Dynamic scaling in spin glasses

TL;DR

This paper demonstrates a strict power-law decay in spin autocorrelation functions at the glass transition in spin glasses, confirming critical scaling relations through new and revisited experimental data, with implications for broader glassy systems.

Contribution

It provides the first experimental verification of a scaling relation involving independent exponents in spin glasses, using new NSE data and analysis of historical results.

Findings

Power-law decay of spin autocorrelation at T_g

Verification of scaling relation with independent exponents

Non-exponential relaxation above the phase transition

Abstract

We present new Neutron Spin Echo (NSE) results and a revisited analysis of historical data on spin glasses, which reveal a pure power-law time decay of the spin autocorrelation function $s(Q,t) = S(Q,t)/S(Q)$ at the glass temperature $T_g$, each power law exponent being in excellent agreement with that calculated from dynamic and static critical exponents deduced from macroscopic susceptibility measurements made on a quite different time scale. It is the first time that this scaling relation involving exponents of different physical quantities determined by completely independent experimental methods is stringently verified experimentally in a spin glass. As spin glasses are a subgroup of the vast family of glassy systems also comprising structural glasses, other non-crystalline systems living matter the observed strict critical scaling behaviour is important. Above the phase transition the strikingly non-exponential relaxation, best fitted by the Ogielski (power-law times stretched exponential) function, appears as an intrinsic, homogeneous feature of spin glasses.