Particle contamination control

Characterization of particle generation sources

With the growing number of transistors in chips following Moore’s law, the size of device features is decreasing. This imposes increasing demands on contamination control in semiconductor manufacturing equipment. The existence of particle contamination negatively affects production yield, i.e. particles on a silicon wafer surface during chip production can cause defects in the produced devices. Particles can originate from several sources, like air-borne contamination from the environment. Particles can also be generated inside the semiconductor manufacturing equipment itself, due to e.g. material deformation and sliding contacts.

Figure 1: Examples of particle contamination in nano-structure manufacturing. (Source: Marc Verschuuren et al., SCIL Nanoimprint Solutions)

Semiconductor manufacturing equipment

In the past decades, Philips Engineering Solutions has built up extensive knowledge and experience in development of equipment requiring minimal particle generation. Essential aspects are bearing concepts, material and coating combinations, and cable solutions.

In a recent project, a mechatronic module has been designed with very stringent requirements on particle contamination. Next to theoretical predictions, more insight in real-life behavior of particle generation was required. Therefore, Philips Engineering Solutions has developed dedicated test equipment for the characterization of particle generation by common wear mechanisms such as sliding, micro-slip, elastic deformation and impact.

Particle generation testing for multiple applications

The test equipment features a clean air environment better than ISO Class 3 (ISO 14644) using a down flow cabinet with ULPA filters, and uses optical, airborne particle detection. By matching the airborne particle detection with the clean air down flow rate, local particle generation phenomena can be characterized.

Particle generation sources can be characterized for the amount of generated particles as well as particle size (from 10 nm to 30 µm). Optionally, generated particles can also be sampled and analyzed with respect to their material composition.

The developed test equipment can be used for multiple applications, like semiconductor equipment, medical equipment and consumer products.

Figure 2: Test equipment for characterization of particle generation and wear

Results

The pictures below show example results of particle generation characterization. Figure 3 shows a typical time series of counted particles from a wear mechanism. During the first hour, the tested wear mechanism is idle (switched off) and only background noise is measured. After ~1 hour, the wear mechanism is switched on, the effect of which can be clearly observed by the amount of counted particles. After ~160 minutes, the mechanism was switched off and on, consequently, the amount of counted particles drops to background noise level.

Figure 3: Time series of counted particles (>10 nm)/cubic foot/minute

In Figure 4, multiple measurement results, like shown in Figure 3, are combined in a box plot, to be able to compare different wear mechanisms, load cases or, in this case, materials. In this example, ‘Material 3’ has substantially higher particle generation, under equal load conditions, than the other materials.

Figure 4: Comparison of particle generation by different wear mechanisms

The testing of particle generation has enabled the development team to optimize the mechanical design and material selection. Furthermore, the customer is provided with insights in particle generation behavior and particle size distribution.