Is it possible to know the kind of alumina phase I have using EDS?
From my, relatively brief, experience with EDX, I want to say no, it is not possible. The reason is that EDX looks simply at atomic transitions, so the only thing you can make sense of is what atoms are present. In the case of different phases of the same material, it will always be the same atoms that are present.
I suppose if one were really good, however, then he/she could MAYBE correlate relative peak intensities and/or "erroneous" peaks (i.e. peaks that are identified as other, non-present elements) to different phases. But again, from my point of view, that's a big maybe.
Separate names with a comma.