A brief review of graphene hot-electron transistors appeared on the arXiv in September. These devices are bipolar rather than field-effect transistors. They are designed with an insulated base, so that cold electrons are blocked while hot electrons are passed. Simulations promise fast (~1 THz) operation. This is not reliably achieved in experiment, probably because the graphene does not integrate well with many silicon-based manufacturing flows. For example, the graphene must be separately grown and then transferred onto the device substrate, it is difficult to non-destructively yet reliably grow other material layers on top of graphene, and reliable contacting is not yet achievable. Nonetheless, it is interesting to see what can be done and how the resulting devices compare to model predictions.
Reference: S Vaziri, et al., “Going ballistic: Graphene hot electron transistors”, arXiv:1509.01025.