I am trying to have a optical system where from a single mode fiber at a wavelength of 532nm, we shape the beam into an ellipse of 1um by 2.5um waist radius at the image plane. For the beam to match acousto-optical deflector’s aperture dimensions (5mmx13mm) with an apodization of 2.5, we collimate the beam with a singlet, then expand it in one direction with a Galilean telescope with cylindrical lenses and then have a telecentric relay system (with a black box objective) that gets the required demagnification at the image plane. 

I have attached a draft system but the performance is not at good as we need it to be. I have two questions:

• How can I make the crosstalk/intensity seen by anything over 3.5um away from the center smaller? (Currently it’s around 10^-(3.3) and we would like it to be 10^-4)
• I noticed that when I measure the size of the rays right before the objective in layout, it isn’t consistent with the beam size at the image. From the PSF, it seems like the 1/e^2 beam waist radius is around 2um, but according to the Gaussian beam propagation equations if the 1/e^2 beam radius is 10mm at the objective (EFFL=57.6mm) the beam waist radius should be 1um at the image - instead of 2um. This maybe be just due to a discrepancy between what the rays in the layout is showing and the actual 1/e^2 envelope - in that case what do these outer rays show in the layout?

Thank you so much for the help, I am still a beginner with optics and Zemax so please let me know if anything about this question makes little sense.