Silicon Nitride grown by which one has less optical loss? Is stoichiometric Si3N4 better in quality than SixNy of PECVD? If yes, then why would someone use PECVD for SiN on a substrate that can stand the temperature of LPCVD?
Hi ! At first, what do you mean "less lossy"...? In what terms 'lossy'...? The quality of LPCVD Si3N4 is the best among others from the point of view of electrical behaviour (highest breakdown voltage values, smallest traps charge), chemical stability and resistance, etc. By fabrication of Si3N4 the one rather does not take into account the substrate, but the FUNCTION of the nitride, e.g. if you need a passivation layer you will use the PECVD (faster and more 'economic' process), if you need a gate dielectric (arther in a gate stack not as a sigle layer) or mask material for e.g. etching, then you will use LPCVD. However, this is a longer story... Regards !
Thanks! Edited.. I am looking for optical loss like absorption and scattering. I guess LPCVD grown SiN should have less optical loss. But I am looking for an evidence. Have found some comparisons of LPCVD and PECVD grown SiN but not a definite answer as to which one is better. Loss is the only concern.
As far as I know, both are relatively good. Often the pinhole defects contribute to most of the losses. In general process optimization is more important than LPCVD or PECVD for Si3N4 waveguides.
HOWEVER, in the visible wavelength range SiNx has some photo-luminescence this can be detrimental for experiments if you are trying to measure very weak signals. This is a more complicated matter and this is something I'm still looking into myself.
For now I can tell you that if you have the option you should go for LPCVD as this should have less absorption then PECVD, however there has been research on tackling the absorption.
Adding to the above colleagues who gave a satisfactory answer, if the SiN is ideally constructed it will have only the well known polarization loss at the high frequency. In case of deviation from the ideal structure because of the deposition method and conditions there will be additional losses which may be in form of dc electrical condition. May be also some additional polarization losses because the presence of other materials other than SiN for example excess of silicon because of non stoichiometry.
So as low pressure chemical process is less defective SiN as it runs at higher temperature it s expected to have less loss.
This is a qualitative answer. For quantitative answer you have to run experiments.