In [1]:
import numpy as np
from scipy import linalg
import plotly.plotly as py
import plotly.graph_objs as go
from plotly import tools
from plotly.offline import init_notebook_mode, iplot
from ipywidgets import interactive
init_notebook_mode()
def fermi_gas_1D(n=1):
n_down = (n + 1) // 2
n_up = n // 2
momenta = np.linspace(-10, 10, 101)
energies = momenta**2
data = [go.Scatter(x=momenta, y=energies, hoverinfo='none', line={'color':'black'})]
discrete_momenta = np.arange(-19, 20)
discrete_energies = discrete_momenta**2
sort_order = np.argsort(discrete_energies)
discrete_momenta = discrete_momenta[sort_order]
discrete_energies = discrete_energies[sort_order]
data.append(go.Scatter(x=discrete_momenta[:n_down], y=discrete_energies[:n_down] - 2,
marker=go.Marker(color='red', symbol='circle'), line={'width':0}))
data.append(go.Scatter(x=discrete_momenta[:n_up], y=discrete_energies[:n_up] + 3,
marker=go.Marker(color='blue', symbol='circle'), line={'width':0}))
layout = go.Layout(
showlegend=False,
autosize=False,
width=600,
height=300,
xaxis={'title':'k'},
yaxis={'title':'E'})
fig = go.Figure(data=data, layout=layout)
iplot(fig, show_link=False)
fermi_gas_1D = interactive(fermi_gas_1D, n=(1, 39))
def plot_dos(dim=1, mu=1., T=0.001):
E = np.linspace(0, 2, 300)
n = E**((dim - 2) / 2) / (np.exp((E - mu) / T) + 1)
data = [go.Scatter(x=E, y=n, hoverinfo='none', line={'color':'black'})]
layout = go.Layout(
showlegend=False,
autosize=False,
width=600,
height=300,
xaxis={'title':'E'},
yaxis={'title':'n'})
iplot(go.Figure(data=data, layout=layout), show_link=False)
dos = interactive((lambda dim=1., mu=1.: plot_dos(dim, mu)), dim=(1, 3), mu=(0., 2.))
dos_finite = interactive(plot_dos, dim=(1, 3), mu=(0., 2.), T=(0.001, .1, 0.01))
NB: Electrons are my favorite particle
Q: What is similar between electrons and phonons?
Focus on differences first, forget the lattice exists (imagine the nuclei form an average potential for electrons).
Electrons obey Schrödinger equation:
$$ H = -\frac{\hbar^2}{2m}\left(\frac{\partial^2}{\partial x^2} + \frac{\partial^2}{\partial y^2} + \frac{\partial^2}{\partial z^2}\right) $$Take $L\times L \times L$ box with periodic boundary conditions: $\psi(L, y, z) = \psi(0, y, z)$, same for $y$ and $z$.
Wave vectors are now quantized: $k_{x,y,z} = 2\pi n_{x, y, z}/ L$
...and energies as well: $E(n_x, n_y, n_z) = \frac{(2\pi\hbar)^2}{mL^2}(n_x^2 + n_y^2 + n_z^2)$
In [2]:
fermi_gas_1D
So:
In [3]:
dos
Fermi-Dirac distribution (check wiki: WP:Fermi Distribution):
$ f(E, T) = \frac{1}{e^{(E - \mu)/k_\text{B}T} + 1}$ ($\mu$ is chemical potential, energy cost for adding an extra electron)
At finite temperature $n(E) = g(E)\times f(E, T)$
In [4]:
dos_finite
$n \approx (1/2) g(E_F) \times k_B T$ electrons have their energy increased by $\delta E \sim k_B T$
$$E(T) - E(0) \approx g(E_F) (k_b T)^2$$So: $$ C_{v, e} \approx g(E_F) k_b^2 T = \ldots = \frac{3}{2} N k_B \frac{T}{T_F}$$
($T_F = E_F/ k_B$ is Fermi temperature)
At room temperature $C_{v, f} = 3Nk_B \gg C_{v, e}$.
At $T \to 0$, $C_{v, f} \sim T^3$, while $C_{v, f} \sim T$.