Free algebras and free groups in Ore extensions and free group algebras in division rings. (Q267405)

The \textit{free subalgebra conjecture} is a major open problem in ring theory, asserting that the only obstruction for a division algebra \(D\) to not contain a free subalgebra over its center is being locally PI (namely, that every finitely generated subalgebra of \(D\) is PI). This problem is considered very deep and difficult, and related to many important problems such as the Kurosh problem. It therefore makes sense to investigate it for certain classes of division rings. This paper constitutes an important step towards a better understanding of this conjecture, clarifying the situation for an important family of division algebras, namely \textit{Ore extensions} over fields (of characteristic zero). These are algebras of the form \(D=K(x;\sigma,\delta)\), where \(K\) is a field of characteristic zero, \(\sigma\) an automorphism of \(K\) and \(\delta\) a \(\sigma\)-derivation (i.e. \(\delta(xy)=\sigma(x)\delta(y)+\delta(x)y\)). The main results are: {\parindent=0.4cm \begin{itemize}\item[{\(\bullet\)}] \(D\) always satisfies the free subalgebra conjecture; \item[{\(\bullet\)}] If \(K\) is finitely generated over some base field \(k\) fixed by \(\sigma\) then \(K(x;\sigma)\) contains a free subalgebra if and only if \(\sigma\) has infinite order, and \(K(x;\delta)\) contains a free subalgebra if and only if \(\delta\neq 0\); \item[{\(\bullet\)}] For \(D\) an arbitrary division algebra, if \(D^\times\) contains a solvable, non locally abelian-by-finite subgroup then \(D\) contains a free subalgebra if the characteristic of the center is \(0\), and contains a free group algebra if the center is uncountable (without restrictions on the characteristic); \item[{\(\bullet\)}] If the fixed subfield of \(K\) by \(\sigma\) has infinite transcendence degree (over its prime subfield) then \(K(x;\sigma)\) contains a free subgroup. \end{itemize}} The paper therefore provides improvements both for the free subalgebra conjecture and for Lichtman's question (when does \(D^\times\) contain a free subgroup?) for division algebras of the form \(K(x;\sigma,\delta)\), and closes a naturally interesting gap left in previous work by Bell and Rogalski (namely, the case where \(\sigma\) is not `geometric'). The proof use tools from dynamics on algebraic varieties, \(p\)-adic analytic interpretation of orbits of birational maps as well as combinatorial algebraic and group theoretic arguments.