The invariants of the second symmetric power representation of \(\mathrm{SL}_2 (\mathbb F_q)\) (Q635462)

Calculating the invariants of the action of \(\mathrm{SL}_2(\mathbb C)\) on a binary form was a central problem in classical invariant theory. In modern terminology, this amounts to computing a generating set for the ring of invariants for a symmetric power of the defining module for \(\mathrm{SL}_2(\mathbb C)\). It is natural to consider the analogous problem for \(\mathrm{SL}_2(\mathbb F)\) with \(\mathbb F\neq \mathbb C\). Consider the generic binary quadratic form over a field \(\mathbb F\) of characteristic \(p>2: a_0X^2+2a_1XY+a_2Y^2\). Then we have a left action of the general linear group \(GL_2(\mathbb F)\) on the second symmetric power \(V=\mathrm{Span}_{\mathbb F}[Y^2,2XY,X^2]\). Next, for \(q=p^n\) we have \(\mathbb F_q\subseteq\mathbb F\) and \(\mathrm{SL}_2(\mathbb F_q)\leq GL_2(\mathbb F)\). The main goal of the present paper is to describe \(\mathbb F[V]^{\mathrm{SL}_2(\mathbb F_q)}\). As the authors admit, this work generalizes and was inspired by Dickson's solution to the \(q=p\) case [\textit{L. E. Dickson}, On invariants and the theory of numbers. The Madison Colloquium (1913, Part 1, Lecture II, Sections 8-9) AMS, reprinted by New York: Dover Publications, Inc. (1966; Zbl 0139.26603)]. More specifically, the main results are described in the following lines. For any \(c\in\mathbb F_q\) denote by \(\sigma_c=\{\alpha_{i,j}\}\) the matrix with entries \(\alpha_{11}=\alpha_{22}=1,\alpha_{12}=c\) and \(\alpha_{21}=0\). Let \(P\) denote the subgroup \(\{\sigma_c| c\in\mathbb F_q\}\). \(P\) is a Sylow \(p\)-subgroup of \(\mathrm{SL}_2(\mathbb F_q)\). The orbit products \(\beta=\prod a_1P=\prod_{c\in\mathbb F_q}(a_1+ca_0)=a_1^q-a_0^{q-1}a_1\) and, for \(k\in\mathbb F_q\), \(\gamma_k=\prod(a_2-ka_0)P=\prod_{c\in\mathbb F_q}(a_2+2ca_1+(c^2-k)a_0)\) are \(P\)-invariant. The discriminant \(\Delta=a_1^2-a_0a_2\) is a well-known \(\mathrm{SL}_2(\mathbb F_q)\)-invariant. In Section 2, the authors show that \(\mathbb F[V]^P\) is the hypersurface generated by \(a_0,\Delta,\beta,\gamma_0\) subject to the relation \(\beta^2=a_0^q\gamma_0+\Delta(\Delta^{\frac{q-1}{2}}-a_0^{q-1})^2\). Let \(\mathcal Q\) the set of quadratic residues in \(\mathbb F_q\) and let \(\overline{\mathcal Q}\) denote the set of quadratic non-residues. Define \(\Gamma=\prod_{k\in\overline{\mathcal Q}}\gamma_k, B=\beta\prod_{k\in\mathcal Q}\gamma_k\) and \(J=a_0\gamma_0\). In Section 3 the authors prove their main result that \(\mathbb F[V]^{\mathrm{SL}_2(\mathbb F_q)}\) is the hypersurface generated by \(\Delta,J,\Gamma,B\) subject to a relation of the form \(B^2=\Delta^q\Gamma^2+J\Phi(\Delta,J,\Gamma)\) for some polynomial \(\Phi\).