Quotient sheaves of algebraic supergroups are superschemes (Q411756)

The article generalizes the theory of algebraic groups to the super context from a functorial point of view.This means that the authors study the quotient sheaf \(\widetilde{G/H}\) of an algebraic supergroup \(G\) by a closed supersubgroup \(H\). Most of the text is used to explain the scheme theory and the geometric invariant theory in the ``super'' setting. The ground field \(K\) is assumed to have characteristic different from \(2\), and the vector spaces graded by \(\mathbb Z_2\) form a symmetric tensor category \(\text{SMod}_K\). Objects from this category get the prefix ``super'', e.g. superalgebra, Hopf superalgebra and so on. All superalgebras including the Hopf superalgebras are assumed to be supercommutative, in the meaning that \(ab=(-1)^{|a||b|}ba\), were \(a\) and \(b\) are homogeneous elements of degree \(|a|\) and \(|b|\) respectively. A \(K\)-functor (resp. a supergroup) is a set-valued (resp. group-valued) functor defined on the category \(\text{SAlg}_K\) of superalgebras. The sheafification of \(K\)-functors includes the subclasses of affine superschemes, contained in superschemes, contained in \(K\)-sheaves. By an algebraic supergroup one understands an algebraic affine supergroup, which is a supergroup \(G\) represented by a finitely generated Hopf superalgebra \(K[G]\). A closed supersubgroup of \(G\) is a supergroup \(H\) represented by a quotient Hopf superalgebra of \(K[G]\). For such \(G\), \(H\), the functor that associated the set \(G(R)/H(R)\) of right cosets to a superalgebra \(R\) is called the naive quotient, and is denoted \((G/H)_{(n)}\). The sheafification of this functor is denoted by \(\widetilde{G/H}\). It is proved that \(\widetilde{G/H}\) is an affine supergroup if \(H\) is normal in \(G\), and this article considers the situation when \(H\) is not necessarily normal in \(G\). If the functor is restricted to the category \(\text{Alg}_K\) of purely even algebras (the ordinary commutative algebras) we get the ordinary algebraic group denoted \(G_{\text{res}}\) with its closed subgroup \(H_{\text{res}}\). The article considers the relation between \(\widetilde{G/H}\) and the quotient \(\widetilde{G_{\text{res}}/H_{\text{res}}}\) in the classical situation. The main questions the article is about to answer are the following: 1) is \(\widetilde{G/H}\) necessarily a superscheme?, and 2) is \(\widetilde{G/H}\) affine when the algebraic group \(H_{\text{res}}\) is geometrically reductive? These questions are asked by Brundan, who give another definition of the quotient \(G/H\), and the authors finally ask under which conditions their quotient coincides with the Brundan quotient. The main result of the article is the affirmative answer of question 1): Let \(G\) be an algebraic supergroup, and let \(H\) be a closed supergroup. Then the \(K\)-sheaf \(\widetilde{G/H}\) is a Noetherian superscheme. This result also holds in the situation with the sheafification \(\widetilde{H\backslash G}\) of the left coset functor. The Brundan quotient of \(G\) by \(H\) is a pair \((X,\pi)\) of a Noetherian superscheme \(X\) and a morphism \(\pi:G\rightarrow X\) with given properties. It is proved that the quotient in the main theorem above fulfills this, so that also question 3) is answered affirmatively. The second question is proven to be true in the introductory parts of the article. It is already proved by Demazure and Gabriel that \(\widetilde{G_{\text{res}}/H_{\text{res}}}\) is a Noetherian scheme, and the authors proof reduces to the this result by the relations to \(\widetilde{G/H}\). The method uses both Hopf algebraic and geometric methods, the geometric methods is represented by a comparison theorem generalizing the analogous classical comparison theorem of Demazure and Gabriel to the super context. In addition to the functorial approach, it is used a supergeometry through geometric superspaces which are topological spaces with structure sheaves of superalgebras. Then the comparison theorem shows that the two approaches are equivalent in some sense. The article uses most of its space to give meaning to the above. Direct limits are defined to construct geometric superspaces from \(K\)-functors, where basic conditions on super(co)algebras and super(co)modules are considered. Then geometric superspaces can be defined by direct limits applied on the \(K\)-functors. This is straight forward, but there is a need to prove all results in the noncommutative situation. The above mentioned comparison theorem is proved, and to be able to define the quotient superschemes, that is its open affines, the supergrassmannian is defined. The quotient dur sheaf \(\widetilde{\widetilde{X/G}}\) associated to an affine superscheme \(X\) with the action of an affine supergroup \(G\) is studied, and it is applied to the case \(U/H\) where \(U\) is an affine open supersubscheme of \(G\) which is stable under right multiplication with \(H\). An important result states necessary and sufficient conditions for the action of \(G\) on \(X\) to be free so that \(X/G\) is affine. The proof of this result uses the bosonization technique of braided Yetter-Drinfeld modules to a more general situation than actually needed. The article defines the ``super'' setting. This is, roughly speaking, the category of \(\mathbb Z_2\)-graded objects. Then the most needed algebraic geometric theory is generalized to this noncommutative setting. This includes Direct limits, supermodules, and supercomodules, all treated in the functorial setting by the so called \(K\)-functors (groups) to the category of sets (groups). Most of the concepts are relatively easily translated to the super setting. Of particular interest is the comparison theorem, geometric superspaces, the supergrassmannian and the superinvariant theory. The bozonization technique involves braided groups and Hopf algebraic methods. All in all the article is a nice generalization to the super setting. The printing is a bit disturbing, for instance, the ``less-that-or-equal''-sign is substituted by ``6'', and the square at the end of proofs are replaced by ``2''. The authors are probably not to blame for this, and the article is a very nice base for the algebraic geometric study of supergeometry. Also, it gives proper hints in the noncommutative setting.