ABSTRACT
Current interest in molecular electronics [1,2] and spintronics [3,4] is largely driven by expectations that these emerging technologies can be used to overcome bottlenecks of standard silicon technology, most importantly, scaling in terms of speed and power dissipation. In particular, it has been frequently envisaged in the past that molecules can be used as nanoelectronics components able to complement/ replace standard silicon CMOS technology [1,2] on the way down to 10 nm circuit components. The first speculations about molecular electronic devices (diodes and rectifiers) were apparently made in the mid-1970s [5]. That original suggestion of a molecular rectifier has generated a large interest in the field and a flurry of suggestions of various molecular electronics components, especially coupled with premature
<
estimates that silicon-based technology cannot scale to below 1 µm feature size. The Aviram-Ratner Donor-insulator-Acceptor construct TTF
––
s
—TCNQ (D
––
s
—
A
, see details below), where carriers were supposed to tunnel asymmetrically in two directions through an insulating saturated molecular ‘‘bridge,” has never materialized, in spite of extensive experimental effort over a few decades [6]. The end result in some cases appears to be a slightly electrically anisotropic ‘‘insulator,” rather than a diode, unsuitable as a replacement for silicon devices. This comes about because in order to assemble a reasonable quality monolayer of these molecules in a Langmuir-Blodgett trough (avoiding defects that will short the device after electrode deposition), one needs to attach a long ‘‘tail” molecule C18 [ (CH
)18] that can produce enough of van der Waals force to keep molecules together, but C18 is a wide-band insulator with a bandgap eV. The outcome of these studies may have been anticipated, but if one were able to assemble the Aviram-Ratner molecules without the tail, they could not rectify anyway. Indeed, a recent ab-initio study [7] of D
s
A
prospective molecule showed no appreciable asymmetry of its I-V curve. The molecule was envisaged by Ellenbogen and Love [8] as a 4-phenyl ring Tour wire with a dimethylene-insulating bridge in the middle directly connected to Au electrodes via thiol groups. Donor-acceptor asymmetry was produced by side and moieties, which is a frequent motif in molecular devices using the Tour wires. The reason for poor rectification is simple: the bridge is too short; it is a transparent piece of one-dimensional insulator, whereas the applied field is three dimensional and it cannot be screened efficiently with an appreciable voltage drop on the insulating group in this geometry. Although there is only 0.7 eV energy separation between levels on the D and A groups, one needs about a 4 eV bias to align them and get a relatively small current because total resonant transparency is practically impossible to achieve. Remember that the model calculation implied an ideal coupling to electrodes, which is impossible in reality and which is known to dramatically change the current through the molecule (see below). We shall discuss below some possible alternatives to this approach.
