3 files changed, 506 insertions, 0 deletions
diff --git a/figs/anim_laser/Makefile b/figs/anim_laser/Makefile
new file mode 100644
index 0000000..9b66e37
--- /dev/null
+++ b/figs/anim_laser/Makefile
@@ -0,0 +1,12 @@
+PROJECTNAME=animate
+
+all: animate.dat
+
+run: animate.dat
+	python3 animate.py animate.dat &
+
+animate.dat:
+	julia laser.jl > animate.dat
+
+clean:
+	rm -f animate.dat
diff --git a/figs/anim_laser/animate.py b/figs/anim_laser/animate.py
new file mode 100644
index 0000000..1e6995a
--- /dev/null
+++ b/figs/anim_laser/animate.py
@@ -0,0 +1,103 @@
+from matplotlib import pyplot as pl
+from matplotlib import animation
+import sys
+
+# read data
+# time dependent data
+frames=[]
+## asymptotic data (located in the first block)
+#asym=[]
+infile=open(sys.argv[1],'r')
+row=[]
+for line in infile:
+    if line=='\n':
+        frames.append(row)
+        row=[]
+    else:
+        dat=[]
+        for n in line.split():
+            dat.append(float(n))
+        row.append(dat)
+infile.close()
+
+
+# set up plot
+fig = pl.figure()
+pl.subplot(211)
+axr=fig.gca()
+asym_rho, = axr.plot([],[],linewidth=3.5,color='#00FF00')
+cn_rho, = axr.plot([],[],color='#FF0000')
+
+pl.subplot(212)
+axJ=fig.gca()
+asym_J, = axJ.plot([],[],linewidth=3.5,color='#00FF00')
+cn_J, = axJ.plot([],[],color='#FF0000')
+
+# plot ranges
+xmax=0
+maxyr=0
+maxyJ=0
+for frame in frames:
+    for i in range(len(frame)):
+        if frame[i][1]>xmax:
+            xmax=frame[i][1]
+        if frame[i][2]>maxyr:
+            maxyr=frame[i][2]
+        if frame[i][3]>maxyr:
+            maxyr=frame[i][3]
+        if frame[i][4]>maxyJ:
+            maxyJ=frame[i][4]
+        if frame[i][5]>maxyJ:
+            maxyJ=frame[i][5]
+xmin=0
+minyr=0
+minyJ=0
+for frame in frames:
+    for i in range(len(frame)):
+        if frame[i][1]<xmin:
+            xmin=frame[i][1]
+        if frame[i][2]<minyr:
+            minyr=frame[i][2]
+        if frame[i][3]<minyr:
+            minyr=frame[i][3]
+        if frame[i][4]<minyJ:
+            minyJ=frame[i][4]
+        if frame[i][5]<minyJ:
+            minyJ=frame[i][5]
+
+# animate
+def init_plot():
+    axr.set_ylim(minyr,maxyr)
+    axr.set_xlim(xmin,xmax)
+    axJ.set_ylim(minyJ,maxyJ)
+    axJ.set_xlim(xmin,xmax)
+
+    axr.vlines(0,minyr,maxyr,linestyles="dotted")
+    axJ.vlines(0,minyJ,maxyJ,linestyles="dotted")
+    axJ.hlines(0,xmin,xmax,linestyles="dotted")
+def update(frame):
+    axr.set_title("t=% .3f fs" % (frame[0][0]))
+    xdata=[]
+    ydata=[]
+    asym_data=[]
+    cn_data=[]
+    for i in range(len(frame)):
+        xdata.append(frame[i][1])
+        asym_data.append(frame[i][2])
+        cn_data.append(frame[i][3])
+    asym_rho.set_data(xdata,asym_data)
+    cn_rho.set_data(xdata,cn_data)
+
+    xdata=[]
+    ydata=[]
+    asym_data=[]
+    cn_data=[]
+    for i in range(len(frame)):
+        xdata.append(frame[i][1])
+        asym_data.append(frame[i][4])
+        cn_data.append(frame[i][5])
+    asym_J.set_data(xdata,asym_data)
+    cn_J.set_data(xdata,cn_data)
+anim = animation.FuncAnimation(fig, update, frames=frames, blit=False, interval=100, repeat=True, init_func=init_plot)
+#anim.save('laser_schrodinger.m4v', fps=10, extra_args=['-vcodec', 'libx264'])
+pl.show()
diff --git a/figs/anim_laser/laser.jl b/figs/anim_laser/laser.jl
new file mode 100644
index 0000000..78cabff
--- /dev/null
+++ b/figs/anim_laser/laser.jl
@@ -0,0 +1,391 @@
+using FFTW
+using Printf
+using SpecialFunctions
+using LinearAlgebra
+
+# numerical values
+hbar=6.58e-16 # eV.s
+m=9.11e-31 # kg
+V_=6.6 # eV
+I=1e17 # W/m^2
+EF=3.3 # eV
+omega_=1.5498*4 # eV
+
+# dimensionless quantities
+V=1
+k=sqrt(EF/V_)
+epsilon=sqrt(hbar^3*I/(8*m*V_*omega_^2))
+omega=omega_/V_
+
+#epsilon=8*epsilon
+#omega=4*omega
+
+nu=60
+ell=20
+
+# for Crank-Nicolson
+xmin_cn=-100
+xmax_cn=100
+nx_cn=4000
+tmax=10
+# mu=dt/dx^2
+mu=0.1
+nt=floor(Int,tmax/(mu*((xmax_cn-xmin_cn)/nx_cn)^2))
+# only save up to memt time steps
+memt=10000
+
+
+# asymptotic value of \varphi_t(x)
+function phi_asym(x,T,gkz,k,epsilon,omega,V,nu,ell)
+  # compute residues
+  res=residue(x,T,gkz,k,epsilon,omega,V,nu)
+  times=compute_times(omega,nu,ell)
+  N=(2*nu+1)*(2*ell+1)
+  fun=zeros(Complex{Float64},N)
+  for m in 1:N
+    for n in -nu:nu
+      fun[m]+=-1im*res[n+nu+1]*exp(-1im*(k*k+n*omega)*times[m])
+    end
+  end
+  return fun
+end
+# asymptotic value of \partial\varphi_t(x)
+function dphi_asym(x,T,gkz,k,epsilon,omega,V,nu,ell)
+  # compute residues
+  res=dresidue(x,T,gkz,k,epsilon,omega,V,nu)
+  times=compute_times(omega,nu,ell)
+  N=(2*nu+1)*(2*ell+1)
+  fun=zeros(Complex{Float64},N)
+  for m in 1:N
+    for n in -nu:nu
+      fun[m]+=-1im*res[n+nu+1]*exp(-1im*(k*k+n*omega)*times[m])
+    end
+  end
+  return fun
+end
+
+# return times
+function compute_times(omega,nu,ell)
+  N=(2*nu+1)*(2*ell+1)
+  times=zeros(N)
+  for j in 1:N
+    times[j]=pi/(omega*(nu+0.5))*(j-1)
+  end
+  return times
+end
+
+# \kappa_m^{(\sigma)}
+function K(m,sigma,V,k,epsilon,omega)
+  return pow(V+2*epsilon*epsilon-k*k-sigma-m*omega,0.5,pi/2)
+end
+# mathfrak q_k
+function Q(k,V)
+  return sqrt(V-k*k)
+end
+
+# T_0
+function T0(k,V)
+  return 2im*k/(1im*k-sqrt(V-k*k))
+end
+
+# compute g
+function g(kappa,epsilon,omega,N)
+  B=zeros(Complex{Float64},N)
+  # prepare vector of B's
+  for i in 1:N
+    t=2*pi/(omega*N)*(i-1)
+    B[i]=exp(1im*epsilon*epsilon/omega*sin(2*omega*t)-kappa*4*epsilon/omega*cos(omega*t))
+  end
+  # take the fft
+  fft!(B)
+  # correct by a factor 1/N
+  for i in 1:N
+    B[i]=B[i]/N
+  end
+  return(B)
+end
+
+# index of m-n
+function mmn(n,nu)
+  return (n>=0 ? n+1 : n+4*nu+2)
+end
+
+# compute residues at -ik^2-in\omega
+function residue(x,T,gkz,k,epsilon,omega,V,nu)
+  # compute residue
+  out=zeros(Complex{Float64},2*nu+1)
+  if x>=0
+    for n in -nu:nu
+      for m in -nu:nu
+	out[n+nu+1]+=1im*gkz[m+nu+1][mmn(n-m,nu)]*exp(-K(m,0,V,k,epsilon,omega)*x)*T[m+nu+1]
+      end
+    end
+  else
+    for n in -nu:nu
+      out[n+nu+1]=(n==0 ? 1im*(exp(1im*k*x)-exp(-1im*k*x)) : 0)
+      for m in -nu:nu
+	out[n+nu+1]+=1im*T[m+nu+1]*gkz[m+nu+1][mmn(n-m,nu)]*exp(-1im*x*pow(k*k+n*omega,0.5,-pi/2))
+      end
+    end
+  end
+
+  return out
+end
+# compute \partial residues at -ik^2-in\omega
+function dresidue(x,T,gkz,k,epsilon,omega,V,nu)
+  # compute residue
+  out=zeros(Complex{Float64},2*nu+1)
+  if x>=0
+    for n in -nu:nu
+      for m in -nu:nu
+	out[n+nu+1]+=-K(m,0,V,k,epsilon,omega)*1im*gkz[m+nu+1][mmn(n-m,nu)]*exp(-K(m,0,V,k,epsilon,omega)*x)*T[m+nu+1]
+      end
+    end
+  else
+    for n in -nu:nu
+      out[n+nu+1]=(n==0 ? -k*(exp(1im*k*x)+exp(-1im*k*x)) : 0)
+      for m in -nu:nu
+	out[n+nu+1]+=pow(k*k+n*omega,0.5,-pi/2)*T[m+nu+1]*gkz[m+nu+1][mmn(n-m,nu)]*exp(-1im*x*pow(k*k+n*omega,0.5,-pi/2))
+      end
+    end
+  end
+
+  return out
+end
+
+
+# matching condition for residue
+function residue_matching(k,epsilon,omega,V,nu)
+  # preliminary: g^{(\kappa_m^{(0)})}
+  # g^{(\kappa_m^{(0)})}_n=gkz[m+nu+1][n+1] if n>=0 and gkz[m+nu+1][n+4*nu+2] if n<0
+  gkz=Array{Array{Complex{Float64}},1}(undef,2*nu+1)
+  for m in -nu:nu
+    gkz[m+nu+1]=g(K(m,0,V,k,epsilon,omega),epsilon,omega,4*nu+1)
+  end
+
+  # solve matching condition
+  G=zeros(Complex{Float64},2*nu+1,2*nu+1)
+  v=zeros(Complex{Float64},2*nu+1)
+  for n in -nu:nu
+    for m in -nu:nu
+      G[n+nu+1,m+nu+1]=((K(m,0,V,k,epsilon,omega)-1im*pow(k*k+n*omega,0.5,-pi/2))*gkz[m+nu+1][mmn(n-m,nu)]-epsilon*(gkz[m+nu+1][mmn(n-m+1,nu)]-gkz[m+nu+1][mmn(n-m-1,nu)]))
+    end
+  end
+  v[nu+1]=-2im*k
+  
+  return (G\v,gkz)
+end
+
+
+# fractional power with an arbitrary branch cut
+function pow(x,a,cut)
+  if(angle(x)/cut<=1)
+    return(abs(x)^a*exp(1im*angle(x)*a))
+  else
+    return(abs(x)^a*exp(1im*(angle(x)-sign(cut)*2*pi)*a))
+  end
+end
+
+# current
+function J(phi,dphi,x,epsilon,omega,t)
+  if x>=0
+    return(2*imag(conj(phi)*(dphi+2im*epsilon*sin(omega*t)*phi)))
+  else
+    return(2*imag(conj(phi)*dphi))
+  end
+end
+
+# compute \psi_t(x) using Crank-Nicolson
+function psi_cn(V,omega,epsilon,k,xs,ts,memt)
+  nx=length(xs)
+  nt=length(ts)
+
+  # length of output vector
+  len=min(nt,memt)
+  # how often to write to output
+  freq=ceil(Int,nt/memt)
+
+  # initialize
+  psis=zeros(Complex{Float64},nx,len)
+  times=zeros(Float64,len)
+
+  # previous time step
+  psi_prev=zeros(Complex{Float64},nx)
+
+  # init
+  for i in 1:nx
+    times[1]=0
+    psis[i,1]=(xs[i]<0 ? exp(1im*k*xs[i])+(T0(k,V)-1)*exp(-1im*k*xs[i]) : T0(k,V)*exp(-Q(k,V)*xs[i]))
+    psi_prev[i]=psis[i,1]
+  end
+  # the boundary condition is that psi at the boundary is constant (nothing needs to be done for this)
+
+  # matrix structure of the Crank-Nicolson algorithm
+  # diagonals of M
+  M0=zeros(Complex{Float64},nx-2)
+  Mp=zeros(Complex{Float64},nx-3)
+  Mm=zeros(Complex{Float64},nx-3)
+  v=zeros(Complex{Float64},nx-2)
+
+  for j in 1:nt-1
+    # print progress
+    progress(j,nt-1,1000)
+
+    # t_j
+    t=ts[j]
+    # t_{j+1}
+    tp=ts[j+1]
+    dt=tp-t
+    for mu in 2:nx-1
+      # x_mu
+      x=xs[mu]
+      dx2=(xs[mu+1]-xs[mu])*(xs[mu]-xs[mu-1])
+      # tridiagonal matrix
+      M0[mu-1]=1im/dt-1/dx2-(x<0 ? 0 : (V-2*epsilon*omega*cos(omega*tp)*x)/2)
+      if mu<nx-1
+	Mp[mu-1]=1/(2*dx2)
+      end
+      if mu>2
+	Mm[mu-2]=1/(2*dx2)
+      end
+      # right side
+      v[mu-1]=1im*psi_prev[mu]/dt-(psi_prev[mu+1]+psi_prev[mu-1]-2*psi_prev[mu])/(2*dx2)+(x<0 ? 0 : (V-2*epsilon*omega*cos(omega*t)*x)*psi_prev[mu]/2)
+      # correct for boundary conditions
+      # assumes the boundary condition is constant in time!
+      if mu==2
+	v[mu-1]-=psi_prev[1]/(2*dx2)
+      end
+      if mu==nx-1
+	v[mu-1]-=psi_prev[nx]/(2*dx2)
+      end
+    end
+
+    M=Tridiagonal(Mm,M0,Mp)
+
+    # copy to psi_prev
+    psi_prev[2:nx-1]=M\v
+
+    # write to output
+    if j%freq==0
+      psis[:,Int(j/freq)+1]=psi_prev
+      times[Int(j/freq)+1]=tp
+    end
+  end
+
+  return (times,psis)
+end
+
+# returns the times in b that are closest to those in a
+# assumes the vectors are ordered
+function nearest_times(a,b)
+  out=zeros(Int,length(a))
+  pointer=1
+  for i in 1:length(a)
+    if pointer==length(b)
+      out[i]=length(b)
+    end
+    for j in pointer:length(b)-1
+      if b[j+1]>a[i]
+	out[i]=(abs(b[j+1]-a[i])<abs(b[j]-a[i]) ? j+1 : j)
+	pointer=out[i]
+	break
+      end
+      # all remaining points are beyond the range
+      if j==length(b)-1
+	out[i]=length(b)
+	pointer=out[i]
+      end
+    end
+  end
+  return out
+end
+
+# space derivative using Crank Nicolson
+function deriv_cn(phi,i,xs)
+  if i==1
+    return (phi[i+1]-phi[i])/(xs[i+1]-xs[i])
+  end
+  if i==length(phi)
+    return (phi[i]-phi[i-1])/(xs[i]-xs[i-1])
+  end
+  return (phi[i+1]-phi[i])/(xs[i+1]-xs[i])/2 + (phi[i]-phi[i-1])/(xs[i]-xs[i-1])/2
+end
+
+################################################
+
+
+# print progress
+function progress(j,tot,freq)
+  if (j-1)%ceil(Int,tot/freq)==0
+    if j>1
+      @printf(stderr,"\r")
+    end
+    @printf(stderr,"%d/%d",j,tot)
+  end
+  if j==tot
+    @printf(stderr,"%d/%d\n",j,tot)
+  end
+end
+
+# print animation data using Crank Nicolson and compare to Laplace
+function print_anim_cn_cmp()
+  @printf(stderr,"epsilon=% .8e omega=% .8e k=% .8e V=% .8e z=% .8e T=% .8e gamma=% .8e\n",epsilon,omega,k,V,4*sqrt(nu)*epsilon/sqrt(omega),2*pi/(omega_/hbar)*1e15,sqrt((V-k^2)/(4*epsilon^2)))
+  xmin=-10
+  xmax=10
+  nx=200
+
+  # array of times for the asymptote
+  times=compute_times(omega,nu,ell)
+
+  # positions at which to compute phi
+  xs=Array(0:nx_cn)*(xmax_cn-xmin_cn)/nx_cn.+xmin_cn
+  # times at which to compute phi
+  ts=Array(0:nt-1)*tmax/nt
+
+  # compute matching condition
+  (T,gkz)=residue_matching(k,epsilon,omega,V,nu)
+
+  # compute wave function
+  (t_psi,psi)=psi_cn(V,omega,epsilon,k,xs,ts,memt)
+
+  # nearest times using Crank Nicolson
+  indices_t=nearest_times(times,t_psi)
+  indices_x=nearest_times(Array(0:nx-1)*(xmax-xmin)/nx.+xmin,xs)
+
+  ps_cn=Array{Array{Complex{Float64},1}}(undef,nx)
+  ps_asym=Array{Array{Complex{Float64},1}}(undef,nx)
+  dps_cn=Array{Array{Complex{Float64},1}}(undef,nx)
+  dps_asym=Array{Array{Complex{Float64},1}}(undef,nx)
+  for i in 1:nx
+    progress(i,nx,1000)
+    x=xmin+(xmax-xmin)*(i-1)/nx
+    ps_cn[i]=Array{Complex{Float64}}(undef,length(times))
+    dps_cn[i]=Array{Complex{Float64}}(undef,length(times))
+    for j in 1:length(times)
+      # phase change in magnetic gauge
+      phase=(x<0 ? 1 : exp(-2im*epsilon*sin(omega*times[j])*x))
+      ps_cn[i][j]=phase*psi[indices_x[i],indices_t[j]]
+      dps_cn[i][j]=phase*deriv_cn(psi[:,indices_t[j]],indices_x[i],xs)-(xs[indices_x[i]]<0 ? 0 : 2im*epsilon*sin(omega*times[j])*psi[indices_x[i],indices_t[j]])
+    end
+    ps_asym[i]=phi_asym(x,T,gkz,k,epsilon,omega,V,nu,ell)
+    dps_asym[i]=dphi_asym(x,T,gkz,k,epsilon,omega,V,nu,ell)
+  end
+
+  # print values at each time
+  for j in 1:length(times)
+    if times[j]>tmax
+      break
+    end
+    for i in 1:nx
+      x=(xmin+(xmax-xmin)*i/nx)
+      # remove point at 0
+      if i!=Int(nx/2)
+	@printf("% .8e % .8e % .8e % .8e % .8e % 8e\n",times[j]*hbar/V_*1e15,x*sqrt(hbar^2*1.6e-19/(2*m*V_))*1e9,abs(ps_asym[i][j])^2,abs(ps_cn[i][j])^2,J(ps_asym[i][j],dps_asym[i][j],x,epsilon,omega,times[j]),J(ps_cn[i][j],dps_cn[i][j],x,epsilon,omega,times[j]))
+      end
+    end
+    print('\n')
+  end
+end
+
+
+print_anim_cn_cmp()