Curious congruences for cyclotomic polynomials II (Q6183517): Difference between revisions

This paper is a subsequent research of the paper by \textit{S. Akiyama} and \textit{H. Kaneko} [Res. Number Theory 8, No. 4, Paper No. 102, 10 p. (2022; Zbl 1503.11007)]. The authors of the current paper prove the conjecture proposed by Akiyama and Kaneko and also improve their result. Let \(\Phi_n (x)\) be the \(n\)-th cyclotomic polynomial. The degree of \(\Phi_n (x)\) is given by Euler's totient function \(\phi (n)\). Define the polynomial \(F_k (x_1, \ldots , x_k) \in \mathbb{Q} [x_1, \ldots , x_k]\) by the generating function: \[ \sum_{k=0}^{\infty} F_k (x_1, \ldots, x_k) \frac{t^k}{k!} =(1+t)^{x_1} \exp \left( 2 \sum_{\nu=1}^{\infty} \frac{B_{2\nu}}{(2\nu) !} (-\log (1+t))^{2\nu} x_{2\nu }\right) \] as per Lemma 3.1 in this paper, where \(B_{2\nu}\) is the \(2\nu\)-th Bernoulli number. The authors prove that \(F_{2k+1} (x_{1}, \ldots ,x_{2k+1})\) is divisible by \(x_1-k\) in \(\mathbb{Q} [x_1, \ldots, x_{2k+1}]\) for every non-negative integer \(k\) as it was conjectured by Akiyama and Kaneko. There is a close relation between the polynomial \(F_k\) and the derivative of cyclotomic polynomial found by \textit{D. H. Lehmer} [J. Math. Anal. Appl. 15, 105--117 (1966; Zbl 0168.29304)]: \[ \frac{ \Phi_n^{(k)}(1)}{\Phi_n (1)} = F_k \left(\frac{\phi(n)}{2}, \frac{J_2 (n)}{4}, \ldots, \frac{J_k (n)}{2k} \right), \] where \(J_k (n) = n^k \prod_{p \mid n} (1-p^{-k})\) is Jordan's totient function. By proving a certain integral property of \(F_k\) (Theorem 2.5), the authors succeed to show: (i) \(\Phi_n' (1)/\Phi_n (1) = \phi (n)/2\); (ii) \(\Phi_n^{(3)} (1) /\Phi_n (1) \) is divisible by \(\phi (n)/2-1\); (iii) \(\Phi_n^{(2k+1)} (1) /\Phi_n (1) \) is divisible by \(\phi (n)/2-2k\) for \(k \ge 2\). These are improvements of the result by Akiyama and Kaneko.