Conduction Pathway of π-Stacked Ethylbenzene Molecular Wires on Si(100)
journal contributionposted on 12.08.2009, 00:00 by Manuel Smeu, Robert A. Wolkow, Hong Guo
One of the most important challenges of molecular electronics is to enable systematic fabrication of molecular functional components on well-characterized solid-state substrates in a controlled manner. Recently, experimental techniques were developed to achieve such fabrication where lines of π-stacked ethylbenzene molecules are induced to self-assemble on an H-terminated Si(100) surface at precise locations and along precise directions. In this work, we theoretically analyze charge transport properties of these ethylbenzene wires using a state-of-the-art first-principles technique where density functional theory (DFT) is used within the nonequilibrium Green’s function formalism (NEGF). Our device model consists of ethylbenzene stacks bonded to an H-terminated Si(100) surface and bridging two metal leads. The electron transmission spectrum and its associated scattering states as well as the resistance of the molecular wire are determined by the self-consistent NEGF-DFT formalism. The transmission spectrum has a resonance nature for energies around the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of the ethylbenzene wires. However, near the Fermi level of the device, which sits inside the HOMO−LUMO gap, the Si substrate is found to play an important role in providing additional pathways for conduction. It has emerged that, within our model system, the transmission peak nearest to the Fermi level corresponds to transport through the Si substrate and not the π-stacked molecular line. The low-bias resistance R is found to increase exponentially with the length of the molecular line n, as R ∼ eβn, indicating a tunneling behavior in conduction. We further found that the exponential scaling has two regimes characterized by two different scaling parameters β: a high value for conduction through the molecular stack in short lines and a lower value for conduction through the substrate in longer lines. Our results suggest that when the conduction of molecular wires bonded to semiconductor substrates is theoretically analyzed, conduction pathways through the substrate need to be taken into account.