A geometric variational problem with logarithmic-quadratic interaction

From MaRDI portal





The authors consider the functional \(\mathcal{J}(\Omega )=\mathcal{P}(\Omega )+\frac{\gamma }{2}\int_{\Omega }\int_{\Omega }K(\left\vert x-y\right\vert )dxdy\), defined on the admissible class \(\mathcal{A}=\{\Omega \subset \mathbb{R}^{2}:\Omega \) is Lebesgue measurable and \(\left\vert \Omega \right\vert =m\}\), \(m>0\). \(\mathcal{P}(\Omega )\) denotes for the perimeter of \(\Omega \), defined, if \(\Omega \) is merely measurable, through \(\mathcal{P }(\Omega )=\sup\{\int_{\Omega }\operatorname{div}g(x)dx:g\in C_{0}^{1}(\mathbb{R}^{2}, \mathbb{R}^{2})\), \(\left\vert g(x)\right\vert \leq 1\) \(\forall x\in \mathbb{R }^{2}\}\), \(\gamma >0\), and \(K\) is the function given by \(K(t)=\frac{1}{2\pi } \log\frac{1}{t}+t^{2}\), \(t>0\). They explain how this problem is associated with the analysis of nanostructures of diblock copolymers, especially in the case of a single disk \(B_{\rho }\) in \(\mathbb{R}^{2}\) satisfying the above area constraint \(\pi \rho ^{2}=m\). The authors prove that a disk is a stationary point of \(\mathcal{J}\).\ Introducing the spaces \(\mathcal{X} =\{\phi \in W^{2,2}(S^{1}):\int_{0}^{2\pi }\phi (\theta )d\theta =0\}\), \( \mathcal{Y}=\{\phi \in W^{1,2}(S^{1}):\int_{0}^{2\pi }\phi (\theta )d\theta =0\}\), and \(\mathcal{Z}=\{\phi \in L^{2}(S^{1}):\int_{0}^{2\pi }\phi (\theta )d\theta =0\}\), where \(S^{1}\) is the unit circle in \(\mathbb{R}^{2}\) centered at the origin, they define \(\mathcal{J}(\phi )=\mathcal{J}(\Omega _{\phi })\) where \(\Omega _{\phi }=\cup _{\theta \in S^{1}}\{te^{i\theta }:t\in \lbrack 0,(\rho ^{2}+2\phi (\theta ))^{1/2}]\}\), and they compute the second variation of \(\mathcal{J}\) and its eigenvalues \(\lambda (n)\). In the case of a single disk, they analyze the stability of \(B_{\rho }\) in terms of \(\rho \) and \(\gamma \). In the case of multiple disks assemblies, the area constraint is \(\left\vert \Omega \right\vert =N\pi \rho ^{2}\), \(N\geq 2\), \( \rho >0\), and the main result proves that for each \(\eta >0\), there exists \( \delta >0\), depending on \(N\) and \(\eta \) only, such that if \(\rho <\delta \), \(\gamma \rho ^{3}<12-\eta \), \(\gamma \rho ^{3}log\frac{1}{\rho }>1+\eta \), there exists a stationary point of \(\mathcal{J}\) satisfying the area constraint. Moreover, this stationary point is the union of \(N\) disconnected components, and each component is close to a disc of radius \(\rho \) centered at \(\xi _{\rho ,j}\), \(j=1,2,\ldots ,N\). As \(\rho \rightarrow 0\), any accumulation point of \((\xi _{\rho ,1},\xi _{\rho ,2},\ldots ,\xi _{\rho ,N}) \) is a global minimum of the function \(F_{N}(\xi _{1},\xi _{2},\ldots ,\xi _{N})=\sum_{j=1}^{N}\sum_{k=1,\ldots ,N,k\neq j}K(\left\vert \xi _{j}-\xi _{k}\right\vert )\). This stationary point is stable in some sense. For the proof, the authors introduce appropriate spaces \(\mathcal{X}\), \(\mathcal{Y}\) , and \(\mathcal{Z}\). They compute the first and second derivatives of \( \mathcal{J}\) and they build pseudo-solutions to the equations \(\mathcal{J} ^{\prime }=0\) and \(\mathcal{J}^{\prime \prime }=0\). The paper ends with some numerical simulations in the cases \(N=13\)\ or \(N=100\).



Cites work









This page was built for publication: A geometric variational problem with logarithmic-quadratic interaction

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6980675)