Schottky Diode - Construction

Construction

A metal–semiconductor junction is formed between a metal and a semiconductor, creating a Schottky barrier (instead of a semiconductor–semiconductor junction as in conventional diodes). Typical metals used are molybdenum, platinum, chromium or tungsten, and certain silicides, e.g. palladium silicide and platinum silicide; and the semiconductor would typically be N-type silicon. The metal side acts as the anode and N-type semiconductor acts as the cathode of the diode. This Schottky barrier results in both very fast switching and low forward voltage drop.

The choice of the combination of the metal and semiconductor determines the forward voltage of the diode. Both N and P type semiconductors show the Schottky effect; the P-type has typically much lower forward voltage. As the reverse leakage current increases dramatically with lowering the forward voltage, it can not be too low; the usually employed range is about 0.5-0.7 V and P-type semiconductors are employed only rarely.Titanium silicide and other refractory silicides, which are able to withstand the temperatures needed for source/drain annealing in CMOS processes, usually have too low forward voltage to be useful; processes using these silicides therefore usually do not offer Schottky diodes.

With increased doping level of the semiconductor the width of the depletion region drops. Below certain width the charge carriers can tunnel through the depletion region. At very high doping levels the junction does not behave as a rectifier anymore and becomes an ohmic contact. This can be used for simultaneous formation of ohmic contacts and diodes, as diodes form between the silicide and lightly doped N region and ohmic contacts form between the silicide and a heavily doped N or P region. Lightly doped P regions pose a problem as the resulting contact has too high resistance for a good ohmic contact and too low forward voltage and too high reverse leakage to be a good diode.

As the edges of the Schottky contact are fairly sharp, high electric field gradient occurs around them which limits the reverse breakdown voltage. Various strategies are used, from guard rings to overlaps of metallization to spread out the field gradient. The guard rings consume valuable die area and are used primarily for large higher-voltage diodes, while overlapping metallization is employed primarily with smaller, low-voltage diodes.

Schottky diodes are often used as antisaturation clamps on transistors. Palladium silicide ones are excellent here due to their lower forward voltage (which has to be lower than the forward voltage of the base-collector junction); platinum silicide ones have forward voltage closer to that and require more attention in layout. The Schottky temperature coefficient is lower than the coefficient of the B-C junction, which limits the use of PtSi at higher temperatures.

For power schottky diodes the parasitic resistances of the buried N+ layer and the epitaxial N layer become important. The resistance of the epitaxial layer is more important here than for a transistor as the current has to cross its entire thickness. It however serves as a distributed ballasting resistor over the entire area of the junction and prevents localized thermal runaway under usual conditions.

In comparison with the power P-N diodes the Schottky diodes are less rugged. The junction lies in direct contact to the thermally sensitive metallization, a Schottky diode can therefore dissipate less power than an equivalent-size PN one with deep-buried junction, before failing - especially during reverse breakdown. The relative advantage of lower forward voltage of Schottky diodes is diminished at higher forward currents, where the voltage drop is dominated by the series resistance.