Solid-state physics

Doping the undopable

Article metrics

Impurities that increase the number of electron carriers are essential in most bulk semiconductors. Introducing such foreign atoms into semiconductor nanocrystals is fiddly, and requires exact knowledge of the material's surface.

Almost a hundred years after the construction of the first ‘bulk’(macroscopic) semiconductor device, Erwin et al. (Doping semiconductor nanocrystals)1 present a mechanism to control the inclusion of transition-metal impurities in semiconductor nanocrystals — impurity inclusion is the process known as doping. This advance could allow the electronic and optical properties of nanocrystals to be engineered for applications ranging from solar cells to electronic devices that function using electron spin, rather than electric charge. An impurity introduced through doping could, for example, be used to inject a localized spin into one nanocrystal in an array, its interaction with other spin carriers forming the basis of a ‘spintronic’ device.

The physical effect that underlies the work of Erwin et al. is known as quantum confinement — the quantization, or splitting, at the nanoscale, of the continuum of electronic energy states present in a bulk crystal, such that the energy levels of semiconductor nanocrystals resemble those of giant molecules. This effect was discovered more than 20 years ago2,3, almost simultaneously by groups in the United States and Russia working respectively on lead sulphide and cadmium sulphide. These materials — compounds of elements from groups II and VI of the periodic table — are very similar to the crystals of galena, a naturally occurring form of lead sulphide, that Ferdinand Braun used in 1907 to build the first solid-state rectifier, ushering in the era of bulk semiconductor devices.

Bulk semiconductors are ubiquitous in device applications, because their properties may change when the number of active electrons (those that are free to move within the material and contribute to conduction) is modified — for example by doping with external impurities. As a result of quantum confinement, semiconductor nanocrystals not only possess markedly different optical properties from those of the bulk material, but they can also become extremely sensitive to doping. Exploiting this sensitivity can allow their physical and chemical properties to be controlled with atomic-scale precision, and can result in materials tailored to possess specific properties. Producing new materials ‘atom by atom’ is a revolution anticipated by Richard Feynman almost 50 years ago4 that is still in the making and represents a highly active field of interdisciplinary research.

Despite decades of experience in doping bulk semiconductors — to build transistors, for example — doping nanocrystals has proved difficult. One explanation for this is the possible existence of intrinsic self-purification processes that could hamper the introduction of defects at the nanoscale. Also, depending on the preparation conditions, II–VI nanocrystals doped with transition metals may suffer from low crystallinity — that is, irregularities in their lattice structure. Nevertheless, there has been significant progress in recent years in doping II–VI nanocrystal solids and free-standing clusters5,6.

Erwin et al.1 suggest that some of the difficulties encountered in nanodoping are due to the fact that the mechanisms of impurity incorporation in bulk materials and at the nanoscale are profoundly different. At the macroscopic scale, thermodynamics provides the fundamental constraint on the amount of one solid that may be incorporated into another, yet the degree of doping achieved so far at the nanoscale is much lower than the thermodynamic limit. Thus, thermodynamic considerations seem to be irrelevant to impurity incorporation at the nanoscale. Rather, say Erwin et al., it is kinetics that plays a key role — in particular, surface kinetics.

According to their model, an impurity present when the nanocrystal is synthesized can find its way in only if it can bind to the nanocrystal surface for a comparable time to that required for the crystal to grow in solution. The ability to dope and so modify a nanocrystal does not therefore stem from the equilibrium thermal diffusion of the ‘guest’ atom, as in a bulk solid, but rather from the binding energy of the guest atom to specific surface facets. In turn, the strength of this binding depends on the morphology of the nanocrystal surface and on the surfactants — molecules that are present in the chemical solution in which the nanocrystal is synthesized and which may bind to or interact with the nanocrystal surface.

Confirmation that, at least in the case of II–VI nanocrystals, the surface binding energy is indeed the protagonist in the incorporation of impurities comes from a specific experiment1 that nicely shows the progress made in the field of nanoscale manipulation. Using an appropriate core seed, Erwin et al.1 grew a cadmium selenide (CdSe) shell with the desired lattice structure (Fig. 1) — a cubic lattice with a zinc blende structure, rather than the hexagonal lattice of the more usually adopted wurtzite structure. This CdSe shell had the surface morphology to which, according to calculations, an impurity of the transition metal manganese would best stick. In this way, the authors managed to use manganese to dope a previously undopable CdSe nanocrystal.

Figure 1: Nanoscale doping.
figure1

Semiconductor nanocrystals such as those investigated by Erwin et al.1 can be engineered at the microscopic scale by the incorporation of impurities (doping). The main image is a ball-and-stick representation of cadmium selenide nanoparticles immersed in solution; the inset shows details of the surface structure interacting with model surfactants and with an impurity (purple sphere). Erwin et al. show that the incorporation of the impurity in a nanocrystal during growth is possible, but depends on the strength of its binding with surface atoms.

Binding energies between the nanocrystal and surfactants have also been found7 to play a key role in determining the shape of CdSe nanostructures, in particular whether they are rods or spheres. Defining the relationship between the microscopic structure and composition of a semiconductor nanocrystal and its function requires complex analysis. For this, ab initio simulations such as those on which Erwin and colleagues' experiment was based can prove most useful.

Surface morphology, structure and kinetics — identified by Erwin et al. as crucial to the doping of nanocrystals — are dominant in many other nanoscale phenomena. Examples are phase transformations8, the optical absorption and emission of group IV nanostructures, and the field of nanomechanics. This highlights some of the challenges of nanoscience research, where ‘every atom counts’. At the nanoscale, details of the atomic structure (such as the surface structure) are often important, there are no known scalable models, and one must resort to the basic equations of quantum mechanics to investigate nanostructures. In addition, many of the processes occurring at the nanoscale are not in thermodynamic equilibrium, and thus simple thermodynamic considerations do not apply. Now we have a demonstration that, at least in some cases, these challenging problems are tractable.

References

  1. 1

    Erwin, S. C. et al. Nature 436, 91–94 (2005).

  2. 2

    Rossetti, R. & Brus, L. J. Phys. Chem. 86, 172–177 (1982).

  3. 3

    Efros, Al. L. Fiz. Tekh. Poluprovodn. 16, 1209–1214 (1982); Sov. Phys. Semicond. 16, 772–775 (1982).

  4. 4

    Feynman, R. http://www.zyvex.com/nanotech/feynman.html

  5. 5

    Yu, D., Philippe, W. & Guyot-Sionnest, P. Science 300, 1277–1280 (2003).

  6. 6

    Raola, O. E. & Strouse, G. F. Nano Lett. 2, 1443–1447 (2002).

  7. 7

    Puzder, A. et al. Nano Lett. 4, 2361–2365 (2004).

  8. 8

    Huang, F. & Banfield, J. F. J. Am. Chem. Soc. 127, 4523–4529 (2005).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Galli, G. Doping the undopable. Nature 436, 32–33 (2005) doi:10.1038/436032a

Download citation

Further reading

Comments

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.