Crossover of the coarsening rates in demixing of binary viscous liquids (Q2443500)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Crossover of the coarsening rates in demixing of binary viscous liquids |
scientific article |
Statements
Crossover of the coarsening rates in demixing of binary viscous liquids (English)
0 references
7 April 2014
0 references
A lower bound for coarsening rates is derived for a phase-field model describing the demixing process of binary viscous liquids which shows a crossover between the diffusion-dominated regime and the convection-dominated regime. Let \(m\) be the order parameter which characterizes the local composition of the mixture, the pure phases corresponding to \(m=1\) and \(m=-1\), respectively. Denoting the fluid velocity by \(u\), the evolution of \((m,u)\) is given by \[ \partial_t m - \lambda \Delta\mu + \mathrm{div}(mu) = 0 \text{ in } (0,\infty)\times\Omega \] and \[ -\Delta u + \nabla p = - m \nabla \mu\;, \quad \operatorname{div}(u) = 0\text{ in }(0,\infty)\times\Omega, \] where \(\mu = - \Delta m - 2(1-m^2) m\) is the chemical potential and \(p\) the pressure. Both \(m\) and \(u\) are periodic in space and the cell of periodicity \(\Omega\subset \mathbb R^d\), \(d\geq 2\), is assumed to be large. Introducing the energy \[ E(m) = \frac{1}{|\Omega|} \int_\Omega \left( \frac{|\nabla m|^2}{2} + \frac{(1-m^2)^2}{2} \right)\;dx \] and a physical length \[ L(m) = d_c\left( \max\{m,0\} , \max\{-m,0\} \right) \] defined with the Wasserstein distance \(d_c\) associated to a suitably chosen cost function \(c\), the following result is established: if \(T\) and \(m(0)\) are such that \[ \frac{1}{|\Omega|} \int_\Omega m(0,x)\;dx = 0\text{ and }T \gg \lambda^{1/2} L(m(0))^3\;, \] then \[ \int_0^T \max\left\{ \lambda^{1/2} E(m(t))^2 , E(m(t)) \right\}\;dt \geq C \int_0^T \min\left\{ \frac{\lambda^{1/2}}{(\lambda t)^{2/3}} , \frac{1}{t} \right\}\;dt. \] In particular, in the diffusive regime corresponding to \(E(m)\gg \lambda^{-1/2}\) and \(t\ll \lambda^{1/2}\), the above result is a time-averaged version of the expected coarsening rate \(E(m) \geq C (\lambda t)^{-1/3}\).
0 references
coarsening rates
0 references
Wasserstein distance
0 references
phase-field model
0 references