Current-voltage Characteristics of Molecular Conductors: Two versus Three Terminal

This paper presents a computational study of the feasibility of a molecular transistor. We address the question of whether a “rigid molecule” (one which does not deform in an external field) used as the conducting channel in a standard three-terminal MOSFET configuration can offer any performance advantage relative to a standard silicon MOSFET. Our self-consistent solution of coupled quantum transport and Poisson’s equation shows that the presence of a gate terminal induces asymmetry in the (otherwise symmetric) two terminal current-voltage (IV) characteristic. This asymmetry arises because the molecular density of states (DOS) is fixed relative to the source Fermi level for the three terminal case, whereas it is roughly halfway between the source and drain Fermi levels for the two terminal case. We show that just as in a conventional MOSFET, the gate oxide thickness needs to be much smaller (one tenth) than the channel length (length of the molecule) for the gate control to be effective. Furthermore, we show that a rigid molecule is a poor switch with a temperature independent subthreshold slope larger than 60 mV/decade, because the metal-induced gap states in the channel prevent it from turning off abruptly. In short, no particular advantage is gained by using a molecular conductor for an FET channel, except that due to the reduced dielectric constant of the molecular environment.