Bounded solutions and their asymptotics for a doubly nonlinear Cahn-Hilliard system (Q2180853): Difference between revisions

The authors consider a bounded and smooth domain \(\Omega \subset \mathbb{R}^{3}\) and the modified Cahn-Hilliard problem \(\partial _{t}u-\Delta \mu =0\), \(\mu \in \varepsilon \partial _{t}u+\beta (\partial _{t}u)-\delta \Delta u+\psi ^{\prime }(u)+g\) in \(\Omega \times (0,T)\). The boundary conditions \(\partial _{n}u=0\), \(\mu =0\) are imposed on \(\partial \Omega \times (0,T)\), and \(u\) starts from an initial condition \(u_{0}\in \{v\in H^{2}(\Omega):\partial _{n}v=0\) on \(\partial \Omega \}\) at \(t=0\). Here \(\varepsilon \) and \(\delta \) are positive parameters, \(\beta \) is a maximal monotone graph, \(\psi ^{\prime }\) is the derivative of a possibly non-convex potential \(\psi \in C^{2}(a,b)\), with \(-\infty \leq a < b\leq +\infty \), which satisfies \(\psi(r)\geq 0\) on \((a,b)\), \(\lim_{r\rightarrow a^{+}}\psi ^{\prime }(r)=-\infty \), \(\lim_{r\rightarrow b^{-}}\psi ^{\prime }(r)=+\infty \), \(\psi ^{\prime \prime}(r)>-K\) on \((a,b)\), and \(g\in H^{1}(0,T;L^{2}(\Omega ))\cap L^{\infty}(\Omega \times (0,T))\) is a forcing term. The main purpose of the paper is to describe the asymptotic behavior of the solution when \(\varepsilon \) or \(\delta \) go to 0. The authors first prove the existence of a unique weak solution \((u_{\varepsilon \delta },\mu _{\varepsilon \delta },\xi_{\varepsilon \delta })\) to this problem with \(\xi _{\varepsilon \delta }\in\beta (\partial _{t}u_{\varepsilon \delta })\). They here extend a previous existence result they proved in [Commun. Pure Appl. Anal. 17, No. 3, 1001--1022 (2018; Zbl 1393.35235)]. They prove that this solution continuously depends on the data \(u_{0}\) and \(g\). The first convergence result is obtained when considering \(\delta >0\) fixed and letting \(\varepsilon \rightarrow 0\).\ Under some hypotheses on the data, the authors prove that the solution \((u_{\varepsilon \delta },\mu _{\varepsilon \delta},\xi _{\varepsilon \delta })\) converges, up to some subsequence and in some weak sense, to a weak solution \((u,\mu ,\xi )\) of the problem \(\partial_{t}u(t)-\Delta \mu (t)=0\), \(\mu (t)=\xi (t)-\delta \Delta u(t)+\psi^{\prime }(u(t))+g(t)\), for a.e. \(t\in (0,T)\), \(\xi \in \beta (\partial_{t}u)\) a.e. in \(\Omega \times (0,T)\), with the initial condition \(u(0)=u_{0}\). Taking now \(\varepsilon >0\) fixed and assuming further hypotheses on the data, the authors prove that the solution \((u_{\varepsilon \delta },\mu_{\varepsilon \delta },\xi _{\varepsilon \delta })\) converges, up to some subsequence and in some weak sense, to a weak solution \((u,\mu ,\xi )\) of the problem \(\partial _{t}u(t)-\Delta \mu (t)=0\), \(\mu (t)=\varepsilon\partial _{t}u(t)+\xi (t)+\psi ^{\prime }(u(t))+g(t)\), for a.e. \(t\in (0,T)\), \(\xi \in \beta (\partial _{t}u)\) a.e. in \(\Omega \times (0,T)\), with the initial condition \(u(0)=u_{0}\). For the proof of the existence result, the authors introduce an approximate problem replacing the source term \(g\) by a smooth function \(g_{\lambda }\) and introducing a regularizing term in the second equation of the problem. They then prove uniform estimates on the solution of this approximate solution which allow to pass to the limit when \(\lambda \rightarrow 0\). For the proof of the convergence results, the authors introduce the Yosida approximation of the maximal monotone graph and they derive estimates.