Irrationality of zeros of polygamma functions

From MaRDI portal





For any positive integer \(k\), let \(\psi_k(z):=(-1)^{k+1}k!\sum_{n=0}^\infty 1/(n+z)^{k+1}\) denote the \(k\)-th polygamma function. The author proves the following results: -- For any positive integer \(k\), \(\psi_{2k+1}\) has no real zeros. -- For any positive integer \(k\), the real zeros of \(\psi_{2k}\) lie in the intervals \((-m+1/2, -m+114/227)\) for all positive integer \(m\). -- \(\psi_1\) has no zeros in \(\mathbb Z[i]\) and no zeros in the half-plane \(\text{Re}(z)\ge 0\). The proofs of these assertions follows from a careful numerical analysis of the series that define the \(\psi_k\). The author also proves that, for any given integer \(q\ge 3\), the dimensions over \(\mathbb Q\) of the \(\mathbb Q\)-vector spaces spanned by the sets \[ \{\psi_{2k+1}(a/q) : 0 \le k \le N, 1\le a \le q, (a,q)=1\} \] and \[ \{1, \psi_{2k}(a/q) : 1 \le k \le N, 1\le a \le q, (a,q)=1\} \] are \(\gg N\) and \(\gg \log(N)\), respectively.











This page was built for publication: Irrationality of zeros of polygamma functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6894854)