Paper is being written and will be submitted shortly but below is some data which is expected to accompany the publication.
After reviewing the literature you’ll notice that very few groups are actually able to achieve very low free carrier concentrations (<5·1015 cm-3) and even fewer groups are able to reproduce this reliably. One of the problems is that while growing GaN by MOCVD (and other means) you’ll incorporate defects into the bulk such as hydrogen, carbon, oxygen, VGa, VN, and then complexes with each other just to name a few. These defects are considered to capture free carriers so the approximation of n=ND no longer holds and its more like n=(ND-NA) where the acceptor concentration depends on the hard to control impurity density.
This work increases the MOCVD TMGa flow during GaN growth to increase the growth rate but due to the increase in compensation, the Silane is also increased proportionally.
Below you can see the effect on the free carrier concentration and SSPC signal of samples grown with 1x, 2x, and 3x TMGA and Silane flows relative to each other.