[AMOS] Estimate Missing Values using FIML and export them to SPSS


I have a dataset with 1000 cases und 100 variables.
In 200 cases the last 20 variables are missing.
I want to use AMOS to do a FIML estimate of the missing data and export the datasheet to SPSS for further analysis.

I added every observed variable to the model in AMOS, but I didn't add any paths or covariances. I activated "Estimate means and intercepts" and choosed "Maximum likelihood".
The only way to export the data seemed to be "Amos Data Imputation" und "Stochastic regression imputation. AMOS created a single file with 10,000 cases (10 complete Datasets).

Is this the right way to do a FIML estimate of the missing data?
How do I handle the resulting file in SPSS - the number of cases has been changend!