generally, the reason why FWHM in DSC is wider is that the crystallization is slower and spreaded over a greater range of temperature, probably due to a more complicated initial amorphous structure. By the way, you could try to repeat DSC analysis changing the heating rate and see the effect of it.
Thanks Luca for your comment. In this case we've introduced Holmium oxide content replacing equivalent amount of Lanthanum oxide and seen that FWHM of crystallisation peak decreases with Holmium oxide content.