Contamination and airflow distributions in a non-unidirectional airflow cleanroom
09/01/2000
Airflow
by Chun-Hung Tsai
Recirculation zones and turbulence intensities play important roles in the contamination distributions in the cleanroom
The airflow characteristics in a non-unidirectional flow cleanroom are complex, compared with a unidirectional flow cleanroom. The large recirculation zone and strong turbulent behavior are typical airflow characteristics in non-unidirectional cleanrooms.1, 2 However, the effects of the airflow characteristics on the contamination dispersion in such cleanrooms are still unclear. According to previous studies, the particles might be trapped and accumulated in the recirculation zones.3, 4 Yet, how do the airflow characteristics affect the spread of particles? In addition, are the different particle sizes affected to the same extent?
 Figure 1. The ISO Class 5 (Class 100) non-unidirectional airflow cleanroom. |
This research is conducted in an ISO Class 5 (Class 100) non-unidirectional airflow cleanroom, which is used for the color filter process in thin-film transistor/liquid crystal display manufacturing. In some troublesome situations, the particles generated within the cleanroom cannot be eliminated by the airflow. By measuring the real-time data of the three-dimensional airflow velocities, the airflow characteristics such as velocity components, turbulence intensities and flow patterns can be analyzed. In addition to the airflow measurement, the size and the number distributions of particles in the cleanroom are also taken in the same locations as the velocity measurement. This research allows contamination distributions to be correlated with airflow characteristics, providing an understanding of particle spreading.
Measuring instrumentation
The layout of the ISO Class 5 (Class 100) non-unidirectional airflow cleanroom is shown in Figure 1. The average air change rate is 300 air changes per hour. Sources of contamination are identified in this research, as is the relationship between the contamination distributions and the airflow characteristics by measuring and analyzing the real-time data.
The airflow characteristics are measured with the three-dimensional ultrasonic anemometer (KAIJO FA600). The span between a pair of emitter-receiver heads of the measuring probe is 5 cm. The velocity in each direction is calculated by measuring the time it takes for a sonic pulse propagating from the emitter to reach the receiver head. The sampling rate of the measurement is 20 Hz, and it takes about 30 seconds to obtain a measurement at each position.
The contamination level is measured by the particle counter, PMS µLPC-110, which detects particles in the 0.1- to 5.0-microns range. The light source is a 2mW He-Ne laser tube.
Airflow characteristics
The instantaneous velocity components, Ui, Vi and Wi corresponding to the x, y and z directions, are recorded by the three-dimensional ultrasonic anemometer. The average velocities, -U, -V and -W, are then obtained by averaging the collected data in the measuring time (equation 1).
 |
null
The turbulence intensity is an index of the flow fluctuation and fluid transportation.
The definition of the turbulence intensity is shown in equations 2 and 3.5
 |
null
 |
null
The airflow distributions and turbulence intensity of the measuring plane A in this cleanroom are shown in Figure 2. The lengths of the arrows represent the magnitudes of velocities, and the radii of the circles represent the magnitude of the turbulence intensity at each measuring point in the cleanroom. The airflow beneath the center of the ULPA filter was directed uniformly downward until it hit the surface of the process equipment.
There are three obvious recirculation zones in the figure, both edges of the ceiling and the region between the ULPA filters.
 Figure 2. Distributions of airflow and turbulence intensity in measuring plane A. |
In the recirculation zone, the direction of the airflow changed significantly and formed a close airflow region. The phenomenon is more essential near the surface of the chemical station because the equipment obstructed the downward airflow and therefore generated the secondary recirculation zone. In such a recirculation zone, the particles generated occasionally by personnel and equipment might be trapped and accumulate in this region causing a serious contamination problem. However, the recirculation zone induced by the non-uniform airflow is a common phenomenon that can be seen in every non-unidirectional airflow cleanroom as well as in some unidirectional airflow cleanrooms.
On the other hand, higher turbulence intensity usually implies higher particle transportation. It is widely known that turbulent flow usually occurs in high Reynolds numbers as well as in separated flow. The airflow velocity in the cleanrooms is relatively small and has often been thought to belong to the laminar flow characteristics. However, in non-unidirectional airflow cleanrooms, it behaves much like the turbulent flow.
 Figure 3. Distributions of airflow and turbulence intensity in measuring plane B. |
Figure 2 shows the turbulence intensities beneath the ULPA filter to be very small. This indicates that airflow entering the cleanroom from the ULPA filters behaves much like the laminar flow, no matter if it is in unidirectional or non-unidirectional airflow cleanrooms. However, the turbulence intensities in the recirculation zones (the edges under the ceiling and the region between the ULPA filters) are much larger than other regions. The strong turbulence in the recirculation would enhance the propagation of the particles. Moreover, the recirculation zone also is potentially a high contamination area as mentioned above. The two effects make the recirculation zone a potentially harmful area to the product, especially near the chemical station.
The airflow distributions and turbulence intensity of the measuring plane B in this cleanroom are shown in Figure 3. The situations in this measuring plane are similar to those in plane A. Because the aligner covers a large area, the measurement can only be done in the other areas.
Contamination distributions
Figure 4 shows a particle concentration size of 0.1 micron in measuring plane A. The radii of the circles in the figure represent the number of 0.1-micron particles at each measuring position in the cleanroom. In the recirculation zone near the surface of the chemical station, it is seriously contaminated by the small particles. In addition, the strong turbulence effect intensifies the particles to propagate to the chemical station. That will destroy the yield of the products in this cleaning process.
 Figure 4. Particle concentration of small (0.1 micron) particles in measuring plane A. |
Near the ground, the particle concentrations are the highest region in the plane. This is not caused directly by the airflow characteristics. Rather, it implies the ground is a contamination source and the airflow from the ULPA filter cannot bring the particles out of the cleanroom. The amounts of 0.1-micron particles in the recirculation zone between the ULPA filters are also great because the particles are blown by the airflow from the ground and the chemical station and then are trapped in this region. The airflow characteristics play an important role in the particle dispersion.
 Figure 5. Particle concentration of large (1-5 microns) particles in measuring plane A. |
The distributions of the large particles, ranging from 1 to 5 microns, are shown in Figure 5. The amount of large particles is minimal because they are not allowed in the cleanroom. Large particles still exist near the surface of the chemical station and ground. The reasons are the same as for the small particles, although the numbers are less. The only position that has a great amount of large particles is under the lamp. This may be caused by the lamp's dirty screws.
The particle dispersion is significantly affected by the aerodynamic characteristics. However, the airflow behaviors related to the recirculation zone could not be clarified by a simple physical quantity. The turbulence intensity is an important index of flow fluctuation and fluid mixing, therefore, it is thought to be a possible index of particle dispersion.
 Figure 6. Relationship of particle concentration and turbulence intensity under ceiling. |
Figure 6 shows the relationship between turbulence intensity and contamination level near the ceiling height in the non-unidirectional airflow cleanroom. The higher turbulence intensity tends to accompany higher particle concentrations. There are greater amounts of smaller particles than larger ones. Of course, as mentioned previously, the turbulence intensity is not the only parameter that affects the particle dispersion. The airflow motion and location of the contamination source also affect particle distributions in cleanrooms. Therefore the deviations of this correlation are quite great. Nevertheless, this correlation still is helpful in understanding the relationship between particle dispersion and aerodynamic characteristics.
 Figure 7. Relationship between particle concentration and turbulence intensity near the ground. |
The relationship between turbulence intensity and contamination level near the ground in the non-unidirectional airflow cleanroom is shown in Figure 7. Note that the ground is a dirty area, so the correlation between the turbulence intensity and particle dispersion is not very obvious. Particle concentrations in the range of the turbulence intensity are almost the same. The figure shows that near the ground, gravity has more of an effect on particle dispersion than aerodynamic characteristics do.
Conclusion
This research correlates contamination distributions with airflow characteristics. From this analysis, the fundamental understanding in particle dispersion caused by aerodynamic characteristics can be realized in the non-unidirectional airflow cleanroom.
Chun-Hung Tsai is a section manager in the facilities department in the Electronics Research & Service Organization, Industrial Technology Research Institute. He received his doctorate in power mechanical engineering from the National Tsing Hua University in 1996. His experience includes facilities engineering, cleanroom technologies and contamination control.
References
- M. Seymour, F. Baban and D. Cansdale, "Airflow Modeling Applications in the Semiconductor Industry", Proceedings of the IEST, 1997.
- R. H. F. Pao, Fluid Dynamics, 1967.
- C. H. Tsai , J. H. Lu and H. H. Li, "Contamination Dispersion and Control in the interface between the Class 10 and 1000 Cleanroom," International Conference on Aerosol Technology/Environmental Measurement and Control, 1997
- C. H. Tsai and C. Y. Huang, "The Contamination Control for the Color Filter Coater in the LCD Process in a Cleanroom," Proceedings of the International Conference on Aerosol Technology/Environmental Measurement and Control, 1997.
- F. M. White, Fluid Mechanics, 1986.