The Waring-Goldbach problem for cubes

From MaRDI portal





\textit{L.~K.~Hua} [Q. J. Math., Oxf. Ser. 9, 68-80 (1938; Zbl 0018.29404)] showed that large numbers are sums of nine cubes of primes. Let \(E(N)\) denote the number of numbers not exceeding~\(N\), are \(\not\equiv 0,\pm 2\) (mod~9), and are not sums of five cubes of primes. In the same paper Hua also proved that \(E(N)\ll N(\log N)^{-A}\) for any fixed positive~\(A\). Here the author improves on the upper estimate to \(E(N)\ll N^{152/153}\). The proof is based on the circle method, with major arcs so long that the Siegel-Walfisz theorem no longer applies. There are thus difficulties concerned with the distribution of the zeros of Dirichlet \(L\)-functions, and this is overcome by applying a method of \textit{J. Y. Liu, N. C. Liu} and \textit{T.~Zhan} [Monatsh. Math. 128, 283-313 (1999; Zbl 0940.11047)]. It is interesting that the possible existence of a Siegel zero has no special influence in the situation of the present problem, so that the Deuring-Heilbronn phenomenon can be avoided.











This page was built for publication: The Waring-Goldbach problem for cubes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4498545)