Bandgaps make semiconductors work--they are the energy region where no electron can exist--measured by the electron-volts needed to make an electron jump from the valence band (where they are trapped by the nucleus) to the conduction band where they are free to conduct electricity. An insulator has a very wide bandgap that under normal operating conditions is never jumped. Conductor has no bandgap (or a miniscule one) so that electricity is free to move in them for even the tiniest voltage offset. Semiconductors have various sized bandgaps which give them their abilities, such as to turn on and off a transistor, or in optical materials to emit photons to shed the energy required to make an electron fall back after jumping its bandgap. Since graphene has no bandgap, it is a semi-metal conductor with extremely high electron mobility due to its extremely uniform honeycomb structure that effectively makes electrons lose their mass. But to make it into a semiconductor you need to either add defects, use passivation, doping, nanoscale perforations, make "ragged" edges, use more than one layer slightly offset from each other, or even expose graphene to humidity.
Forgive me for being naiive, but I thought one of the great strengths of graphene was its native zero bandgap condition. I've seen numerous articles discussing the strengths of graphene as a material that can be processed into both P and N junctions in the same piece of substrate. It sounds like this research is trying to optimize substrates for a single type of junction. Am I misunderstanding the issue here?