2 give

cSCT∣λ=1/2={49.0338,λ6=6π2,48.8231,λ6=12π2.(36)\left.\frac{c_{S}}{C_{T}}\right|_{\lambda=1/2} = \begin{cases} 49.0338, & \lambda_{6}=6\pi^{2},\\ 48.8231, & \lambda_{6}=12\pi^{2}. \end{cases} \tag*{(36)}

The stability conditions and the full (λ,λ6)(\lambda,\lambda_{6}) dependence are given in the Supplemental Material, using Ref. [27].

Finite-NN interacting CFTs can also be tested using critical Casimir amplitudes. For a three-dimensional classical critical system on a periodic slab, the excess free energy per transverse area is Δ/L2\Delta/L^{2}. This is the Casimir energy of Eq. (7) with Δ=−εvac\Delta=-\varepsilon_{\mathrm{vac}}, so

cS=−3Δ.(37)c_{S}=-3\Delta. \tag*{(37)}

Periodic-slab Monte Carlo amplitudes for the Ising and O(N)O(N) models, combined with bootstrap determinations of CTC_{T}, give the endpoint ratios in Table I. These are finite-NN, strongly interacting, nonholographic tests of Eq. (1). For O(4)O(4) and O(6)O(6), CTC_{T} is taken at its bootstrap lower bound [28], so these two ratios are upper estimates.

Published thermal and stress-tensor 1/N1/N expansions [7, 29, 30, 31, 32] give (cS/CT)O(N)=48.3376+1.5557/N+O(N−2)(c_{S}/C_{T})_{O(N)}=48.3376+1.5557/N+O(N^{-2}); the Gross–Neveu and conformal QED3_{3} endpoints also satisfy Eq. (1) in their controlled large-NfN_{f} regimes (Supplemental Material).

Table I summarizes the tests.

Discussion. The free scalar’s behavior near x=1x=1 and as x→∞x\to\infty provides a partial explanation of its conjectured role in setting the upper bound on cS/CTc_{S}/C_{T}. Near x=1x=1, its ϕ2\phi^{2} boundary term in the modular Hamiltonian lifts the

TheorycS/CTc_{S}/C_{T}Test
free scalar60.422saturation
critical O(N)O(N), N=∞N=\infty48.338full
sextic line, N=∞N=\infty52.869–60.422full
Einstein holography45.336full
free Dirac45.316full
3d Ising [33, 34]51.08(13)endpoint
O(2)O(2) [33, 35]50.84(27)endpoint
O(3)O(3) [33, 36]50.32(23)endpoint
O(4)O(4) [28, 33]50.05(24)endpoint
O(6)O(6) [28, 33]49.66(29)endpoint
RB CS-matter45.316–60.422endpoint
other CS-matter<60.422<60.422endpoint
Gross–Neveu45.32+0.08/Nf45.32+0.08/N_{f}endpoint
QED3_{3}45.32+0.90/Nf45.32+0.90/N_{f}endpoint

Table 1. Thermal ratios and status of the heat-capacity test. “Full” refers to the x>1x>1 response on the indicated branch; “endpoint” tests only the flat-space limit.

thermal curvature above the ordinary collider bound. At large xx, the conformal coupling cancels the H2H^{2} spectral gap, leaving a gapless bosonic spectrum and the logarithm in Eq. (29). Interactions generate a thermal mass in the large-NN bosonic examples, removing the latter effect.

The CS examples in Eq. (36) show why the local collider argument, Eq. (19), cannot determine the global bound: two theories with identical leading CTC_{T} and separated-point ⟨TTT⟩\langle TTT\rangle can have different thermal coefficients. Thus, control of the full thermal response requires information beyond those local data.

Thermodynamic positivity gives Ψ′(x)≥0\Psi'(x)\geq0 because heat capacity is a Hamiltonian variance, but does not bound this positive quantity from above in units of CTC_{T}. The cycle-exchange interpretation in Eq. (8) suggests investigating whether reflection positivity, modular covariance, or a vacuum variational principle on S1×RS^{1}\times\mathbb{R} can constrain the normalized Casimir amplitude.

The stronger pointwise conjecture is vulnerable to changes of thermal phase: a first-order transition at x>1x>1 with positive latent heat would produce a positive delta function in Ψ′(x)\Psi'(x) and violate Eq. (13), even if the flat-space bound survives. To our knowledge, no such transition has been established for a unitary three-dimensional CFT on S1×H2S^{1}\times H^{2} with vacuum Rényi boundary conditions. Further tests could include a first-principles CS-matter calculation on this space.

The flat-space bound may extend to d≥3d\geq3. Near four dimensions, every real multiscalar fixed point with a stable symmetric thermal state obeys its at leading O(ε)O(\varepsilon) (Supplemental Material). Static Matsubara modes give an opposing O(ε3/2)O(\varepsilon^{3/2}) correction, which is subleading at sufficiently small ε\varepsilon. More seriously, persistent thermal symmetry breaking can contribute with the opposite sign already at O(ε)O(\varepsilon) [37]. At the ordered O(1)×O(n)O(1)\times O(n) biconical fixed points, this contribution cancels only part of the O(ε)O(\varepsilon) reduction, so the bound still holds at this order for every finite nn (Supplemental Material). The cancellation becomes complete as n→∞n\to\infty: an emergent flat direction leaves the thermal coefficient equal to that of free scalars at leading order, though the CFT remains interacting [37, 38]. No general inequality for thermally ordered phases, or higher-dimensional hyperbolic envelope, has been established.

Acknowledgments

This work was supported by STFC grants ST/X000753/1 and UKRI14231. As is supported by the Royal Society under grant URF\R1\211417. OpenAI ChatGPT (GPT-5.6, 6.1 Sol and 6 Astra), Anthropic Claude (Opus 5.5, Fable 5.1) assisted with literature synthesis, proof attempts, cross-checks, code, figures, and revision. The authors directed and checked the analysis and take responsibility for the manuscript.

Data availability. Published Monte Carlo and bootstrap inputs are cited. Numerical data and code for the calculations performed here are available from the authors.

References

Supplemental Material

Methods

The idea for this Letter emerged from the following prompt that Herzog issued to GPT-5.6 Sol and an analogous one to Claude Opus 5.5:

I would like you to write for me a short article that could be published in Physical Review Letters. I would like you to choose the question, do the analysis and write up the results. The constraints I put on the question are 1) that it should be in the hep-th set of topics (so that I have a hope of evaluating the arguments and results); 2) the answer can be written up in a short format of 5 pages or less; 3) the result is of significance enough that it can be published in PRL. I want to break up the task into several steps. The first is to find a good question, so give me three candidates. I’m also going to involve Claude in the process. Claude is going to find three candidates too, and the next step will be to evaluate each others’ candidate questions.

The resulting six ideas were then rank ordered by the agents and consensus was reached. Ideas one and three were tossed out after initial explorations, and the second idea [main text, (1)], proposed by Claude, was used for the letter. After ChatGPT pointed out the relevance of the sphere free energy conjecture [2], Claude proposed the extended conjecture on S1×H2S^1 \times H^2.

All of the calculations were carried out by agents, and the initial draft of the manuscript was also prepared by agents, within a few hours. The resulting prose was often insufficiently detailed, and key references were often missing. GPT-6 Astra and Claude Fable 5.1 were recruited to check the examples, generate a few further examples and try, unsuccessfully, to find a proof of the conjecture. After two weeks of work, we verified all of the computations by hand or using Mathematica, clarified many of the explanations, and suggested a few key additional details. For example, we suggested the inclusion of [17], the conjecture in terms of CFT data ((S3–S6)) and the section on multiscalar theories.

Despite the essential use of AI in all aspects of manuscript preparation, we feel that the result is reliable and of sufficient interest to be published.

Thermal OPE and spectral reformulations

For a scalar primary O\mathcal{O} of dimension Δ\Delta, its thermal two-point function on Sβ1×R2S^1_\beta\times\mathbb{R}^2 admits an OPE in which the stress tensor appears with a coefficient conventionally denoted aTa_T. The Ward identity fixes the zero-temperature OOT\mathcal{O}\mathcal{O}T coupling, while the finite-temperature one-point function of TμνT_{\mu\nu} is determined by the thermal pressure. In the conventions of Ref. [39], in d=3d=3 one obtains

aT=Δ4πcSCT.(38)a_T = \frac{\Delta}{4\pi}\frac{c_S}{C_T}. \tag*{(38)}

The conjectured bound cS/CT≤16πζ(3)c_S/C_T \leq16\pi\zeta(3) is therefore equivalent to

aT≤4ζ(3)Δ.(39)a_T \leq4\zeta(3)\Delta. \tag*{(39)}

For the free scalar with Δ=1/2\Delta= 1/2, the image sum on the thermal circle gives coefficients proportional to ζ(J+1)\zeta(J+1); the spin-two term reproduces aT=2ζ(3)a_T = 2\zeta(3) and saturates this bound. This is not a proof: generic thermal OPE coefficients are not sign definite, and the discontinuity entering the thermal Lorentzian inversion formula [39] does not provide an obvious positive measure.

The thermal coefficient is itself spectral CFT data. Place the theory on Sβ1×S2S^1_\beta\times S^2, with S2S^2 of unit radius. Its finite-temperature one-point functions are Boltzmann-weighted sums of diagonal matrix elements over local operators [40, 41]. For the integrated stress tensor, the Ward identity relates this one-point function to the Hamiltonian; only the energy spectrum survives. In the high-temperature limit,

cS=34πlim⁡β→0β2log⁡∑Oe−βΔO,(40)c_S = \frac{3}{4\pi}\lim_{\beta\to0}\beta^2 \log\sum_{\mathcal{O}} e^{-\beta\Delta_{\mathcal{O}}}, \tag*{(40)}

where the sum includes all local operators, descendants, and spin components, with multiplicity. Thus, the flat-space conjecture can be written as

lim⁡β→0β2log⁡∑Oe−βΔO≤64π2ζ(3)3CT=2ζ(3)CTCT,s.(41)\lim_{\beta\to0}\beta^2\log\sum_{\mathcal{O}}e^{-\beta\Delta_{\mathcal{O}}}\leq\frac{64\pi^2\zeta(3)}{3}C_T=2\zeta(3)\frac{C_T}{C_{T,s}}. \tag*{(41)}

Descendants multiply the contribution of a primary by at most (1−e−β)−3(1-e^{-\beta})^{-3}, changing the logarithm only by O(log⁡β)O(\log\beta). Accordingly, the leading bound is unchanged if the sum is restricted to primaries, weighted by their spin degeneracy 2J+12J+1.

Let N(E)N(E) count local operators, with multiplicity, of scaling dimension at most EE. Positivity gives Z(β)≥N(E)e−βEZ(\beta)\geq N(E)e^{-\beta E}, and hence

log⁡N(E)≤inf⁡β>0[βE+log⁡Z(β)],Z(β)=∑Oe−βΔO.(42)\log N(E)\leq\inf_{\beta>0}\left[\beta E+\log Z(\beta)\right],\qquad Z(\beta)=\sum_{\mathcal{O}}e^{-\beta\Delta_{\mathcal{O}}}. \tag*{(42)}

If log⁡Z(β)∼A/β2\log Z(\beta)\sim A/\beta^2, the minimizing β\beta is asymptotically (2A/E)1/3(2A/E)^{1/3}, giving a coefficient 32(2A)1/3\frac{3}{2}(2A)^{1/3} for E2/3E^{2/3}. Conversely, the Laplace representation Z(β)=β∫0∞e−βEN(E) dEZ(\beta)=\beta\int_0^\infty e^{-\beta E}N(E)\,\mathrm{d}E recovers the corresponding upper bound on AA by a saddle-point estimate. Equation (S4) is consequently equivalent, at the level of upper asymptotic growth bounds, to

lim sup⁡E→∞log⁡N(E)E2/3≤32(4ζ(3)CTCT,s)1/3.(43)\limsup_{E\to\infty}\frac{\log N(E)}{E^{2/3}}\leq\frac{3}{2}\left(4\zeta(3)\frac{C_T}{C_{T,s}}\right)^{1/3}. \tag*{(43)}

Thus, no unitary CFT is conjectured to have more rapid asymptotic growth of local operators at fixed CTC_T than the free scalar. In two dimensions, modular invariance fixes the analogous growth in terms of the central charge. No corresponding modular constraint on S1×S2S^1\times S^2 is presently known.

Planar Einstein gravity with scalar hair

Consider four-dimensional Einstein gravity with scalar fields ϕa\phi^a and positive-definite sigma-model metric GabG_{ab}, normalized by

S=116πG∫d4x−g[R−12Gab(ϕ)∂μϕa∂μϕb−V(ϕ)].(44)S=\frac{1}{16\pi G}\int\mathrm{d}^4x\sqrt{-g}\left[R-\frac{1}{2}G_{ab}(\phi)\partial_\mu\phi^a\partial^\mu\phi^b-V(\phi)\right]. \tag*{(44)}

Assume a source-free asymptotically AdS vacuum of radius LL. For a homogeneous planar black brane write ds2=e2A(r)[−h(r)dt2+dx⃗ 2]+dr2/h(r)\mathrm{d}s^2=e^{2A(r)}[-h(r)\mathrm{d}t^2+\mathrm{d}\vec{x}^{\,2}]+\mathrm{d}r^2/h(r), with ϕa=ϕa(r)\phi^a=\phi^a(r) and primes denoting rr derivatives. The radial Einstein equations imply

A′′=−14Gabϕa′ϕb′≤0,h′′+3A′h′=0.(45)A^{\prime\prime}=-\frac{1}{4}G_{ab}\phi^{a\prime}\phi^{b\prime}\leq0,\qquad h^{\prime\prime}+3A^\prime h^\prime=0. \tag*{(45)}

The first relation is the four-dimensional scalar specialization of the warp-factor concavity in Eqs. (4.5) and (4.8), (4.9) of Freedman et al. [42]; the second and the finite-temperature generalization of the first are analogous to Eqs. (25a), (25b) of Gubser and Nellore [43], which use five bulk dimensions and a different radial gauge. In particular, (e3Ah′)′=0(e^{3A}h^\prime)^\prime=0. With A′(∞)=1/LA^\prime(\infty)=1/L, the former implies A(r)≥Ah+(r−rh)/LA(r)\geq A_h+(r-r_h)/L. If J=∫rh∞e−3A(r)drJ=\int_{r_h}^{\infty}e^{-3A(r)}\mathrm{d}r, then J≤Le−3Ah/3J\leq Le^{-3A_h}/3, and

T=e−2Ah4πJ≥3eAh4πL,s=e2Ah4G.(46)T=\frac{e^{-2A_h}}{4\pi J}\geq\frac{3e^{A_h}}{4\pi L},\qquad s=\frac{e^{2A_h}}{4G}. \tag*{(46)}

Consequently cS=s/T2≤4π2L2/(9G)c_S=s/T^2\leq4\pi^2L^2/(9G). Using the Einstein value CT=3L2/(π3G)C_T=3L^2/(\pi^3G) yields

cSCT≤4π527.(47)\frac{c_S}{C_T}\leq\frac{4\pi^5}{27}. \tag*{(47)}

Equality requires vanishing scalar gradients. This concerns only regular, isotropic, source-free planar solutions and the flat-space endpoint, not hyperbolic-horizon transitions.

Analytic hyperbolic envelope for free and large-NN fields

We give analytic arguments for the full-curve comparisons in the Letter, first for free fields and Einstein holography, and then for interacting scalar saddles.

Free Dirac versus free scalar. Set wB(z)=z3ez/(ez−1)2w_B(z)=z^3e^z/(e^z-1)^2 and wF(z)=z3ez/(ez+1)2w_F(z)=z^3e^z/(e^z+1)^2. Differentiating the free-field spectral integrals gives

Ψs′(x)=16xπ4∫0∞dz wB(z)tanh⁡xz2,ΨF′(x)=16xπ4∫0∞dz wF(z)coth⁡xz2.(48)\Psi_s'(x)=\frac{16x}{\pi^4}\int_0^\infty dz\,w_B(z)\tanh\frac{xz}{2},\qquad\Psi_F'(x)=\frac{16x}{\pi^4}\int_0^\infty dz\,w_F(z)\coth\frac{xz}{2}. \tag*{(48)}

Writing t=tanh⁡(xz/2)t=\tanh(xz/2) and t1=tanh⁡(z/2)t_1=\tanh(z/2), one finds

Ψs′(x)−ΨF′(x)=4xπ4∫0∞dz z3(t2−t12)sinh⁡2(z/2)t>0(x>1),(49)\Psi_s'(x)-\Psi_F'(x)=\frac{4x}{\pi^4}\int_0^\infty dz\,\frac{z^3(t^2-t_1^2)}{\sinh^2(z/2)t}>0\qquad(x>1), \tag*{(49)}

using only that tanh⁡\tanh is increasing.

Einstein holography versus free scalar. Let as=96ζ(3)/π4=1.1847…a_s=96\zeta(3)/\pi^4=1.1847\ldots be the asymptotic slope of Ψs′\Psi_s'. Since ∫0∞wB dz=6ζ(3)\int_0^\infty w_B\,dz=6\zeta(3), writing tanh⁡(xz/2)=1−2/(exz+1)\tanh(xz/2)=1-2/(e^{xz}+1) and substituting v=xzv=xz gives

Ψs′(x)=asx−32π4∫0∞dv wB(v/x)ev+1.(50)\Psi_s'(x)=a_sx-\frac{32}{\pi^4}\int_0^\infty dv\,\frac{w_B(v/x)}{e^v+1}. \tag*{(50)}

Differentiating and returning to z=v/xz=v/x,

Ψs′′(x)=as+32π4G(x),G(x)=∫0∞dz zwB′(z)exz+1.(51)\Psi_s''(x)=a_s+\frac{32}{\pi^4}G(x),\qquad G(x)=\int_0^\infty dz\,\frac{zw_B'(z)}{e^{xz}+1}. \tag*{(51)}

Since (log⁡wB)′=3/z−coth⁡(z/2)(\log w_B)'=3/z-\coth(z/2) decreases strictly, wBw_B has a single maximum at some z∗z_*. So zwB′zw_B' is positive for z<z∗z<z_* and negative for z>z∗z>z_*. For x≥1x\ge1 the ratio r(z)=(ez+1)/(exz+1)r(z)=(e^z+1)/(e^{xz}+1) is positive and non-increasing, because ey/(ey+1)e^y/(e^y+1) increases with yy. Splitting the integral at z∗z_* therefore gives G(x)≥r(z∗)G(1)G(x)\ge r(z_*)G(1). Direct evaluation gives G(1)=19π4/480−3ζ(3)=0.2496…G(1)=19\pi^4/480-3\zeta(3)=0.2496\ldots, which is equivalent to Ψs′′(1)=19/15\Psi_s''(1)=19/15. Hence Ψs′′(x)>as\Psi_s''(x)>a_s for x≥1x\ge1, and integrating from Ψs′(1)=1\Psi_s'(1)=1 gives

Ψs′(x)≥1+as(x−1),x≥1.(52)\Psi_s'(x)\ge1+a_s(x-1),\qquad x\ge1. \tag*{(52)}

For Einstein holography ΨEin=y2\Psi_{\mathrm{Ein}}=y^2 with y=(x+x2+3)/3y=(x+\sqrt{x^2+3})/3. Setting x=3sinh⁡θx=\sqrt{3}\sinh\theta and v=e−2θv=e^{-2\theta} gives

ΨEin′′(x)=891+3v(1+v)3≤89,(53)\Psi_{\mathrm{Ein}}''(x)=\frac{8}{9}\frac{1+3v}{(1+v)^3}\le\frac{8}{9}, \tag*{(53)}

because (1+v)3≥1+3v(1+v)^3\ge1+3v. Integrating from ΨEin′(1)=1\Psi_{\mathrm{Ein}}'(1)=1 gives ΨEin′(x)≤1+89(x−1)\Psi_{\mathrm{Ein}}'(x)\le1+\frac{8}{9}(x-1). Since as>8/9a_s>8/9,

Ψs′(x)−ΨEin′(x)≥(as−89)(x−1)>0,x>1.(54)\Psi_s'(x)-\Psi_{\mathrm{Ein}}'(x)\ge\left(a_s-\frac{8}{9}\right)(x-1)>0,\qquad x>1. \tag*{(54)}

This assumes that the hyperbolic black hole without scalar hair is the dominant saddle for x≥1x\ge1.

Interacting scalar saddles and monotone thermal mass. For the critical O(N)O(N) model and the large-NN sextic scalar line we assume that the homogeneous, vacuum-connected saddle gives the leading-NN thermodynamics on S1×H2S^1\times H^2 for x≥1x\ge1, with the standard replica boundary conditions. The following argument establishes the envelope on those saddles, not the exclusion of competing phases.

Write ω=p2+m2\omega=\sqrt{p^2+m^2}. The subtracted tadpole integral in the Letter’s critical-O(N)O(N) gap equation is

I(m,x)=∫0∞dp[ptanh⁡(πp)ωcoth⁡(πωx)−1].(55)\mathcal{I}(m,x)=\int_0^\infty dp\left[\frac{p\tanh(\pi p)}{\omega}\coth\left(\frac{\pi\omega}{x}\right)-1\right]. \tag*{(55)}

The critical model obeys I(m,x)=0\mathcal{I}(m,x)=0, whereas on the sextic line m=αI(m,x)m=\alpha\mathcal{I}(m,x) with 0≤α≤10\le\alpha\le1. For m>0m>0, differentiation at fixed mm or xx, respectively, gives

∂xI=πx2∫0∞dp ptanh⁡(πp)csch⁡2(πωx)>0,(56)\partial_x\mathcal{I}=\frac{\pi}{x^2}\int_0^\infty dp\,p\tanh(\pi p)\operatorname{csch}^2\left(\frac{\pi\omega}{x}\right)>0, \tag*{(56)}
∂mI=−m∫0∞dp ptanh⁡(πp)[coth⁡(πω/x)ω3+πxω2csch⁡2(πωx)]<0.(57)\partial_m\mathcal{I}=-m\int_0^\infty dp\,p\tanh(\pi p)\left[\frac{\coth(\pi\omega/x)}{\omega^3}+\frac{\pi}{x\omega^2}\operatorname{csch}^2\left(\frac{\pi\omega}{x}\right)\right]<0. \tag*{(57)}

Thus, I\mathcal{I} increases with xx and decreases with mm; the latter monotonicity also holds at m=0m = 0 by continuity. At x=1x = 1, I(0,1)=0\mathcal{I}(0, 1) = 0 because tanh⁡(πp)coth⁡(πp)=1\tanh(\pi p)\coth(\pi p) = 1. For x>1x > 1, I(0,x)>0\mathcal{I}(0, x) > 0, since the hyperbolic cotangent is strictly decreasing in its argument. Finally, at fixed xx, I(m,x)=−m+O(m−1)\mathcal{I}(m, x) = -m + O(m^{-1}) as m→∞m \to\infty. Indeed, separating tanh⁡(πp)=1+[tanh⁡(πp)−1]\tanh(\pi p) = 1 + [\tanh(\pi p) - 1] and coth⁡u=1+2/(e2u−1)\coth u = 1 + 2/(e^{2u} - 1), the first term integrates to ∫0∞(p/ω−1) dp=−m\int_0^\infty(p/\omega- 1)\,dp = -m; the curvature correction is −1/(24m)+O(m−3)-1/(24m) + O(m^{-3}), while the thermal correction is exponentially small at fixed xx. The critical gap equation therefore has a unique root m(x)>0m(x) > 0; the sextic equation has a unique root for each α>0\alpha> 0, since m−αI(m,x)m - \alpha\mathcal{I}(m, x) increases strictly from a negative value to a positive one. Implicit differentiation now establishes

mcrit′(x)=−∂xI∂mI>0,mα′(x)=α∂xI1−α∂mI>0,(x>1).(58)m_{\mathrm{crit}}'(x) = -\frac{\partial_x\mathcal{I}}{\partial_m\mathcal{I}} > 0,\qquad m_\alpha'(x) = \frac{\alpha\partial_x\mathcal{I}}{1-\alpha\partial_m\mathcal{I}} > 0,\qquad(x > 1). \tag*{(58)}

The m→0+m \to0^+ derivative requires care: the integral is not uniformly differentiable at p=m=0p = m = 0. Its one-sided expansion near (m,x)=(0,1)(m, x) = (0, 1) is given below in the sextic-curvature section.

For z=2πp2+m2/xz = 2\pi\sqrt{p^2 + m^2}/x, the single-mode heat capacity cB(z)=z2/[4sinh⁡2(z/2)]c_B(z) = z^2/[4\sinh^2(z/2)] is strictly decreasing in z>0z > 0. Indeed, dlog⁡cB/dz=2/z−coth⁡(z/2)<0d\log c_B/dz = 2/z - \coth(z/2) < 0. The bosonic entropy ΦB(x;m)\Phi_B(x; m) defined in the Letter obeys

∂xΦB=1x∫0∞dp2π ptanh⁡(πp)cB(z),(59)\partial_x\Phi_B = \frac{1}{x}\int_0^\infty\frac{dp}{2\pi}\,p\tanh(\pi p)c_B(z), \tag*{(59)}
∂mΦB=−m∫0∞dp2πptanh⁡(πp)p2+m2cB(z)≤0.(60)\partial_m\Phi_B = -m\int_0^\infty\frac{dp}{2\pi}\frac{p\tanh(\pi p)}{p^2 + m^2}c_B(z) \le0. \tag*{(60)}

At fixed xx, ∂xΦB(x;m)≤∂xΦB(x;0)\partial_x\Phi_B(x; m) \le\partial_x\Phi_B(x; 0). Therefore, along the saddle,

ddxΦB(x;m(x))=∂xΦB+m′(x)∂mΦB≤∂xΦB(x;0).(61)\frac{d}{dx}\Phi_B(x; m(x)) = \partial_x\Phi_B + m'(x)\partial_m\Phi_B \le\partial_x\Phi_B(x; 0). \tag*{(61)}

With CT=NCT,s+O(N0)C_T = NC_{T,s} + O(N^0), this is precisely Ψ′(x)≤Ψs′(x)\Psi'(x) \le\Psi_s'(x) for x>1x > 1, with strict inequality for positive interaction. At α=0\alpha= 0 the free scalar saturates.

Curvature along the large-NN sextic scalar line

The strict large-NN O(N)O(N) sextic theory with interaction g6(ϕ⋅ϕ)3/(6N2)g_6(\phi\cdot\phi)^3/(6N^2) has a stable conformal line 0≤α=g6/(4π)≤10 \le\alpha= \sqrt{g_6}/(4\pi) \le1. Let mm be the hyperbolic dimensionless thermal mass, determined for x≥1x \ge1 by m=αI(m,x)m = \alpha\mathcal{I}(m, x) with I\mathcal{I} defined in Eq. (55). Near the conformal temperature, set δ=x−1>0\delta= x - 1 > 0. The joint small-mm, small-δ\delta expansion on this branch is

I(m,1+δ)=π8δ−π2m+o(∣δ∣+m).(62)\mathcal{I}(m, 1 + \delta) = \frac{\pi}{8}\delta- \frac{\pi}{2}m + o(|\delta| + m). \tag*{(62)}

Consequently,

m=πα4(2+πα)δ+o(δ).(63)m = \frac{\pi\alpha}{4(2 + \pi\alpha)}\delta+ o(\delta). \tag*{(63)}

To see the resulting change in entropy, the derivative of ΦB\Phi_B above gives

−lim⁡m→0+∂mΦB(1;m)m=∫0∞dp2πtanh⁡(πp)pcB(2πp)=π16.(64)-\lim_{m\to0^+}\frac{\partial_m\Phi_B(1; m)}{m} = \int_0^\infty\frac{dp}{2\pi}\frac{\tanh(\pi p)}{p}c_B(2\pi p) = \frac{\pi}{16}. \tag*{(64)}

Using Ψ=(128/π)ΦB\Psi= (128/\pi)\Phi_B at leading NN therefore gives Ψα(x)−Ψs(x)=−4m2+o(δ2)\Psi_\alpha(x) - \Psi_s(x) = -4m^2 + o(\delta^2). It follows that

Ψα′′(1+)=1915−π2α22(2+πα)2.(65)\Psi_\alpha''(1^+) = \frac{19}{15} - \frac{\pi^2\alpha^2}{2(2 + \pi\alpha)^2}. \tag*{(65)}

At the Bardeen–Moshe–Bander endpoint α=1\alpha= 1,

ΨBMB′′(1+)=1915−π22(2+π)2=1.0799967….(66)\Psi_{\mathrm{BMB}}''(1^+) = \frac{19}{15} - \frac{\pi^2}{2(2 + \pi)^2} = 1.0799967\ldots. \tag*{(66)}

The curvature decreases monotonically along the sextic line, despite the presence of a dimension-one scalar bilinear at leading large NN.

Next-to-leading flat-space endpoint checks

To combine published thermal free energies with stress-tensor normalizations, we use s=cST2s=c_{S}T^{2} and f=−cST3/3f=-c_{S}T^{3}/3. The next-to-leading thermal coefficients are given in Refs. [29, 30], while CTC_{T} is given in Refs. [7, 31, 32]. For the scalar O(N)O(N) Wilson–Fisher CFT,

cSCT=48.3376+1.5557N+O(N−2).(67)\frac{c_{S}}{C_{T}}=48.3376+\frac{1.5557}{N}+O(N^{-2}). \tag*{(67)}

For the critical Gross–Neveu and zero-CS-level conformal QED3_3 theories, let NfN_{f} count two-component complex Dirac fermions. The fermionic CTC_{T} formulas of Refs. [31, 32] use N=2NfN=2N_{f}. Combining the results yields

cSCT∣GN=45.3165+0.0755Nf+O(Nf−2),cSCT∣QED3=45.3165+0.8994Nf+O(Nf−2).(68)\left.\frac{c_{S}}{C_{T}}\right|_{\mathrm{GN}}=45.3165+\frac{0.0755}{N_{f}}+O(N_{f}^{-2}),\qquad\left.\frac{c_{S}}{C_{T}}\right|_{\mathrm{QED}_{3}}=45.3165+\frac{0.8994}{N_{f}}+O(N_{f}^{-2}). \tag*{(68)}

These are endpoint checks in the controlled large-flavor regime, not finite-NfN_{f} bounds or hyperbolic-response calculations.

Multiscalar fixed points in d=4−εd=4-\varepsilon

The flat-space conjecture is not necessarily special to three dimensions. With s=cSTd−1s=c_{S}T^{d-1}, the free real scalar has (cS/CT)s=4(d−1)πd/2ζ(d)/Γ(d/2)(c_{S}/C_{T})_{s}=4(d-1)\pi^{d/2}\zeta(d)/\Gamma(d/2) (equal to 2π6/152\pi^{6}/15 in d=4d=4). We ask whether this free-scalar value bounds cS/CTc_{S}/C_{T} for unitary CFTs in d≥3d\geq3, and test this possibility near four dimensions using general quartic scalar fixed points. Two effects can oppose the conjectured inequality. One is an infrared contribution at O(ε3/2)O(\varepsilon^{3/2}) from the static Matsubara modes, defined below. The other is thermal ordering, already at O(ε)O(\varepsilon). Our considerations in this section concern only the flat-space coefficient, not the full hyperbolic-space envelope.

For NN real scalars, let V=16π24!λijklϕiϕjϕkϕlV=\frac{16\pi^{2}}{4!}\lambda_{ijkl}\phi_{i}\phi_{j}\phi_{k}\phi_{l}, with real, fully symmetric λijkl\lambda_{ijkl}. The one-loop fixed-point equation is [44, 45]

ελijkl=λijmnλmnkl+λikmnλmnjl+λilmnλmnjk.(69)\varepsilon\lambda_{ijkl}=\lambda_{ijmn}\lambda_{mnkl}+\lambda_{ikmn}\lambda_{mnjl}+\lambda_{ilmn}\lambda_{mnjk}. \tag*{(69)}

The leading interaction correction to the flat-space thermal free-energy density is 2π2λiijj(T2/12)22\pi^{2}\lambda_{iijj}(T^{2}/12)^{2} [46]. Relative to NN free scalars in the same dimension, the symmetric thermal saddle gives

cScS,free=1−54Nλiijj+O(ε3/2).(70)\frac{c_{S}}{c_{S,\mathrm{free}}}=1-\frac{5}{4N}\lambda_{iijj}+O(\varepsilon^{3/2}). \tag*{(70)}

At one loop the fields acquire a thermal mass matrix Mij2=2π2T23BijM_{ij}^{2}=\frac{2\pi^{2}T^{2}}{3}B_{ij}, where Bij=λijkkB_{ij}=\lambda_{ijkk}. Equation (S33) presumes that BB is positive semidefinite, B⪰0B\succeq0, so that the symmetric saddle is locally stable. Since CT/(NCT,s)=1+O(ε2)C_{T}/(NC_{T,s})=1+O(\varepsilon^{2}) [45], the same leading correction applies to (cS/CT)/(cS/CT)free(c_{S}/C_{T})/(c_{S}/C_{T})_{\mathrm{free}}. The trace λiijj\lambda_{iijj} is also the sum of the one-loop anomalous dimensions of scalar bilinears [47, 48].

Contracting Eq. (69) yields the positive sum

ελiijj=BijBij+2λijklλijkl>0(71)\varepsilon\lambda_{iijj}=B_{ij}B_{ij}+2\lambda_{ijkl}\lambda_{ijkl}>0 \tag*{(71)}

for every non-Gaussian real fixed point. Moreover, BijBij≥(λiijj)2/NB_{ij}B_{ij}\geq(\lambda_{iijj})^{2}/N and λijklλijkl≥3(λiijj)2/[N(N+2)]\lambda_{ijkl}\lambda_{ijkl}\geq3(\lambda_{iijj})^{2}/[N(N+2)]. Thus λiijj≤N(N+2)ε/(N+8)\lambda_{iijj}\leq N(N+2)\varepsilon/(N+8) [47, 48], saturated by the O(N)O(N) Wilson–Fisher fixed point. Consequently,

1−5N+24N+8ε≤cScS,free<1at leading order in ε.(72)1-\frac{5N+2}{4N+8}\varepsilon\leq\frac{c_{S}}{c_{S,\mathrm{free}}}<1\qquad\text{at leading order in }\varepsilon. \tag*{(72)}

The free scalar therefore maximizes the thermal coefficient among symmetric-saddle fixed points to this order, while the O(N)O(N) model has the largest leading reduction.

The next correction has the opposite sign. It comes from the static Matsubara modes, which are the components of ϕi\phi_i that are constant around the thermal circle. These modes are massless at tree level, so M2M^{2} is their entire mass. When B⪰0B\succeq0 they contribute +563tr⁡B3/2+\frac{5\sqrt{6}}{3}\operatorname{tr}B^{3/2} to Eq. (70). For the O(N)O(N) model, comparison of the periodic Casimir amplitude in Eq. (19) of Ref. [49] with the free amplitude in the same dimension gives

cSO(N)cS,free=1−54qε+563(qε)3/2+O(ε2),q=N+2N+8.(73)\frac{c_{S}^{O(N)}}{c_{S,\mathrm{free}}}=1-\frac{5}{4}q\varepsilon+\frac{5\sqrt{6}}{3}(q\varepsilon)^{3/2}+O(\varepsilon^{2}),\qquad q=\frac{N+2}{N+8}. \tag*{(73)}

The dimension-dependent free-field terms in that reference cancel upon taking the ratio. At fixed NN, the positive O(ϵ3/2)O(\epsilon^{3/2}) term cannot overwhelm the negative O(ϵ)O(\epsilon) term for sufficiently small ϵ\epsilon. Extrapolation to ϵ=1\epsilon= 1 is unreliable: the Ising truncation gives 1−0.42+0.79≃1.371 - 0.42 + 0.79 \simeq1.37, whereas the three-dimensional Monte Carlo value is cS/cS,free≃0.800c_S/c_{S,\mathrm{free}} \simeq0.800 [33].

Thermal order. A negative eigenvalue of BB makes the symmetric saddle unstable, and the dimensionally reduced theory on Rd−1\mathbb{R}^{d-1} may then enter an ordered phase [37, 50]. The zero-temperature quartic potential is still nonnegative at a real one-loop fixed point, since Eq. (69) implies

ϵλijklvivjvkvl=3(λijmnvivj)(λklmnvkvl)≥0(74)\epsilon\lambda_{ijkl}v_i v_j v_k v_l = 3(\lambda_{ijmn}v_i v_j)(\lambda_{klmn}v_k v_l) \ge0 \tag*{(74)}

for all real viv_i [44]. The leading free energy must be evaluated at the thermal minimum [37, 46]:

cScS,free=1−54Nλiijj−30NW+O(ϵ3/2),W=min⁡ϕU(ϕ)≤0.(75)\frac{c_S}{c_{S,\mathrm{free}}} = 1 - \frac{5}{4N}\lambda_{iijj} - \frac{30}{N}W + O(\epsilon^{3/2}), \qquad W = \min_{\phi} U(\phi) \le0. \tag*{(75)}

Here U(ϕ)=Bijϕiϕj/T2+2λijklϕiϕjϕkϕl/T4U(\phi) = B_{ij}\phi_i\phi_j/T^2 + 2\lambda_{ijkl}\phi_i\phi_j\phi_k\phi_l/T^4, and π23T4U(ϕ)\frac{\pi^2}{3}T^4U(\phi) is the leading effective potential of the static modes. Ordering increases cSc_S already at O(ϵ)O(\epsilon), so the argument leading to Eq. (72) does not cover ordered phases. (75) can be written as

cScS,free=1−54Nλiijj(1−Q)+O(ϵ3/2),Q=−24Wλiijj.(76)\frac{c_S}{c_{S,\mathrm{free}}} = 1 - \frac{5}{4N}\lambda_{iijj}(1-Q) + O(\epsilon^{3/2}), \qquad Q = -\frac{24W}{\lambda_{iijj}}. \tag*{(76)}

Here QQ is the fraction of the leading interaction correction that ordering cancels. So cSc_S stays below its free value at O(ϵ)O(\epsilon) if and only if Q<1Q < 1.

As an example, split the N=n+1N = n + 1 scalars into a single scalar σ\sigma and an O(n)O(n) vector ϕ\phi. The O(1)×O(n)O(1) \times O(n)-invariant quartic potential is

V(σ,ϕ)=2π2ϵ[aσ4+b(ϕ2)2+2gσ2ϕ2].(77)V(\sigma,\phi) = 2\pi^2\epsilon\left[a\sigma^4 + b(\phi^2)^2 + 2g\sigma^2\phi^2\right]. \tag*{(77)}

The couplings (a,b,g)(a,b,g) are the (α,β,γ)(\alpha,\beta,\gamma) of Ref. [37] with m=1m = 1, and Eq. (69) reduces to their Eqs. (3.6)–(3.8). The matrix BB is diagonal, with entry ϵb1\epsilon b_1 for σ\sigma and ϵb2\epsilon b_2 for each component of ϕ\phi, where b1=3a+ngb_1 = 3a + ng and b2=(n+2)b+gb_2 = (n + 2)b + g. For n≥18n \ge18 the only fixed point with g≠0g \ne0, apart from the O(n+1)O(n + 1) one, is the biconical fixed point. It has b1<0<b2b_1 < 0 < b_2. So σ\sigma condenses and breaks the Z2\mathbb{Z}_2 symmetry σ→−σ\sigma\to-\sigma at every temperature. The minimum of UU lies at ϕ=0\phi= 0 and

σ2=−b1T212a,\sigma^2 = -\frac{b_1T^2}{12a},

so W=−ϵb12/(24a)W = -\epsilon b_1^2/(24a). With λiijj=tr⁡B=ϵ(b1+nb2)\lambda_{iijj} = \operatorname{tr} B = \epsilon(b_1 + nb_2), this gives Q=b12/[a(b1+nb2)]Q = b_1^2/[a(b_1 + nb_2)]. The fixed-point equations simplify this to

1−Q=ϵ[6(n−1)a−8ng]λiijj.(78)1 - Q = \frac{\epsilon[6(n - 1)a - 8ng]}{\lambda_{iijj}}. \tag*{(78)}

They also give a(1−9a)=ng2a(1 - 9a) = ng^2, so a>0a > 0. Then b1<0b_1 < 0 requires g<0g < 0. Since λiijj>0\lambda_{iijj} > 0 by Eq. (71), we have Q<1Q < 1 for every nn. Numerically, Q≃0.512Q \simeq0.512 at n=100n = 100 and Q≃0.787Q \simeq0.787 at n=300n = 300. Solving the fixed-point equations for every nn from 18 to 10710^7 shows that QQ increases with nn. So ordering cancels a growing fraction of the correction, but never all of it at finite nn.

At large nn the fixed point approaches (a,b,g)∼(9,1,−3)/n(a,b,g) \sim(9,1,-3)/n, and Eq. (77) becomes

V(σ,ϕ)∼2π2ϵn(3σ2−ϕ2)2,(79)V(\sigma,\phi) \sim\frac{2\pi^2\epsilon}{n}(3\sigma^2 - \phi^2)^2, \tag*{(79)}

whose zero-temperature flat direction 3σ2=ϕ23\sigma^2 = \phi^2, related to the large-rank moduli space of Ref. [37], lets the condensate cancel the leading thermal masses of the nn vector fields, in line with 1−Q=7/n+O(n−2)1 - Q = 7/n + O(n^{-2}) from Eq. (78). The independent large-nn analysis of Ref. [38], Eq. (5.8) shows that the thermal coefficient then equals that of nn free scalars at leading order throughout 3<d<43 < d < 4, saturating the cS/cTc_S/c_T bound although the CFT remains interacting, and that next-to-leading corrections lift the flat direction and select the Z2\mathbb{Z}_2-ordered vacuum found above.

The dimensional reduction also cautions against continuing the ordered saddle to d=3d = 3. Continuous finite-temperature symmetry breaking is forbidden on flat R2\mathbb{R}^2, but discrete ordering remains possible. Smolkin and Yung find that the Z2\mathbb{Z}_2 ordering of the large-nn biconical model persists to arbitrarily high temperature in d=3d = 3 [38]. None of these flat-space results yet establishes an inequality for the response on S1×Hd−1S^1 \times H^{d-1}, where entangling-surface terms involving scalar bilinears may also require care.

Planar Chern–Simons-matter theories

Common planar endpoint bound. Combining Eqs. (7.1), (7.4), (7.6), (7.9), and (7.11) of Aharony et al. [25] with the planar stress-tensor normalization [26] gives, for each of the regular/critical boson and fermion theories,

cSCT=16πG(μ,φ),G(μ,φ)=112∫μ∞t(t2−μ2)cosh⁡t−cos⁡φ dt(80)\frac{c_S}{C_T}=16\pi\mathcal{G}(\mu,\varphi),\qquad\mathcal{G}(\mu,\varphi)=\frac{1}{12}\int_{\mu}^{\infty}\frac{t(t^2-\mu^2)}{\cosh t-\cos\varphi}\,dt \tag*{(80)}

where μ≥0\mu\geq0 is the magnitude of the dimensionless thermal mass. For bosons φ=π∣λ∣\varphi=\pi|\lambda|, while for fermions φ=π(1−∣λ∣)\varphi=\pi(1-|\lambda|). The integral is our rewriting of the polylogarithmic function in Eq. (7.1) of Ref. [25]:

G(μ,φ)=1sin⁡φIm⁡[μ23Li⁡2(z)+μLi⁡3(z)+Li⁡4(z)],z=e−μ+iφ.(81)\mathcal{G}(\mu,\varphi)=\frac{1}{\sin\varphi}\operatorname{Im}\left[\frac{\mu^2}{3}\operatorname{Li}_2(z)+\mu\operatorname{Li}_3(z)+\operatorname{Li}_4(z)\right],\qquad z=e^{-\mu+i\varphi}. \tag*{(81)}

This follows by inserting 1/(cosh⁡t−cos⁡φ)=(2/sin⁡φ)∑n≥1e−ntsin⁡(nφ)1/(\cosh t-\cos\varphi)=(2/\sin\varphi)\sum_{n\geq1}e^{-nt}\sin(n\varphi) and integrating term by term. We assume the standard planar uniform-holonomy thermodynamic saddle. For 0<φ<π0<\varphi<\pi the integrand is nonnegative, and the function G\mathcal{G} decreases when either μ\mu or φ\varphi is increased:

∂μG=−μ6∫μ∞t dtcosh⁡t−cos⁡φ≤0,(82)\partial_\mu\mathcal{G}=-\frac{\mu}{6}\int_{\mu}^{\infty}\frac{t\,dt}{\cosh t-\cos\varphi}\leq0, \tag*{(82)}
∂φG=−sin⁡φ12∫μ∞t(t2−μ2) dt(cosh⁡t−cos⁡φ)2<0.(83)\partial_\varphi\mathcal{G}=-\frac{\sin\varphi}{12}\int_{\mu}^{\infty}\frac{t(t^2-\mu^2)\,dt}{(\cosh t-\cos\varphi)^2}<0. \tag*{(83)}

As G(0,0)=ζ(3)\mathcal{G}(0,0)=\zeta(3), the endpoint bound cS/CT≤16πζ(3)c_S/C_T\leq16\pi\zeta(3) follows for all four theories, with equality only in the free-scalar limit. This is a flat-space endpoint result, not a calculation of the full S1×H2S^1\times H^2 response. It assumes the standard planar free energies and the appropriate physical thermal saddles; it does not exclude arbitrary off-shell or inhomogeneous phases.

Regular-boson sextic line. We illustrate the bound over two couplings in the regular-boson (RB) U(N)U(N) theory and check the physical saddle using the phase analysis of Dey et al. [27]. The scalar interaction is

δL=λ66N2(ϕ†ϕ)3,λ=Nk,0≤λ<1,(84)\delta\mathcal{L}=\frac{\lambda_6}{6N^2}(\phi^\dagger\phi)^3,\qquad\lambda=\frac{N}{k},\qquad0\leq\lambda<1, \tag*{(84)}

where kk is the renormalized Chern–Simons level. We use parity to restrict to nonnegative λ\lambda. At strict N=∞N=\infty the sextic coupling is marginal. We work at the conformal point with vanishing relevant mass and quartic deformations.

Thermal mass and ungauged limit. Set

λ^=(λ2+λ68π2)1/2,μ≡MT,z=e−μ+iπλ.(85)\widehat{\lambda}=\left(\lambda^2+\frac{\lambda_6}{8\pi^2}\right)^{1/2},\qquad\mu\equiv\frac{M}{T},\qquad z=e^{-\mu+i\pi\lambda}. \tag*{(85)}

Here MM denotes the self-consistent thermal mass of the fundamental boson in the unHiggsed saddle on Sβ1×R2S^1_\beta\times\mathbb{R}^2 with T=1/βT=1/\beta. The positive phase in zz matches the convention used in G\mathcal{G} above. For the unHiggsed thermal saddle, Eq. (7.10) of Ref. [25] is equivalent to

(1+2λ^)μ=2πλIm⁡Li⁡2(z).(86)\left(1+\frac{2}{\widehat{\lambda}}\right)\mu=\frac{2}{\pi\lambda}\operatorname{Im}\operatorname{Li}_2(z). \tag*{(86)}

The λ→0\lambda\to0 limit is taken continuously. At the conformal point, μ\mu depends on the couplings but not on temperature; the regular-boson ratio is therefore 16πG(μ,πλ)16\pi\mathcal{G}(\mu,\pi\lambda) from the common expression above.

The ungauged limit provides a consistency check against the sextic scalar theory discussed earlier. Write Nr=2NcN_r=2N_c for the number of real scalar components corresponding to NcN_c complex scalars. With ϕ†ϕ=φ2/2\phi^\dagger\phi=\varphi^2/2, comparison of g6(φ2)3/(6Nr2)g_6(\varphi^2)^3/(6N_r^2) with the interaction above gives λ6=2g6\lambda_6=2g_6. Thus, for α=g6/(4π)\alpha=\sqrt{g_6}/(4\pi), one has λ^→2α\widehat{\lambda}\to2\alpha as λ→0\lambda\to0. Using Im⁡Li⁡2(e−μ+iπλ)=−πλlog⁡(1−e−μ)+O(λ3)\operatorname{Im}\operatorname{Li}_2(e^{-\mu+i\pi\lambda})=-\pi\lambda\log(1-e^{-\mu})+O(\lambda^3), Eq. (86) becomes

(1+1α)μ=−2log⁡(1−e−μ).(87)\left(1+\frac{1}{\alpha}\right)\mu=-2\log(1-e^{-\mu}). \tag*{(87)}

It agrees with the x→∞x\to\infty limit of the hyperbolic sextic gap equation m=αI(m,x)m=\alpha I(m,x), identifying μ=2πm/x\mu=2\pi m/x. In this limit, I(m,x)=−m−(x/π)log⁡(1−e−μ)+o(x)I(m,x)=-m-(x/\pi)\log(1-e^{-\mu})+o(x). At α=1\alpha=1 the solution is μ=log⁡2\mu=\log2, reproducing the Bardeen–Moshe–Bander endpoint.

Stable couplings and phases. For 0<λ<10 < \lambda< 1 define the sextic parameter used in Ref. [27],

x6=λ624π2λ2−1.(88)x_6 = \frac{\lambda_6}{24\pi^2\lambda^2} - 1 . \tag*{(88)}

Requiring the zero-temperature large-NN effective potential to be bounded below gives ϕ1≤x6≤ϕ2\phi_1 \le x_6 \le\phi_2, with

ϕ1=43[1(2−λ)2−1],ϕ2=43[1λ2−1].(89)\phi_1 = \frac{4}{3}\left[\frac{1}{(2-\lambda)^2}-1\right], \qquad\phi_2 = \frac{4}{3}\left[\frac{1}{\lambda^2}-1\right] . \tag*{(89)}

Equivalently, writing u≡λ6/(8π2)u \equiv\lambda_6/(8\pi^2), the domain is

λ2[4(2−λ)2−1]≤u≤4−λ2.(90)\lambda^2\left[\frac{4}{(2-\lambda)^2}-1\right]\le u\le4-\lambda^2 . \tag*{(90)}

We use the positive-mass unHiggsed solution of Eq. (86); its selection over the Higgsed branch is justified below.

Numerical check and contour map. Solving Eq. (86) and evaluating 16πG(μ,πλ)16\pi G(\mu,\pi\lambda) gives at λ=1/2\lambda=1/2:

λ6\lambda_6x6x_6μ=M/T\mu=M/TcS/cTc_S/c_T
6π26\pi^2000.2986044450.29860444549.033767749.0337677
12π212\pi^2110.3426255250.34262552548.823112548.8231125

Table S54.

These theories have equal planar CTC_T and separated-point ⟨TTT⟩\langle TTT\rangle at fixed λ\lambda, but different thermal coefficients: the sextic interaction changes the thermal mass while not changing the leading spinning-current two- and three-point functions [26]. The entire sampled stable domain lies below the proposed scalar upper bound.

The limiting values identify familiar theories. At (λ,λ6)=(0,0)(\lambda,\lambda_6)=(0,0), the massless free scalar gives cS/CT=16πζ(3)c_S/C_T=16\pi\zeta(3); at (0,32π2)(0,32\pi^2), the ungauged Bardeen–Moshe–Bander sextic endpoint gives cS/CT=14πζ(3)c_S/C_T=14\pi\zeta(3). At the narrowing tip (λ,λ6)→(1,24π2)(\lambda,\lambda_6)\to(1,24\pi^2), the ratio approaches the free-Dirac value 12πζ(3)12\pi\zeta(3).

Exclusion of a competing Higgsed saddle. We use the positive-mass unHiggsed solution of Eq. (86) within this domain. To check whether it competes with a Higgsed phase, we specialize the finite-temperature saddle equations of Ref. [27], Sec. 4, to zero mass and quartic couplings. Write c>0c>0 for the dimensionless thermal mass in a putative Higgsed saddle (this is the c^B\hat c_B of Ref. [27]), and denote the normalized Higgs condensate by σ=2πv2/(λT)>0\sigma=2\pi v^2/(\lambda T)>0. The bosonic thermal tadpole, averaged over the gauge holonomy, is

S(c)=12∫−ππdθ ρ(θ)[log⁡(2sinh⁡c+iθ2)+log⁡(2sinh⁡c−iθ2)],=c2−1πλIm⁡Li⁡2(e−c+iπλ),(91)\begin{aligned} \mathcal{S}(c)=\frac{1}{2}\int_{-\pi}^{\pi}d\theta\,\rho(\theta)\left[\log\left(2\sinh\frac{c+i\theta}{2}\right)+\log\left(2\sinh\frac{c-i\theta}{2}\right)\right], \\ &=\frac{c}{2}-\frac{1}{\pi\lambda}\operatorname{Im}\operatorname{Li}_2\left(e^{-c+i\pi\lambda}\right), \tag*{(91)} \end{aligned}

where ρ(θ)=1/(2πλ)\rho(\theta)=1/(2\pi\lambda) for ∣θ∣≤πλ|\theta|\le\pi\lambda and vanishes otherwise. The second line follows by integrating the logarithms over the uniform holonomy distribution.

Dey et al. introduce an off-shell free energy depending on cc, σ\sigma, and an auxiliary tadpole variable S~\widetilde{\mathcal{S}}. Its stationarity conditions, Eqs. (4.3)–(4.5) of Ref. [27], set S~=S(c)\widetilde{\mathcal{S}}=\mathcal{S}(c) on-shell and separate the two branches through

(S~+σ)[−c+λ(S~+σ)]=0.(92)(\widetilde{\mathcal{S}}+\sigma)\left[-c+\lambda(\widetilde{\mathcal{S}}+\sigma)\right]=0 . \tag*{(92)}

The first factor describes the unHiggsed branch; on the Higgsed branch the second factor vanishes. The resulting first condition below selects the Higgsed branch [Eq. (4.5) of Ref. [27]], while the second follows from the Higgs-field tadpole condition [Eq. (3.30) of Ref. [27]] at vanishing mass and quartic deformations:

c=λ(S(c)+σ),c2=λ2(4+3x6)σ2.(93)c=\lambda\left(\mathcal{S}(c)+\sigma\right), \qquad c^2=\lambda^2(4+3x_6)\sigma^2 . \tag*{(93)}

Because the Higgs condensate has σ>0\sigma>0, these equations imply

λS(c)c=1−14+3x6.(94)\frac{\lambda\mathcal{S}(c)}{c}=1-\frac{1}{\sqrt{4+3x_6}} . \tag*{(94)}

For 0<λ<10<\lambda<1 and c>0c>0, the imaginary part of the dilogarithm in Eq. (91) is positive, so the left-hand side of Eq. (94) is strictly smaller than λ/2\lambda/2. On the other hand x6≥ϕ1x_6\ge\phi_1 implies 4+3x6≥2/(2−λ)\sqrt{4+3x_6}\ge2/(2-\lambda), hence the right-hand side is at least λ/2\lambda/2. Thus, no positive-mass Higgsed saddle competes at the conformal point within the stable interval. This statement concerns the standard planar uniform-holonomy construction, not the general finite-NN phase diagram.

Figure S1. Planar RB Chern–Simons thermal ratio cS/CTc_S/C_T, obtained by solving Eq. (86) over the stable domain (S53). The axes are λ=N/k\lambda= N/k and u=λ6/(8π2)u = \lambda_6/(8\pi^2); the black boundaries are the large-NN stability limits, and grey denotes couplings outside them. Labelled points A and B are the benchmarks in Eq. (S54). The λ=0\lambda= 0 edge contains the free scalar and the Bardeen–Moshe–Bander (BMB) sextic endpoint. The limiting tip λ→1\lambda\to1 approaches the free-fermion thermal ratio. Values shown are for the planar uniform-holonomy saddle.

References

  1. [1]P. Kovtun and A. Ritz, Universal conductivity and central charges, Phys. Rev. D 78, 066009 (2008), arXiv:0806.0110.DOI
  2. [2]P. Bueno, H. Casini, O. Lasso Andino, and J. Moreno, Conformal bounds in three dimensions from entanglement entropy, Phys. Rev. Lett. 131, 171601 (2023), arXiv:2307.05164.DOI
  3. [3]I. R. Klebanov and A. A. Tseytlin, Entropy of near-extremal black p-branes, Nucl. Phys. B 475, 164 (1996), arXiv:hep-th/9604089.DOI
  4. [4]S. S. Gubser and I. R. Klebanov, Absorption by branes and Schwinger terms in the world volume theory, Phys. Lett. B 413, 41 (1997), arXiv:hep-th/9708005.
  5. [5]G. Anastasiou, I. J. Araya, A. Argandoña, and R. Olea, CFT correlators from shape deformations in cubic curvature gravity, J. High Energy Phys. 11, 031 (2022), arXiv:2208.00093.
  6. [6]S. Sachdev, Polylogarithm identities in a conformal field theory in three dimensions, Phys. Lett. B 309, 285 (1993), arXiv:hep-th/9305131.DOI
  7. [7]A. C. Petkou, CT and CJ up to next-to-leading order in 1/N in the conformally invariant O(N) vector model for 2 < d < 4, Phys. Lett. B 359, 101 (1995), arXiv:hep-th/9506116.
  8. [8]E. Shaghoulian, Modular forms and a generalized Cardy formula in higher dimensions, Phys. Rev. D 93, 126005 (2016), arXiv:1508.02728.
  9. [9]N. Bobev, J. Hong, and V. Reys, Holographic thermal observables and M2-branes, J. High Energy Phys. 12, 054 (2023), arXiv:2309.06469.
  10. [10]H. Casini, M. Huerta, and R. C. Myers, Towards a derivation of holographic entanglement entropy, J. High Energy Phys. 05, 036 (2011), arXiv:1102.0440.
  11. [11]E. Perlmutter, A universal feature of CFT Rényi entropy, J. High Energy Phys. 03, 117 (2014), arXiv:1308.1083.DOI
  12. [12]P. Bueno, R. C. Myers, and W. Witczak-Krempa, Universal corner entanglement from twist operators, J. High Energy Phys. 09, 091 (2015), arXiv:1507.06997.DOI
  13. [13]Y. Kusuki, H. Ooguri, and S. Pal, Universality of Rényi entropy in conformal field theory, Phys. Rev. Lett. 135, 061603 (2025), arXiv:2503.24353.
  14. [14]L.-Y. Hung, R. C. Myers, and M. Smolkin, Twist operators in higher dimensions, J. High Energy Phys. 10, 178 (2014), arXiv:1407.6429.DOI
  15. [15]D. M. Hofman and J. Maldacena, Conformal collider physics: Energy and charge correlations, J. High Energy Phys. 05, 012 (2008), arXiv:0803.1467.
  16. [16]S. Dutta Chowdhury, J. R. David, and S. Prakash, Constraints on parity violating conformal field theories in d = 3, J. High Energy Phys. 11, 171 (2017), arXiv:1707.03007.
  17. [17]C. P. Herzog and T. Nishioka, The edge of entanglement: getting the boundary right for non-minimally coupled scalar fields, J. High Energy Phys. 12, 138 (2016), arXiv:1610.02261.
  18. [18]J. Lee, A. Lewkowycz, E. Perlmutter, and B. R. Safdi, Rényi entropy, stationarity, and entanglement of the conformal scalar, J. High Energy Phys. 03, 075 (2015), arXiv:1407.7816.
  19. [19]I. R. Klebanov, S. S. Pufu, S. Sachdev, and B. R. Safdi, Rényi entropies for free field theories, J. High Energy Phys. 04, 074 (2012), arXiv:1111.6290.DOI
  20. [20]L.-Y. Hung, R. C. Myers, M. Smolkin, and A. Yale, Holographic calculations of Rényi entropy, J. High Energy Phys. 12, 047 (2011), arXiv:1110.1084.DOI
  21. [21]A. Mukherjee, Rényi entropy of the critical O(N) model (2015), arXiv:1512.01226.arxiv.org/abs/1512.01226
  22. [22]W. A. Bardeen, M. Moshe, and M. Bander, Spontaneous breaking of scale invariance and the ultraviolet fixed point in O(N)-symmetric (φ^3) theory, Phys. Rev. Lett. 52, 1188 (1984).
  23. [23]M. Goykhman, V. Rosenhaus, and M. Smolkin, The background field method and critical vector models, J. High Energy Phys. 02, 074 (2021), arXiv:2009.13137.
  24. [24]C. Fleming, B. Delamotte, and S. Yabunaka, Finite N origin of the Bardeen–Moshe–Bander phenomenon and its extension at N = ∞ by singular fixed points, Phys. Rev. D 102, 065008 (2020), arXiv:2001.07682.
  25. [25]O. Aharony, S. Giombi, G. Gur-Ari, J. Maldacena, and R. Yacoby, The thermal free energy in large N Chern–Simons-matter theories, J. High Energy Phys. 03, 121 (2013), arXiv:1211.4843.
  26. [26]O. Aharony, G. Gur-Ari, and R. Yacoby, Correlation functions of large N Chern–Simons-matter theories and bosonization in three dimensions, J. High Energy Phys. 12, 028 (2012), arXiv:1207.4593.
  27. [27]A. Dey, I. Halder, S. Jain, L. Janagal, S. Minwalla, and N. Prabhakar, Duality and an exact Landau–Ginzburg potential for quasi-bosonic Chern–Simons-matter theories, J. High Energy Phys. 11, 020 (2018), arXiv:1808.04415.
  28. [28]F. Kos, D. Poland, and D. Simmons-Duffin, Bootstrapping the O(N) vector models, J. High Energy Phys. 06, 091 (2014), arXiv:1307.6856.DOI
  29. [29]A. V. Chubukov, S. Sachdev, and J. Ye, Theory of two-dimensional quantum Heisenberg antiferromagnets with a nearly critical ground state, Phys. Rev. B 49, 11919 (1994), arXiv:cond-mat/9304046.
  30. [30]O. Diatlyk, F. K. Popov, and Y. Wang, Beyond N = ∞ in large N conformal vector models at finite temperature, J. High Energy Phys. 08, 219 (2024), arXiv:2309.02347.
  31. [31]K. Diab, L. Fei, S. Giombi, I. R. Klebanov, and G. Tarnopolsky, On CJ and CT in the Gross–Neveu and O(N) models, J. Phys. A 49, 405402 (2016), arXiv:1601.07198.
  32. [32]S. Giombi, G. Tarnopolsky, and I. R. Klebanov, On CJ and CT in conformal QED, J. High Energy Phys. 08, 156 (2016), arXiv:1602.01076.
  33. [33]A. Bulbulian, M. Caselle, A. Nada, and M. Panero, Casimir effect in critical O(N) models from nonequilibrium Monte Carlo simulations, Phys. Rev. E 112, 064126 (2025), arXiv:2505.20403.
  34. [34]C.-H. Chang, V. Dommes, R. S. Erramilli, A. Homrich, P. Kravchuk, A. Liu, M. S. Mitchell, D. Poland, and D. Simmons-Duffin, Bootstrapping the 3d Ising stress tensor, J. High Energy Phys. 03, 136 (2025), arXiv:2411.15300.DOI
  35. [35]M. Reehorst, E. Trevisani, and A. Vichi, Mixed scalar-current bootstrap in three dimensions, J. High Energy Phys. 12, 156 (2020), arXiv:1911.05747.DOI
  36. [36]S. M. Chester, W. Landry, J. Liu, D. Poland, D. Simmons-Duffin, N. Su, and A. Vichi, Bootstrapping Heisenberg magnets and their cubic instability, Phys. Rev. D 104, 105013 (2021), arXiv:2011.14647.
  37. [37]N. Chai, S. Chaudhuri, C. Choi, Z. Komargodski, E. Rabinovici, and M. Smolkin, Thermal Order in Conformal Theories, Phys. Rev. D 102, 065014 (2020), arXiv:2005.03676.DOI
  38. [38]M. Smolkin and L. Yung, Thermal Order in the Biconical Model (2026), arXiv:2608.02720 [hep-th].arxiv.org/abs/2608.02720
  39. [39]L. Iliesiu, M. Koloğlu, R. Mahajan, E. Perlmutter, and D. Simmons-Duffin, The conformal bootstrap at finite temperature, J. High Energy Phys. 10, 070 (2018), arXiv:1802.10266.DOI
  40. [40]S. El-Showk and K. Papadodimas, Emergent spacetime and holographic CFTs, J. High Energy Phys. 10, 106 (2012), arXiv:1101.4163.
  41. [41]N. Benjamin, J. Lee, H. Ooguri, and D. Simmons-Duffin, Universal asymptotics for high energy CFT data, J. High Energy Phys. 03, 115 (2024), arXiv:2306.08031.DOI
  42. [42]D. Z. Freedman, S. S. Gubser, K. Pilch, and N. P. Warner, Renormalization group flows from holography: Supersymmetry and a c-theorem, Adv. Theor. Math. Phys. 3, 363 (1999), arXiv:hep-th/9904017.arxiv.org/abs/hep-th/9904017
  43. [43]S. S. Gubser and A. Nellore, Mimicking the QCD equation of state with a dual black hole, Phys. Rev. D 78, 086007 (2008), arXiv:0804.0434.DOI
  44. [44]S. Rychkov and A. Stergiou, General Properties of Multi-scalar RG Flows in d = 4 − ε, SciPost Phys. 6, 008 (2019), arXiv:1810.10541.DOI
  45. [45]H. Osborn and A. Stergiou, Seeking fixed points in multiple coupling scalar theories in the ε expansion, J. High Energy Phys. 05, 051 (2018), arXiv:1707.06165.DOI
  46. [46]L. Dolan and R. Jackiw, Symmetry behavior at finite temperature, Phys. Rev. D 9, 3320 (1974).DOI
  47. [47]H. Osborn and A. Stergiou, Heavy handed quest for fixed points in multiple coupling scalar theories in the ε expansion, J. High Energy Phys. 04, 128 (2021), arXiv:2010.15915.DOI
  48. [48]M. Hogervorst and C. Toldo, Bounds on multiscalar CFTs in the ε expansion, J. High Energy Phys. 04, 068 (2021), arXiv:2010.16222.DOI
  49. [49]H. W. Diehl, D. Grüneberg, and M. A. Shpot, Fluctuation-induced forces in periodic slabs: Breakdown of epsilon expansion at the bulk critical point and revised field theory, Europhys. Lett. 75, 241 (2006), arXiv:cond-mat/0605293.DOI
  50. [50]S. Weinberg, Gauge and global symmetries at high temperature, Phys. Rev. D 9, 3357 (1974).DOI

Paper details

vlnc:W_s4x2up4w
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

GPT-5.6, 6.1 Sol, GPT-6 Astra (bare), Claude Opus 5.5 (bare), Claude Fable 5.1 (bare)

View statement
Looking for related papers…
Contents