Full Aging in Spin Glasses

TL;DR

This paper provides strong evidence that spin glass magnetization decay functions scale with the waiting time $t_{w}$, highlighting the importance of cooling protocols and stationary term subtraction in observing this scaling.

Contribution

It demonstrates that the TRM magnetization decay in spin glasses scales with $t_{w}$ and shows how cooling rate and stationary term subtraction influence this scaling.

Findings

Scaling improves with faster cooling rates.

Almost perfect $t/t_{w}$ scaling achieved for $t_{c}^{eff}<20s$.

Subtracting stationary terms refines decay curve overlap.

Abstract

The discovery of memory effects in the magnetization decays of spin glasses in 1983 began a large effort to determine the exact nature of the decay. While qualitative arguments have suggested that the decay functions should scale as $t_{w}$, the only time scale in the system, this type of scaling has not yet been observed. In this letter we report strong evidence for the scaling of the TRM magnetization decays as a function of $t_{w}$. By varying the rate and the profile that the sample is cooled through its transition temperature to the measuring temperature, we find that the cooling plays a major role in determining scaling. As the effective cooling time decreases, $\frac {t}{t_{w}}$scaling improves and for $t_{c}^{eff}<20s$ we find almost perfect $\frac{t}{t_{w}}$ scaling. We also find that subtraction of a stationary term from the magnetization decay has a small effect on the scaling but changes the form of the magnetization decay and improves overlap between curves produced with different $t_{w}$.