Endomorphisms of quantized Weyl algebras. (Q657632)

Let \(A_n\) be the \(n\)-th Weyl algebra, i.e., the associative algebra generated over \(\mathbb C\) by the commuting variables \(x_1,\dots,x_n\) and the \(n\) partial derivatives \(\partial_i=\partial/\partial x_i\), with the natural commutator relations between them. The classical Dixmier conjecture \(\text{D}_n\) (open for all \(n>0\)) states that \(\Aut(A_n)=\text{End}(A_n)\), i.e., every algebra endomorphism of \(A_n\) is an automorphism. It is known that \(\text{D}_n\) implies the \(n\)-th Jacobian conjecture \(\text{J}_n\) (for the endomorphisms of \(\mathbb C[x_1,\dots,x_n]\) with invertible Jacobian matrix and open for all \(n>1\)). It is also known that \(\text{J}_{2n}\) implies \(\text{D}_n\). Let \(P_n=\mathbb C[r_1,\dots,r_n,s_1,\dots,s_n]\) be the polynomial algebra in \(2n\) variables with its canonical Poisson bracket \(\{\,,\,\}\), where \(\{r_i,s_j\}=\delta_{ij}\), \(\{r_i,r_j\}=\{s_i,s_j\}=0\). Conjecturally \(\Aut(P_n)=\text{End}(P_n)\) (where the endomorphisms preserve both the multiplication and the Poisson bracket in \(P_n\)) and this is equivalent to \(\text{J}_n\). \textit{A. Belov-Kanel} and \textit{M. Kontsevich} [Lett. Math. Phys. 74, No. 2, 181-199 (2005; Zbl 1081.16031)] stated the \(\text{BK}_n\)-conjecture that the groups \(\Aut(A_n)\) and \(\Aut(P_n)\) are canonically isomorphic. This is known for \(n=1\) only. In the proof of the equivalence between the Dixmier and Jacobian conjectures and the approach of Belov-Kanel and Kontsevich to their own conjecture an essential moment is that the Weyl algebra in positive characteristic is Azumaya and the conditions which guarantee that an endomorphism of an Azumaya algebra is an automorphism. In the paper under review the author proposes that instead of using Weyl algebras in positive characteristic to use quantized Weyl algebras defined over \(\mathbb C\) at roots of unity. The author discusses the disadvantages and the advantages of the root of unity method to the \(\text{BK}_n\)-conjecture versus the reduction modulo prime method. The main advantage is that one never leaves the base field \(\mathbb C\). As an illustration the case \(n=1\) is considered in detail.