Stationary semiconductor equations modeling avalanche generation (Q1917986)

The paper deals with a mixed boundary value problem for the system of elliptic equations modelling the semiconductor avalanche devices, where the generation of charged particles is controlled by avalanche due to impact ionization. The following system of PDEs for the electron and hole densities \(n\), \(p\) and the electrostatic potential \(\psi\) is considered \[ -\text{div }J_1= \alpha_1(\nabla\psi)|J_1|+ \alpha_2(\nabla\psi)|J_2|,\;-\text{div }J_2= \alpha_1(\nabla\psi)|J_1|+ \alpha_2(\nabla\psi)|J_2|,\tag{1} \] \[ -\Delta\psi= p-n+f,\tag{2} \] where \(J_1=D_1\nabla n-\mu_1n\nabla\psi\) is the conduction current density of electrons, \(J_2=-(D_2\nabla p+\mu_2 p\nabla\psi)\) is the conduction current density of holes. Here \(D_1\), \(D_2\) are the diffusion coefficients of electrons and holes, respectively, while \(\mu_1\), \(\mu_2\) are their mobility coefficients. These coefficients are generally field dependent, but in the present paper \(D_i\) and \(\mu_i\) are assumed to be positive constants \((i=1,2)\). System (1)--(2) is considered in a bounded domain \(\Omega\subset\mathbb{R}^d\) (\(d=2\) when the semiconductor device may be regarded as an intrinsically two-dimensional structure, or \(d=3\) in the general case). The boundary of \(\Omega\), \(\partial\Omega\) is supposed to be Lipschitzian, \(\partial\Omega=\Gamma_0\cup\Gamma_1\), \(\Gamma_0\) open in \(\partial\Omega\). Then the boundary conditions on \(n\), \(p\) and \(\psi\) are as follows: \[ n=n_0,\;p=p_0,\;\psi=\psi_0\text{ on }\Gamma_0,\;J_1\cdot\nu= J_2\cdot\nu= \partial\psi/\partial\nu= 0\text{ on }\Gamma_1,\tag{3} \] where \(\nu\) denotes the unit outward normal along \(\partial\Omega\). The functions \(n_0\), \(p_0\), and \(\psi_0\) represent the prescribed boundary values of densities \(n\), \(p\) and the electrostatic potential \(\psi\) at the Ohmic contact \(\Gamma_0\). The authors' main result is the proof of the existence of a weak solution to (1)--(3).