4

In solid-state physics at $T=0 K$ considering the semi-classical approximation for electrons, I understand physically that if a band is completely filled, the electrons can't move and there is no conduction, so there is no current. And of course, if the band is empty there is impossible to have a current because there are no electrons that can move. But I cannot understand why the mathematical condition for a completely filled or empty band is

$$ \iiint_{Brillouin Zone} \frac{2 \cdot d^3 \vec{k}}{(2\pi)^3} \cdot v_g(\vec{k})=0$$

where $v_g(\vec{k})=\vec{\nabla}_{\vec{k}} \left[ \epsilon(\vec{k}) \right]$ is the group velocity.

Roger V.
  • 58,522

1 Answers1

7

$\mathbf{v}_g(\mathbf{k})$ is the velocity of electron in state $\mathbf{k}$. The full current carried by all the electrons is obtained by integrating over all the filled states of the Brillouin zone: $$ \mathbf{j}\propto \int_{\text{BZ}}\mathbf{v}_g(\mathbf{k})n(\mathbf{k})d^3\mathbf{k}, $$ In many calculations $n(\mathbf{k})$ is just the Fermi function $$ n(\mathbf{k}) =\frac{1}{e^{\beta(E(\mathbf{k})-\mu)}+1}, $$ however, when we speak about a completely filled band at zero temperature, the occupation factor is $1$ everywhere in the Brillouin zone, i.e., out integral becomes simply $$ \mathbf{j}\propto \int_{\text{BZ}}\mathbf{v}_g(\mathbf{k})d^3\mathbf{k} = 0, $$ since, if it were non-zero, we would have a non-zero current.

Example
Let us consider a one-dimensional tight-bidning chain with the dispersion law $$ E(k)=-\Delta\cos(ka), $$ where $2\Delta$ is the band width and $a$ is the lattice spacing.
The Brillouin zone is $$ -\frac{\pi}{a}\leq k \leq \frac{\pi}{a}, $$ whereas the electron group velocity is $$ v_g=-\frac{1}{\hbar}\frac{\partial E(k)}{\partial k}= \frac{\Delta a}{\hbar}\sin(k a) $$ The integral $$ \int_{BZ} v_g(k)dk=\int_{-\frac{\pi}{a}}^{\frac{\pi}{a}}\frac{\Delta a}{\hbar}\sin(k a) dk $$ is zero, since it is an integral over an odd function in symmetric limits.

Roger V.
  • 58,522
  • Could you please explain with math (or a simple example) why if the band is completely filled the integral of the velocity over the Brillouin zone is zero? The main problem is that if someone give me any expression for the energy of the band/velocity of the wave, I'm not sure about the limits of the integrals to check if this is zero or non-zero – user239504 Feb 04 '22 at 14:00
  • 1
    @user239504 I added an example. – Roger V. Feb 04 '22 at 14:22
  • Sorry to add more questions, but just to fully understand it, in this particular example at T=0K it doesn't matter if the band is full or not right? For example, the limits of integration are for a specific energy $E_0$ in the range $(-k_0, +k_0)$, so the current is always zero by symmetry, or am I missing something? – user239504 Feb 04 '22 at 18:21
  • 2
    @user239504 For example, the limits of integration are for a specific energy E0 in the range (−k0,+k0), so the current is always zero by symmetry-That happens at zero electric field, and there the current is obviously zero. When you apply an electric field, the filling factor slightly changes (approximately, the states between $(-k_0+\epsilon,k_0+\epsilon)$ remain filled), and there, the current is non-zero for a partially filled band. However, for a filled band, the electrons cannot redistribute themselves, and the current continues to be a zero even in a non-zero electric field. – Archisman Panigrahi Feb 04 '22 at 19:08