Unipotent elements of finite groups of Lie type and realization fields of their complex representations. (Q1421806): Difference between revisions

Let \(G\) be a finite group of Lie type in prime characteristic \(p\). In the study of the representations of \(G\), the unipotent elements of \(G\) have special importance. This paper is mainly concerned with the rationality properties of the unipotent elements. By this, we mean the following. Let \(u\) be a unipotent element of \(G\), in other words, an element of \(p\)-power order. For what integers \(r\) coprime to \(p\) is \(u\) conjugate to \(u^r\)? If \(u\) is conjugate to all such powers \(u^r\), we say that \(u\) is rational. A stricter condition involves the notion of strong rationality. We say that \(u\) is strongly rational if whenever \(u\) centralizes a semisimple element \(s\), \(u\) and all its powers \(u^r\), as above, are conjugate in \(C_G(s)\). Rationality properties of unipotent elements in groups of Lie type have received little attention, partly because the main theorems about conjugacy in algebraic groups refer to semisimple elements and also because the properties of unipotent classes are very sensitive to the goodness or badness of the prime \(p\) (the bad primes are found among 2, 3, 5 and 7). This study of the rationality of unipotent classes is motivated by a desire to investigate certain ramification properties of the irreducible characters of \(G\). Thus, let \(\chi\) be an irreducible character of \(G\) and let \(\mathbb{Q}(\chi)\) denote the field generated over the rational numbers by the values of \(\chi\). Define the field \(\mathbb{Q}^p(\chi)\) by \(\mathbb{Q}^p(\chi)=\mathbb{Q}(\chi)\cap\mathbb{Q}(\eta)\), where \(\eta\) is a primitive root of unity whose order is a sufficiently large power of \(p\). In this paper, the authors prove the remarkable result that \(|\mathbb{Q}^p(\chi):\mathbb{Q}|\leq 2\), except for certain \(\chi\) which arise when \(G\) is of type \(E_7\) or \(E_8\) and \(p\) is either 2 or 5. They also investigate which groups \(G\) have the property that \(\mathbb{Q}^p(\chi)=\mathbb{Q}\) for all \(\chi\). By restricting to the case that \(G\) is a finite connected reductive group, the authors obtain an exhaustive list of groups for which each unipotent element is strongly rational. Such groups have the property that \(\mathbb{Q}^p(\chi)=\mathbb{Q}\) for all \(\chi\). The list is given in Theorem 1.9 of the paper but is too long to include here. (We remark that \(G\) is said to be a finite connected reductive group if \(G={\mathcal G}^F\) is the subgroup of fixed points of a Frobenius map \(F\) acting on a connected reductive algebraic group \(\mathcal G\) in characteristic \(p\).) Another question that the authors wish to investigate is the following. Let \(\mathbb{K}=\mathbb{Q}(\varepsilon)\), where \(\varepsilon\) is a primitive root of unity of order equal to the \(p'\)-part of \(|G|\). When can each complex irreducible representation of \(\mathcal G\) be realized in the field \(\mathbb{K}\) or in some small extension of it? They obtain the following general result. Suppose that \(G\) is a finite connected reductive group over a field of characteristic \(p\), where \(p\) is almost good for \(G\). Then every complex irreducible representation of \(G\) can be realized over \(\mathbb{K}\) if \(p=2\) and over \(\mathbb{K}(\sqrt p)\) if \(p\) is odd. The condition that \(p\) is almost good for \(G\) is designed to rule out certain configurations related to the occurrence of groups of type \(E_8\), \(E_7\), \(F_4\) and Suzuki groups as composition factors of \(G\). There are for example exceptions to the realizability theorem just mentioned when \(p=5\) and \(G\) is of type \(E_8\). In conclusion, we have only attempted to give some flavour of the large body of detailed information obtained by the authors in a 64 page paper. For anyone who wishes to study rationality problems relating to finite groups of Lie type, especially where the natural prime \(p\) is concerned, this paper is essential reading.