Anomalous Tunneling in Carbon/Alkane/TiO 2 / Gold Molecular Electronic Junctions: Energy Level Alignment at the Metal/Semiconductor Interface Haijun Yan †,‡ and Richard L. McCreery* ,‡,§ Department of Chemistry, The Ohio State University, 100 West 18th Avenue, Columbus, Ohio 43210, National Institute for Nanotechnology, National Research Council of Canada, Edmonton, Alberta, Canada T6G 2M9, and Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2 ABSTRACT Carbon/TiO 2 /gold electronic junctions show slightly asymmetric electronic behavior, with higher current observed in current density (J)/voltage (V) curves when carbon is biased negative with respect to the gold top contact. When a ∼1-nm-thick alkane film is deposited between the carbon and TiO 2 , resulting in a carbon/alkane/TiO 2 /gold junction, the current increases significantly for negative bias and decreases for positive bias, thus creating a much less symmetric J/V response. Similar results were obtained when SiO 2 was substituted for the alkane layer, but Al 2 O 3 did not produce the effect. The observation that, by the addition of an insulating material between carbon and TiO 2 , the junction becomes more conductive is unexpected and counterintuitive. Kelvin probe measurements revealed that while the apparent work function of the pyrolyzed photoresist film electrode is modulated by surface dipoles of different surface-bound molecular layers, the anomalous effect is independent of the direction of the surface dipole. We propose that by using a nanometer-thick film with a low dielectric constant as an insertion layer, most of the applied potential is dropped across this thin film, thus permitting alignment between the carbon Fermi level and the TiO 2 conduction band. Provided that the alkane layer is sufficiently thin, electrons can directly tunnel from carbon to the TiO 2 conduction band. Therefore, the electron injection barrier at the carbon/TiO 2 interface is effectively reduced by this energy-level alignment, resulting in an increased current when carbon is biased negative. The modulation of injection barriers by a low-κ molecular layer should be generally applicable to a variety of materials used in micro- and nanoelectronic fabrication. KEYWORDS: molecular electronics • injection barrier • titanium dioxide • work function • monolayer • tunneling • electron transport INTRODUCTION I nterfacial energetics play a central role in the physics of various micro- and nanoelectronic devices, including organic light-emitting diodes (OLEDs), polymer light- emitting diodes (PLEDs), organic field-effect transistors (OFETs), organic solar cells, and molecular electronic junc- tions (1-3). In many cases, the performance or character- istics of such electronic devices are greatly affected or controlled by interfacial charge injection barriers (4-18) at the interface between the contact electrode and the active layer. The problem is illustrated in Figure 1, for the case of a metal/TiO 2 /metal junction, but similar issues apply to injection into molecular layers with highest occupied (HOMO) and lowest unoccupied (LUMO) molecular orbitals. Figure 1A shows that a hole injection barrier (Φ h ) can be defined as Φ h ) E f - E HOMO (1) where E HOMO can be replaced by E VB for a semiconductor. Likewise, the electron injection barrier (Φ e ) is defined as Φ e ) E LUMO - E f (2) where E LUMO can be replaced by E CB for a semiconductor. In the example shown in Figure 1B, the electron injection barrier, Φ e , prevents electron injection into the conduction band (CB) until the applied bias is large enough to overcome the barrier by field emission, thermionic emission, or tun- neling. Both the electron and hole injection barriers can play dominant roles in charge transport in organic and molecular electronic devices and may ultimately determine the per- formance of OLEDs, PLEDs, and OFETs. Several techniques have been employed to modulate the charge injection barrier in micro- and nanoelectronic devices through interface engineering, either by physical/chemical treatment of the contact electrode surface by mechanical polishing (19), oxygen plasma treatment (5, 20), and chemi- cal treatment with acids and bases (21), etc., or by deposi- tion of an additional “intermediate layer” between the contact electrode and the active layer (4, 10-12, 22-38). Although these interface engineering techniques have im- proved interfacial charge injection and enhanced the device * Corresponding author. Tel.: 780-641-1760. E-mail: richard.mccreery@ ualberta.ca. Received for review October 15, 2008 and accepted December 29, 2008 † The Ohio State University. ‡ National Research Council of Canada. § University of Alberta. DOI: 10.1021/am800126v © 2009 American Chemical Society ARTICLE www.acsami.org VOL. 1 • NO. 2 • 443–451 • 2009 443 Published on Web 01/23/2009