Exponential convergence to equilibrium in a Poisson-Nernst-Planck-type system with nonlinear diffusion (Q256252)

Exponential convergence to the equilibrium is shown for non-negative solutions \((u,v)\) to a Poisson-Nernst-Planck system with confinement in \(\mathbb{R}^3\), provided that the strength of the electric field is small enough. More precisely, given \(\varepsilon>0\) and two confining potentials \(U\) and \(V\), the system under study reads \[ \begin{align*}{ \partial_t u & = \mathrm{div}\left( u \nabla (2u + U + \varepsilon \psi) \right)\ , \quad (t,x)\in (0,\infty)\times\mathbb{R}^3\ , \cr \partial_t v & = \mathrm{div}\left( v \nabla (2v + V - \varepsilon \psi) \right)\ , \quad (t,x)\in (0,\infty)\times\mathbb{R}^3\ , }\end{align*} \] where \(\psi = G * (u-v)\) and \(G\) is the Poisson kernel in \(\mathbb{R}^3\) (\(G(x) = 1/(4\pi|x|)\), \(x\neq 0\)). The external potentials \(U\) and \(V\) are both uniformly convex and grow at most quadratically at infinity. The system is supplemented with non-negative initial conditions \((u_0,v_0)\in L^2(\mathbb{R}^3;\mathbb{R}^2)\) having unit mass \(\|u_0\|_1=\|v_0\|_1=1\). As shown in [\textit{D. Kinderlehrer}, \textit{L. Monsaingeon} and \textit{X. Xu}, ``A Wasserstein gradient flow approach to Poisson-Nernst-Planck equations'', \url{arXiv:1501.04437}], this system is a gradient flow with respect to the \(2\)-Wasserstein distance in \(\mathcal{P}_2\times \mathcal{P}_2\) for the energy \[ \mathcal{E}[u,v] = \int_{\mathbb{R}^3} \left[ u^2 + v^2 + u U + v V + \frac{\varepsilon}{2} |\nabla\psi|^2 \right]\;dx\;, \] where \(\mathcal{P}_2\) is the space of probability measures on \(\mathbb{R}^3\) with finite second moment. In the paper under review, it is proved that \(\mathcal{E}\) has a unique minimizer \((u_\infty,v_\infty)\in W^{1,2}(\mathbb{R}^3;\mathbb{R}^2)\) for all \(\varepsilon>0\) and that there is \(\Lambda_\varepsilon>0\) such that \[ \| u(t) - u_\infty\|_2 + \|v(t) - v_\infty\|_2 \leq C(u_0,v_0) e^{-\Lambda_\varepsilon t}\;, \quad t>0\;, \] provided \(\varepsilon>0\) is sufficiently small. The proof of the convergence result relies on a further use of the gradient flow structure: the functional associated to the uncoupled system corresponding to \(\varepsilon=0\) enjoys suitable convexity properties and, for \(\varepsilon\) small enough, a well-chosen perturbation of it retains these properties.