Actually it depends on your substrate too that which method you are going to use .As Cross sectional TEM can be used for measuring film thickness but its not applicable for all kind of substrates. Miss Lourdes you can measure your film thickness using XRR(x-ray reflectivity) and also by using profilometer but it all depends upon your film thickness as for very thin films profilometer is not an adequate tool.
XRR reflectivity can measure the thickness but it has to be used with a caution. Unless your film is a single layer, you'll have to model XRR response to match your experimental data to deduce the film thickness.
As others mentioned, there are other profiling techniques which you can utilize such as cross-sectional SEM, cross-sectional TEM, AFM (if you film is very thin, within a few nm), optical profilometers, spectroscopic ellipsometry, and profilometers with physical stylus (e.g: Veeco Dektak lines).
All these techniques have their pros and cons thus it all comes down to what technique will suit your film the best (as other have mentioned above).
In principle, specular XRR and XRD are actually same. Both can get Bragg pattern of your film (if your sample is sort of lamella structure like mine), and refine the data by means of form factor, thickness. Form factor magnitude can be estimated from the reflected intensity and qz (angle). The point is you should be aware of your sample structure to design the fitting routine.
XRR is the general term used for measuring thickness of films ranging from amorphous to crystalline to epitaxial. The measurements are generally made by examining the long range order signal close to the I0 (incident beam vector). Correct? XRR, SAXS, GIXRD, WAXS etc. fall in to the super set of XRD, I'd think.
Please review the following post for super lattice epitaxial film thickness and make suggestions when convenient.
Here is some recent XRR data that we acquired at MIT using the NIST2000SRM.
XRR is truly "grazing", isn't it? That is why the certitude & precision is so high. Angular range includes 0.40 to 1.40 degree 2-Theta at a 0.75 Arcsec resolution! The Pixel size of 29um in this case was tantamount to a 26 Arcsec "slit" at a SDD (sample to detector distance) of 232 mm. The challenge usually is due to the large sample coverage at that shallow an angle. The values are an average over the sampling area. However, XRR is non-destructive, non-contact, non-intrusive, in situ and real time irrespective of the film chemistry. Including monocrystalline, epitaxial, polycrystalline, amorphous, polymeric, metallic, semiconducting, superconducting, insulating, etc.