A remark on estimates of bilinear forms of gradients in Hardy space (Q676767): Difference between revisions

Let \(H^1({\mathbb{R}}^2)\) denote the Sobolev space consisting of functions \(f\in L^2({\mathbb{R}}^2)\) with derivatives \(f_{x_1},f_{x_2}\in L^2({\mathbb{R}}^2)\). Let \(\eta\) be in \(C_0^\infty({\mathbb{R}}^2)\) with \(\text{ supp} \eta\in B(0,1)\) and \(\int_{{\mathbb{R}}^2}\eta(x)dx=1\), where \(B(x,r)\) denotes the closed ball of radius \(r\) centered at \(x\in{\mathbb{R}}^2\). The normed local Hardy space \(h^1({\mathbb{R}}^2)\) is defined by \[ h^1({\mathbb{R}}^2)=\{f\in L^1({\mathbb{R}}^2):f^{**}\in L^1({\mathbb{R}}^2)\} \] with the norm \[ |f|_{h^1({\mathbb{R}}^2)}=|f^{**}|_{L^1({\mathbb{R}}^2)}, \] where \[ f^{**}(x)=\sup_{0<r<1} \left|r^{-2}\int_{{\mathbb{R}}^2} \eta\left({x-y}\over{r}\right)f(y)dy\right|. \] Let \(\psi^\varepsilon(x)=\psi(x/\varepsilon)\), \[ T=\{\omega\in C_0^\infty({\mathbb{R}}^2):\omega(x_1,x_2)\geq 0\;\text{ for } x_2\geq 0, \;\omega(x_1,x_2)=-\omega(x_1,-x_2)\}, \] \[ S=\left\{{1\over 2\pi}\int_{{\mathbb{R}}^2}\omega(y)\log|x-y|dy:\omega\in T\right\}. \] It is proved that there exists a sequence \(\{\psi^\varepsilon\}_{0<\varepsilon<1}\) in \(S\) such that \[ \sup_{0<\varepsilon<1} |\psi^\varepsilon|_{H^1({\mathbb{R}}^2)}<\infty \] and \[ \lim_{\varepsilon\to 0} |\varphi\{(\psi_{x_1}^\varepsilon)^2- (\psi_{x_2}^\varepsilon)^2\}|_{h^1({\mathbb{R}}^2)} =\infty, \] where \(\varphi\in C_0^\infty({\mathbb{R}}^2), 0\leq\varphi\leq 1, \varphi|B(0,1/8)\equiv 1\) and \(\text{ supp} \varphi\in B(0,1/4)\). This gives a counterexample for an estimate by \textit{L. C. Evans} and \textit{S. Müller} [J. Am. Math. Soc. 7, 199-219 (1994; Zbl 0802.35120)] of local Hardy space norms for certain quadratic functions of \(\psi_{x_1},\psi_{x_2}\).