Issue



Nanotechnology: Microelectronics' nanotechnology future


01/01/2000







Pieter "Pete" Burggraaf, Senior Technical Editor

Special Section - Technology Outlook

Increasingly, various forms of "nanotechnology" are being envisioned for a broad range of new frontiers in high technology disciplines and manufacturing, including microelectronics.

Nanotechnology was created by the IC revolution and its miniaturization requirements. Perhaps it is appropriate to trace its origins back to Binning and Rhorer's Nobel-Prize-winning work on the development of scanning tunneling microscopy (STM) in 1986 [1] and their subsequent observations of the atomic structure of surfaces.

Also of significance in early work, Don Eigler and associates at IBM were able to use scanning tunneling microscopy to laboriously manipulate single atoms into 3-D atomic structures built to human specifications for the first time [2].

Other important early work included fabrication of films, clusters, spheres, and tubes with atomic precision using molecular beam epitaxy [3], molecular self-assembly [4, 5], and recombinant DNA technology [6].

As we round the corner into the 21st century, nanotechnology, more specifically nanoelectronics, is poised to fill some of the predicted limitations in IC technology outlined in the 1999 International Technology Roadmap for Semiconductors.

Yet, there is still no consensus on a revolutionary, relatively inexpensive route for producing the required smaller IC feature sizes. An economic route to feature sizes beyond 100nm has been a primary motivation for research on the nanometer frontier [7].

Potential nanotechnology routes specifically for IC fabrication include:

  • Greater application of molecular beam and organometallic vapor-phase epitaxial deposition methods. These are capable of arbitrary thicknesses and composition with an accuracy of one atomic layer along the direction of growth, enabling the development of high-performance devices such as quantum-well lasers, high-electron mobility transistors, and resonant tunneling diodes [7].

  • The extraordinary potential for chemical synthesis or the spontaneous self-assembly of molecular clusters from simple reagents in solution for the production of 3-D nanostructures or quantum dots of arbitrary diameter. Molecular self-assembly works well to define nanometer-scale structures reproducibly and spontaneously at room temperature and to reject defects according to size, because fabrication is thermodynamically driven (see "Molecular-scale device progress" on p. 63). Efforts to control the process are progressing with the application of atom optics to manipulate, deflect, and focus atoms using forces that develop from spatial undulations in a light field [7].

  • The emergence of tiny, single-electron transistors where quantum effects of electricity predominate classical charge transport. Here, quantum mechanical tunneling through a classically forbidden energy barrier is regulated by the quantized charge of a single electron (see "Single-electron transistor research" on p. 66).

As circuit integration progresses from today's ICs with 11 million transistors to ICs with 1-2 billion transistors by the year 2010, it is likely that ICs will eventually have functional throughput equivalent to the human brain. One industry nanotechnology tracker recently told Solid State Technology, "While the human brain has 1015 synapses operating effectively with a 1kHz clock frequency at a 0.1% activity level, a silicon multiprocessing unit (MPU) in the year 2011, fabricated using 50nm line rules, will incorporate 1.4 billion transistors operating with a 10GHz clock frequency at a 1% activity level. The functional throughput of an MPU, defined as the product of the number of gates, the clock frequency, and the activity, is therefore expected to exceed that of the human brain, and yet cost <$1000 to manufacture, if current cost projections are accurate. These projections are not fanciful; they are grounded in thousands of man-years of physical science research and development, and are based simply on a practical recasting of the MOSFET transistor as a nanotransistor with nanometer-scale dimensions."

Molecular-scale device progress*

While molecular-based nanoelectronics remains very much a laboratory project, significant progress is being made. A combination of advanced synthetic chemistry, novel nanofabrication techniques, and a better understanding of through-bond electronic transport has led to the first molecular-scale device demonstrations and the first electrical measurements of these systems a quarter century after they were first proposed in 1974. Researchers have, for example, fabricated simple diode structures arising from different contact potentials on a molecule, made the first measurements of through-bond (Schottky barriers, effectively) conductance, and begun early investigation of switching behavior in these systems. But significant challenges remain. For example, understanding contacts to molecular-scale systems is critical to future device development and exploitation of various metal-molecule contact systems is currently needed. The present challenge in these systems is to develop an ohmic contact technology so that long-range electron transfer can be explored.

*Source: M.A. Reed, "Prospects for molecular-scale devices," presented at IEDM 1999.

Single-electron transistor research

In the realm of nanoscale ICs, conventional semiconductor device technologies are increasingly inadequate because so few electrons are involved in device action (i.e., voltage and power consumption) that statistical fluctuations in electron numbers create insurmountable problems (e.g., too much noise, etc.). Accordingly, much of the current work on silicon-based nanoscale devices revolves around so-called single-electron transistors (SETs) that operate with just one electron and can be the basis of both memory and logic circuits.

A review of related presentations at the annual International Electron Devices Meeting (IEDM) shows various methods of confining single electrons, including use of oxide-nitride tunneling dielectrics, germanium nanocrystals, and silicon triton island structures. Nanocrystal memories, for example, are CMOS-compatible, floating gate memory structures, where the floating gate region is a scaled medium that stores discrete and small numbers of electrons.

Reports at last month's 1999 IEDM show that significant progress is being made. For example, work at the University of Cambridge has produced SETs fabricated by methods compatible with MOS processing in SOI (SOISETs). These SETs are being investigated for memory cell applications. Briefly described, the SOISET is a highly doped (1019-1020 P/cm3) silicon nanowire þ50nm wide, 30nm thick, and 0.5-1µm long. Ends of the wire hold the source and drain and the length holds closely spaced gates. Under controlled conditions, single-electron transport in the nanowire is through a series of islands and tunneling barriers (i.e., "multiple tunnel junctions"). So far, a memory cell version of this SET has shown 10ns response.

In other work related to logic applications of SETs reported at the 1999 IEDM, engineers at NTT in Japan have fabricated a complementary single-electron inverter, which operates similarly to CMOS-MOSFET inverters in an extremely small area (i.e., two SETs in 200nm by 100nm) on an SOI substrate. They have demonstrated gains "larger than unity at 27 Kelvin."

As can be imagined, development of SETs brings with it a whole new series of issues. For example, coulomb oscillations must be measured. In addition, researchers now have to account for quantum mechanical threshold voltage shift, an effect that was first observed by engineers at the University of Tokyo (as reported at 1999 IEDM). Quantum mechanical effects will play an important role in the transport properties of future nanoelectronic devices.

References

  1. G. Binning, H. Rhorer, "Nobel Prize 1986," Sci. Am., 253, 50, 1985.
  2. D.M. Eigler et al., Nature, 344, 524, 1990.
  3. H. Sakaki, Nanotechnology, Springer Verlag, pp. 207-256, 1999.
  4. L. Brus, Nanotechnology, Springer Verlag, pp. 257-284, 1999.
  5. M.S. Dresselhaus et al., Nanotechnology, Springer Verlag, pp. 285-330, 1999.
  6. T.J. Deming et al., Nanotechnology, Springer Verlag, pp. 371-402, 1999.
  7. G. Timp, Nanotechnology, Springer Verlag, p. 2, 1999.