The \(q\)-unit circle: the unit circle in prime characteristics and its properties (Q2422199)

The analogies between number fields and function fields is remarkably deeper between number fields and global function fields and much more deep between cyclotomic number fields and cyclotomic function fields (Carlitz module). The paper under review gives one step further into this analogy. Let \(\mu_n=\{\xi\in {\mathbb C}\mid \xi^n=1\}\) and \({\mathbb M}:=\bigcup_{n\in{\mathbb N}}\mu_n\). Then \({\mathbb M}\) gives the maximal abelian extension of \({\mathbb Q}\), that is, \({\mathbb Q}(\zeta_{\infty}):={\mathbb Q}({\mathbb M})\) (Kronecker-Weber theorem). L. Carlitz found that there exists an object in congruence function fields, the one that now is called the \textit{Carlitz module}, that is totally analogous to the number field case. Let \(K={\mathbb F}_q(T)\) be a rational congruence function field where \({\mathbb F}_q\) is the finite field of \(q\) elements and let \(R_T={\mathbb F}_q[T]\). The Carlitz action of \(R_T\) on a fixed algebraic closure \(\bar{K}\) of \(K\), denoted by D. Hayes as \(u^M\), is an additive action analogue to the action of \({\mathbb Z}\) on \(\bar{\mathbb Q}^*\): \(n\circ x:=x^n\) in the number field case. Hayes showed that there exists the corresponding Kronecker-Weber theorem in the function field case. For \(M\in R_T\setminus\{0\}\), let \(\Lambda_M:=\{u\in\bar{K}\mid u^M=0\}\) and \({\mathbb T}_q:=\bigcup_{M\in R_T\setminus\{0\}}\Lambda_M\). Let \(K_M:=K(\Lambda_M)\). Then, if \(A_T:=K({\mathbb T}_q)\), the maximal abelian extension of \(K\) is \(A_T A_{1/T}\). Since \({\mathbb R}\) is the completion of \({\mathbb Q}\) at the infinite prime, \(K_{\infty}\) the completion of \(K\) at the infinite prime, is called the \textit{real line} in the function field case. Furthermore, the behavior of the infinite prime of \(R_T\) is the same as that of the infinite prime of \({\mathbb Z}\). Namely, the latter decomposes totally in the real subfield \({\mathbb Q}(\zeta_n)^+={\mathbb Q}(\zeta_n)\cap {\mathbb R}\) and ramifies fully in \({\mathbb Q}(\zeta_n)/{\mathbb Q}(\zeta_n)^+\). The former is totally decomposed in the real subfield \(K_M^+=K_M\cap K_{\infty}\) and fully ramified in \(K_M/ K_M^+\) (for \(q\neq 2\)). Thus \({\mathbb T}_q\) is the function field analogue of \({\mathbb M}\). In this very interesting paper, the author defines the analogue of \(S^1=\{\xi\in{\mathbb C}\mid |\xi|=1\}\) in the function field case. In the number field case, we have that the closure of \({\mathbb M}\) in \({\mathbb C}\) is \(S^1\). We have \({\mathbb Q}(\zeta_{\infty})^+:=\bigcup_{n\in{\mathbb N}}{\mathbb Q}(\zeta_n)^+\subseteq{\mathbb R}\) and \(A_T^+:=\bigcup_{M\in R_T\setminus\{0\}}K_M^+\subseteq K_{\infty}\) and also \(A_T\subseteq K_{\infty}(\sqrt[q-1]{-1/T})\) as the analogue of the fact \({\mathbb C}={\mathbb R}(\sqrt{-1})\). The analogue of \(S^1\) in the function field case is \({\mathbb S}_q\), ``\textit{the \(q\)-unit circle}'' defined by the author as the completion of \({\mathbb T}_q\) at a fixed infinite prime of \(A_T\). Equivalently, \({\mathbb S}_q\) is the closure of \({\mathbb T}_q\) in \(K_{\infty} (\sqrt[q-1]{-1/T})\). In the number field case we have \(\mu_n\cap {\mathbb R}\subseteq \{\pm 1\}\) and correspondingly in the function field case we have \(\Lambda_M\cap K_{\infty}=\{0\}\). In this work, the author proves many results which are analogue to the characteristic \(0\) case. Among them, it is shown the following. All roots of \(q\)-unity lie in \({\mathbb S}_q\) and are dense in it; \({\mathbb S}_q\) is compact; \({\mathbb S}_q\cap K_{\infty}=\{0\}\). It is also shown that \({\mathbb S}_q\) is an additive group, an \(R_T\)-module and an \({\mathbb F}_q\)-vector space; \(K_{\infty}({\mathbb S}_q)=K_{\infty}(\sqrt[q-1]{-1/T})\) is a \(K_{\infty}\)-vector space of dimension \(q-1\). In addition, it is defined the group of multiplicative Möbius transformations of \({\mathbb S}_q\); it is defined the extension of the Carlitz action as a natural action of \(K_{\infty}\) on \({\mathbb T}_q\), \(u^N\) for \(u\in{\mathbb T}_q\) and \(N\in K_{\infty}\) (\(u^N\) not necessarily belongs to \({\mathbb T}_q\)) such that this action defines a map from \({\mathbb S}_q\) into itself, that is, this is an action of exponentiation on \({\mathbb S}_q\) by any real number. Furthermore, it is shown that if \(\lambda \in {\mathbb S}_q\) is integral over \(R_T\) and all conjugates of \(\lambda\) over \(K\) are contained in \({\mathbb S}_q\), then \(\lambda \in {\mathbb T}_q\); for any measure \(\mu\) on \(K_{\infty}({\mathbb S}_q)\) such that \(\mu(C)<\infty\) for every compact subset \(C\), we have \(\mu({\mathbb S}_q)=0\). In the topological aspect, it is defined a natural hyperbolic plane of finite dimension associated with \({\mathbb S}_q\). It is also shown that there exist modular forms and Eisenstein series which are consistent with the lattice structure generated by \({\mathbb S}_q\). All these and much more is the content of the paper. As concluding remarks, the author poses several open questions and points out that there should be a similar theory for Drinfeld modules or rank larger than one and asks if there exist a similar analogy to \(S^n\), \(n>1\) in the function field case.