Computing \(n\)-th roots in \(\text{SL}_{2}\) and Fibonacci polynomials (Q2183150)

The general theme of this paper is the study of images of word maps on groups. The conjecture due to Ore made in 1951 asserting that every element of a nonabelian finite simple group is a commutator was proved just a decade back. We mention in passing that the analogous question conjectured by Thompson asserting that in any nonabelian, finite simple group \(G\) there exists a conjugacy class \(C\) so that \(C.C =G\) is still open. J-P. Serre showed in a finitely generated pro-\(p\) group \(G\), how to express the elements of \(G'\) as bounded products of commutators, thereby proving that subgroups of finite index are automatically open. This was generalized to general finitely generated profinite groups spectacularly by Nikolov and Segal. They invented a method which is analogous to the Hensel lemma to solve `Waring type' problems on groups. Since then several important works have appeared due to the efforts of O'Brien, Guralnick, Larsen, Liebeck, Lubotzky, Shalev and Tiep among others. Techniques ranging from algebraic geometry to analytic number theory and character theory have led to deep results. For instance,Larsen and Shalev proved that for any positive integer \(d\) and two nontrivial words \(w_1, w_2\), there exists a positive integer \(N\) depending on them so that for any simply connected almost simple algebraic group \(G\) of dimension \(d\) over a finite field \(F\) with the finite, simple group \(S= G(F)/\mathrm{center}\) having order \(\geq N\), we have \(w_1(S) w_2(S) = S\). This enables them to prove for any nonabelian finite simple group \(G\) that for three nontrivial words \(w_1,w_2,w_3\), one has \(G = w_1(G) w_2(G) w_3(G)\). Apart from character theory and analytic number theoretic estimates, they use Deligne's theorem on Riemann Hypothesis for varieties over finite fields. Later, a number of authors have derived probabilistic results again using technically powerful methods. To point out a very recent result [Ann. Math. (2) 190, No. 2, 561--608 (2019; Zbl 1448.20063)], in a tour de force, \textit{M. Larsen} et al. have shown that every word of the form \(w_1 w_2\) where \(w_1, w_2\) are non-trivial words in disjoint sets of variables, induces almost uniform distributions on finite simple groups with respect to the \(L^1\) norm. They prove and use geometric results such as: For every \(l\), there exists an explicit \(N = O(l^4)\), such that if \(w_1, \dots, w_N\) are non-trivial words of lengths at most \(l\) in disjoint sets of variables, then the word map induced on a semisimple algebraic group over an arbitrary field by the product \(w_1 \cdots w_N\) is a flat morphism. In the present paper, the authors use elementary methods to explicitly obtain results on \(\mathrm{SL}_2(k)\) and \(\mathrm{SL}_1(D)\) for an arbitrary field \(k\) of characteristic different from \(2\) and an arbitrary quaternion division algebra \(D\) over such a field \(k\). More precisely, they rephrase the condition that an element of \(\mathrm{SL}_2(k)\) is an \(n\)-th power as the solvability of certain equations over \(k\) which are in terms of generalized Fibonacci polynomials. The general criteria are exploited in the case of certain \(n\) to obtain explicitly the number \(c(n,q)\) of conjugacy classes (as well as the number \(s(n,q)\) of elements) in \(\mathrm{SL}_2(\mathbb{F}_q)\) which are \(n\)-th powers. From these results, they are able to obtain different, elementary proofs of Waring-type problems proved earlier in much more generality albeit by highly technical methods. For instance, they show: Let \(l\) be an odd prime and let \(q\) be odd. Then, the word map \(X^lY^l\) is surjective on \(\mathrm{SL}_2(\mathbb{F}_q)\) excepting the case \(l=q=3\). This case is indeed an exception. The above result is shown to hold for general \(k\) of characteristic different from \(2\); viz., the word \(X^2Y^2\) is surjective on \(\mathrm{SL}_2(k)\) and the word \(X^3Y^3\) is surjective on \(\mathrm{SL}_2(k)\) if \(\operatorname{char}k \neq 2,3\). The more interesting case they tackle is the anisotropic case. They prove: Let \(G = \mathrm{SL}_1(D)\) for a quaternion division algebra over an arbitrary field \(k\) of characteristic different from \(2\). If \(2\) is a square in \(k\), then the word map \(X^2Y^2\) is surjective on \(G\) if and only if \(-1\) is a square in \(k\). If \(-1\) is a square in \(k\), then the word \(X^nY^n\) for \(n\) odd is surjective.