A weak \(L^2\) estimate for a maximal dyadic sum operator on \(\mathbb R^n\) (Q1409653)

In [Acta Math. 116, 135--137 (1966; Zbl 0144.06402)], \textit{L. Carleson} proved his celebrated theorem on almost everywhere convergence of Fourier series of square functions on \(\mathbb{R}\). This was followed by a new proof given by \textit{C. Fefferman} in [Ann. Math. (2) 98, 551--571 (1973; Zbl 0268.42009)]. The techniques used by Fefferman have become known as time-frequency analysis and have found wide application in harmonic analysis. In particular, Lacey and Thiele have refined and extended these ideas in their pioneering work on the bilinear Hilbert transform on \(\mathbb{R}\). In [Math. Res. Lett. 7, 361--370 (2000; Zbl 0966.42009)], using their new techniques, \textit{M. Lacey} and \textit{Ch. Thiele} obtained a new proof of Carleson's theorem by proving a weak-type \((2,2)\) inequality of the maximal operator \[ Cf(x)= \sup_N \Biggl| \int^N_{-\infty}\widehat f(\xi)\,e^{2\pi ix\xi}\,dx\Biggr|. \] This paper is a generalization of these techniques (time-frequency analysis) to higher dimensions. In particular, the authors obtain a weak-type \((2,2)\) estimate for a maximal dyadic sum operator on \(\mathbb{R}^n\), that is a higher dimension analogue of \(Cf(x)\). As an application, the authors obtain a new proof of Sjölin's theorem on weak \(L^2\) estimate for the maximal conjugate Calderón-Zygmund operator on \(\mathbb{R}^n\).