A fundamental goal of molecular electronics is to design useful transport signatures into solid-state devices through control over the molecular components. In principle, theory and experiment should unite to provide a “chemical intuition” that guides 1) the direction of molecular design and synthesis and 2) the choice of the device materials and architectures that are employed in molecular electronics. To date, however, the field has advanced largely through the development of phenomenological—rather than quantitative—models of molecular devices. For example, traditional models for correlating properties, such as intermolecular charge transfer with molecular structure (electron-transfer theory) are designed for the solution phase. Similarly, such structure–property relationships can be tested by using a variety of sophisticated spectroscopic techniques, such as optical measurements and multidimensional NMR spectroscopy, which have been designed to work in solution on statistical numbers of molecules. Consider the case of a single molecule bridging two electrodes. Any theory of such a system must consider the molecule and the electrode as a single nonseparable entity. The objectives of the theory would be 1) to predict the alignment of the Fermi levels of the electrodes with the molecular orbital energies in the molecule, 2) to forecast the voltage-dependent electron conductance across the molecule–electrode interfaces of the device junction, and 3) to provide fundamental insight that can be fed back into molecular and device design for optimization of properties. Ideal experiments should be able to provide a robust test of such predictions. This theory does not exist yet, and, until recently, there were very few experiments that could be used to guide or test such a theory. The situation is now changing, as increasingly reliable measurements of molecular electronic devices have been carried out in many laboratories. In particular, single-molecule devices are emerging as a powerful high-resolution spectroscopic tool for molecular electronics. Early experiments utilized two-terminal break junctions in which one or a few molecules were suspended across a mechanically adjusted electrode gap. More recent experiments have employed an electrical break junction, together with a gate electrode, to form a three-terminal device (3TD). The gate can be utilized to correlate the molecular energy levels with the Fermi energies of the electrodes, and thus somewhat normalize out the device-todevice fluctuations that are observed in the two-electrode measurements. Herein we utilize single-molecule 3TDs with a Pt breakjunction source and drain electrodes to investigate four molecules—namely, the bistable [2]rotaxane AR and SR and the corresponding dumbbell-shaped compounds AD and SD, from which the mechanically, redox-switchable rotaxanes were obtained (Scheme 1) by a template-directed protocol that clips a tetracationic ring around the tetrathiafulvalene (TTF) units in the dumbbell-shaped compounds. Our goal was to address the nature of the molecule–electrode contact. We had previously found that similar amphiphilic, bistable [2]rotaxanes and bistable [2]catenanes—rendered amphiphilic with dimiristoylphosphatidic acid (DMPA) counterions—could be utilized as solid-state, bistable switches in devices in which a Langmuir–Blodgett (LB) monolayer is sandwiched between a Si bottom electrode—passivated with the native oxide—and a Ti/Al top electrode. In the case of bistable [2]catenanes that were specially designed to interact noncovalently with the side walls of single-walled carbon nanotubes (SWNTs), we have also found that usable switching devices could be fabricated using a semiconducting SWNT bottom electrode. Various control molecules, including the dumbbell precursors of the bistable [2]rotaxanes and nondegenerate [2]catenanes, were also explored, and none of these controls yielded a switch. Thus, we attributed the bistable character of working solid-state devices as arising from hysteretical mechanical motions in these redox-switchable interlocked molecules. 12] Conversely, we have never observed switching signatures attributable to a molecular bistability when we have utilized a metal (Au or Pt) bottom electrode. One possible reason for the difference between Group IV (Si or C) electrode materials and transition metal electrodes is that the increased ionic character of the organic/ metal interface leads to Schottky-like barriers to charge-flow, and that those barriers dominate the device characteristics. We report here that the device characteristics of 3TDs containing bistable [2]rotaxanes and Pt electrodes are extremely sensitive to the chemical nature of the molecule–electrode contact, but are much less sensitive to the details of the molecular structure away from those contacts. This result has strong implications for the design of molecular electronic devices. While the bistable [2]rotaxaneAR can form “chemical” bonds between the terminal, incipient five-membered disulfide ring and a Pt electrode, the tetraarylmethane stopper at the other end ofAR will, at most, become physisorbed onto platinum, thus the molecule–electrode contacts are asymmetric. In the case of the bistable [2]rotaxane SR, both ends [*] Prof. J. R. Heath, Dr. H. Yu, Dr. Y. Luo, K. Beverly Caltech Chemistry, MC 127-72 1200 East California Boulevard, Pasadena, CA 91125 (USA) Fax: (+1)310-206-4038 E-mail: heath@caltech.edu
Read full abstract
Nov 26, 2021
Yanan Wei + 6
Open Access