Yes, that's what my intuition tells me too. But I think the radius of the source beam would also be a factor.
The problem is, all I could find after a couple of hours of research, is the Fraunhöfer diffraction formula (
Fraunhofer diffraction - Wikipedia, the free encyclopedia), which can be used to calculate intensity and spread of the interference pattern of a 1D-single-slit scenario. But I have absolutely no idea how this already very complicated piece of mathematics could be translated to suit a 2D-multi-slit scenario, i.e. diffraction grating.
I happen to be getting my PhD in something related to diffraction, specifically, manufacturing complex grating structures, and I can tell you that quantifying scientifically the intensity of the diffracted beams is very hard to obtain even for 2D gratings. There is software out there but it is really expensive and my boss won't even purchase that.
What I can tell you though, from personal experience fabricating 1D and 2D gratings is that the 2D gratings behave in the same way as 1D gratings but with powers reduced, in terms of the orders, it is correct to assume that the intensity drops as the order # goes up. Yes, the main beam will have most of the power, but that depends on how efficient the grating is. Grating efficiency is very hard to understand and model, especially for 2D gratings and complex structures the equations are just too mathematically complex to even bother solving. The efficiency of the gratings depends on several factors, like wavelength, shape of the grooves, depth of the grooves, etc.
To clarify something, any sort of periodic array of elements is itself a diffraction grating, even a 1D set of lines. The number and type of periodicity translates into the shape of the Fraunhofer diffraction pattern; for example a set of parallel lines gives a set of dots in a line. Lines in two dimensions (or any sort of arrangement with 2 periodicities) gives a diffraction pattern with 2 periodicities as well, usually an arrangement of dots in a square like fashion, much like you are describing.
Getting an equation specific for an NxN matrix is something I haven't looked up, but there are articles like
this one that have information on this.
My best recommendation would be to just measure the power of the reflected beam, which should be the highest, and adjust your power accordingly so it does not exceed 5 mW. Or you could measure the first order spots (closest to the 0th order, the reflected/transmitted beam), and make sure THOSE don't exceed 5 mW and then block the main one.
PS: sorry for bringing a thread back from the dead, but I read diffraction and that is one thing I am proud for knowing about, because very few people do.
