The symmetry of finite group schemes, Watanabe type theorem, and the a-invariant of the ring of invariants
Abstract
Let k be a field, and G be a k-group scheme of finite type. Let Gad be the k-scheme G with the adjoint action of G. We call λG,G=H0(Spec k,e*(ωGad)) the Knop character of G, where e:Spec k→ Gad is the unit element, and ωGad is the G-canonical module. We prove that λG,G is trivial in the following cases: (1) G is finite, and k[G]* is a symmetric algebra; (2) G is finite and \'etale; (3) G is finite and constant; (4) G is smooth and connected reductive; (5) G is abelian; (6) G is finite, and the identity component G of G is linearly reductive; (7) G is finite and linearly reductive. Let V be a small G-module of dimension n<∞. We assume that λG,G is trivial. Let H= Gm be the one-dimensional torus, and let V be of degree one as an H-module so that S=Sym V* is a G-algebra generated by degree one elements, where G=G× H. We set A=SG. Then we have (i) ωAωSG as (H,A)-modules; (ii) a(A)≤ -n in general, where a(A) denotes the a-invariant. Moreover, the following are equivalent: (1) The action G→GL(V) factors through SL(V); (2) ωS S(-n) as ( G,S)-modules; (3) ωS S as (G,S)-modules; (4) ωA A(-n) as (H,A)-modules; (5) A is quasi-Gorenstein; (6) A is quasi-Gorenstein and a(A)=-n; (7) a(A)=-n. This partly generalizes recent results of Liedtke--Yasuda arXiv:2304.14711v2 and Goel--Jeffries--Singh arXiv:2306.14279v1.
Turn this paper into a lesson
ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.