I'm looking for detailed information on how the thickness of the mounted powder sample influences the resolution/peak shape in a powder xrd ( Bragg-Brentano geometry).
Excluding the problems of flat specimen, the main effect is a shift of the peaks: the subsurface is off focus with respect to the surface. Actually you should also have some asymmetry due to absorption, so the larger the absorption, the smaller the problem. In that case roughness might start playing a role
Will it also broaden the peaks? since there are several "focal planes" i.e. the volume from which diffraction occurs is very expanded. please see my crude sketch.
If you are interested from the experimental point of view, you will need to take a NIST standard and chek the goniometer and optics alignment, systematic errors, geometry abberations of the x-ray machine. The nature of the sample (metallic, insulator,..) could be very important for absoption effects.
The picture is for parallel beam and not Bragg-Brentano. The asymmetry I was talking about of course adds to the side of the existing peak so it gives also some additional broadening. Remember the scales as the penetration can even be lower than the error in positioning the specimen on the axis of the machine that in any case is aligned using a real specimen.. that had its absorption.
The assymetry will be more severe at low 2-theta values, especially for samples that have low absorption. As noted above, you also have to account for other errors due to the instrument (mis) alignment, slit width, etc. You can reduce the peak broadening from the sample by using a zero-background sample holder with a very shallow powder depth.
You can obtain custom-sized zero background quartz sample holder inserts from The Gem Dugout in Pennsylvania. We have purchased inserts from them in the past. It helps if you send them one of your sample holders so they can properly fit their inserts.
The phenomenon becomes even problematic for multiphase material of high and low-absorbing phases which perhaps form clusters (particles) of smaller grains. Then the assumption of an homogenized sample is no more fulfilled. The determined phase fraction is shifted to the high-absorbing phase.