We are trying to find a supplier that can grow 200 nm stoichiometric LPCVD Si3N4 on quartz or fused silica substrates and so far no one agreed to do it. They say the film will peel off. What I do not understand is this, we got Si3N4 on oxidized silicon substrates before, so adhesion between Si3N4 and oxide was fine enough so that it can hold the stress on Si3N4 film, what makes the difference when the oxide is 500 micron thick instead of 3 micron oxidized silicon wafers? Does it bow? What is the problem? They say, they can grow silicon rich low stress, but then the waveguides that will be made out of this thin film layer will be absorptive.