Zoran Skoda comonad for comodule algebra

Let an algebra EE has two HH-comodule structures, ρ 1,ρ 2\rho_1,\rho_2 with isomorphism of corresponding comonads α:G 1→G 2\alpha: G_1\to G_2. This induces an isomorphism of kk-modules α E:E⊗H→E⊗H\alpha_E:E\otimes H\to E\otimes H underlying G 1(E)→G 2(E)G_1(E)\to G_2(E), and let α E(1⊗1)=z\alpha_E(1\otimes 1) = z. Then for all e∈Ee\in E, taking into account that α E\alpha_E is a morphism of EE-modules,

α E(ρ 1(e))=α E(e▹ 1(1⊗1))=e▹ 2α E(1⊗1)=ρ 2(e)z \alpha_E(\rho_1(e)) = \alpha_E(e\triangleright_1 (1\otimes 1)) = e\triangleright_2 \alpha_E(1\otimes 1) = \rho_2(e) z

As the natural transformation has to commute with the counit of the comonad, we read in particular that (id M⊗ϵ)α M=id M⊗ϵ(id_M \otimes \epsilon) \alpha_M = id_M \otimes \epsilon on M⊗HM\otimes H for each EE-module MM. In particular, z−1∈Ker(id E⊗ϵ)z-1 \in Ker(id_E \otimes \epsilon). On the other hand, α E\alpha_E has to correctly commute with id⊗Δid \otimes \Delta, what imposes that if z=∑z 1⊗z 2z = \sum z^1 \otimes z^2, then 1⊗1↦(α E⊗id H)(1⊗Δ(1))=∑z 1⊗z 2⊗11 \otimes 1 \mapsto (\alpha_E\otimes id_H)(1\otimes \Delta(1)) = \sum z^1 \otimes z^2\otimes 1, but also 1⊗1↦(id E⊗Δ)α E(1⊗1)=(id E⊗Δ)(z)=∑z 1⊗z (1) 2⊗z (2) 21\otimes 1\mapsto (id_E\otimes\Delta)\alpha_E(1\otimes 1) = (id_E \otimes \Delta)(z) = \sum z^1 \otimes z^2_{(1)}\otimes z^2_{(2)}. Act with id⊗ϵ⊗id\id \otimes \epsilon\otimes \id to obtain (∑z 1ϵ(z 2))⊗1=∑z 1⊗z 2=z(\sum z^1 \epsilon(z^2)) \otimes 1 = \sum z_1 \otimes z_2 = z. Thus z=z′⊗1z = z' \otimes 1 what with z−1∈Ker(id E⊗ϵ)z-1 \in Ker(id_E \otimes \epsilon) implies z′=1z' = 1; hence also z=1z = 1, therefore α E(ρ 1(e))=ρ 2(e)\alpha_E(\rho_1(e)) = \rho_2(e).

Created on January 7, 2020 at 20:37:50. See the history of this page for a list of all contributions to it.