The question is why we have very small crystal size in XRD but in SEM we can observe big faceted crystals, For example perhaps the faceted particles may not be single crystal?
You misunderstand the measure achieved by XRD. This is not crystal size. There is a relationship between Bragg peak breadth and average coherent scattering domain size. This has been called 'crystallite' size which causes interminable confusion. There are many similar topics on this site where this is dicussed, so best to use the search tool to find out more.
it is not crystal size, but crystallite size what you get from XRD! You can calculate Particle size from SEM, which is totally different from Crystallite size. what i can understand from your question, your material is polycrystalline. so no question of single crystal
to add onto above answers, for sure your crystal size (or more preferably called particle size, formed by 2+ crystals) estimated from SEM/TEM/DLS etc would be larger than your crystallite size (individual crystal) estimated from XRD (Scherrer equation).
You may find our published article detailing both crystallite sizes (XRD) and particles sizes (SEM) of TiO2 nanoparticles and to see the difference in both.