In that case you might go for the third option I mentioned (interferometry/ellipsometry); at least that's how people in our institute measure the thickness of their coatings.
The thickness of a known oxide can be measured by ellipsometer to measure the thickness of oxide layers. For more details of the method you can follow the link:Article Fundamentals and applications of spectroscopic ellipsometry
The other method is an electrical method where you supply the layer by a lower metal contact and an upper metal contact and just measure the capacitance C of resulting parallel plate capacitor. Then use the formula C= epsilon A/ d , where A is the area of the capacitor, d its thickness and epsilon its dielectric material.
Knowing all quantities except d one can determine d.
You can bevel the edge of the layer on the substrate and use either the optical microscope or the electron microscope to measure the length of the beveled layer from which one can determine the layer thickness if the bevel angle is known.
You can find chapters on measuring the material characteristics in the book vlsi technology by s m Sze.
If you have access to a XPS equipment , and you have a ultra-thin film, lets say up to 5-6 nm , you can use the techinque AR-XPS (Angle Resolved XPS) , it has the benefits that is very sensible to the surface