Asymptotic long-range order for the XY-model on random geometric graphs

Abstract: We study the classical $XY$-model on random geometric graphs $\mathcal{G}<em>{n, \varepsilon}$, which are obtained by sampling $n \in \mathbb{N}$ independent points in a finite domain $Ω\subset \mathbb{R}<sup>d$, $d \geq 2$, and connecting two points by and edge if their distance is of order $\varepsilon &gt; 0$. We refer to $\mathcal{G}</em>{n, \varepsilon}$ as the random environment. Letting $\varepsilon \to 0$ as $n \to \infty$ at a sufficiently slow rate, these graphs capture the geometry of $Ω$. Denoting the inverse temperature by $β$, we show that in the limit $β\to \infty$ at a rate depending on $n$ and $\varepsilon$, the $XY$-model on $\mathcal{G}_{n, \varepsilon}$ exhibits long range order in the sense that we prove a lower bound away from zero on the two-point function. Our result is quenched in the random environment: long-range order holds with large probability, converging to one as $n \to \infty$. To prove the statement, we show that with high probability the environment is sufficiently regular to apply a convexity argument and the Brascamp--Lieb inequality.