Abstract:
Spin glasses are paradigmatic examples of systems with quenched
disorder. At sufficiently low temperature, Tc, they are thought to
pass from a paramagnetic phase to an ordered, spin-glass phase.
Despite thirty years of study, however, the nature of this possible
spin-glass phase remains controversial. Experiment can provide no easy
answers since at temperatures near and below Tc, spin glass
dynamics slows dramatically and, consequently, all macroscopic spin
glass systems are far from equilibrium.
But as time passes spatial correlations must evolve in a system that
begins in a random initial condition, so that the length scale,
L_{eq}(t_w), over which spin glasses are locally equilibrated must
grow with waiting time, t_w. In many classical systems, we
understand the growth of such domains of local order. In Ising
ferromagnets, for example, they correspond roughly to regions of
up and down order, in which spins have correlations
characteristic of infinite equilibrated systems. In spin glasses, on
the other hand, development of local equilibrium is ill-understood,
even qualitatively. What is the form of spin-glass order that
develops? How many different long-time states can be observed in a
given system region of scale-L? How do these relate to any putative
infinite system equilibrium order in the first place? I will introduce
a framework for posing and then addressing these questions and will
provide a limited set of answers.