Since its invention in 1948, the transistor has revolutionized everyday life. The electronics revolution is based on miniaturization of transistors; smaller transistors are faster, and denser circuitry has more functionality. Transistors in the present generation of integrated-circuit chips have sizes of ≈0.18 µm, and the electronics industry has completed development of 0.13-µm transistors, which will enter production within the next few years. Industry researchers are now working to reduce transistor sizes below 0.1 µm — a thousandth of the width of a human hair. However, studies indicate that the miniaturization of silicon transistors will soon reach its limit.

Figure 1. Transistors Have Been Miniaturized, starting from the macroscopicscale, where thousands of dopant atoms can be regarded as being in a smooth distribution that determines the switch-on voltage. At the mesoscopic scale,transistors have different switch-on voltages, depending on precisely where the dopant atoms are located. To prevent such unit-to-unit variations, the next step in miniaturization will not be accomplished by making conventional transistors smaller: instead, transistors with dimensions <1 nm will be made from chains of precisely positioned atoms.

For further progress in microelectronics, it is necessary to turn to nanotechnology. Rather than continuing to miniaturize transistors to a point where they become unreliable, nanotechnology offers the new approach of building devices on the atomic scale. One vision for the next generation of miniature electronic circuitry is that of atomic chain electronics; according to this vision, each device is composed of atoms aligned on top of a substrate surface in a regular pattern. The Atomic Chain Electronics Project (ACEP) — part of a nanotechnology group at Ames Research Center — has been developing the theory for understanding atomic chain devices, and a patent for atomic-chain electronics has been filed and is now pending.

Figure 2. A Semiconducting Chain of Germanium Atoms electrically insulated from other atoms present is formed on top of a (100) silicon substrate, the unused regions of which are hydrogenated.

The use of dopants is critical to the functionality of transistors. Dopants are impurities intentionally added to the semiconducting transistor channel to raise or lower the device switch-on voltage. Typically, a macroscopic transistor (one with a size of the order of 10–6 m) contains several thousand dopant atoms that, on the scale of the overall device, appear to be smeared out as a dopant “jelly,” as illustrated in the top part of Figure 1. Because of the large number of dopant atoms, the precise location of each dopant atom is not very important to the functioning of the transistor. However, when the size of a transistor is reduced to the range of 10–8 to 10–7 m, as illustrated in the middle part of Figure 1, the number of dopant atoms is less than about 100, in which case the position of each dopant atom does matter. Current manufacturing techniques do not provide the means to control the locations of the few dopant atoms precisely, and as a result, small variations in functioning occur among transistors. Such variations are fatal when millions or billions of transistors are integrated in a computer chip, because the variations can cascade from one device to the next, eventually giving rise to a malfunction. A solution of this aspect of the miniaturization problem, devised by the ACEP, is to create all the device structures with atomic chains laid out in a regular precise pattern by anchoring atoms to a substrate, as illustrated in the bottom part of Figure 1.

A second critical aspect of the functioning of a transistor is gain: the output of the transistor should be a magnified version of its input. ACEP research has led to the conclusion that in designing atomic-chain transistors to produce gain, one should exploit the field effect, which is the only mechanism verified experimentally so far at the atomic scale. Semiconductors conduct current only when a gate voltage is (is not) applied — field effect. Metals always conduct current. A field-effect transistor uses a semiconductor for the channel and a metal for electrodes, and this is how it achieves large gain. Thus, it is necessary to devise chains with semiconductor and metal properties. First, the ACEP team tackled a simple problem theoretically: If one can arrange silicon atoms along a line floating in air, is this chain semiconducting? The answer is surprising: although bulk or thin-film silicon is semiconducting, an isolated chain of silicon atoms is always metallic. Fortunately, ACEP research also leads to the finding that a magnesium chain is semiconducting, even though bulk magnesium is metallic.

Of course, it is not possible to float atoms in air: it is necessary to place them on top of a substrate. Surface atoms of the substrate attract atoms in the chain and hold them at fixed positions. The surface of a substrate can be made to have atomic-scale corrugations, and these can be used to create a precise pattern for atomic-chain electronics. However, the attractive force between the substrate and chain atoms that are simply placed in the corrugations is too weak to secure atoms reliably. ACEP research has led to the conclusion that a chemical bonding scheme is needed to secure chain atoms reliably, and to the further conclusion that the properties of a chain are strongly influenced by the substrate material and surface orientation when chemical bonding is used. Indeed, the same atomic chain can be either metallic or semiconducting, depending on the number of chemical bonds between a chain atom and the substrate atoms.

A further concern is that electrons traveling along an atomic chain can detour into the substrate through chemical bonds and possibly leave through a neighboring chain, resulting in the short-circuiting of two atomic chains. Such short circuits are fatal for electronic applications and must be avoided. ACEP research has clarified the conditions under which the substrate surface can be made electrically insulating in a chemical-bonding scheme. Among the findings of this research are that silicon and germanium crystals can be good insulating substrates.

Thus far, the ACEP team has made the first step toward the development of atomic-chain electronic devices. On top of a silicon substrate, it is possible to attach germanium atoms with two chemical bonds each, as shown in Figure 2. The resulting structure is a semiconducting chain on an insulating substrate. Nearby chemical bonds on the substrate are saturated with hydrogen atoms, so that they are electronically inactive. Such a chain can be a component of a field-effect transistor on an atomic scale.

This work was done by Toshishige Yamada of Computer Sciences Corp. for Ames Research Center. For further information, access the Technical Support Package (TSP) free on-line at www.nasatech.com/tsp  under the Electronic Components and Systems category.

This invention is owned by NASA, and a patent application has been filed. Inquiries concerning nonexclusive or exclusive license for its commercial development should be addressed to

the Patent Counsel
Ames Research Center
(650) 604- 5104.

Refer to ARC-14246.