Prescribing traces of primitive elements in finite fields (Q6996924)

This is a small paper on the existence of primitive elements of a finite field \(E\), whose traces over a family of subfields \(F_i\) are prescribed. The paper is an extended work of previous studies of the author in this matter. In this paper, the author shows that, up to a natural exception, these elements exist. The paper is organized as follows: Section 1 is a introductory section on the existence of primitive elements in finite field extensions whose trace over the base field is prescribed. The authors, in previous works, using character sums and Gauss sums, showed that these primitives can be found except in very few cases. The authors also explore a generalization of this problem, considering the existence of primitive elements with prescribed traces over several intermediate extensions. In contrast to the single trace case, natural restrictions appear for this case. In fact, if one prescribes a trace over a specific extension, the traces of all its subextensions are already determined. So, it is necessary to assume that no intermediate extension is contained into another. In this section, Definition 1 of \(d\)-admissible condition is a trace condition on the subextensions fields that meets this property. \N\NTheorem 1 guarantees the existence of primitive elements with prescribed trace values provided that these values meet the \(d\)-admissible condition. This theorem provides a nice range where we can prescribe traces of primitive elements but it is ineffective in some cases. Section 1 ends with a discussion of theorem 2, the main result of this paper. Theorem 2 amends the weakness of theorem 1, providing an asymptotic general existence theorem on primitive elements with prescribed traces, with the unique exception of cases with quadratic extensions. \N\NSection 2 provides the preliminaries about \(\mathbb F_q\)-affine spaces, its degrees, primitive elements on affine spaces, module structure of a finite field, generators polynomials through this module structure, and interpretation of the trace functions through the module structure of finite fields. \N\NSection 3, the last section, is dedicated to the proof of theorem 2. The authors show that, if \(n\) is fixed and \(q\) is large enough, the \(\mathbb F_q\)-affine subspaces of \(\mathbb F_{q^n}\) satisfying a generic algebraic condition (given in a previous paper of the author) contain a large number of primitive elements. For this proof, most of the work concentrates on proving that the \(\mathbb F_q\)-affine space comprising the elements with prescribed traces satisfies the algebraic conditions mentioned. This is obtained essentially by using the module structure of finite fields through normal bases.\N\NFor the entire collection see [Zbl 1559.11005].