Gaussian Free Field in the background of correlated random clusters, formed by metallic nanoparticles

TL;DR

This study investigates how metallic nanoparticles influence the electrostatic potential in disordered 2D media, revealing a softening effect and scale-invariant properties linked to the Ising model correlations.

Contribution

It introduces a model connecting metallic nanoparticle correlations with Gaussian free fields and critical exponents, providing new insights into electrostatic potential fluctuations.

Findings

MNPs soften the random potential and reduce fluctuations.

Electrostatic correlation functions remain logarithmic across temperatures.

Critical exponents vary with temperature, linked to Ising model correlations.

Abstract

The effect of metallic nano-particles (MNPs) on the electrostatic potential of a disordered 2D dielectric media is considered. The disorder in the media is assumed to be white-noise Coulomb impurities with normal distribution. To realize the correlations between the MNPs we have used the Ising model with an artificial temperature $T$ that controls the number of MNPs as well as their correlations. In the $T\rightarrow 0$ limit, one retrieves the Gaussian free field (GFF), and in the finite temperature the problem is equivalent to a GFF in iso-potential islands. The problem is argued to be equivalent to a scale-invariant random surface with some critical exponents which vary with $T$ and correspondingly are correlation-dependent. Two type of observables have been considered: local and global quantities. We have observed that the MNPs soften the random potential and reduce its statistical fluctuations. This softening is observed in the local as well as the geometrical quantities. The correlation function of the electrostatic and its total variance are observed to be logarithmic just like the GFF, i.e. the roughness exponent remains zero for all temperatures, whereas the proportionality constants scale with $T-T_c$. The fractal dimension of iso-potential lines ($D_f$), the exponent of the distribution function of the gyration radius ($\tau_r$), and the loop lengths ($\tau_l$), and also the exponent of the loop Green function $x_l$ change in terms of $T-T_c$ in a power-law fashion, with some critical exponents reported in the text. Importantly we have observed that $D_f(T)-D_f(T_c)\sim\frac{1}{\sqrt{\xi(T)}}$, in which $\xi(T)$ is the spin correlation length in the Ising model.