What does non-Gaussian mean?
What does non-Gaussian mean?
-
α2=⟨x4⟩3⟨x2⟩−1 is the "non-Gaussian parameter", used as a measure of dynamical heterogeneities
-
For a Gaussian distribution, ⟨x4⟩=3⟨x2⟩, so α2=0
-
Can be calculated from particle-tracking
What does non-Gaussian mean?
-
A random walk (diffusion) is Gaussian at any given time:
-
P(x,t)=12√πDt−ex24Dt
-
-
So if the particles aren’t diffusing, then α2≠0
-
For caged particles, α2≈−15
-
-
Closely related to caging and cage-breaking behavior
-
In a cage, particles move very small distances
-
Cage-breaking would involve much larger jumps
-
The distribution of "step sizes" would be very non-Gaussian
-
What does non-Gaussian mean?
-
α2 measures how "not gaussian" the distribution is
Diffusion |
Mixed |
Caging |
α2=0 |
α2>0 |
α2=−15 |
Data
What kind of step distributions do we actually get?
What kind of step distributions do we actually get?
What kind of step distributions do we actually get?
The Non-Gaussian Parameter α2
System Size Dependence of α2
Lines drawn are A(ϕ⋆−ϕ)n, and ϕ⋆ is fitted
System Size Dependence of α2
α₂ goes to ∞?
-
Provocative, but inconclusive
-
ϕ⋆=0.600±0.001 is an unusual density
-
There is less than two orders of magnitude on this plot
-
3.5 is a long ways from ∞
-
Maximizing α2
Maximizing α2
-
Start with the sum of two gaussians P(r)∝Ar2σ2e−r2σ2+Br2e−r2
-
Increasing σ while decreasing AB gives a larger α2
-
More specifically: For a given σ, AB=σ21+σ2 yields the maximum α2=(δ2−1)24δ2
-
Back to the Step Distributions
-
As we increase density, we get an increased separation
-
As time varies, the ratio AB varies
Approximating α2 with Aging
-
Prepare a state at ϕ0=0.55 at equilibrium
-
Fast quench it to some density ϕ
-
Calculate max as a function of time
Aging \alpha_2
Cartoon
|
Simulation
|
Close to Jamming
Close to Jamming

-
Glassy behavior starts at \phi \approx 0.55 – 0.59
-
Jamming is at \phi \approx 0.64
-
What happens if we go really close to jamming?
-
Permanent Caging
-
Floaters
-
Step Distributions Close to Jamming
Close to Jamming: |
Glassy: |
Peak at 10^{-2} : Tight Cages |
Peak at 10^{-1} : Caged Floaters |
Peak at 10^{-1} : Caged Particles |
Peak at 10^{0} : Rearrangements |
Step Distributions Close to Jamming
Solid: Backbone |
Dashed: Floaters |
Thin Dotted: Both |
Rearrangements
-
Each dot represents a single system
-
At 1000 timepoints over the course of the simulation, a snapshot was taken, and the structure was "minimized" to find the inherent structures visited
-
Systems seem able to access either a very limited number of inherent structures, or a very large number
Other Directions
Fitting The Step Distributions
-
Fit the step distributions to the sum of two gaussians
-
Figure out how that scales with time and ϕ
-
This is hard.
That’s all.
Thanks!
-
Corey O’Hern, Mark Shattuck, Christine Jacobs-Wagner
-
Brad Parry, Ivan Surovtsev, Eric Dufresne, and everyone I talked to
-
Sackler, PEB, and HHMI