next up previous contents
Next: E. Diffraction in Far Up: Dissertation Rainer Minixhofer Previous: C. General Algorithm for


D. From Boltzmann Distribution to Drift-Diffusion Current Equations

We consider a steady state situation and, for simplicity, a one-dimenstional geometry [208]. With the use of a relaxation time approximation as in (3.2) the BOLTZMANN equation becomes

$\displaystyle \frac{e F}{m^*} \frac{\partial f(v,x)}{\partial v} + v \frac{\partial f(v,x)}{\partial x} = \frac{f_{eq}(v,x)-f(v,x)}{\tau}$ (D.1)

Here the relation $ m^* \vec{v} = \vec{p} = \hbar
\vec{k}$ was used, which is valid for a parabolic energy band. Note that the charge $ e$ has to be taken with the proper sign of the particle (positive for holes and negative for electrons). A general definition of current density is given by

$\displaystyle J(x)=e \int v f(v,x) dv$ (D.2)

where the integral on the right hand side represents the first moment of the distribution function. This definition of current can be related to (D.1). After multiplying both sides by $ v$ and integrating over v one gets

$\displaystyle \frac{e F}{m^*} \int v \frac{\partial f(v,x)}{\partial v} dv + \i...
...brace{\int v f_{eq}(v,x) dv}_0 - \int v f(v,x) dv\right] = -\frac{J(x)}{e \tau}$ (D.3)

since the function $ v f_{eq}(v,x)$ is odd in $ v$, and its integral is therefore zero. Thus, one has from (D.3)

$\displaystyle J(x) = -e \frac{e \tau}{m^*} F \int v \frac{\partial f}{\partial v} dv - e \tau \frac{d}{dx} \int v^2 f(v,x) dv$ (D.4)

Integrating by parts yields

$\displaystyle \int v \frac{\partial f}{\partial v} dv = \underbrace{\left[v f(v,x)\right]_{-\infty}^{\infty}}_0 - \int f(v,x) dv = -n(x)$ (D.5)

and

$\displaystyle \int v^2 f(v,x) dv = n(x) \langle v^2 \rangle$ (D.6)

can be written, where $ \langle v^2 \rangle$ is the average of the square of the velocity defined as

$\displaystyle \langle v^2 \rangle = \frac{1}{n} \int v^2 f(v,x) dv$ (D.7)

Because of the equipartition theorem, for a purely one-dimensional treatment, the $ -\frac{3}{2}$ exponent in (3.3) may be replaced with $ -\frac{1}{2}$, while the appropriate thermal kinetic energy becomes $ \frac{k_B T}{2}$ instead of $ \frac{3 k_B T}{2}$.
The drift-diffusion equations are derived introducing the mobility $ \mu = \frac{e \tau}{m^*}$ and replacing $ \langle v^2 \rangle$ with its average equilibrium value $ \frac{k_B T}{m^*}$, therefore neglecting thermal effects. The diffusion coefficient $ D = \frac{\mu
k_B T_0}{e}$ (Einstein's relation) is also introduced, and the resulting drift-diffusion current is

$\displaystyle J_n = q n(x) \mu_n F(x) + q D_n \frac{dn}{dx}$    
$\displaystyle J_p = q p(x) \mu_p F(x) - q D_p \frac{dp}{dx}$ (D.8)

where q is used to indicate the absolute value of the electronic charge. Although no direct assumptions on the non-equilibrium distribution function $ f(v,x)$ was made in the derivation of (D.8), the choice of equilibrium (thermal) velocity means that the drift-diffusion equations are only valid for very small perturbations of the equilibrium state (low fields). The validity of the drift-diffusion equations is empirically extended by introducing field-dependent mobility $ \mu(F)$ and diffusion coefficient $ D(F)$, obtained from experimental measurements.
next up previous contents
Next: E. Diffraction in Far Up: Dissertation Rainer Minixhofer Previous: C. General Algorithm for

R. Minixhofer: Integrating Technology Simulation into the Semiconductor Manufacturing Environment