Algebraic supergroups of Cartan type (Q741373)

There is a classification theorem for complex finite dimensional simple Lie algebras, and a classification of all complex finite dimensional simple Lie superalgebras. The last objects fits into two disjoint families: The ones of classical type, that is strict super-analogues of simple, finite dimensional, complex Lie algebras and the ones of Cartan type, which are simple but infinite dimensional. This classification result can be used as a basis for the classification, existence, and construction of simple Lie supergroups and simple algebraic supergroups. The problems above for Lie supergroups are solved by a superanalogue of Lie's Third Theorem, the problems in the algebraic situation remains. In the standard situation, a constructive procedure giving all finite dimensional connected simple algebraic groups was given by Chevalley. Starting with a (complex) f.d. simple Lie algebra \(\mathfrak g\), a faithful \(\mathfrak g\)-module \(V\), a group of requested type is eventually constructed as a subgroup of \(\text{GL}(V)\). In particular, applying this construction gives all connected algebraic groups whose tangent Lie algebra is finite dimensional and simple. This method can be extended to the framework of reductive \(\mathbb Z\)-group schemes. By analogy, Chevalley's method is used in known work to the finite dimensional simple Lie superalgebras of classical type, to construct connected, algebraic supergroup-schemes (over \(\mathbb Z\)) which integrate such Lie superalgebras. In this article the author implements Chevalley's idea to simple Lie superalgebras of Cartan type. The main result is a constructive procedure for connected, algebraic supergroup-schemes (over any ring) whose tangent Lie superalgebra is simple of Cartan type. A seceond result in the article is a uniqueness theorem for algebraic supergroups given above. The author starts of with a finite dimensional Lie superalgebra \(\mathfrak g\) of Cartan type. Then a detailed description of the root spaces with respect to a fixed Cartan subalgebra is given and the key notion of \textit{Chevalley basis} is introduced. The existence of such a Chevalley basis is proved, and a PBW-theorem for the Kostant \(\mathbb Z\)-form of the universal enveloping superalgebra of \(\mathfrak g\) is given. The next step is to choose a faithful \(\mathfrak g\)-module \(V\), and to show that there is a lattice \(M\) in \(V\) fixed by the Kostant superalgebra and by a certain integral form \(\mathfrak g_V\) of \(\mathfrak g\). Let \((\text{salg}_{\Bbbk}\) be the category of commutative \(\Bbbk\)-superalgebras, and consider the functor \(G_V:(\text{salg}_{\Bbbk})\rightarrow(\text{groups})\) given as follows: For \(A\in(\text{salg}_{\Bbbk})\), \(G_V(A)\) is the subgroup of \(\text{GL}(A\otimes_{\mathbb Z} M)\) generated by homogeneous one-parameter subgroups associated with the root vectors and with the toral elements in a Chevalley basis. Finally, \(\mathbb G_V\) is the sheafification of \(G_V\). The commutation relations among the generators gives a factorization of \(\mathbb G_V\) into a direct product of representable algebraic superschemes. Then \(\mathbb G_V\) is itself representable, which defines it as an \textit{affine} algebraic supergroup. This construction of \(\mathbb G_V\) yields an existence theorem of a supergroup having \(\mathfrak g_V\) as tangent Lie superalgebra. Also, the author proves the corresponding uniqueness theorem, that every such supergroup is isomorphic to some \(\mathbb G_V\). As an example, the author constructs \(\mathbb G_V\) for \(\mathfrak g\) of type \(W(n)\) and \(V\) the defining representation; that is the Grassmann algebra with \(n\) odd indeterminates, and \(W(n)\) the algebra of superderivations. The article is self-contained and detailed. It starts with all the necessary preliminaries on superalgebras, superspaces, and supergroups. It defines Lie superalgebras and functions with values in Lie superalgebras, and it gives the definition and properties of the two types of Lie superalgebras: Cartan type and classical type. For the first type, \(W(n)=\text{Der}_{\mathbb K}(\Lambda(n))\), the set of \(\mathbb K\)-(super)derivations of \(\Lambda(n)\), is important, and its Lie structure and its properties are given. Other necessary concepts are Cartan subalgebras, \(\mathfrak g\)-modules, roots, and root spaces; all are thoroughly defined and introduced. After the introductory chapters follow the integral structures with the main objective: To extend the classical notion of Chevalley bases for (semi)-simple Lie algebras. The Kostant superalgebra is defined, and explicit computations are used to prove results needed in the rest of the text, in particular in the proof of Kostant's PBW theorem: The Kostant superalgebra \(K_{\mathbb Z}(\mathfrak g)\) is a free \(\mathbb Z\)-module, that is, for any given total order \(\preceq\) of the set \(\tilde{\Delta}\coprod\{1,\dots,n\}\) a \(\mathbb Z\)-basis of \(K_{\mathbb Z}(\mathfrak g)\) is the set \(\mathcal B\) of ordered PBW-like monomials, i.e. all products without repetitions of factors of type \(X_{\tilde\alpha}^{\ell_{\tilde\alpha}},(H_i,n_i),X_{\tilde\gamma}\) with \(\tilde\alpha\in\tilde\Delta_{\overline 0}\), \(i\in\{1,\dots,n\}\), \(\tilde\gamma\in\tilde\Delta_{\overline 1}\), and \(\ell_{\tilde\alpha}\), \(n_i\in\mathbb N\) - taken in the right order with respect to \(\preceq\). Algebraic supergroups \(G_V\) of Cartan type are defined and generalized from the classical situation. This gives a sheaf functor \(\mathbb G_V:(\text{salg})_{\Bbbk}\rightarrow(\text{groups})\) which is a sheaf functor such that \(\mathbb G_V(A)=G_V(A)\) when \(A\in(\text{salg})_{\Bbbk}\). These functors are then eventually proved to be the affine algebraic supergroups of Cartan type, the main object of interest in the rest of this article: Let \(\mathbb G_V\) be the affine supergroup of Cartan type built upon \(\mathfrak g\) and the \(\mathfrak g\)-module \(V\). Then \(\text{Lie}(\mathbb G_V)\) is quasi-representable, and actually representable, namely \(\text{Lie}(\mathbb G_V)=\mathcal L_{\mathfrak g_{V,\Bbbk}}\) as functors from \((\text{salg})_{\Bbbk}\) to \((\text{Lie}_{\Bbbk})\). The final chapter gives even more detailed results on the classical case. The article is a good, advanced and technical, but self-contained study of the subject of Lie superalgebras. There is a lot of interesting and important generalizations, and the article, which is an attribute Pierre Cartier on his 80th birthday, is well worth working through.