Constraint correlation functions of the Ising model in the scaling limit

Ivan Balog, Adam Rançon · 2025

Read Constraint correlation functions of the Ising model in the scaling limit by Ivan Balog, Adam Rançon (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-2509-08557.txt.

Introduction

How does a chain of magnets behave when its total magnetization is held fixed? Ivan Balog and 1 co-author study that question for the one-dimensional Ising model in a 2025 paper. They focus on the scaling limit near absolute zero.

In momentum space, the correlation function shows surprising swings as magnetization changes. The swings repeat with a period tied to momentum, and the authors link them to domain walls, the borders between magnetized regions. That pattern differs sharply from behavior in a steady magnetic field. It also helps explain recent computer simulations of the two-dimensional model.

Catalog record

CREATORIvan Balog, Adam Rançon
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-08557

Browse the full library · About the Tower of Babel

Full text

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

Constraint correlation functions of the Ising model in the scaling limit Ivan Balog, Adam Rançon Abstract We study the correlation function of the one-dimensional Ising model at fixed magnetization. Focusing on the scaling limit close to the zero-temperature fixed point, we show that this correlation function, in momentum space, exhibits surprising oscillations as a function of the magnetization. We show that these oscillations have a period inversely proportional to the momentum and give an interpretation in terms of domain walls. This is in sharp contrast with the behavior of the correlation function in constant magnetic fields, and sheds light on recent results obtained by Monte Carlo simulations of the critical two-dimensional Ising model at fixed magnetization. I Introduction The Ising model is a cornerstone of statistical physics, thanks to its ability to capture the most stringent aspects of phase transitions despite its apparent simplicity. Of foremost importance, it can be solved exactly for arbitrary temperature and zero magnetic field in two dimensions (2D), and arbitrary temperature and (constant) field in one dimension (1D) McCoy and Wu 1973 . The latter case only displays an ordering transition at zero temperature but is nevertheless of interest, as it is governed at low temperature by localized excitations (domain walls, kinks or instantons). Those excitations, appearing in a variety of classical and quantum problems, are typically non-perturbative and are usually treated using ad hoc methods, e.g. introducing droplets Bruce and Wallace 1981 ; Bruce and Wallace 1983 or using instantonic calculations Rulquin et al. 2016 . This calls for developing more generic methods, for instance, building on approximation schemes of the Functional Renormalization Group (FRG) Dupuis et al. 2021 , as was undertaken in Farkaš et al. 2023 ; Balog et al. 2025a . Understanding the relation between FRG and exact results for the 1D Ising model close to its zero-temperature transition is thus of interest, to better understand how such excitations can be identified in FRG calculations. Another interesting aspect of the Ising model is that it allows for studying the emergence of collective phenomena that take place when the correlation length \xi (of a hypothetical infinite system) is much larger than the actual system size N . This happens for instance close to a second-order phase transition where universality is expected to appear. One such universal property is given by the probability distribution function (PDF) of the magnetization, i.e. the sum of all the spins, which becomes non-Gaussian due to the strong correlations between the spins, yet universal and depending only on the ratio N/\xi and the boundary conditions. The PDF of the Ising universality class has been studied in 2D and 3D both numerically Binder 1981a ; Binder 1981b ; Bruce and Wilding 1992 ; Nicolaides and Bruce 1988 ; Tsypin 1994 ; Tsypin and Blöte 2000 ; Malakis et al. 2014 ; Xu et al. 2020 and using perturbative RG approaches Bruce et al. 1979 ; Eisenriegler and Tomaschitz 1987 ; Esser et al. 1995 ; Bruce 1997 ; Rudnick et al. 1998 ; Sahu et al. 2025 ; Sahu 2025 , while that of the 1D Ising model has been computed for different boundary conditions using various methods Antal et al. 2004 ; García-Pelayo 2009 ; Wang et al. 2017 ; Dantchev and Rudnick 2022 ; Dantchev et al. 2023 ; Dantchev et al. 2024 ; Stepanyan et al. 2024 . Recently, a FRG implementation for the PDF’s calculation has been devised Balog et al. 2022 ; Balog et al. 2025b ; Rançon et al. 2025 . In this framework, a functional equation is solved to obtain the log of the PDF, called the rate function in probability theory, and the constraint effective potential in field theory. This equation is naturally expressed in terms of the “constraint correlation functions”, i.e. the correlation functions between spins at fixed magnetization. We have studied these correlation functions at the critical point of the 2D and 3D Ising model both with FRG and Monte Carlo simulations in Rose et al. 2025 . There, we observed in our 2D Monte Carlo data surprising oscillations in the magnetization dependence of the constraint correlation functions (in momentum space and with periodic boundary conditions). To better understand the origin of this non-trivial behavior, we study here the constraint correlation function of the 1D Ising model in the scaling regime close to its zero-temperature critical point. In the limit N/\xi\to 0 , we do find the same kind of oscillations. We show that they are induced by the imposed magnetization, which creates a new length scale associated with the typical size of the spin domains. This is in sharp contrast with the correlation function in constant field. The manuscript is organized as follows. In Sec. II we define and compute the PDF and the constraint correlation function for the finite-size 1D Ising model. In Sec. III we compute the scaling limit of the constraint correlation function, and analyze its non-trivial behavior as a function of the magnetization. Finally, in Sec. IV , we discuss our results and in particular their implications for functional methods such as FRG. Some technical details about the calculations are given in the appendices. II Finite-size constraint correlation function We are interested in the correlation function of the 1D Ising model in the “canonical” ensemble with fixed magnetization \sum_{i}\sigma_{i}=S . Here and below, the sum is over the N sites of the lattice, and we assume periodic boundary conditions. It is useful to first define the constraint partition function Z_{N}(S)=\langle\delta_{\sum_{i}\sigma_{i},S}\rangle, (1) where the thermal average is over the usual Boltzmann weight \langle\ldots\rangle=\sum_{\{\sigma_{i}\}}\ldots e^{-H[\sigma,0]}/\mathcal{Z}_{N}(0). (2) The Hamiltonian is that of the ferromagnetic nearest-neighbor 1D Ising model in a magnetic field H[\sigma,h]=-K\sum_{i}\sigma_{i}\sigma_{i+1}-h\sum_{i}\sigma_{i}, (3) and \mathcal{Z}_{N}(h)=\sum_{\{\sigma_{i}\}}e^{-H[\sigma,h]} is the (unconstrained) partition function. Of course, Z_{N}(S)=P_{N}(S) , the PDF of the magnetization that has been studied extensively in Antal et al. 2004 ; García-Pelayo 2009 ; Wang et al. 2017 ; Dantchev and Rudnick 2022 ; Dantchev et al. 2023 ; Dantchev et al. 2024 ; Stepanyan et al. 2024 . The magnetization S can take N possible values, from -N to N with increments by 2 . Furthermore, the PDF is normalized, i.e. \sum_{S}Z_{N}(S)=1 . The constraint correlation function between two spins at position i and j , the main object of our study, is defined as G_{N}(r;S)=\langle\sigma_{i}\sigma_{j}\delta_{\sum_{i}\sigma_{i},S}\rangle/Z_{N}(S), (4) which is a function of r=|i-j| by invariance per translation induced by the periodic boundary conditions. Note that it obeys the sum-rule \sum_{r=1}^{N}G_{N}(r;S)=S^{2}/N , since \langle\sigma_{j}\delta_{\sum_{i}\sigma_{i},S}\rangle/Z_{N}(S)=S/N by invariance by translation. The constraint partition function as well as \mathbf{G}_{N}(r;S)=\langle\sigma_{i}\sigma_{j}\delta_{\sum_{i}\sigma_{i},S}\rangle can be computed using a transfer matrix approach as follows, see Wang et al. 2017 ; Dantchev and Rudnick 2022 ; Dantchev et al. 2023 ; Stepanyan et al. 2024 for similar calculations of the former. First, we exponentiate the constraint using \delta_{n,m}=\int_{-\pi}^{\pi}\frac{d\theta}{2\pi}e^{i(n-m)} , and the constraint partition function can be rewritten as Z_{N}(S)=\int_{-\pi}^{\pi}\frac{d\theta}{2\pi}e^{-iS\theta}\langle e^{i\theta\sum_{i}\sigma_{i}}\rangle=\int_{-\pi}^{\pi}\frac{d\theta}{2\pi}e^{-iS\theta}\frac{\mathcal{Z}_{N}(h=i\theta)}{\mathcal{Z}_{N}(0)}. (5) Both Z_{N}(S) and G_{N}(r;S) can be computed from the unconstrained partition function and correlation function, and we recall now these standard results, see ( McCoy and Wu 1973 , Chap. III) for an introduction to the method. The transfer matrix is T(h)=\begin{pmatrix}e^{K+h}&&e^{-K}\\ e^{-K}&&e^{K-h}\end{pmatrix}, (6) with eigenvalues \lambda_{\pm}(h)=e^{K}\left(\cosh(h)\pm\sqrt{e^{-4K}+\sinh^{2}(h)}\right), (7) and eigenvectors \begin{split}u_{+}&=\begin{pmatrix}\cos(\phi/2)\\ \sin(\phi/2)\end{pmatrix},\\ u_{-}&=\begin{pmatrix}-\sin(\phi/2)\\ \cos(\phi/2)\end{pmatrix},\end{split} (8)

Read the complete text