Search CTRL + K

Lecture 05 - Various Applications

Let us consider the Heisenberg Spin chain equation:

Su=S×Sxx

If we identify with its tangent S→T then we obtain:

Tu=T×Tss

If we work on the Binormal frame, recall:

(TNB)s=(0κ0−κ0τ0τ0)(TNB)

Then we obtain an equation that is both a binormal and normal flow:

Tu=T×(κN)=κsB+κτN

If we use the following compatibility condition:

Tus=TsuNus=NsuBus=Bsu

We obtain the following result (Da Rios System):

κu=−2κsτ−κτsτu=[κssκ−τ2]s−κκs

Thus if we use our (favourite) Hasimoto map we obtain our NLSE again:

iψu=ψss+12|ψ|2ψ
'Recipe'

Physical Equation ↔FrenetSerret Moving Curve ↔mapHasimoto NLSE(+family)

Inhomogeneous Heisenberg Spin Chain

H=∑fi(s)Si⋅Si+1, where fi is a function of s.
Continuum case:

St=(fS×Sx)x↔Tu=(fT×Ts)s=(fκ)sB−fκτN

('Homework') We will obtain after some computation:

iψu+(fψ)ss+2ψ(f|ψ|2+∫−∞∞f(s′)|ψ|2ds′)=0

This is a generalization of the NLSE.

Linear Induction Approximation (LIA)

Suppose vij is a velocity vector for a small neighborhood around a point γi, with k is the vortex strength:

γij=γi−γjvij=−k4π∫|γij|−3∂γij∂si×γij dsj

LIA concept in the limiting case of a vortex filament of infinitesimal core size (long distance effects are considered)

γij=γij(ξ,t)=γi(si,t)−γj(si+ξ,t)γij=a1ξ+a2ξ2+...

where a1=∂γij∂ξ and a2=12∂2γij∂ξ.

∂γij∂si=∂γij∂ξ∂γij∂si×γij=a1×a2ξ2+O(ξ3)|γij|=γij⋅γij=(a1ξ+a2xi2+...)⋅(a1ξ+a2xi2+...)=|a1||ξ|1+2a1⋅a2|a1|ξ+...

Thus we obtain:

vij=k4π∫[a1×a2|a1|31ξ+O(1)]ds=k4πa1×a2|a1|3log(1|ξ|)+O(1)

Now we let Γ=k4Π|a1|3log(1/ξ) then the structure of our velocity is simply vij=Γγξ×γξξ.