The Hybrid Solution For Molecular Electronics

The availability of molecular devices able to perform logical operations'71 does not guarantee per se the possibility of molecular electronics, unless one is able to arrange them in an ordered and accessible way and to probe their states. The fundamental limits posed by quantum mechanics suggest that the connection of the microscopic world of molecules to the macroscopic world of humans or machines will require a stage of amplification of the signal at the mesoscopic level. Because single devices of microelectronic ICs are mesoscopic in nature and may be operated to behave as measurement apparatuses, the most promising architecture is the hybrid one, in which a very dense array of molecular devices is hosted in a silicon-based microelectronic IC, which provides the functions for addressing, reading, and writing, in addition to the input-output power stage.

This point deserves a particular discussion. According to the idea of Bohr,'11] eventually formalized in the theory of Daneri, Loinger, and Prosperi'12] the detection of the microscopic state of a system requires the interaction of the microscopic system with a macroscopic apparatus. To act as a measurement device, the composite (microscopic + macroscopic) system must have so many states as the number of microscopic states and must have ergodic properties such as to reproduce the probability distribution of the microscopic state alone.

The analysis of ideal experiments (such as the detection of a particle in a Wilson chamber or in a Geiger counter) shows indeed that typical situations of nuclear physics, in which the microscopic state is characterized by a large energy (compared with the thermal unit kBT, with kB the Boltzmann constant and T the apparatus temperature) concentrated on a few degrees of freedom, are actually described by the above picture.'131 That certain ICs may operate as measuring devices is demonstrated by the detection of single photons through p-i-n diodes supplied at a voltage just above the breakdown[14] and of alpha particles through the soft errors of dynamic random access memories.[15] The first example is particularly interesting because it applies to a situation in which the event to be detected (the presence of a photon) has an energy (say, 1 eV) not exceedingly higher than the thermal unit (say, 25meV). This occurrence thus suggests the possibility of using suitable microelectronic devices for the detection of the state (e.g., the oxidation number) of microscopic systems (e.g., molecules) hosted on the IC.

Actually, scanned probe techniques [such as the scanning tunneling microscope (STM)] have already been proven to have the necessary sensitivity to image single electrons in nanostructures[16] (although "they almost invariably alter the properties of the system they are measuring''[17]). Even assuming the ability to reduce the STM size in the micrometer length scale (to be compatible with the IC circuitry), its use for the detection and modification of the electronic state of an ordered array of molecules (considered as a molecular memory) will require dramatic changes in circuit architecture. In present ICs, all bits are matrix-ordered and are singularly accessible by specifying row and column; the use of STM as sensing-writing element suggests the need of organizing the memory as a shift register.

The STM is not the unique apparatus able to detect the electronic state of single molecules; in recent years, the single electron transistor has been proven to work. Such a transistor, based on the concept of Coulomb blockade[18,19] has the advantage over STM to be producible with the methods of the IC technology and not to require dramatic changes in circuit architecture.

This state of affairs suggests that molecular electronics will necessarily involve a combination of microtechnologies and nanotechnologies[20] and identifies in self-assembly (i.e., the deposition of functional molecules in wanted regions of the microelectronic device) and in grafting (i.e., the selective bonding of certain groups of the molecules to electrodes in the microelectronic regions) the key problems of molecular electronics. Actually, the few examples of logic devices prepared exploiting the electronic properties of molecules are hybrid devices with photolithographically defined electrodes.[21-23]

Determining the conductance of individual molecules is a difficult task. The case of DNA is emblematic of this difficulty—depending on the study, DNA is indeed said to be an insulator, semiconductor, conductor, or proximity-induced superconductor.[24] The usual approach to allow electrical testing of organic molecules is to terminate them with thiol (-SH) groups, which behave as alligator clips to gold. Gold electrodes with parallel-plate geometry are now currently prepared down to separations of 4 nm. While useful for the determination of the electrical characteristics of the molecules (allowing the assessment of their conduction mechanism), this arrangement is, however, largely incompatible with silicon-based ICs, for which gold should be avoided.

Although it may seem difficult to insert functional molecules between two electrodes with a controlled separation in the nanometer length scale (say, 3nm), a solution to this problem, fully compatible with the IC technology, has recently been proposed.[25] The MOS transistor provides indeed a parallel-plane geometry with silicon plates separated by just the gate-oxide thickness—3 nm in the current technology. To allow the insertion of functional molecules therein, it is sufficient to define the gate geometry with an etch sequence formed by a plasma etch of the unmasked polysilicon followed by a wet (aqueous HF) etch of the underlying SiO2. This process, initiated with a "vertical'' attack, does not complete with the total dissolution of the unmasked SiO2 but rather proceeds laterally overetching a portion of the SiO2 covered by the polysilicon. The width of the overetched region (the "recessed region'') is controlled by the exposure of the structure to the etching solution. Of course, this is only one of the several possibilities offered by microelectronics, and it is expected that the region hosting the molecular electronics will be obtained exploiting features (such as overetches) hitherto ignored (or even demolished) in IC processing.

Accepting this view, the development of any nano-technology for data processing does necessarily require the ability to control a number of IC surfaces to allow self-assembly, while still providing the functions of power supply, addressing, sensing, etc. To a large extent, the materials of microelectronics are limited to Si, SiO2, Si3N4, Al, and Ti; of them, only silicon is single crystalline and is thus the most natural candidate as the seed for ordered organic layers.

Solar Power

Solar Power

Start Saving On Your Electricity Bills Using The Power of the Sun And Other Natural Resources!

Get My Free Ebook

Post a comment