A quotient stack related to the Weyl algebra. (Q408499)

Let \(k\) be a field of characteristic \(0\). The Weyl algebra is the ring \(A=k\langle x,y\rangle/(xy-yx-1)\). A \(\mathbb Z\)-grading is imposed on it by setting \(\deg x=1\) and \(\deg y=-1\). There is an isomorphism between \(A\) and the ring of differential operators with polynomial coefficients on the affine line \(\text{Spec\,}k[t]\) that is given by sending \(x\) to multiplication by \(t\) and \(y\) to \(-d/dt\). The main result in the article is that the category \(\text{Gr\,}A\) of \(\mathbb Z\)-graded \(A\)-modules is equivalent to the category of quasi-coherent sheaves on a quotient stack \(\mathcal X\) whose coarse moduli space is the affine line \(\text{Spec\,}k[t]\), and whose stacky structure consists of stacky points \(B\mathbb Z_2\) supported at each integer point \(n\in\mathbb Z\subset\mathbb A^1_k\): \(\text{Gr\,}A\equiv\text{Qcoh\,}\mathcal X\). \(\mathcal X\) is constructed as follows: Let \(\mathbb Z_{\text{fin}}\) be the group of finite subsets of \(\mathbb Z\) with group operation given by exclusive or. Let \(G\) be the affine group scheme whose coordinate ring is the group algebra \(k\mathbb Z_{\text{fin}}\) with its usual Hopf algebra structure. \(\mathbb Z_{\text{fin}}\) is \(2\)-torsion and is generated by the singleton sets \(\{n\}\). A \(k\)-valued point \(g\in G\) thus corresponds to a function \(\mathbb Z\to\{\pm 1\}\), \(n\mapsto g(\{n\}):=g_n\). An action of \(G\) on the ring \(C:=k[z][\sqrt{z-n}\mid n\in\mathbb Z]\) is given by \(g\cdot\sqrt{z-n}:=g_n\sqrt{z-n}\), defining the stack theoretic quotient \(\mathcal X:=\left[\frac{\text{Spec\,}C}{G}\right]\). The invariant ring of \(C\) is \(C^G=C_\emptyset=k[x_n^2]\) for all \(n\). Therefore, its coarse moduli space is the affine line \(\text{Spec\,}k[x_0^2]=k[z]\), which is an algebraic stack whose diagonal is locally of finite type but not quasi-compact. The \(k\)-valued points in \(\text{Spec\,}C\) are tuples \((a_i)\in k^{\mathbb Z}\), \(a_1\in k\), such a point corresponding to the maximal ideal \(\sum_{i\in\mathbb Z}(x_i-a_i)\). The relations for \(C\) imply that at most one \(a_i\) is zero. The isotropy group at such a point is isomorphic to \(\mathbb Z_2\). Such points are those where \(x_0^2\) takes an integer value. Hence all the stacky structure on \(\mathcal X\) occurs at the integer points \(x_0^2=n\), \(n\in\mathbb Z\), on the coarse moduli space \(\text{Spec\,}k[x_0^2]\). The action of \(G\) on \(C\) corresponds to the \(\mathbb Z_{\text{fin}}\)-grading of \(C\) given by \(\deg\sqrt{z-n}=\{n\}\). A standard result for quotient stacks says that \(\text{Qcoh\,}\mathcal X\) is equivalent to the category of \(G\)-equivariant sheaves on \(\text{Spec\,}C\), or equivalently, there is an equivalence \(\text{Qcoh\,}\mathcal X\equiv\text{Gr}(C,\mathbb Z_{\text{fin}})\), where \(\text{Gr}(C,\mathbb Z_{\text{fin}})\) is the category of \(\mathbb Z_{\text{fin}}\) graded \(C\)-modules. The author notes that \(C\) is isomorphic to the polynomial ring \(k[x_n\mid n\in\mathbb Z]\) modulo the relations \(x_n^2+n=x_m^2+m\) for all \(m\) and \(n\), with the grading given by \(\deg x_n:=\{n\}\), and the isomorphism given by \(z\leftrightarrow x_0^2\), and \(\sqrt{z-n}\leftrightarrow x_n\). The author uses results of Sierra's work on the graded Weyl algebra. In this picture, each geometric point represents a simple graded \(A\)-module: If \(\lambda\in k\setminus\mathbb Z\) there is a single simple graded \(A\)-module up to isomorphism, namely \(A/(xy-\lambda)A\); if \(n\in\mathbb Z\) there are two simple modules \[ X(n)=\left(\frac{A}{xA}\right)(n)\text{ and }Y(n)=\left(\frac{A}{yA}\right)(n-1). \] The isomorphism between \(A\) and \(\mathcal D(\mathbb A^1)\) may be chosen so that \(Y(1)\) corresponds to the natural module \(k[t]\) and \(X(1)\) corresponds to the module \(k[t,t^{-1}]/k[t]\). There is a non-split extension of \(X(n)\) by \(Y(n)\), and a non-split extension of \(Y(n)\) by \(X(n)\) for each \(n\). Having obtained the equivalence between \(\text{Gr}(A)\) and \(\text{Qcoh\,}\mathcal X\) or, equivalently, with \(\text{Gr}(C,\mathbb Z_{\text{fin}})\), alternative proofs of many of Sierra's results can be given by transferring results from \(\text{Gr}(C,\mathbb Z_{\text{fin}})\) to \(\text{Gr}(A)\). This is effective because the fact that \(C\) is a graded principal ideal domain makes the study of its graded modules easier. The author studies the correspondence between the different categories in a very thorough and direct way. The article is a really good example of methods and results obtained from noncommutative geometry, nontrivial isotropy groups and the relation to quotient stacks. Also, the article is basically self contained, except for the results from Sierra's work where the author analyzes her results. Both directly and indirectly, the article contains basic concepts of most of the noncommutative geometries developed the last decade.