Molecular Electronics

Introduction

You may recall from your high school biology that we employ molecular electronics in our bodies. Every time that a signal is transmitted in our nervous system, an electric pulse travels down the axon, releases a chemical messenger that in turn creates an electric pulse in the next nerve cell, all the way into our brain. You may say that, in fact, we are happy for using molecular electronics! What generates the electric pulse? How is it transmitted through the axon of the nerve cell? How is the signal amplified? All of these questions are relatively well understood by biologists and biophysicists. We know that molecules act as gates and switches, as rectifiers and amplifiers, and the brain itself functions as an electronic network . More over, when light falls on the cones and rods in our retina it interacts with the molecular rhodopsin to create electric pulse, similar to a semiconductor silicone photodetector. This molecular process in the eye follows a circular set of reactions that brings back the rhodopsin molecule back to its initial state ready to create more electric pulses from impinging light.

But the molecular electronics of our biology is very different from the electronics in our today's technology (which here I refer to as "computer electronics"). For one, the computer electronics is based on solid state electronics, while the biological one is fluid based (we are mostly made of water). Also, our computer electronics is exceptionally faster than the biological one. These differences have distinct advantages and shortcoming. So, why should we investigate the possibility of using molecular electronics for our computers?

There are four reasons often sited for this. The first is size. Our today's technology has made it possible to construct 5 nm diameter nanowires separated by about 15 NM, to form a lattice. If we were to create a transistor from these solid state nanowires, we would need to dope them with impurity atoms whose size is comparable to the lattice constant itself. So, a nanowire that is a micron long could receive 10 to 20 dopant atoms, resulting in about one impurity per transistor. This low number would make our nanotransitor to behave nonstatistically. The second reason is fabrication costs. Present methods of creating nanoscale electronics involves the top-down approach. There are often about 30 steps even in the lithography production process. Molecular beam epitaxy and ion beam lithography are far more costly than the lithography techniques that have allowed for mass production of microelectronics. Another limiting feature of traditional means of creating nanometer size circuitry is related to energy consumption per unit and thus the heat transfer issues. Finally, the problems of cross talk and quantum noise becomes important as the circuit size approaches the manometer scale. For these reasons it is desirable to investigate the possibilities of new techniques and approaches, perhaps, based on molecular circuits in the fashion of biological systems.

Molecular electronics clearly has the advantage of size. The components of these circuits are molecules, so the circuit size would inherently range between 1 to 100 NM We know from our current study of molecules that their interactions are specific. As a result, assembly of the components could use this as a "recognition" feature and thus produce more cost effective assemblies. A third feature of molecules that may make them attractive for circuitry is that some have more than one stable configuration. These variations in the isomers, similar to those in the rhodopsin, can be employed for converting different signals (as is done in the vision by converting light energy into electric pulses). Another important feature of such a circuit would be its speed. In terms of charge transport molecular circuits could far out perform solid state devices. Finally, through proper synthesis designers of these circuits could tailor the molecules interaction to affect transport properties, its magnetic or optical characteristics, or how it binds to other components of the circuit.

The first formal proposal for development of molecular electronics was made in 1974 by Mark Ratner and Avi Aviram. They proposed the possible production of a molecular rectifier based on an asymmetric tunneling junction. But it took several more years until in late 1990s Mark Reed and Jim Tour succeeded in demonstrating molecular circuitry in the laboratory. They showed that a monolayer of about 1000 molecules exhibited the phenomenon of negative differential resistance. In the last decade or so this field has received a great deal of attention, but it is still in its infancy and most likely there will be another one or two decades before it can reach commercial viability at the level of today's microelectronics. Below, we will examine some of the features of molecular electronics that have been examine in research laboratories and have lead to promising new applications.

Charge transport

In today's commercial electronics the primary mechanism for transfer of information is charge transport. Electric current that flows in the metallic conductors of the circuit are switched and manipulated by the transistors to create the computer. In the case of optoelectronic devices the charge is converted into light energy, but in these devices also the charge transport plays a major role. In either case the charge transport corresponds to motion of electrons in the conduction band or/and motion of holes in the valence band. In molecular electronics very different mechanisms could lead to charge transport. One of these is coherent tunneling (we have already investigated this in the section dealing with Q-dots). The bond electron in the Q-dot can tunnel out of it and jump onto a nearby junction electrode and visa versa. This process, unlike that in bulk electronics, is non-ohmic. In other molecule type, called a donor-bridge-electron acceptor (DBA) the donor and acceptors are part of the same molecule, but they are separated by a "bridge" part. While thermal electrons can jump from the donor site into the acceptor and lead to (hopping) ohmic charge transfer there are other mechanisms that create a non-ohmic flow. In one of these, a so called superexchange, an electron that tunnels from one nearby electrode to the acceptor site, due to the presence of a bias potential, could coherently transfer to the donor before tunneling to the other electrode. Similarly holes could do the same in the opposite direction. All three of these transports is finally responsible for conduction across a DBA junction. One of the interesting aspects of a DBA is that the molecule itself could be designed with asymmetry so that it requires no external bias voltage for its energy to have an asymmetry. Also, it turns out that the hoping mechanism (donor to acceptor flow) and the superexchange rates are dependent on the molecular size. For example, for a short bridge of 0.5 to 1 NM the superexchange rate dominates the hopping rate. These rates reverse for longer bridges. So, by designing the suitable bridge lengths the transfer rates could be adjusted via synthetic variations.

Some molecules change functionality by changing their shapes. This is called isomerization. By applying external fields isomeric molecules could change functionality, which makes them good candidates for use as switches and other active devices.

Electrode

In the case of bulk electronics the role of electrodes is simply to allow for ohmic transport of electric current. Typical electrodes are metallic"wires" that connect individual components (switches) to each other and to the outside. In the case of molecular electronics electrodes can play a far more significant role. The molecule-electrode interface could significantly alter the charge transport or the molecular functionality altogether. But in most of today's constructs this is excluded in the design in favor of simple pragmatic approach. This is because currently we know little about these interactions.

DNA

Thanks to years of biochemistry investigations we know a great deal about DNA. The protein molecules can self-assemble. These large molecules can also recognize other molecules and self-orient to bond to them. It turns out that many of these functionalities can only be made possible when the molecular size is large. Recently investigators have studied the phenomena of charge transfer in DNA molecules extensively. These studies suggest that the mechanisms for charge transfer in DNA are indeed very varied. But little is understood about electrical transport in DNA junctions. Current speculations predict that DNA-junction can act as a wide band-gap semiconductor. But a great deal of research is most likely necessary to better understand these properties.

Circuit Architecture

Traditional bulk circuits use macroscopic architecture and interconnects. The only difference between these and their older macrocircuits is that in the VLSI designs the architecture can be three dimensional, as opposed to the earlier printed circuit designs. Nevertheless, these architectures are simply standard electrical circuits. In contrast, molecular circuits have to contend with the fact that their circuit elements are simply molecules that can interact with other elements once they are close enough. For this reason and for the reason of self assembly - a very cost effective assembly feature- crossbar and multiplexes are the most likely architecture that would be useful in molecular electronics. In the crossbars individual molecules are sandwiched in between nanowires that crisscross each other. Researches have already constructed crossbar molecular circuits. While the current efforts intend to create patchwork of different cross bars tied together so as to create an efficient computational unit.

The Future

Many of current studies aimed at understanding the behavior of molecules in a single circuit tend to suggest that the variations of the results of different measurements may be a consequence of variation in the junctions. It is difficult to determine that repeated applications of the same measurement create identical "circuits". This is in contrast with the solid-state devices that show little or no variation from device to device. In addition, these single molecular measurements show a temperature dependence of physical effects that make it hard to ascribe clear relations between measurable quantities, say voltage and current. Recently, however, research groups have begun a new approach to these measurements. Instead of experimenting with single molecular devices they are building devices using self-assembled monolayers of the molecules. In one of these studies the research team headed by Mark Reed has determined tunneling to be the primary mechanism for charge transport in a self assembled monolayers of alkanethiols. (Alkanethiols are large HOMOLUMO - Highest Occupied Molecular Orbit/Lowest Occupied Molecular Orbit - gap molecules that have short molecular lengths.) These measurements show far more consistent and repeatable sets of data, than previous ones that studied with one molecule at a time.

There are other avenues that people are currently investigating. One of these, for example, uses low temperature STM manipulation of individual CO molecules on a gold surface. The tunneling of the molecules between different binding configurations are used as bits for computation. In these molecular cascades the relative position of the molecules is the key for computation. In another approach a molecular-based solid state switch is reconfigured using voltages. This device could be used in memory applications. By far all of these approaches require much further work before molecular electronics begins to compete with bulk electronics. In the process, we hope to learn a great deal not only about synthesis, but also about new science.

Last Modified June 28, 2005 malekis@union.edu