Semiconductor Technology from A to Z

Everything about semiconductors and wafer fabrication

1. Contacting doped semiconductors

After the transistors have been fabricated in the silicon substrate, they have to be connected to one another by means of electrical contacts. On the one hand, the gate electrode is contacted to control the transistor; on the other hand, the doped source and drain regions, through which the current flows, have to be addressed. Problems arise here at the contact areas where the metallization meets the silicon, since, depending on the doping type of source and drain, there is either a lack of electrons (p-doped) or a surplus of electrons (n-doped).

The Fermi level plays an important role here. The Fermi level is the energy level up to which electrons are still present at absolute zero temperature (-273.15 °C). In conductors, electrons occupy the valence band as well as the energetically higher conduction band, so the Fermi level lies at the level of the conduction band. As an illustration, consider the surface of a lake: the water molecules beneath it represent the electrons, which extend up to the surface – the Fermi level.

Fermi level in metals

In doped semiconductors, foreign atoms are present in the crystal lattice as donors or acceptors. In n-doped semiconductors, the Fermi level lies close to the conduction band edge, since the donor atoms can provide free electrons even with only a small input of energy. Correspondingly, in a p-doped semiconductor the Fermi level lies close to the valence band edge, since electrons from the valence band of the silicon crystal can easily be captured by the acceptor atom (see chapter Doping).

Fermi level in doped semiconductors

2. n-type semiconductor contact

Whether and in which direction electrons flow at the contact is determined by the work function of the two materials – the energy required to release an electron from the Fermi level out of the material. If the work function of the metal is greater than that of the n-type semiconductor (as is the case, for example, with aluminum on silicon), the Fermi level of the metal lies energetically lower: upon contact, electrons flow from the silicon into the metal, since they can occupy energetically more favorable states there. In the opposite case – work function of the metal smaller than that of the n-type semiconductor – no barrier forms; instead, an ohmic contact is established right from the start.

Band diagram before contact

Thus, the probability of electrons being present in the conduction band of the semiconductor decreases: the distance between the conduction band edge and the Fermi level – which describes the highest energy state still occupied by electrons – increases at the interface.

Due to the departed negative charge carriers, positive donors (e.g. phosphorus ions) remain behind, and a space charge region is formed. The bending of the conduction band illustrates the voltage barrier (Schottky barrier) that the remaining electrons in the n-type conductor must overcome in order to flow into the metal.

When metal and semiconductor come into contact, the Fermi levels align through diffusion processes; in the region of the interface, the Fermi level is constant.

Band diagram after contact

The width w of the depletion zone depends on the strength of the doping. The electrons that have migrated out of the semiconductor generate a negative space charge in the metal, which is confined to the surface region.

This metal-semiconductor contact exhibits a nonlinear current-voltage characteristic, known as a Schottky diode. Electrons can overcome this barrier either through thermal energy from outside or by "tunneling" through it under an applied electric field (according to quantum theory, a particle can cross a region in which it cannot exist for energetic reasons by, to put it in simplified terms, briefly borrowing energy to overcome the barrier and then returning that energy: the tunneling effect). This effect can also be observed with aluminum. Since aluminum always oxidizes at the surface, two aluminum surfaces placed against each other would have an insulating effect. However, a current flow can be observed, which is based on the tunneling effect.

Depending on the application, one may want to produce this diode effect or prevent it. To create an ohmic contact, i.e. a contact without this potential barrier, the contact area can be heavily doped (n+ doping) so that the depletion zone becomes very thin and the metal-semiconductor contact exhibits a linear current-voltage relationship as a result of the tunneling effect.

Band diagram after n⁺ doping

Since aluminum is incorporated into silicon as an electron acceptor (it accepts electrons), forming a p-doping at the interface, an ohmic contact results with p-doped silicon. In an n-doped region, however, the aluminum causes a doping reversal, so that a p-n junction is formed here: a diode. There are two ways to avoid this:

  • the n-doped region is doped so heavily that the aluminum only weakens it but does not reverse it
  • an intermediate layer of titanium, chromium, or palladium prevents the re-doping of the n-doped regions

To improve the contacts, metal silicides (metals combined with silicon) can also be applied to the contact surface.

In contrast to the diode at the p-n junction, where the switching speed relies on the diffusion of electrons, Schottky diodes have very short switching times. They are therefore suitable as protection diodes to absorb voltage spikes.

3. p-type semiconductor contact

In metal-p-type semiconductor contacts, an exchange of charge carriers between the metal and the semiconductor results in a downward band bending; holes in the semiconductor recombine with electrons from the metal. Due to the reduction of the hole concentration, a negative space charge region forms in the semiconductor crystal. The distance between the valence band edge and the Fermi level – representative of the highest occupied states by electrons – increases at the interface, and the resulting potential barrier at the valence band edge prevents further movement of the holes, which – complementary to electrons – seek to occupy the energetically highest states.

Band diagram after metal–p-type semiconductor contact

Without an external voltage, the diffusion processes come to a halt. Here too, the Fermi levels align with each other in thermodynamic equilibrium.

4. Band model of a p-n junction

From the fact that the Fermi level must be constant (otherwise electrons would flow to locations with a lower Fermi level, occupy free states there, and thereby raise the Fermi level again), a bending of the bands also results at the p-n junction. This illustrates the space charge region that forms as a result of the departed majority charge carriers and the remaining fixed dopant atoms; in other words, the potential barrier that, in the equilibrium state (without an applied voltage), prevents further diffusion of electrons and holes into the respective other crystal. In silicon, the diffusion voltage needed to overcome this potential drop is about 0.7 V.

Band diagram at the p-n junction