James P. Sethna: Research 1990  1994
In this past five years, I've been mostly interested in the
dynamics of large interacting systems, with and without disorder.
I've gradually moved from glassy systems without intrinsic
disorder, through the dynamics of disordered systems undergoing
martensitic transformations, and am now making progress on understanding
the dynamics of hysteresis loops and void formation in electromigration.
I've also always had a compelling interest in understanding data
and fixing problems: my colleagues have led me into many fascinating
problems over the past few years.
Immersed in a rich scientific environment, I've been drawn into several
other projects through collaborations and conversations with my
colleagues; I've grouped these into two broad categories. In addition,
there are three topics on which my students and I have made substantial
progress: slow relaxation, tweed in martensites, and hysteresis.
I. Talking with Experimentalists
Grappling with the real world is not only our job, but is also the source
of our inspiration.

Quasiparticle Bound States at a Superconducting Vortex Line:
Harald Hess's ZeroBias Tunneling Peak. (Shore, Huang, Dorsey;
41)

Level Repulsion and Inhomogeneous Broadening:
Ambrose and Sievers Find a Cusp.
(52)

FiniteSize Fluctuations in Sliding ChargeDensity Waves:
Rob Thorne's Skepticism Is Vindicated. (Myers;
63, 64)

Void Dynamics and Electromigration:
Moeckly and Buhrman's HighT_c River Deltas. (Wickham;
74)

Null Results:
Our Model for Bodenschatz's Defect Chaos Doesn't Correlate. (Roberts;
72)
II. Talking with Theorists

Crack Growth Laws from Symmetry:
Wash, Ingraffea, and Desert Floral Foam.(Hodgdon;
57)

Negative Specific Heat:
The Danes Melt a Copper Cluster.
(65)

Atomic Tunneling from a STM/AFM Tip:
Mark Stiles' Phonons give an Ohmic Coupling.(Louis;
73)
III. Slow Relaxation
Several years ago, I spent a few months with Joel Shore and
Ming Huang months working out some of Daniel Fisher's ideas on a scaling
theory for the glass transition. Barrier heights were supposed to
diverge as the ``ideal glass transition'' temperature $T_0$ was
approached, and the resulting slowdown of the dynamics led to the
diverging viscosities characteristic of glass formers. Abstract thought
proving inadequate, Joel and I started trying to find a tangible model
system for studying diverging barrier heights, and Shelly Shumway and I
started thinking about ways to accelerate more traditional glass
simulations.

Logarithmic Growth in a Quenched Ising Model.
(Shore, Holzer;
50, 55, 59)

Evolving Enzymes:
Accelerating Relaxation in Glassy Systems. (Shumway;
51, 54)
Despite this promising beginning, when students graduate their projects
usually leave with them. I got dragged into some other directions, which
frankly were more fun...
The field of disordered systems is predicated on the assumption that by
studying some obscure, technologically irrelevant systems (spin glasses,
chargedensity waves, ...) we can find general principles which will
apply to practical problems. I like our work on tweed, because we've
mapped a practical problem in technologically important shapememory
alloys into the least likely of theoretical models: the infiniterange
spinglass, known and loved by practitioners of the black art of
replica symmetry breaking.

What is Tweed?.

Tweed is a Spin Glass.
V.
Hysteresis and Hierarchies:
Dynamics of DisorderDriven FirstOrder Phase Transformations
(Dahmen, Perkovic';
66, 68, 69,
75)
Even though everything is made of atoms that are 10^{8} cm and wiggle
at 10^{13} Hz, interesting things still happen on human length and
time scales. Nineteenth century physics mastered hydrodynamics and
elastic theory  describing systems where the fluctuations and the
discreteness don't impinge on the behavior. In the seventies and the
eighties, we mastered critical phenomena in equilibrium systems: near a
continuous phase transition, the rapid fluctuations on the microscale
propagate upward to human length and time scales, if one is near the
critical temperature. Much attention has been focused in recent times
on systems where the dynamics has events on all length and time scales:
crackling noises, avalanches, earthquakes, ... We have been working on
a class of models exhibiting
hysteresis, where avalanches occur on
large scales because we are in the neighborhood of a critical point.
We're trying to blame avalanches on plain old criticality.

The ReturnPoint Memory Effect, or Wiping Out.

The Critical Point.

The Epsilon Expansion.
Links Back

Entertaining Science done at

LASSP.
Last modified: January 15, 1995
James P. Sethna, sethna@lassp.cornell.edu
Statistical Mechanics: Entropy, Order Parameters, and Complexity,
now available at
Oxford University Press
(USA,
Europe).