Hello everyone
I'm trying to grow Si3N4 (or SiO2) layers (100-200 nm) to insulate metallic
layers. The Si3N4 is deposited by RF reactive sputtering of a silicon
target in a nitrogen/argon atmosphere (the sample can not be heated).
Unfortunately, the layers are not that insulating and contain defects.
Measured in the plane, the resistance of the silicon nitride is "infinite"
but the resistance for a current flowing perpendicularly to the film is not
always infinite. For a same sample, some device are well isolated but
others are not. The yield seems to vary with the growth condition (better
at higher pressures particularly). Have anybody already faced this problem?
Are they any papers published on the subject? Is there any "recipe" to
improve the yield (I've heard about doing the growth in two steps and bring
back the sample to the air between the two steps)

Best regards, Francois