•  first of all, you should define your XRD spectrum is HRXRD or Powder XRD spectrum, I would guess you are talking about the powder XRD? in powder XRD spectrum case, the FWHM of the peaks are followed the Debye–Scherrer equation, which you can calculate the grain size of your sample, therefore, the composition can will gives smaller grain size, the orientation of your sample can be checked in JCPD database, if your sample is single crystalline material you should really think about the strain and defects effect on HRXRD spectrum. hope that helps you

Hi...Ali, ...In addition to previous answer, The difference of peaks (bar ZnO and ZnO/TiO2 , which sharp and broad respectively) . It means the size of particle in bar ZnO is relatively bigger than that of in ZnO/TiO2. In other words, nanosize distribution in bar ZnO is wider than that of in ZnO/TiO2. It can imply that the preparation nano-ZnO (disperse in TiO2) will give a better result (Bar form of ZnO in TiO2 is shorter).

Dear Ali,

There are several reasons for peak brodening in XRD. You can find several interesting discussion on RG about this subject.

One of the reasons would be the internal stress. Other is, as indicated previously, the mean crystallite size which can be estimated from the breath of the diffraction pattern. 

@Xiaoyi : it is not Debye–Scherrer equation but only Scherrer equation

thank you very much my colleagues for your answers. your answer is very useful for me

Most crystallites have a mosiac structure and it is the average size of the mozaic blocks, that determines the peak width.  The crystallite size is usually much bigger than the mosaic block size.