A free boundary problem modeling electrostatic MEMS: I. Linear bending effects (Q464144)

The authors consider the problem of an unknown function \(\psi(t,x,z)\) and a free boundary \(z=u(t,x)\) under the conditions \[ \begin{aligned} \gamma^2 \partial_t^2u + &\partial_t u +\beta\partial_x^4 u - \tau \partial^2_x u\\ = -\lambda\big(\varepsilon^2|\partial_x \psi(t,x,u(t,x)) & |^2+|\partial_z \psi(t,x,u(t,x))|^2\big),\;t>0, \;x\in (-1,1);\\ u(t,\pm 1)=\beta &\partial_x u(t,\pm 1)=0,\;t>0;\\ u(0,x)=u^0(x),\;&\gamma^2\partial_tu(0,t)=\gamma^2u^1(x), \;x\in (-1,1);\\ \varepsilon^2\partial_x^2 \psi+\partial_z^2 &\psi=0, \;(x,z) \in \Omega(u(t)), \, t>0;\\ \psi(t,x,z)=\frac{1+z}{1+u(t,x)},& \;(x,z) \in \partial\Omega(u(t)), \, t>0,\end{aligned} \] where \(\Omega(u(t))=\{(x,z): x\in (-1,1)\), \(z\in (-1,u(t,x))\}\) and \(\gamma\), \(\tau \geq 0\), \(\lambda\), \(\beta\), \(\varepsilon >0\). For the above problem they prove the existence of a unique solution with \(\gamma =0\) and \(\gamma > 0\) and \(\lambda > 0\) for \(t\in [0,T_m)\) such that \(0<T_m<\infty\) and under some condition on \(u(t)\) \ \(T_m=\infty\). Moreover, they prove the existence of a steady state for sufficiently small \(\lambda\) and the non-existence of a steady state for sufficiently small \(\varepsilon \) and large \(\lambda=\lambda(\varepsilon)\).