A thought experiment: (note that this comes from a student with only amateur experience in solid state physics) suppose you could shortcircuit a pn diode... ie you just link a wire between both ends, or maybe even bend the pn diode in a ring (so you'd have a n-surface between two p-surfaces, and a p-surface between two n-surfaces). What would you get? Based on that second idea, I was thinking something among the lines of : you'd get two depletion areas at both ends, with a width that's equal to half of the width you'd get from a single end connection. I'm having problems with realizing what the differences in characteristics are between the 'internal potential' of the diode, evoked by the difference in free electron concentration (I hope that's an ok interpretation) , and the 'external potential', evoked by the source. I understand how the external potential can 'amplify' or 'lessen' the diode effect - that's somewhat like simple charge induction - however, when I try to combine the idea of the diode and the external potential as in above experiment, I get stuck. Maybe it would help if someone showed me convincingly why a diode can't be used as a power source.