Minimal polynomials and distinctness of Kloosterman sums (Q1891283)

Let \(q\) be a power of a prime \(p\), and let \(\text{Kl}_n (q,a)\) be the usual multiple Kloosterman sum generated by the rational function \(x_1+ \cdots+ x_n+ a/(x_1 \dots x_n)\) over \(\mathbb{F}_q\). Let \(\xi_p\) be a primitive \(p\)-th root of unity, and let \(H\) be the group of automorphisms \(\theta\) of \(\mathbb{Q} (\xi_p)\) for which \(\theta^{(n+1, p-1)} =1\). Then \(\text{Kl}_n (q,a)\) lies in the fixed field of \(\mathbb{Q} (\xi_p)\) under \(H\). The first theorem of the paper shows that \(\text{Kl}_n (q,a)\) in fact generates this fixed field, providing that \(\text{trace} (a)\neq 0\). A stronger result is also proved. The argument, which uses Stickelberger's theorem, improves a result of \textit{B. Fisher} [Math. Ann. 301, 485-505 (1995; Zbl 0816.11065)]. The author also shows that if \(p\) is large enough (explicitly described in the paper) then \(\text{Kl}_n (q,a)\neq \text{Kl}_n (q,b)\) unless \(a= b^{p^j}\) for some \(j\).