On the sharpness of Mockenhaupt's restriction theorem (Q359543)

Let \(\alpha\in (0,1)\) and \(1<p<\frac{4}{\alpha}-2\). The authors construct a probabilistic measure \(\mu\) on \([0,1]\) supported on a set of dimension \(\alpha\), and a sequence of functions \(\{f_j\}\) on \([0,1]\) such that \(\mu(B(x,r))\leq C_{1}r^{\alpha}\) for all \(x\in{\mathbb R}\) and \(r>0\), \(|\widehat{\mu}(\xi)|\leq C_{2}(1+|\xi|)^{-\beta/2}\) for all \(\xi\in{\mathbb R}\) and every \(\beta<\alpha\) (with \(C_{2}\) depending on \(\beta\)), and \[ \frac{\|\widehat{f_{j}d\mu}\|_{L^{p}({\mathbb R})}}{\|f_{j}\|_{L^{2}(d\mu)}} \to \infty. \] The existence of this measure and sequence of functions means that the range of \textit{G. Mockenhaupt}'s restriction theorem for Salem sets [Geom. Funct. Anal. 10, No. 6, 1579--1587 (2000; Zbl 0974.42013)], with the endpoint estimate due to \textit{J.-G. Bak} and \textit{A. Seeger} [Math. Res. Lett. 18, No. 4, 767--781 (2011; Zbl 1271.42014)], is optimal.