300-mm Imperative

Characterizing FOUPs and evaluating their ability to prevent wafer contamination

Kirk Mikkelsen and Tracy Niebeling, Entegris

Both seal and filter capabilities affect the performance of FOUPs, but tests show that well-designed pods can protect 300-mm wafers from exposure to particles.

As the semiconductor industry's contamination control requirements have become increasingly stringent, the environments used to transport and store process wafers have also evolved. The open cassettes that were the transport method of choice in typical 200-mm wafer fabs are being superseded by sophisticated automated systems based on front-opening unified pods (FOUPs), which were developed in response to the move to 300-mm wafers. These passive minienvironments, which can hold a process lot of 25 wafers, have three critical performance functions. First, each FOUP must provide precise accessibility for reliable, high-throughput wafer transfer. Second, it must protect wafers from contamination and damage. And finally, it must be capable of interoperability with multiple process tools and automated material-handling systems.

This article focuses on the second critical performance function, that of protecting wafers from particle generation during the FOUPs' normal functions of wafer transport, wafer storage, and interfacing with process and handling equipment. An introductory section on FOUP design considerations is followed by a discussion of using conductance to characterize a pod's resistance to air- and particle flow. The results of FOUP storage, overhead transport, and loadport tests that evaluated pod performance in terms of particle adders are then presented, and finally, some suggestions for additional testing are proposed.

FOUP Design Considerations

In an open wafer fab or inside a process tool chamber, wafers are kept clean by a vertical laminar flow of filtered air. The inside of a FOUP, however, is a very different environment, because normally it contains no flowing air. And while the air in a process tool chamber is typically at a pressure higher than that of the surrounding fab, the inside of a FOUP is not actively pressurized when its door is closed. The atmosphere inside a FOUP consists primarily of air that flows in from process chambers during wafer transfer steps, along with some air that enters through a breather filter during pressure equalization of the closed FOUP.

Although it would seem logical that a FOUP should have a hermetic seal to guarantee that no outside air reaches the wafers inside, practicality demands that they not be designed to be airtight. Achieving such impermeability would be challenging, since the door seal is very long and the available sealing forces are quite small. More importantly, however, if a FOUP were made airtight, a change of the pressure differential between the interior and exterior of the pod could cause the door to lock up, making it difficult, or impossible, to open when connected to a process tool loadport. Pressure differentials as small as 0.2 in.H₂O (0.0074 psi) will generate more than a pound of force holding the door shut. Such a pressure-related seizure can be caused by a simple atmospheric pressure increase, which typically occurs when fair weather moves in after a storm. It can also be caused by a temperature change inside the pod. For example, if a FOUP is filled with hot wafers and then closed, the cooling wafers will cause the inside air to contract, lowering the pressure and causing the door to lock. Thus, FOUPs are usually designed with a pressure-relief feature known as a breather port, which contains a filter that allows air into the pod while trapping any airborne particles. Figure 1 shows a filtered breather port on the bottom of a FOUP, and Table I lists pressure changes and corresponding air-volume changes that occur within FOUPs under normal use conditions.

 

Figure 1: Close-up of a breather port on the bottom of a FOUP. The filter inside the port prevents airborne particles from entering the minienvironment during air pressure equalization.

Condition Change in Pressure (in.H2O) Air Volume Entering Pod (L) Overpressure episode 0.005–0.04 Negligible Transport between cleanrooms with differing pressures 0.1 0.01 Weather event, such as typhoon or hurricane 13.8 0.95 Following hot wafer insertion (180°C initial air temperature inside pod) — 12.7

Table I: Pressure changes and corresponding air-volume changes in FOUPs under various normal use conditions.

When characterizing FOUPs, it is important to consider the performance of both the seal and the breather port. (In this discussion, the term seal encompasses all enclosure features of the FOUP, since most air leakage occurs through the door seal if a FOUP is properly designed.) The goal is to design the seal and breather port so that almost all of the air entering the pod comes through the port's filter rather than through the unfiltered seal, as shown in Figure 2. Filters that have a particle capture efficiency of >99% are widely available. Assuming that the filter media used in a port approaches 100% efficiency, a port-to-seal airflow ratio of 50:1 will result in a net efficiency of 98%, which should provide sufficient protection against contamination, since FOUPs are used within cleanrooms.

 

Figure 2: Schematic showing airflow into a FOUP when the pressure outside the pod (P1) is higher than the pressure within it (P2).

While a highly efficient filter is needed to ensure that airborne particles are trapped, the filter's airflow properties are also an important selection criterion. If a filter does not have sufficient airflow capability to achieve pressure equalization within the pod, airflow past the FOUP door seal will be relatively high. It has been found, for example, that a PTFE membrane filter is a poor choice for FOUP breather ports. This type of filter is very effective at excluding particles, but its airflow capability is so low that many airborne particles pass through the FOUP door seal and are deposited onto the wafers during normal pressure-change events. In contrast, the use of an electrostatic filter results in cleaner wafers, even through the filter media does not have an absolute pore size. Because an electrostatic filter admits more air than a membrane filter, it allows the FOUP to breathe through the filtered port, minimizing the amount of unfiltered air that leaks through the door seal.

FOUP Conductance

A helpful technique for characterizing the leakage rate of a FOUP is to determine the conductance of its door seal and port filter. Commonly used in vacuum applications, conductance is usually expressed in liters per second and can be calculated without specifying pressure using the following general equation:

Q = DP xC

where Q is the mass airflow rate and DP is the pressure differential between the inside and outside of the pod. (Ohm's Law is an electrical equivalent to this equation, with pressure differential being equivalent to voltage and flow rate equivalent to current.) If flow rates (Q_n) are measured over a range of pressure differentials and then plotted versus those differentials (DP_n), conductance (C) will be the slope of a line drawn through the data points. Conductance can also be calculated by dividing the flow rate by the pressure differential at each data point and averaging the values. Measurements should be made under both positive and negative pressure differentials. Because the seals on FOUPs are not airtight, a steady-state method is used for conductance testing: that is, air is introduced into or removed from the pod at a steady rate (see Figure 3). Figure 4 shows an airflow test bench being used with a FOUP.

 

Figure 3: Schematic showing air being introduced into and pumped out of a FOUP during conductance testing.

Figure 4: A FOUP attached to a test bench for conductance testing.

Ideally, a seal should perform better under a negative pressure differential condition (lower internal pressure), which exists when potentially contaminated air is trying to flow into the pod, than it does under a positive pressure differential condition. Figure 5 shows a plot of flow versus pressure for a FOUP that features a door seal that has been especially designed to accommodate negative pressure differentials. To obtain these data, the breather port filter was plugged, so that the resulting FOUP conductance valves represented seal conductance only. In this instance, the conductance was constant (at C = 0.90 L/sec) throughout the measurement range at positive pressure differentials, but the data points deviated from the linear at the highest negative pressure differential (1.2 in.H₂O). This deviation is a result of the flap design of the seal, which is shown in cross section in Figure 6. At high negative pressure differentials, the added pressure on the outside of the seal flap puts more force on the sealing surface, resulting in a better seal (low seal conductance).

 

Figure 5: Plot of airflow rate versus pressure for a FOUP with the filter plugged. In this example, the conductance at positive pressure differentials was constant at 0.90 L/sec.

Figure 6: Cross section of a FOUP door seal designed to handle negative pressure differential conditions.

When conductance measurements have been made for both the seal and a port filter of a given flow diameter, the total conductance of the pod assembly can be calculated using the equation

C_P = C_s + C_f

where C_P is the conductance of the pod assembly, C_s is the conductance of the seal, and C_f is the conductance of the filter. The net particle capture efficiency of the pod assembly also can be calculated using the equation

where C_s is the conductance of the seal, C_P is the conductance of the pod assembly, h_f is the known efficiency rating of the filter, and h_P is the net efficiency of the pod assembly.

Applying conductance data to the design of a FOUP, the selection of an electrostatic filter medium with an area conductance of 22.0 L/sec/cm² and a flow diameter of 2.8 cm will yield a filter with a conductance of 135 L/sec. The ratio of filter conductance to seal conductance is then >100:1. A filter's capture efficiency varies with flow rate, or filter face velocity, but over the typical range of relevant flow rates, the chosen filter's lowest efficiency is 99.9% for 0.3-µm particles. If this filter is placed in the pod with a seal conductance of 0.9 L/sec, which was used in the conductance test depicted in Figure 5, the resulting net pod assembly capture efficiency would be 99.2%.

As described earlier, the pressure inside a FOUP must be equalized when pressure changes occur outside the pod. In some cases, such a pressure change can occur quickly; for example, the pressure difference between one cleanroom and another can be as high as 0.1 in.H₂O, and the pod would experience this change in seconds when it is transported between the two cleanrooms. Atmospheric pressure changes may be greater but usually occur more slowly. A severe weather event such as a typhoon, for instance, will cause a pressure change of 13 in.H₂O over several days. For the pod assembly in the example above, 0.95 L of air would be required to equalize pressure if such an event occurred. Assuming the pod was being used in a Class 10,000 cleanroom, the outside air could contain 30,000 particles ≥0.3 µm (10,000 ≥0.5 µm) per cubic foot, or 1060 particles ≥0.3 µm per liter. However, since the pod assembly has a particle capture efficiency of 99.2%, only 8.5 particles/L would enter the FOUP during the pressure equalization.

FOUP Performance Testing

Passive Storage. To evaluate the performance of a FOUP with a flap-type door seal and an electrostatic filter in the breather port, an F300 autopod from Entegris (Chaska, MN) was loaded with wafers and then stored in a Class 100,000 environment for 72 hours. Low-particle-count wafers were placed in pockets 1, 13, and 25, and clean dummy wafers were loaded into the other 22 pockets. During the storage period, the average airborne particle counts in the room were 627,182 ≥0.03-µm particles and 41,677 ≥0.05-µm particles per cubic foot. Despite the FOUP's long exposure to such a dirty environment, poststorage wafer testing indicated that few particles had entered the pod.

Table II shows the number of ≥0.12-µm particles added to the three test wafers. These numbers are very low and may be at or below the detection limit of the SP1 wafer surface particle counter from KLA-Tencor (San Jose). Hence, determining the number of ≥0.03-µm particles added was definitely beyond the instrument's capability. However, these test results can serve as a starting point for calculating the FOUP's performance capabilities when it is used in cleaner environments. For example, if the pod were used in a Class 10,000 environment, probable particle adders would be only one-tenth of these totals. In addition, the results clearly indicate that a well-designed FOUP can protect wafers from particles even in environments that are not tightly controlled.

 

FOUP Pocket Particles Added 1 5.3 13 12 25 6.0 Average 7.8

Table II: Results of passive storage testing.

Overhead Transport. To test the performance of the FOUP design under more severe conditions, a pod was loaded with wafers and transported for 30 minutes within a Class 100,000 environment. Low-particle-count wafers were placed in pockets 1, 13, and 25, and clean dummy wafers were placed in the remaining pockets. During the test period, the pod was hung from the top robotic flange of the cleanroom's transport system and vibrated at 0.25 g peak and 3 Hz. The ambient environment had average counts of 648,000 ≥0.03-µm particles and 41,100 ≥0.05-µm particles per cubic foot. Table III presents the results of this testing. As with the passive storage test, these data indicate that the FOUP performed its wafer protection function quite well, considering that the exterior environment was Class 100,000. Assuming that all of the particle adders came from the external environment, estimates of the pod's performance in cleaner environments can be made. However, further testing in a more realistic challenge environment of Class 10,000 is needed to determine whether this assumption is valid.

 

FOUP Pocket Particles Added 1 102.7 13 80.7 25 137.3 Average 106.9

Table III: Results of overhead transport testing.

Loadport Testing. During the initial movement to open a FOUP door for wafer transfer, a negative pressure differential is created that is equalized by air entering the pod through the breather port and around the door seal as it pulls away from the shell. Any air that enters the pod through the filtered port reduces the volume and velocity of the unfiltered air that will enter around the seal, thereby also reducing the possibility of airborne particles penetrating the pod's interior. The speed at which the door opens affects how much air will enter via the filter during this operation. That is, the slower the door opens, the more time there will be for air to enter through the filter to offset the volume of air entering around the seal.

To assess the effectiveness of the FOUP design during door opening, a pod was subjected to repeated open-and-close cycles while interfaced with two identical loadports. For both loadports, airborne particle counts were taken inside the pod, just inside the loadport door, and just outside the FOUP shell. Although the ambient air was better than Class 1 and the air within the pod was better than Class 1 when the cycling was not occurring, movement of the loadport door caused the quality of the pod air to decrease to Class 10 during cycling.

During the test, 25 wafers were scanned, placed into the pod, and rescanned after 500, 10,000, and 15,000 cycles. The average number of ≥0.09-µm particles added per wafer per cycle at each of these test intervals is shown in Table IV. These averages are very low, and the counts were near the limit of the wafer scanner's repeatability. Converting the two averages to a per-area basis yields 4.87 x 10^–6 particles/cm² per cycle. To fully assess the filter's positive contribution to achieving these low particle counts, additional testing must be done with the external environment artificially degraded to different levels.

 

Loadport Particles per Wafer per Cycle 5000 Cycles 10,000 Cycles 15,000 Cycles Average 1 0.0095 0.001 0.0008 0.004 2 –0.0003 0.002 0.005 0.002

Table IV: Results of loadport testing.

Future Considerations

The move to 300 mm has been accompanied by the use of closed FOUPs for wafer transport and storage. Protecting the wafers from airborne particles is among the critical functions of these pods, which may incorporate filtered breather ports and flap-type seals to achieve that capability. The concept of conductance has been found helpful in characterizing the pods; and storage, transport, and loadport FOUP testing has found that they admit very low levels of particles. However, additional research is still needed.

In standard 200-mm wafer fabs, open cassettes are used for wafer transport and cleanroom air is maintained at Class 1. To achieve cost savings in 300-mm fabs, where wafers are transported in closed FOUPs, the semiconductor industry would like to keep the ambient air at some cleanliness level less than Class 1, perhaps Class 100, Class 1000, or even Class 10,000. Because there are no industry-standard methods to measure FOUP performance in such environments, a challenge environment test must be developed. Conducting such a test is not as simple as placing the FOUP in a Class 1000 cleanroom and evaluating its capabilities, however. Class 1000 would be the worst-case situation in such a cleanroom; the actual cleanliness levels would be much better for most of a test period. What is needed for a true FOUP challenge is for the outside environment to be controlled at or near the worst-case limit throughout the testing.

The semiconductor industry has also expressed interest in the use of an inert gas purge to ensure effective wafer isolation within FOUPs. The inert gas would be pumped into a closed pod at a pressure slightly higher than ambient so that particle-filled outside air would not permeate the interior as the FOUP is transported or stored. After the purge, the wafers would not only be protected from exposure to particles, they would also be shielded from contact with oxygen and water vapor.

Implementing an inert gas purge in FOUPs will be challenging, however. Inlet and outlet ports will be needed on the pods, and corresponding interface ports will have to be installed at many locations around the fab. There are also important questions that must be addressed through testing, including how often and at what points during the transport process FOUPs will need to be purged to maximize their effectiveness at preventing contamination. A cooperative effort between suppliers of FOUPs, process tools, and automated material-handling systems will be needed in order to adopt and benefit from this proposed design feature.

Acknowledgment

The authors would like to thank Russ Wendt of IBM's Almaden Research Center for the work he performed on FOUPs at International Sematech. His promotion of conductance as a standard method of characterizing FOUPs has benefited the semiconductor industry greatly.

Kirk Mikkelsen is manager of Entegris's product test laboratories, which are in Chaska, MN, and Colorado Springs, CO. The labs are responsible for characterizing the physical, pressure, electrical, thermal, and contamination properties of the company's products. In his 10 years at Entegris, Mikkelsen has also held the positions of senior materials engineer and 200-mm-pod program manager. Previously, he was a research engineer at the University of Minnesota (Minneapolis) and a project engineer at Rheometrics in Piscataway, NJ. He is the author or coauthor of several articles on contamination issues in the semiconductor industry and holds five patents. He has a BS in chemical engineering from the University of Minnesota. (Mikkelsen can be reached at 952/556-8037 or kirk_ mikkelsen@entegris.com.)

Tracy Niebeling has spent 20 years at Entegris, where he has been involved in the design, technical field support, and marketing of wafer management products. He has been active in the SEMI standards development process and has served as cochair of the group's wafer carriers subcommittee. Niebeling holds three patents for wafer-handling technology and has published articles on 300-mm wafer handling. He received a BME from the University of Minnesota (Twin Cities campus) and an MBA from the University of St. Thomas in Minneapolis. (Niebeling can be reached at 952/556-4180 or tracy_niebeling@entegris.com.)