A Renormalization Group Analysis of the Ising Model Coupled to Causal Dynamical Triangulations

R. Barouki, and D. Laurenzano · 2025

Read A Renormalization Group Analysis of the Ising Model Coupled to Causal Dynamical Triangulations by R. Barouki, and D. Laurenzano (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-2504-01134.txt.

Introduction

Ryan Barouki and Davide Laurenzano study a model of matter on a shifting space in a 2025 paper. The model joins the Ising model, a grid of tiny magnets, with causal dynamical triangulations. That second part builds space from triangles with a fixed flow of time. The authors view the whole setup through a math lens called the functional renormalization group.

They hunt for fixed points, spots where the model's rules stop changing as the view zooms out. Known fixed points for plain gravity turn up again. So does one for gravity with a steady push outward. A new fixed point also shows, with three key directions. That count matches the count of main fields in the Ising theory of phase change, which supports their reading of it.

Catalog record

CREATORR. Barouki, and D. Laurenzano
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-01134

Browse the full library · About the Tower of Babel

Full text

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

A Renormalization Group Analysis of the Ising Model Coupled to Causal Dynamical Triangulations R. Barouki, and D. Laurenzano Abstract We analyze the matrix model characterizing the Ising model coupled to Causal Dynamical Triangulations (CDT) from the point of view of the Functional Renormalization Group Equation (FRGE). This model is a dually weighted matrix model, whose Feynman diagrams are in correspondence with discrete triangulations of two-dimensional geometries with a preferred time foliation. In particular, we find the fixed points of the beta-function equations, showing that the number of relevant directions in each case is compatible with the physical interpretation of the CFT living on the fixed points. In addition to recovering the fixed points for topological gravity and pure gravity with a cosmological constant, we find a new fixed point featuring three relevant directions which match the number of primary fields in the Ising CFT. 1 Introduction The problem of finding a consistent theory of quantum gravity has long been a central challenge in theoretical physics. In this paper, we focus on two-dimensional quantum gravity , which can be understood as a theory of random surfaces. An effective way to tackle random surfaces is to discretize them, typically by triangulating —i.e. requiring that all faces of the embedded graph be triangular. The functional integral over the metrics then becomes a sum over random surfaces \int_{\mathcal{M}}\mathcal{D}g~~\rightarrow\sum_{\text{triangulations of }\mathcal{M}} and the Einstein-Hilbert action is also discretized. This was first done by Regge [ 1 ] for numerical General Relativity. This discretization introduces a regulator in the form of the triangle edge lengths, making computations more tractable, after which one can take a continuum limit to recover the original continuous geometry. An ingenious approach to generate all possible triangulations of a surface was first proposed by ’t Hooft in [ 2 ] wherein he introduced the concept of matrix models. He showed that one can use these simple matrix integrals to generate all possible triangulations by essentially generating “fat” Feynman diagrams, where the “fatness” of the graphs creates internal loops that can be thought of as faces. Thus, whereas Gaussian integrals of vector or scalar quantities enumerate objects with vertices and edges (graphs), Gaussian integrals of matrices enumerate objects with vertices, edges and faces (maps/embedded graphs). They have since been used extensively to study many aspects of two-dimensional quantum gravity with some notable recent examples [ 3 , 4 , 5 ] . See [ 6 ] for an older review and [ 7 ] for a more recent one. When summing over a random discrete surface, there is some freedom in deciding which ensemble of graphs to include in the sum. One choice is the ensemble of planar graphs, often called the Euclidean ensemble. While this leads to many interesting results consistent with the continuum perspective of Liouville quantum gravity, it also encounters serious issues. One prominent example is the unexpected value of the Hausdorff dimension, d_{H}=4 , which seems to contradict the naive expectation of d_{H}=2 . These difficulties motivated the development of the Causal Dynamical Triangulations (CDT) program, which imposes a causal structure on the graph ensemble by demanding a global time foliation [ 8 ] . In CDT, edges are split into two types—time-like and space-like—ensuring a clear causal structure. This distinction allows for well-defined Wick rotations so that one can work in Euclidean signature while still preserving essential Lorentzian features such as causality. CDT has been remarkably successful in the non-perturbative study of two-dimensional quantum gravity, with many analytic and numerical results [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 ] . Moreover, considerable progress has been made in four-dimensional CDT using numerical methods [ 19 , 20 , 21 ] . A description of quantum gravity would not be complete without understanding the interplay between matter and geometry. To that end, we study two-dimensional CDT coupled to a simple matter model: the Ising model. Specifically, we adopt a matrix model representation of this coupled Ising–CDT system, as introduced in [ 22 ] , and investigate its fixed-point structure using Functional Renormalization Group (FRG) methods. Fixed points of the FRG represent potential double scaling limits of the matrix model, where N\rightarrow\infty and couplings scale towards a critical value, the scaling being characterized by a critical exponent. Following the approach of [ 23 ] , where FRG techniques were applied to tensor models, we extend the results of Castro and Koslowski [ 24 ] , who first applied FRG methods to the pure CDT matrix model by Benedetti and Henson [ 25 ] . In particular, we demonstrate the existence of fixed points of RG flow in the Ising–CDT matrix model with precisely the right number of relevant directions to describe a continuum Ising CFT. More specifically, we show that the number of relevant directions matches the three primary fields of the Ising CFT [ 26 ] . Moreover, we find other types of fixed points with different numbers of relevant directions. These could, in principle, represent different continuum limits of the discrete theory. One of our main findings is that the parameter space has a rich structure, including segments such that all points on the segment are fixed points of the RG flow. The paper is organized as follows. In Section 2 , we review the main features of the pure CDT and Ising–CDT matrix models. In Section 3 , we set up the Functional Renormalization Group Equation (FRGE) for the Ising–CDT matrix model. Section 4 presents the derivation of the general form of the beta-function equations. Finally, we analyze the fixed-point structure in Section 5 . 2 A Matrix Model for Ising CDT The matrix model for pure CDT was introduced by Benedetti and Henson (BH) [ 25 ] and is defined by the following partition function Z=\int\mathcal{D}A\,\mathcal{D}Be^{-N\text{Tr}[\frac{1}{2}A^{2}+\frac{1}{2}(C^{-1}B)^{2}-gA^{2}B]}\,, with Hermitian matrices A and B , along with the constant matrix C that satisfies the condition \text{Tr}[C^{m}]=N\delta_{m,2}\,. (1) The propagators are given by \langle A_{ij}A_{kl}\rangle=\frac{1}{N}\delta_{il}\delta_{jk},\quad\langle B_{ij}B_{kl}\rangle=\frac{1}{N}C_{il}C_{jk},\quad\langle A_{ij}B_{kl}\rangle=0\,. The two matrices A and B are necessary to capture the two different types of edges in a dual CDT graph: space-like edges represented by the matrix A and time-like edges represented by the matrix B . Note that time-like edges in the dual graph correspond to space-like edges in the triangulation and vice-versa. The structure of the Feynman graphs generated by this model consists of AAB vertices and faces with two time-like edges, with the exception of the initial and final slices of the graph where the faces have 0 time-like edges (see Figure 1 ). Each face with m internal B lines contributes a factor of \text{Tr}[C^{m}] , hence the condition ( 1 ) ensures that each face has either two or zero B -lines (time-like edges). These local requirements are sufficient to generate valid CDT graphs with a global foliation [ 25 ] . Figure 1: An example of a CDT strip and the dual ribbon graph. The red lines are the propagators for the A matrix and the gray dashed lines are the B propagators. The B field can be integrated out to obtain a simpler expression only in terms of the A matrix: Z=\int\mathcal{D}A\,e^{-N\text{Tr}[\frac{1}{2}A^{2}-\frac{g^{2}}{2}(A^{2}C)^{2}]}\,. (2)

Read the complete text