Thanks for your patience on this one! I heard back from my colleague. We notice that the concentration of carriers in your results is extremely low, especially if you compare it to a typical doped semiconductor. So, it is possible that the small variations that you are seeing are not actually accurate results but numerical artifacts. We would like to know more about what kind of effect you are trying to understand so that we can evaluate if CHARGE can simulate it accurately. Also, I would ask you to double check the simulation and make sure that the doping is correctly applied. For this, mesh the simulation region:
Then go the the "grid" result in CHARGE solver object and visualize the result "N" (which shows the doping level).