1: \begin{abstract}

2: We consider the stochastic Cahn-Hilliard equation

3: with additive noise term $\eps^\gamma g\, \dot{W}$ ($\gamma >0$)

4: that scales with the interfacial width parameter $\eps$.

5: %We propose an implicit time semi-discretization for the equation

6: %based on the backward Euler scheme.

7: %

8: We verify strong error estimates for a gradient \revised{flow} structure-inheriting time-implicit discretization, where $\eps^{-1}$

9: only enters {\em polynomially}; the proof is based on

10: {higher-moment estimates for}  iterates, 

11: and a (discrete) spectral estimate for its deterministic counterpart.

12: For $\gamma$ sufficiently large, convergence {in probability} of iterates

13: towards the deterministic Hele-Shaw/Mullins-Sekerka problem

14: in the sharp-interface limit $\eps \rightarrow 0$ is shown.

15: %the numerical solution of the stochastic Cahn-Hilliard equation converges to the solution of the deterministic Hele-Shaw problem

16: %in the sharp-interface limit $\eps \rightarrow 0$.

17: These convergence results are partly generalized to a fully discrete finite element based discretization. 

18: 

19: We complement the theoretical results by computational studies to provide practical evidence

20: concerning the effect of noise (depending on its 'strength' $\gamma$) on the geometric

21: evolution in the sharp-interface limit.

22: For this purpose we compare the simulations with those

23: from a fully discrete finite element numerical scheme for the (stochastic) Mullins-Sekerka problem.

24: The computational results indicate that the limit for $\gamma\geq 1$ is the deterministic problem,

25: and for $\gamma=0$ we obtain agreement with a (new)

26: stochastic version of the Mullins-Sekerka problem. 

27: 

28: 

29: %Using the proposed schemes we perform numerical studies of both models

30: %to examine the convergence behavior of the stochastic Cahn-Hilliard model with 

31: %different scaling of the noise.

32: \end{abstract}

33: