The momentum-space conformal bootstrap in 2d

Marc Gillioz · 2025

Read The momentum-space conformal bootstrap in 2d by Marc Gillioz (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-21227.txt.

Introduction

Marc Gillioz derives in 2025 a new bootstrap equation for two-dimensional conformal field theory from momentum-space data. Standard crossing rules lean on symmetric functions that momentum forms lack. The trick here uses the analytic traits of the commutator to link two expansion forms.

A four-point function with two light-like momenta drives the derivation. The result holds for any theory in one or two dimensions, in an off-balance form. Two projections land on known tool sets, and one closed form here is new. Sample uses show how the equation constrains theory data.

Catalog record

CREATORMarc Gillioz
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE126 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-21227

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The momentum-space conformal bootstrap in 2d Marc Gillioz Abstract A new bootstrap equation in 2-dimensional conformal field theory is derived starting from the momentum-space representation of the correlation functions. Since Wightman functions are not crossing-symmetric, the analyticity properties of the commutator are leveraged instead to obtain a relation between two distinct operator product expansions. The procedure requires evaluating a 4-point function with two light-like momenta. The result is an asymmetric equation valid for arbitrary theories in 1d and 2d. The new crossing equation admits two simple projections onto orthogonal bases of Jacobi polynomials, reproducing known sets of analytic functionals. One of these, with zeros at double-twist dimensions, was known only as a contour integral. The closed-form expression found in this work is new. We provide a few examples of applications of the new crossing equation. 1 Introduction Among all quantum field theories, the ones with conformal symmetry stand out for many reasons. They are found at the end point of renormalization group flows, describe continuous phase transitions, and are dual to gravity in anti de-Sitter space. But maybe most importantly, they can be rigorously defined in terms of an infinite algebra of operators characterized by a scaling dimension and spin quantum numbers. Remarkably, requiring the closure of this algebra is sometimes sufficient for solving an entire conformal field theory — this goes by the name of conformal bootstrap [ 1 , 2 ] . The development of the conformal bootstrap program has been rapid in recent years and some of its results are truly remarkable [ 3 , 4 ] . Nevertheless, our analytical understanding of the mechanisms behind the conformal bootstrap is still incomplete. It is therefore crucial to develop it further [ 5 ] . One direction that has been investigated in recent years is the momentum-space representation of the conformal correlation functions. It was initially noticed that the Fourier transform of Euclidean correlation functions admits interesting representations, both in closed form for 3-point functions [ 6 , 7 ] and in integral form for higher-point functions [ 8 ] . However, the Fourier integral interferes with the nice convergence properties of the operator product expansion (OPE) in Euclidean space, so these representations have not proven useful in the context of the conformal bootstrap. The situation is different in Minkowski space-time, where Wightman functions do admit an OPE both in the position- and momentum-space representations [ 9 , 10 ] . Alas, Wightman functions lack another crucial property required by the bootstrap approach: the permutation symmetry of the correlation functions, synonym with associativity of the operator algebra. For this reason, existing results in Minkowski momentum space rely on the delicate connection between Wightman functions and time-ordered correlators, such as the optical theorem [ 11 , 12 , 13 ] , dispersion formulae [ 14 ] , or LSZ reduction [ 15 ] , but a true conformal bootstrap in momentum space has never been achieved so far. The goal of this work is to propose such a momentum-space bootstrap, working with Wightman functions and replacing the missing permutation symmetry with a property coming from the analyticity of the commutator, a direct consequence of the causality axiom of quantum field theory. The present work focuses on theories with one time and one space directions only. Even though the method introduced below can be directly applied to conformal field theory in higher dimensions, many technicalities are simpler in 2d. Of course, 2d is also the domain in which the traditional bootstrap excels, thanks to powerful techniques involving Virasoro symmetry and its generalizations, or the modular invariance of the torus partition function. Our approach does not rely on such techniques in an essential way, though: it only assumes global conformal invariance in flat space-time, even though the integer spacing of conformal weights is used in some examples given below. Summary of results The calculations in this work all involve correlation functions of local operators characterized by pairs of conformal weight (h,{\bar{h}}) that can in principle be distinct for each operator. However, for the sake of clarity, here we highlight the main results with identical scalar operators. The idea is very similar to the ordinary conformal bootstrap. Starting from the 4-point function of a scalar operator \phi with conformal weights (h_{\phi},h_{\phi}) , we obtain two distinct expansions in conformal blocks that depend on the conformal weights (h,{\bar{h}}) of the exchanged operator \mathcal{O} , multiplied by the square of the OPE coefficients C_{\phi\phi\mathcal{O}} . A specificity of 2-dimensional conformal field theory is that the correlation functions, and hence the conformal blocks, can be factorized into holomorphic (left-moving) and anti-holomorphic (right-moving) parts that each depend on h or {\bar{h}} only. We obtain the equation \sum_{\mathcal{O}}C_{\phi\phi\mathcal{O}}^{2}S_{h}(w)S_{{\bar{h}}}(\bar{w})=\sum_{\mathcal{O}}C_{\phi\phi\mathcal{O}}^{2}T_{h}(w)T_{{\bar{h}}}(\bar{w}),\qquad w,\bar{w}\in(0,1). (1.1) The holomorphic blocks S_{h} and T_{h} are functions of a single kinematic variable w restricted to the interval (0,1) . As we shall see, w is the dimensionless ratio between the (left-moving) momentum of one of the external operators and the center-of-mass energy of the 2\to 2 scattering process. Similarly, the anti-holomorphic blocks S_{{\bar{h}}} and T_{{\bar{h}}} depend on \bar{w} , defined in terms of right-moving momenta. The s-channel conformal block S_{h} is a hypergeometric function multiplied by a power of w and a ratio of gamma functions, S_{h}(w)=\frac{\Gamma(2h)}{\Gamma(h)^{2}\Gamma(2h_{\phi})}w^{2h_{\phi}-1}{}_{2}F_{1}(1-h,h;2h_{\phi};w). (1.2) It differs from the t-channel block, T_{h}(w)=\frac{1}{\Gamma(2h_{\phi}-h)\Gamma(h)}w^{h-1}{}_{2}F_{1}(h-2h_{\phi}+1,h;2h;w). (1.3) The derivation of equations ( 1.1 ), ( 1.2 ), and ( 1.3 ) is the subject of Section 2 below. The two functions S_{h} and T_{h} are reminiscent of the position-space blocks [ 16 , 17 ] and, in fact, a non-trivial connection with them can be established through the analytic functionals of Ref. [ 18 ] . As a reminder, the usual conformal blocks are also hypergeometric functions multiplying non-integer powers of the cross-ratio z . They can be analytically continued to the whole complex plane in z except for two branch cuts at z\in(-\infty,0) and z\in(1,\infty) . We show in Section 3 that eqs. ( 1.2 ) and ( 1.3 ) correspond to a particular integral transform of the discontinuity along either of these branch cuts. With hindsight, it is actually rather simple to recover eq. ( 1.1 ) from the position-space crossing equation once the transform is given. However, it is probable that the transform would have never been found without the momentum-space origin of the equation. The first consequence of this remarkable connection with the Euclidean branch cuts is that the t-channel block ( 1.3 ) has zeros at h=2h_{\phi}+n with integer n\in\mathbb{N} , as visible from the first gamma function in its denominator. This corresponds to the exchange of double-twist operators of the form \phi\partial^{n}\bar{\partial}^{\bar{n}}\phi . The second, unfortunate consequence is that the absolute convergence of the Euclidean conformal block expansion is lost in eq. ( 1.1 ). While the t-channel expansion is absolutely convergent — eq. ( 1.3 ) actually has similar asymptotics as the Euclidean blocks when h\to\infty — the s-channel expansion is only convergent in a distributional sense: it does not converge point-wise at fixed w , but only after integrating against a suitable functions in the domain w,\bar{w}\in(0,1) . Interestingly, there exists a natural basis of functions that give a nice reformulation of the crossing equation ( 1.1 ): after

Read the complete text