Existence and stability of stationary solutions to the compressible Navier-Stokes-Poisson equations (Q2630063)

The authors start with the problem \(\partial _{t}\rho +\operatorname{div}(\rho u)=0\), \(\rho (\partial _{t}u+(u\cdot \nabla )u)+\nabla P(\rho )=\mu \Delta u+(\mu +\lambda )\nabla \operatorname{div}u+\rho \nabla \Phi +\rho f\), \(\Delta \Phi =\rho - \overline{\rho }\), \(\lim_{|x|\rightarrow \infty }\Phi (t,x)=0\), which describes the coupled evolutions of the density \(\rho \), of the momentum \( m=u\rho \) and of the electrostatic potential \(\Phi \) of the electrons. They add the initial conditions \((\rho ,u)(0,x)=(\rho _{0},\frac{m_{0}}{\rho _{0}} )(x)\), which satisfy \((\rho _{0},\frac{m_{0}}{\rho _{0}})(x)\rightarrow ( \overline{\rho },0)\), as \(|x|\rightarrow \infty \). The viscosity coefficients \(\lambda \), \(\mu \) satisfy \(\mu >0\), \(\lambda +\frac{2}{3}\mu \geq 0\). The pressure function \(P=P(\rho )\) is supposed to be a smooth function in a neighborhood of the background doping profile \(\overline{\rho } \) and to satisfy \(P^{\prime }(\overline{\rho })>0\). The external force \(f\) is taken as \(f=\operatorname{div}f_{1}+f_{2}\) with \(\sum_{\nu =1}^{3}\left\| (1+|x|)^{\nu +1}\nabla ^{\nu }f\right\| _{L^{2}}+\left\| (1+|x|)^{3}f\right\| _{L^{\infty }}+\left\| (1+|x|)^{2}f_{1}\right\| _{L^{\infty }}+\left\| f_{2}\right\| _{L^{1}}<c_{0}\varepsilon \), for \(c_{0}>0\) and \(\varepsilon \) small enough. To this evolution problem, the authors associate the steady one, written as: \(\operatorname{div}(\rho _{st}u_{st})=0\), \(\rho _{st}(u_{st}\cdot \nabla )u_{st}-\mu \Delta u_{st}-(\mu +\lambda )\nabla \operatorname{div}u_{st}+\nabla P(\rho _{st})=\rho _{st}\nabla \Phi _{st}+\rho _{st}f\), \(\Delta \Phi _{st}=\rho _{st}-\overline{\rho }\). The first main result of the paper proves the existence of a solution \((\rho _{st},u_{st},\Phi _{st})\) to the steady problem in an appropriate combination of weighted Sobolev spaces in \( \mathbb{R}^{3}\), with a small norm in these spaces. Moreover this solution is unique in the set of small solutions in these spaces. The proof is based on the Banach closed range theorem, the authors introducing an approximate problem, uniform weighted \(L^{2}\) estimates and the contraction mapping principle. The second main result of the paper proves that if \((\rho _{st},u_{st},\Phi _{st})\) is a solution to the steady problem, there exist positive constants \(\varepsilon _{1}\) and \(C\) such that if \(\left\| (\rho _{0}-\rho _{st},u_{0}-u_{st})\right\| _{H^{3}}+\left\| \nabla \Phi _{0}-\nabla \Phi _{st}\right\| _{L^{2}}\leq \varepsilon _{1}\), the non-steady problem admits a unique global solution \((\rho ,u,\nabla \Phi )\) which satisfies global stability estimates near the stationary state. The authors here prove a priori energy estimates for an equivalent system. Finally, the authors prove time decay rates when the initial perturbation belongs to \(\overset{.}{H}^{-s}\) with \(0\leq s<\frac{3}{2}\), analyzing the evolution of negative norms of the solution.