Hi,
Sorry for the delay, I missed this post.
Well, one thing that can happen is the following:
You do have a phase only modulator, but due to the finite numerical aperture of every optical setup, you always end up with some amplitude modulation, at least for the high spatial frequencies.
Consider the following. You display a checkerboard on the SLM. I assume that the SLM is working in pure phase only modulation, that means that the squares are either with a 0 phase or a pi phase. If I use two lenses as a 4-f system, what should I see in the image plane? If everything was perfect, I should see nothing with a CCD camera as it should be a phase only image similar to the one in the SLM plane. But if you do the experiment, you will see black lines at the edges or the squares but the inside of the squares themselves are all white. What happens is that the lenses you use do not have a numerical aperture equal to 1. When you display something with sharp edges, the corresponding high spatial frequencies are diffracted quickly and do not enter your lenses (or whatever optical system you have). So in your object plane, you do not have exactly the same image as in the SLM plane as you filtered out (involuntarily) the high spatial frequencies.
The short story is, when you have high spatial frequencies in a phase only image, such as sharp edges, you will always have a bit of amplitude modulation.
Now, back to the Fresnel lenses. In such pattern, you should not have sharp edges ideally. The phase should gradually go from 0 to 2pi and then to 2pi to 4pi (so 0 to 2pi again) etc... BUT, if you did not calibrated the SLM correctly, this means that you do not know accurately the value corresponding to 2pi, you will have sharp edges. If for example the value you put for the 2 pi pixel value really corresponds to 1.8 pi, the value would go gradually from 0 to 1.8 pi then from 2 pi to 3.8 pi, so you will have a sharp jump from 1.8 pi to 2 pi. These sharp edges will give rise to amplitude modulation in a finite numerical aperture system.
Sebastien