Some light on Littlewood-Paley theory (Q5957326)

If \({\mathcal D}\) is any collection of disjoint intervals in \(\mathbb{R}\) and \(2\leq p<\infty\), \textit{J. L. Rubio de Francia} [Rev. Mat. Iberoam. 1, No. 2, 1-14 (1985; Zbl 0611.42005)] proved that \[ \Biggl\|\biggl(\sum_{I\in{\mathcal D}}|S_I f|^2\biggr)^{1/2}\Biggr\|_{L^p(\mathbb{R})} \leq C_p\|f\|_{L^p(\mathbb{R})} \] for all \(f\in L^p(\mathbb{R})\), where \((S_I f)^\land=\chi_I\widehat f\). He also observed that this is not true if \(1<p<2\). However, he conjectured that one might always have \[ \Biggl\|\biggl(\sum_{I\in{\mathcal D}}|S_I f|^{p'}\biggr)^{1/{p'}}\Biggr\|_{L^p(\mathbb{R})} \leq D_p\|f\|_{L^p(\mathbb{R})} \] for all \(f\in L^p(\mathbb{R})\), where \(p'=p/(p-1)\). In this paper, the authors give a counterexample to indicate Rubio's conjecture cannot hold if \(1<p<2\). Let \({\mathcal D}(\mathbb{R})\) be all integrable functions \(f\) on \(\mathbb{R}\) such that \(\int^{n+1}_n f(x) dx=0\) for all integers \(n\). \textit{M. Cowling, G. Fendler} and \textit{J. J. F. Fournier} in [Math. Ann. 285, No. 2, 333-342 (1989; Zbl 0662.43003)] claimed that the real interpolation space \([{\mathcal D}(\mathbb{R}), L^2(\mathbb{R})]_{\theta, p}\) is the Lebesgue space \(L^p(\mathbb{R})\), where \(1/p=1-\theta/2\). In this paper, the authors also show that this is not true.