An algebraic characterization of the affine three space (Q2074393)

Let \(k\) be a field and \(\mathbb{A}_k^n\) denote the affine \(n\)-space over \(k\). Finding useful characterization of \(\mathbb{A}_k^n\) is an important problem in affine algebraic geometry. While characterizing \(\mathbb{A}_k^1\) is not hard to achieve, it is difficult for \(\mathbb{A}_k^n\), \(n \ge 2\). The problem is open when \(n \ge 4\); and for \(n =2 \ \text{ and} \ 3\), only a few results are achieved. A topological characterization of \(\mathbb{A}_{\mathbb{C}}^2\) was obtained by C.P. \textit{C. P. Ramanujam} [Ann. Math. (2) 94, 69--88 (1971; Zbl 0218.14021)]. Later M. Miyanishi gave an algebraic characterization of \(\mathbb{A}_k^2\) when \(k\) is an algebraically closed field of characteristic zero in [\textit{M. Miyanishi}, J. Math. Kyoto Univ. 15, 169--184 (1975; Zbl 0304.14021)]. Miyanishi's characterization of \(\mathbb{A}_k^2\) played an important role in solving the Zariski Cancellation problem for the affine plane [\textit{T. Fujita}, Proc. Japan Acad., Ser. A 55, 106--110 (1979; Zbl 0444.14026); \textit{M. Miyanishi} and \textit{T. Sugie}, J. Math. Kyoto Univ. 20, 11--42 (1980; Zbl 0445.14017)]. \textit{M. Miyanishi} [Am. J. Math. 106, 1469--1485 (1984; Zbl 0595.14025)] and \textit{S. Kaliman} [Pac. J. Math. 203, No. 1, 161--190 (2002; Zbl 1060.14085)] gave a topological characterizations of \(\mathbb{A}_{\mathbb{C}}^3\). In this article under review, Dasgupta-Gupta gave an algebraic characterization of \(\mathbb{A}_k^{2}\) where \(k\) is a field of characteristic zero and of \(\mathbb{A}_k^3\) where \(k\) is an algebraically closed field of characteristic zero. Their method involves a variant of Makar-Limanov invariant (ML-invariant) which they termed as Makar-Limanov-Freudenburg invariant (MLF-invariant). Before we state the results we define ML-invariant and MLF-invariant. Let \(B\) be an affine \(k\)-domain and \(\text{LND}(B)\) denote the set of locally nilpotent \(k\)-derivations of \(B\). Set \(\text{LND}^*(B ) := \{ D \in \text{LND}(B) \ | \ D(s) =1 \ \text{for some} \ s \in B \}\). The ML-invariant of \(B\), denoted by \(\text{ML}(B)\), is defined as \(\displaystyle \text{ML}(B)= \bigcap_{D \in \text{LND}(B)} \text{Ker} \ D\); and the MLF-invariant of \(B\), denoted by \(\text{ML}^*(B)\), is defined by \(\displaystyle \text{ML}^*(B)= \bigcap_{D \in \text{LND}^*(B)} \text{Ker} \ D\). The main results of the paper under review are the following. Theorem 3.8. Let \(k\) be a field of characteristic zero and \(B\) a two-dimensional affine \(k\)-domain. Then the following are equivalent: \begin{itemize} \item [(I)]\(B=k^{[2]}\). \item [(II)]\(ML^{*}(B)=k\). \item [(III)]\(ML(B)=k\) and \(ML^{*}(B) \neq B\). \end{itemize} Theorem 4.6. Let \(k\) be an algebraically closed field of characteristic zero and \(B\) an affine \(k\)-domain such that \(B\) is a UFD and dim \(B=3\). Then the following are equivalent: \begin{itemize} \item [(I)]\(B=k^{[3]}\). \item [(II)]\(ML^{*}(B)=k\). \item [(III)]\(ML(B)=k\) and \(ML^{*}(B) \neq B\). \end{itemize}