I am doing the following experiment in my lab. I have a collimated laser beam, which is focused by a lens. Then the light is collected by a microscope objective. Also, I use two lenses (L1 and L2) to image the focus plane and back focal plane (BFP) out of the objective. Then I introduce an iris in the imaged BFP to do angular filtering. Consequently, changing the iris diameter I am able to change the numerical aperture (NA) of the focused laser. I have also two camera and can record the images of the BFP and the focus plane.
Now, I use some good lens with high transmission as a focusing lens and record the image of the focus plane. This image is a reference image. Then, I use a microscope objective as a focusing lens and again record the image of the focus plane. Dividing the integrated intensity of the focus plane by the integrated intensity of the reference image, I am able to calculate the efficiency of the focusing objective. Integration occurs over the main Airy disk of the spot. Finally, changing the iris diameter I can plot the efficiency vs. NA of the focusing objective. I measured several objectives and every time get the same tendency: efficiency drops with NA. I wonder why is it so? I could think about spherical aberrations because they affect more with higher NA. However, good objectives are designed to reduce the aberrations. Also note that the spots in the focus are perfect Airy patterns. What could be possible reasons for the tendency of the drop?