nanoparticle size

Im working with a multimodeV AFM in AC mode to characterize the size and shape of Ag nanoparticles. Samples are prepared by dropping the nanoparticles on mica sheet. Nanoparticles are clearly visible and well dispersed without forming large aggregates.   AFM topography images  are showing an average heigth of about 2 nm with a  lateral size of about 15 nm. I supposed that this was due to the tip convolution effect. I used TESP with a nominal apex radius guaranteed <12 nm.

I also performed XRD analisys and I  used the Debye-Scherrer formula to recover the nanoparticle size :  D=0.9*Lambda/(FHWM*cos(Theta)) where lambda is the wave length (0.1541 nm) FWHM is the full width at half maximum of the selected diffraction peak  and Theta is the correspective diffraction angle.   The calculated average value of the particle size by means of XRD seems to be in agreement with the lateral diameter values extrapolated by grain analysis of AFM images but not with the heigths. 
This was surprising me as I supposed AFM height values were more meaningful. Of course I checked with the z-calibration gratings TGZ01 from Mikromash . This calibration test showed the expected step heigths of 20 nm. 

I looked in literature and I found several scientific papers claiming a perfect agreement of XRD  results with AFM lateral size measuraments. But neither an esplicit value of nanoparticles heigths nor the z-scale bar in AFM images was reported.  

Anyone can help me to clarify my doubts?

Thanks in advance Fabrizio