Spontaneous cluster formation in stoichiometric quantum critical systems

Abstract

Metallic systems with magnetic ions embedded which have been prepared to undergo a second-order phase transition at zero Kelvin (called the quantum critical systems) historically appear to fall into two distinct categories: (chemically) heavilydoped systems in which the unusual properties can be attributed to a disorder-induced distribution of Kondo shielding temperatures, and (nearly) stoichiometric systems where the departures from Fermi-liquid theory have been attributed to intrinsic instabilities. We show that this distinction between doped and stoichiometric systems is no longer a clear-cut boundary and that magnetic clusters associated with a distribution of Kondo shielding temperatures found in the heavily doped quantum critical Ce(Fe0:755Ru0:245)2Ge2 are also present in CeRu2Si2, a stoichiometric system close to a quantum critical point. By revisiting published data on CeRu2Si2 and comparing them to the results of Ce(Fe0:755Ru0:245)2Ge2, we show that clusters consisting of Ce-ions with their moments aligned with their neighbors exist in both systems at low temperatures and dominate the macroscopic response of these systems. We show these clusters form from a distribution of Kondo shielding temperatures which naturally arises from a distribution in inter-ionic separations; the chemical doping of ruthenium in Ce(Fe0:755Ru0:245)2Ge2 and zero-point motion in stoichiometric CeRu2Si2. We show that the dominant physics which drives heavily-doped systems, namely spontaneous formation of mag netic clusters, also plays a leading role in the response of homogenous systems. We investigate the presence of these spontaneous clusters in the heavily studied YbRh2Si2 where neutron scattering data have not been published. We show that the specific heat and susceptibility are naturally interpreted when ordered magnetic clusters are taken into account. Thus, cluster formation appears to be ubiquitous at the quantum critical point. We end with potential implications of the findings (that small changes in interionic separations ends up dominating the macroscopic response) to another class of highly correlated electron systems: the cuprate high Tc superconductors. We speculate that ionic displacements associated with hole doping leads to a pathway for the formation of Cooper pairs. We show that this pathway leads to an explanation for the many experimental observations, such as the striped phase and the hourglass dispersion.