Contamination Monitoring
Metallic contamination in semiconductors adversely affects device performance. As linewidths decrease the allowable levels of metal contamination get smaller and smaller. The most commonly occuring forms of metal contamination are either in the middle of the periodic table, called heavy metals, or are in the alkaline metal column, at the left of the periodic table.
Alkaline metal contamination affects oxides, rather than the bulk silicon, and therefore is discussed in the oxide monitoring application section of the website. Heavy metals – iron is by far the most common culprit – tend to diffuse through the semiconductor material, rather than aggregate in the oxide. Heavy metals can be detected via the change they cause in lifetime or diffusion length, as these impurities create energy states in the bandgap of the semiconductor material.
The circled elements in the periodic table below, can be detected via lifetime or diffusion length measurement. (The circled elements in this chart are the ones mentioned in the literature, and it is quite possible that contamination due to additional elements can also be detected by lifetime and diffusion length measurements.)
Lifetime is commonly measured via a technique called microwave-PCD, and diffusion length is commonly measured via a technique commonly referred to as SPV. Both techniques rely on the fact that contamination in a semiconductor creates new allowed energy states in the bandgap. The existence of these allowed energy states increases the probability that an excess carrier in the conduction band will be able to lose the required energy and momentum and return to the valence band. The increased probability of returning to the valence band results in shorter recombination time (lifetime) and shorter diffusion length.
Some heavy metal contaminants, including Fe, Cu, and Cr, can exist in more than one molecular state in silicon. For instance, Fe can exist in p-type silicon as either interstial iron, Fei, or paired with boron atoms, Fe-B. The different molecular states create different energy states, with different effects on lifetime and diffusion length. If a contaminant can exist in two different molecular states and there is a unique process that transforms all of the contaminant from one state to the other – the process uniquely affecting that contaminant – then one can identify the element causing the degradation in lifetime or diffusion length. Further, by comparing lifetime or diffusion length measurements before and after the transformation, one can not only identify the contaminant but also quantify it. Semilab’s Fe measurement capability is an excellent example of this.
The WT-2000 offers maps of lifetime and diffsion length measurements for semiconductor wafers, via a benchtop system. The WT-2500 offers similar capability via FOUP and/or Open Casette system for up to 300mm wafers.










