Read full article onlineFull Article
A switch is an ensemble of pieces whose function is to stop and/or establish the electrical current in a circuit. This 19th century definition of a switch continues to be valid in our days. But because of the requirement of better communication means during the 20th century, the pieces of our modern switches are no more mechanical or electromechanical in nature. In our information society era which uses plenty of communication and computation machines, the many switches constituting those machines are now made of solid-state semiconducting materials. After the relay and the vacuum tube, the microfabricated transistor is the modern reference of an integrated switch at the surface of a silicon wafer. Furthermore, not only electrical current, but also photons flux, magnetic flux, etc. have now their own switches, expanding the possibility of miniaturizing a switch to other types of material, such as dielectrics and ferromagnets.
A human operator was supposed to operate a switch in the 19th century. The increasing complexity of circuitry made this operator to disappear to the benefit of a third electrode per switch. In electronics, a switch is now a three-terminal device. The “on/off” triggering information is generally provided to a given switch by another one inside the circuit, without involving the operator. Taking the transistor as reference for switch performances, there is now a continuous demand for a further reduction of the weight, the size, the power consumption, and the switching time of the solid-state switches.
One solution is to reduce the amount of material required to fabricate a transistor by improving the lithographic technique, engraving a transistor on an always smaller and smaller portion of the surface of a semiconductor wafer. A radically new approach was proposed by Aviram and Ratner in 1974. Instead of continuing the miniaturization of the transistor at the surface of a wafer by inventing new lithography techniques, why not go all the way down to the atomic scale. Here the next big step is to create switches whose pieces are simple organic molecules and even a few atoms bound together to form a single molecule. Those switches are called molecular switches. In principle, they can control an electron flux or, for example, a photon flux in a photonic circuit.
What remains unchanged is the kernel of the definition of a switch: whatever the amount of material assembled or bonded to fabricate a switch (a condensed phase, a macromolecule, a few molecules, a single molecule), a switch must have two states (the “on” and “off” states) well separated to ensure the stability of both the “on” and the “off” states. Behind the scene, there is often an internal physical or chemical variable to describe the state of the material assembled to constitute the switch. This is independent of the physical or chemical effect triggering the switch and whatever the flux to be controlled. Along a reaction coordinate proper to each kind of switch, a double-well energy curve can often be plotted defining the “on,” the “off” states of the switch, and the transition region. But in simple monostable molecular switches, only one well is existing. The well is simply deformed or displaced by the triggering effect. The double-well case gives rise to bistable molecular switches: in a few examples, the same value of the triggering parameter can maintain both the “on” and the “off” state depending on the history of the switching event, i.e., there is a memory effect (note that conventional switches for domestic applications are usually bistable, but electronic switches such as transistors are only monostable). The memory corresponds to a hysteresis effect related to a cooperative phenomenon often obtained with molecular material molecular switches.
In practice, mainly molecular electronic switches are explored. The reason is that the exchange of information between a given molecular switch and other molecular switches in the same circuit requires communication and interconnections means adapted to the size of ultimately a single molecule. Electrons have a much more practical wavelength than photons for this purpose. Furthermore, a tunnel junction is a very useful source of tunneling electrons where a molecule can be positioned on, like a jumper. Of course, such positioning is still very delicate to perform experimentally, taking into account the necessity to know exactly at the atomic scale the conformation and the exact position of the molecule in the junction. The invention of the scanning tunnel microscope (STM) in 1981 arrived exactly on time to show experimentally as early as in 1988 that such electrical interconnection between a single molecule and a macroscopic operator was possible.
Many experiments on molecular switches are still performed by averaging the answer of the molecules over many and even billions of the same type. This increases the signal-to-noise ratio and amplifies the signal to be measured. In those experiments, each molecule is supposed to be noninteracting with the others, so that the switching ability depends only on one class of molecules. Using many instead of one is not only a matter of convenience. This approach can speed up the exploration of new molecular switches, by separating two otherwise intricated problems: obtaining a switching effect and positioning one molecule on a surface.
In this article, we shall consider first “chemical switches,” i.e., systems based on a chemical reaction, or on a cooperative process involving many molecules. Thus although the change in properties between the “on” and “off” states can be spectacular, it is not, in principle, possible to reach the truly monomolecular scale. Then we shall see the case of monomolecular switches, this latter goal being now strongly boosted by technological and fundamental purposes. There will be two variants: first “intrinsic” molecular switches, i.e., molecules which present a special sensitivity to an external perturbation, with a bistable character, and can be studied in solution, in particular for screening purposes; the second variant will be “surface bound” molecular switches, really interconnected to an electrical circuit, which implies a strong interaction with the surface, so that the switching effect comes actually from the molecule–surface ensemble.