News & Views | Published:


How clean is too clean?

Naturevolume 440pages3435 (2006) | Download Citation


Silicon nanowires could form the building-blocks of future electronic devices, but under ultra-clean conditions, regulating their growth is difficult. Is the strictly controlled environment the problem?

At the heart of the success of microelectronic gadgets, from laptop computers to mobile phones and iPods, is the fact that silicon-based electronic devices keep getting smaller. In the most advanced of such devices currently in production, the smallest features that regulate the flow of current are less than 100 nanometres in size. But if the trend to miniaturization of these features is to continue — say, to scales below 10 nanometres — a new basic unit for electronics, other than the conventional silicon wafer, is required. Several candidate materials exist: carbon nanotubes, molecular switches and nanoscale silicon wires are examples. In a paper published online today, Hannon et al.1 report in situ observations under an electron microscope of the growth of silicon nanowires in ultra-clean conditionsFootnote 1. Their observations might cast doubt on the suitability of such nanowires for mass production. They might, on the other hand, simply be telling us that the particular experimental conditions under which these nanowires were made were just too clean.

But first, some background. The idea of using nanowires of conventional semi-conductors such as silicon or gallium arsenide as building-blocks of nanoelectronic or nanophotonic devices was introduced into mainstream research at the end of the 1990s (see, for example, ref. 2). These silicon nanowires are generally grown by the vapour–liquid– solid (VLS) method. In this, a tiny liquid droplet of a metal, such as gold, absorbs silicon from a gaseous precursor, such as silane (SiH4) or disilane (Si2H6), with such efficiency that the gold–silicon alloy droplet becomes supersaturated with silicon. This supersaturation causes a single, cylindrical silicon crystal — the nanowire — to nucleate, with the diameter of the nanowire being determined by the size of the initial gold droplet.

The droplet remains at the tip of the nanowire, which grows steadily outwards from it. If the nanowire lies on the surface of a silicon single-crystal wafer, growth occurs ‘epitaxially’, that is, with the same crystal orientation as the underlying silicon. The fundamentals of this technique for the growth of silicon and other semiconductors were understood more than 40 years ago3,4 for creating objects with diameters larger than 100 nanometres. But it was not until 1992 that the first electronic device on the nanoscale, based on gallium arsenide semiconductor nanowires, came to fruition5.

The droplets at the tips of semiconductor nanowires that grow in parallel have generally been considered to be independent. But Hannon et al.1 show that, under the ultra-clean, high-vacuum conditions of their special electron microscope, larger droplets grow at the expense of smaller ones. These smaller droplets then shrink away, preventing any further nanowire growth from them. This effect, known as Ostwald ripening — sometimes jokingly referred to as the capitalistic principle — is named after Wilhelm Ostwald, 1909 chemistry Nobel laureate, who explained the effect as resulting from a decrease in total surface energy that occurs when atoms are transferred by diffusion processes from smaller to larger crystals6. (For a mathematical treatment of the effect, see refs 7, 8.)

Such an energy-minimizing diffusion transfer requires the efficient transport of atoms between neighbouring gold droplets. This cannot occur through gaseous diffusion above the silicon wafer because of the extremely low vapour pressure of gold; equally, the transport of gold atoms through the bulk of the silicon is also negligible. Hannon et al.1 argue convincingly that the mode of transport is surface diffusion, which requires not only a high diffusivity of gold on the silicon surface, but also a high solubility of gold on the surface or in a thin surface layer (Fig. 1). This would fit with what we know about the growth of silicon nanowires by molecular-beam epitaxy. This is a technique based on the VLS method, but in which silicon is supplied not as a gas but as a directed beam of atoms. Molecular-beam epitaxy usually requires an ultra-clean environment, and here the transport of silicon as well as that of gold occurs through diffusion on the silicon surface, not through the gas9.

Figure 1: Gold migration.
Figure 1

A schematic cross-section of two silicon nanowires growing on a silicon substrate in a silicon-containing vapour (of silane, SiH4, and disilane, Si2H6) by means of the vapour–liquid–solid (VLS) method, as used by Hannon et al.1. a, Two liquid-gold droplets are initially of slightly different size, causing a net diffusion flux of gold atoms from the smaller to the larger gold droplet through the Ostwald ripening mechanism along the silicon surface. b, Later, the smaller gold droplet has shrunk away completely, and its silicon nanowire has stopped growing. The implications of this effect for the mass production of silicon nanowires are potentially immense. (Dimensions not to scale; diameter changes exaggerated.)

Ostwald ripening would make the growth of ordered arrays of millions of essentially identical silicon nanowires — a prerequisite for nanoelectronic applications — exceedingly difficult. Even tiny variations in the size of the gold nucleation droplets would lead to unacceptable variations in the length and diameter of the silicon nanowires. There is, however, an escape clause mentioned by the authors that is also in agreement with our own preliminary experimental results. A tiny amount of oxygen — as is present under most technological growth conditions, but not in the ultra-clean, high-vacuum environment used in the authors' experiments — might efficiently block the diffusion path of gold on the silicon surface. This would render the gold droplets independent of each other, as has been assumed for the past 40 years.

In this way, Hannon and colleagues' results could resemble the oxygen-in-silicon story of the 1970s. At that time, manufacturers of integrated circuits were concerned that silicon crystals contained some oxygen atoms from the crystal-growth process, and pushed the manufacturers of silicon wafers to eliminate these remnants. It was subsequently observed that wafers with negligible oxygen content produced less reliable electronic devices. It turns out that oxygen in fact precipitates onto small regions of the wafers, creating traps for detrimental metallic impurities in a process known as gettering10.

Today, the oxygen content of silicon wafers is exactly specified for the best possible gettering performance. Analogously, the indirect conclusion from the fact that Hannon and colleagues1 observe an effect under highly controlled, low-vacuum conditions that is not observed under less severely controlled conditions could be that a little added oxygen impurity — although not too much — is beneficial for silicon nanowire growth. That would be an unexpected and useful message: sometimes extremely clean is just too clean.


  1. 1.

    *This article and the paper1 concerned were published online on 29 January 2006.


  1. 1

    Hannon, J. B., Kodambaka, S., Ross, F. M. & Tromp, R. M. Nature 440, 69–71 (2006).

  2. 2

    Morales, A. M. & Lieber, C. M. Science 279, 208–211 (1998).

  3. 3

    Wagner, R. S. & Ellis, W. C. Appl. Phys. Lett. 4, 89–90 (1964).

  4. 4

    Givargizov, E. I. Highly Anisotropic Crystals (Reidel, Dordrecht, 1987).

  5. 5

    Haraguchi, K. et al. Appl. Phys. Lett. 60, 745–747 (1992).

  6. 6

    Ostwald, W. Z. Phys. Chem. 34, 495–503 (1900).

  7. 7

    Wagner, C. Z. Elektrochem. 65, 581–591 (1961).

  8. 8

    Lifschitz, I. M. & Slyozov, V. V. J. Phys. Chem. Solids 19, 35–50 (1961).

  9. 9

    Schubert, L. et al. Appl. Phys. Lett. 84, 4968–4970 (2004).

  10. 10

    Tan, T. Y. & Tice, W. Phil. Mag. 34, 615–631 (1976).

Download references

Author information


  1. the Max Planck Institute of Microstructure Physics, Weinberg 2a, Halle, D-06120, Germany

    • Ulrich Gösele


  1. Search for Ulrich Gösele in:

About this article

Publication history


Issue Date


Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing