Evidence for magnetic clusters in stoichiometric quantum critical CeRu$_2$Si$_2$

TL;DR

This paper provides evidence that magnetic clusters are present in stoichiometric CeRu$_2$Si$_2$, challenging the traditional view that such clusters only occur in heavily-doped quantum critical systems, and suggests a unified understanding of quantum criticality.

Contribution

It demonstrates that magnetic clusters exist in a stoichiometric quantum critical system, blurring the distinction between doped and pure systems in quantum critical behavior.

Findings

Magnetic clusters are present in CeRu$_2$Si$_2$ at low temperatures.

Ce-ion moments within clusters align with neighbors.

Cluster formation influences quantum critical phenomena.

Abstract

Systems that have been prepared to undergo a second-order phase transition at zero Kelvin, the so-called quantum critical systems, appear to fall into two categories: (chemically) heavily-doped systems where the unusual properties can be related to a disorder-induced distribution of Kondo shielding temperatures, and (almost) stoichiometric systems where the departures from Fermi-liquid theory have been attributed to intrinsic instabilities. Here we show that this distinction is not as clear cut and that magnetic clusters associated with a distribution of Kondo shielding temperatures are also present in CeRu$_2$Si$_2$, a system close to a quantum critical point. By revisiting published data on this system and comparing them to the results for heavily-doped quantum critical Ce(Ru$_{0.755}$Fe$_{0.245}$)$_2$Ge$_2$, we show that clusters exist in both systems at low temperatures, and that the moments of the Ce-ions within these clusters have all lined up with their neighbors. This implies that the dominant physics that drives heavily-doped systems, namely spontaneous formation of magnetic clusters, should also play a leading role in the response of homogenous systems. This represents a notable departure of how the physics that governs quantum critical points is treated in the literature.