The Lee-Yang property of isotropic vector ferromagnets and lattice fields
Yuri Kozitsky · 2026
Read The Lee-Yang property of isotropic vector ferromagnets and lattice fields by Yuri Kozitsky (2026) 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-2603-18675.txt.
Introduction
Kozitsky proves a Lee-Yang type result for isotropic vector spin models on the integer lattice. The Lee-Yang property says a partition function has only imaginary zeros in the field variable. Many analytic tools for spin models and quantum lattice fields rely on it. Until now, the generalized property was known only in two dimensions. Kozitsky shows it holds for models on Z in all even dimensions D. The result covers both classical rotors and D-component quantum lattice fields.
Catalog record
| CREATOR | Yuri Kozitsky |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29842 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.18675
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The Lee-Yang property of isotropic vector ferromagnets and lattice fields Yuri Kozitsky Abstract . The Lee-Yang property of a given spin model means that its partition function has purely imaginary zeros as a function of an external magnetic field. A similar property is also used in the theory of quantum anharmonic crystals and quantum lattice fields. A number of powerful analytic methods of the mathematical theory of such models employ this property. Its suitable generalization is used in the theory of models of isotropic D D -dimensional spins (rotors) or D D -component quantum lattice fields. So far, the (generalized) Lee-Yang property has been established only for two-dimensional isotropic models. In this work, we prove that isotropic spin and field models living on ℤ \mathds{Z} have this property for all even D D . 1991 Mathematics Subject Classification 82B20; 81T18; 30C15; 32A15 1. Introduction The present research was inspired by an item on the list of open problems in mathematical physics recently published by Barry Simon in [ 19 ] . In Conjecture 9.2.27 on page 570, it is proposed to “Prove a Lee–Yang theorem for isotropic classical D -rotors for D\geq 4 ”. Here we prove this statement for all even D and a particular models of such ‘rotors’ and Euclidean quantum lattice fields. In 1952, two physicists, T. D. Lee and C. N. Yang, proved a statement [ 15 ] , which turned out to be seminal for both mathematical physics [ 1 , 7 , 8 , 9 , 16 , 17 , 19 ] and pure mathematics [ 4 , 5 , 6 , 17 ] . Their original result describing the Ising spin model can be sketched as follows. Consider the variables \sigma_{l} , l=1,\dots,N – spins – taking values \pm 1 , and the function (1.1) Z^{\rm Ising}_{N}(h)=\sum_{\dots\sigma_{l}=\pm 1\dots}\exp\left(h\sum_{1\leq l\leq N}\sigma_{l}+\sum_{1\leq l<l^{\prime}\leq N}J_{ll^{\prime}}\sigma_{l}\sigma_{l^{\prime}}\right). In statistical physics, it is the partition function of a magnet, where J_{ll^{\prime}} are interaction intensities and h stands for an external magnetic field. As a finite sum of exponentials, it can be continued to an exponential-type entire function of h\in\mathds{C} . According to the Lee-Yang theorem, all zeros of Z^{\rm Ising}_{N} lie on the imaginary axis, provided that the interaction is of ferromagnetic type, i.e., J_{ll^{\prime}}\geq 0 for all l,l^{\prime} . Later, this result was extended in different directions; see [ 7 , 8 , 9 , 16 , 17 ] , [ 1 , page 171] , and [ 19 , Chapter 3] , with applications in statistical physics and lattice Euclidean quantum field theory. In particular, the summation in ( 1.1 ) was replaced by integration, and the variables \sigma_{l} were turned into D -vectors. For D=2 , an analog of the Lee-Yang theorem was proved in [ 16 ] , see also [ 19 , Chapter 3] . In this article, we study the function (1.2) Z_{N}(z)=\int\exp\left(\sum_{l=1}^{N}z\cdot\sigma_{l}+J\sum_{l=1}^{N-1}\sigma_{l}\cdot\sigma_{l+1}\right)\prod_{l=1}^{N}\chi(d\sigma_{l}). In the language of statistical physics, the corresponding model lives on the graph \mathds{Z} , which means that J_{ll^{\prime}}=J>0 if |l-l|=1 , and J_{ll^{\prime}}=0 otherwise. In ( 1.2 ), the integral is taken over (\mathds{R}^{D})^{N} , D,N\in\mathds{N} , z=x+iy\in\mathds{C}^{D} , z\cdot\sigma=x\cdot\sigma+iy\cdot\sigma is the corresponding scalar product and \chi is a finite positive measure on \mathds{R}^{D} – single-spin measure – satisfying certain conditions. In particular, we assume that (1.3) \forall a>0\qquad\int_{\mathds{R}^{D}}e^{a\sigma^{2}}\chi(d\sigma)<\infty,\qquad\sigma^{2}:=\sigma\cdot\sigma, and that \chi is isotropic , i.e., it satisfies \chi(UA)=\chi(A) , holding for all orthogonal transforms U\in O(D) and each Borel A\subset\mathds{R}^{D} . An example is the measure supported on the sphere \mathds{S}_{r}\subset\mathds{R}^{D} of radius r>0 , the restriction of which to this sphere is a uniform measure. It corresponds to the aforementioned model of isotropic rotors. Our result is the statement that, for all even D\in\mathds{N} , the function in ( 1.2 ) can be written in the form (1.4) Z_{N}(z)=Z_{N}(0)\prod_{j=1}^{\infty}(1+\gamma_{j,N}z^{2}),\quad z^{2}:=z\cdot z,\quad\gamma_{j,N}>0,\quad\sum_{j}\gamma_{j,N}<\infty, whenever \chi belongs to a subset of the set of isotropic measures on \mathds{R}^{D} , for which ( 1.3 ) and the following condition are verified (1.5) \widehat{\chi}(z):=\int_{\mathds{R}^{D}}e^{z\cdot\sigma}\chi(d\sigma)=C_{\chi}\prod_{j=1}^{\infty}(1+\gamma_{j}z^{2}),\quad\gamma_{j}>0,\quad\sum_{j}\gamma_{j}<\infty. The function \widehat{\chi} will be referred to as the Laplace transform of \chi . The function of z^{2} that appears on the right-hand side of ( 1.5 ) is a Laguerre entire function of the first kind. These are polynomials possessing real non-positive zeros only, or limits of such polynomials, taken in the topology of uniform convergence on compact subsets of \mathds{C} ; see [ 10 , pages 8–23] . In the following, by \mathcal{L} we denote the set of all such functions. Each f\in\mathcal{L} can be written in the form (1.6) f(\zeta)=C\zeta^{m}e^{\alpha\zeta}\prod_{j=1}^{\infty}(1+\gamma_{j}\zeta), with m\in\mathds{N}_{0} , \alpha\geq 0 , and the set \{\gamma_{j}\} (infinite, finite, or empty) that satisfies the condition in ( 1.5 ). In particular, e^{\alpha\zeta}\in\mathcal{L} , for all \alpha\geq 0 . Noteworthy, for D=1 , \varphi_{\chi}(t):=\widehat{\chi}(it) is the Fourier transform of \chi , which by ( 1.5 ) has only real zeros. The study of such functions – Riemannschen ganzen Funktionen – goes back to Riemann; see [ 10 , Chapter 3] . In particular, verifying the famous Riemann hypothesis concerning the zeros of the zeta function amounts to verifying this property for a certain measure; see [ 18 ] . 2. The Result In this section, we first introduce the set of single-spin measures \chi , which we are going to deal with. Then we formulate and discuss our result: the validity of the representation ( 1.4 ) under the condition that the single-spin measure \chi in ( 1.2 ) belongs to this set. 2.1. Strongly isotropic measures Let \mathcal{S}(\mathds{R}^{D}) and \mathcal{S}(\mathds{R}_{+}) be the Schwartz spaces of real-valued test functions defined on \mathds{R}^{D} and \mathds{R}_{+}:=[0,+\infty) , respectively. Let also O(D) denote the group of orthogonal transforms of \mathds{R}^{D} . For U\in O(D) and f\in\mathcal{S}(\mathds{R}^{D}) , we define f_{U} by the equality f_{U}(x)=f(Ux) . Set \mathcal{S}_{D}=\{f\in\mathcal{S}(\mathds{R}^{D}):\forall U\in O(D)\ f_{U}=f\} . It is known, see [ 2 , Theorem 2.1 and Note on page 245] , that for each f\in\mathcal{S}_{D} , there exists a unique \phi\in\mathcal{S}(\mathds{R}_{+}) such that f(x)=\phi(x^{2}) . Moreover, the map f\mapsto\phi is an isomorphism of the corresponding Schwartz spaces. A distribution, T\in\mathcal{S}^{\prime}(\mathds{R}^{D}) , is said to be isotropic if \forall U\in O(D)\quad T(f_{U})=T(f):=\int_{\mathds{R}^{D}}T(x)f(x)dx. By \mathcal{S}^{\prime}_{D} we denote the set of all isotropic T\in\mathcal{S}^{\prime}(\mathds{R}^{D}) . It is known that \mathcal{S}(\mathds{R}_{+}) is weakly dense in \mathcal{S}^{\prime}(\mathds{R}_{+}) . Let \mathcal{T} be the strongest topology on \mathcal{S}^{\prime}(\mathds{R}^{D}) , for which \phi\mapsto f can be extended to a continuous map from \mathcal{S}^{\prime}(\mathds{R}_{+}) to \mathcal{S}^{\prime}(\mathds{R}^{D}) . By \overline{\mathcal{S}_{D}} we denote the closure of \mathcal{S}_{D} in \mathcal{T} . Then, for each T\in\overline{\mathcal{S}_{D}} , one can find \tau\in\mathcal{S}^{\prime}(\mathds{R}_{+}) such that T(x)=\tau(x^{2}) . Obviously, \overline{\mathcal{S}_{D}}\subset\mathcal{S}^{\prime}_{D} . Definition 2.1 . A positive measure, \chi , on \mathds{R}^{D} is said to be strongly isotropic if: (a) it satisfies ( 1.3 ); (b) there exists T\in\overline{\mathcal{S}_{D}} such that (2.1) \chi(d\sigma)=T(\sigma)d\sigma=\tau(\sigma^{2})d\sigma. Furthermore, we say that an isotropic measure has the Lee-Yang property if it satisfies ( 1.5 ). The latter is related to all isotropic measures. The first relevant example is the aforementioned uniform measure on the sphere \mathds{S}_{r}\subset\mathds{R}^{D} , for which T(\sigma)=\delta_{r}(\sigma)=\delta(\sigma^{2}-r) , r>0 , \delta being Dirac’s function. Its Laplace transform is (2.2) \displaystyle\widehat{\chi}(z)=\int e^{z\cdot\sigma}\delta(\sigma^{2}-r)d\sigma=\pi^{D/2}w_{D}(z^{2},r), \displaystyle w_{D}(\zeta,r)=\sum_{n=0}^{\infty}\frac{\zeta^{n}r^{D/2+n-1}}{2^{2n}n!\Gamma(D/2+n)}, where \Gamma is Euler’s \Gamma -function. By means of the Pólya-Schur theorem, see [ 10 , pages