2.4.2.1 The Drift-Diffusion Model

By multiplying (2.2) with the first two moments of the distribution function $ \Phi_0=1$ and $ \Phi_1=\hbar\mathbf{k}$, integration over $ \mathbf{k}$ space, using a parabolic dispersion relation, and applying the macroscopic relaxation-time approximation (RTA) for the integral of the collision operator, the following equation system can be derived

$\displaystyle \ensuremath{{\mathbf{\nabla}}}\cdot \ensuremath{{\mathbf{J_n}}}$ $\displaystyle = \ensuremath {\mathrm{q}}R + \ensuremath {\mathrm{q}}\frac{\partial n}{\partial t}\ ,$ (2.4)
$\displaystyle \ensuremath{{\mathbf{\nabla}}}\cdot \ensuremath{{\mathbf{J_p}}}$ $\displaystyle = - \ensuremath {\mathrm{q}}R - \ensuremath {\mathrm{q}}\frac{\partial p}{\partial t}\ ,$ (2.5)
$\displaystyle \ensuremath{{\mathbf{J_n}}}$ $\displaystyle = \ensuremath {\mathrm{q}}n\mu_n \mathbf{E} + \ensuremath {\mathrm{q}}D_n \ensuremath{{\mathbf{\nabla}}}n\ ,$ (2.6)
$\displaystyle \ensuremath{{\mathbf{J_p}}}$ $\displaystyle = \ensuremath {\mathrm{q}}p\mu_p \mathbf{E} - \ensuremath {\mathrm{q}}D_p \ensuremath{{\mathbf{\nabla}}}p\ .$ (2.7)

In these equations $ \mathbf{J}$ denotes the current density, $ R$ the net recombination rate, $ \mu$ the mobility, $ {\mathbf{E}}$ the electric field, and $ D$ the diffusion coefficient. Together with (2.1), these basic semiconductor equations form the drift-diffusion model which, due to its simplicity, is widely used for the simulation of semiconductor devices.

A. Gehring: Simulation of Tunneling in Semiconductor Devices