Fourier multipliers on the real Hardy spaces (Q284110): Difference between revisions

The author provides a variant of Hytönen's embedding theorem to get some sufficient conditions for a function to become a Fourier multiplier on the real Hardy spaces \(H^p(\mathbb R^n)\) for \(p\in (0,2]\). The main result establishes that if \(m\in L^\infty(\mathbb R^n)\) and there exist \(\rho,C>0\), \(t\in (1,2]\), \(d\in \mathbb N\) and \(l\in \{[n/t]+1,\cdots, dn\}\) such that \(D^\alpha m\in L^1_{\mathrm{loc}}(\mathbb R^n)\) for all multi-indices \(\alpha\) belonging to \(I_{l,d}=\{ \alpha\in\mathbb N_0^n: \|\alpha\|_\infty=\max_{1\leq i\leq n} |\alpha_i|\leq d, \|\alpha\|_1=\sum_{i=1}^n |\alpha_i|\leq l\}\) and satisfies \[ \|\chi_{2^{\mu-1}<|x|\leq 2^{\mu+1}} m\|^\rho_\infty \| 2^{\mu(\|\alpha\|_1- n/t)} \chi_{2^{\mu-1}<|x|\leq 2^{\mu+1}}D^\alpha m\|_t\leq C \] for all \(\mu\in \mathbb Z\) and all multi-indices \(\alpha\in I_{l,d}\), then \(m\) is a Fourier multiplier on \(H^p(\mathbb R^n)\) for all \(p\in (0,2]\) with \(\frac{1}{p}-\frac{1}{2}<\frac{1}{1+p}(\frac{1}{n^2}+\frac{1}{2}-\frac{1}{tn})\). The reader can compare this result with the ones proved by \textit{T. Hytönen} [Math. Nachr. 274--275, 74--103 (2004; Zbl 1112.42007)] and \textit{Yu. S. Kolomoitsev} [Ukr. Math. J. 64, No. 10, 1562--1571 (2013); translation from Ukr. Mat. Zh. 64, No. 10, 1373--1380 (2012; Zbl 1270.42012)]. The result shows that in some cases one can actually weaken the conditions obtained in the previously mentioned papers. Some other possible extensions and several corollaries and results on weighted Hardy spaces are also presented.