PerspectivePhysics

Surface Transfer Doping of Semiconductors

See allHide authors and affiliations

Science  25 Aug 2006:
Vol. 313, Issue 5790, pp. 1057-1058
DOI: 10.1126/science.1127589

“Doping” of semiconductors—that is, the local manipulation of their conductivity—is a key technology for electronic devices. Without doping, for example, a gallium nitride sample larger than the White House would be needed to host a single mobile charge at room temperature; for diamond, not even the volume of the globe would be sufficient. It is only through doping that semiconductors become useful electronic materials. Recent studies have revealed an unconventional way to achieve doping through surface engineering.

Doping of semiconductors is usually achieved by incorporating atoms of appropriate elements into the host lattice of the semiconductor. The dopants either release an excess electron as a free negative charge carrier to the semiconductor (n-type doping) or they consume one more electron for chemical bonding than they brought with them (p-type doping). In the latter case, the “stolen” electron behaves like a positive charge carrier—a hole—in the semiconductor.

The electrons or holes remain weakly bound to the dopants that carry their respective counter charge, and it takes a characteristic activation energy δ to release them as free charge carriers. For effective dopants, this energy is rather small. At room temperature, it is easily supplied by the vibrations of the semiconductor atoms, and one ends up with free charge carriers of one sign and fixed ionized dopants with opposite charge. This situation is sketched for p-type doping in the figure (left panel).

If the dopants are distributed homogeneously in the semiconductor lattice, then so will be the mobile charge carriers. Positive and negative charge cancel, and on average, no electric field acts on the mobile charge carriers. For electronic devices to function, the doping must therefore be inhomogeneous. Inhomogeneous doping results in local variations of electron and hole concentrations, which tend to equalize by diffusion. The result is a delicate equilibrium between charge separation and electric field that determines the electrical response of a device to externally applied voltages. The simplest such device is the p-n junction, which consists of a ptype and an n-type doped layer of the same semiconductor.

In all classical devices, the dopants are impurity atoms introduced into the bulk of the semiconductor. But doping can also be achieved by an electron exchange between a semiconductor and dopants situated at its surface. The surface dopants—below, we will use acceptors for illustration—possess unoccupied molecular orbitals for electrons (UMOs). If the energetically lowest of these orbitals (called LUMO) is close to the valence band maximum of the semiconductor, it will steal an electron from the semiconductor, just as classical acceptors do (see the figure, right panel). As a result, holes will form in the semiconductor, and negative charge will be localized on the surface acceptors (1). This charge separation will automatically establish an electrostatic potential that confines the holes in a perpendicular direction but leaves them free to move parallel to the surface.

This kind of p-type surface transfer doping has recently been demonstrated for fullerene (2) and fluorofullerene molecules (3) serving as surface acceptors on hydrogen-terminated diamond. The hydrogen termination leads to an exceptionally low ionization energy for the diamond; the fullerenes were chosen for their high electron affinities. For C60F48 (4), the activation energy Δ is even negative, and each molecule brought onto the diamond surface creates a hole (1).

The electronic states at the surface are not necessarily associated with molecular adsorbates. In fact, the first observation of p-type surface transfer doping of diamond involved a complex electrochemical system, in which hydrated ions acted as surface acceptors (5). These ions were usually created unintentionally when hydrogen-terminated diamond was exposed to the atmosphere (6, 7). In this electrochemical variant of surface transfer doping, the redox potential of the hydrated ions effectively determines the effective acceptor level of the electronic system (8).

Beyond conventional doping.

This band diagram illustrates classical ptype doping (left) and p-type surface transfer doping (right), using the energy of an electron in free space as a reference (vacuum level). Ec and Ev are the energies of the conduction band minimum and the valence band maximum, respectively.

The balance between electrons localized in acceptor states and free holes in the valence band is expressed by the constant Fermi energy EF. The closer EF is to Ev, the higher the local density of holes. LUMO and HOMO are the lowest unoccupied and highest occupied molecular orbitals of the surface acceptors, respectively. There are several reasons why diamond is particularly susceptible to p-type transfer doping. First, its electron affinity can be tailored to the lowest value of all semiconductors by simple hydrogen termination of the surface bonds. Second, because no solid oxide is present on its surface, intimate contact with surface dopants is possible. Finally, the bulk conductivity is low and will not mask the effect of the transfer doping.

Under similarly favorable conditions, ptype surface transfer doping was very recently observed for silicon (9). In these experiments, the sheet conductivity of very thin silicon layers (10 to 40 nm) on top of a SiO substrate was measured. After appropriate preparation, the Si surface atoms rearrange and form rows of asymmetric dimers. With this reconstruction, unoccupied surface states close to the valence band maximum of silicon are formed; these states play the role of the LUMO, with an activation energy of 0.3 eV. [This energy is called “effective band gap” in (9).]

In the field of carbon nanotubes, surface transfer doping is in fact the method of choice for manipulating electronic conductivity. Nanotubes essentially consist of one or a few rolled-up sheets of graphene, and donors or acceptors are naturally positioned on the surface of these tubes rather than incorporated into the rigid graphene layers. The electrical conductivity of carbon nanotubes changes markedly upon exposure to different gases (10, 11). In some cases, this behavior has shown striking similarities to electrochemical surface transfer doping of diamond (12).

Surface transfer doping thus appears to be the mechanism behind a variety of surface electronic phenomena. When controlled, it may become a valuable tool for engineering micrometer- and nanometer-scale electronic devices.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
View Abstract

Stay Connected to Science

Navigate This Article