1: \begin{abstract}

2: We prove a weak rate of convergence of a fully discrete scheme for

3:   stochastic Cahn--Hilliard equation with  additive noise,

4:   where the spectral Galerkin method is used in space and the backward Euler method is used in time.

5:   Compared with the Allen--Cahn type stochastic partial differential equation, the error analysis here is much more sophisticated due to the presence of the unbounded operator in front of the nonlinear term.

6:   To address such issues, a novel and direct approach has been exploited which does not rely on a Kolmogorov equation but on the integration by parts formula from Malliavin calculus.

7:    To the best of our knowledge, the rates of weak convergence

8:     are revealed in the stochastic Cahn--Hilliard equation setting for the first time.

9: \end{abstract}

10: