Minienvironments are a key component in the manufacturing of 300 mm wafers. By isolating the tool, manufacturers can cost-effectively reach ISO Class 2 or 3 cleanliness levels in critical wafer handling areas. At these cleanliness levels, even with greatly reduced line and die sizes from increasing the number of dies per wafer, higher yields can be achieved allowing a 30+% reduction in the manufacturing cost per chip.

With such significant improvements in cleanliness, some have questioned whether particle monitoring is necessary in minienvironments. This paper summarizes a set of studies recently conducted in 200 and 300 mm minienvironments. Based upon these results, particle monitoring in minienvironments remains an essential component of maximizing yield.

Particle Incidence Studies

In these studies, if the tool, load lock, FOUP, etc. all performed correctly, the minienvironment kept the wafers extremely clean. This did not mean that there were no particles inside the enclosure. Many tools generated high levels of particles during normal operation (see Figure 1)(Download this paper for all tables and figures) (252.1 KB). If all system components performed correctly, then the air handling, positive pressure, etc. kept these particles away from the wafer. Die kills occurred when something malfunctioned. For example, failure of a fan filter unit allowed particles to gather near the wafer, instead of automatically being swept out of the minienvironment.

Figure 1: (Download this paper for all tables and figures) (252.1 KB)

As a part of these studies, minienvironments were monitored for a wide range of tools, for both copper and non-copper processes: wafer/lot sorters, wafer probers, surface scanners, vertical diffusion furnaces, ion implanters, photolithography trackers, film thickness trackers, and chemical vapor deposition tools.

Results showed that:

Approximately 90% of the particles were generated by the tools, often as a normal part of operation.

The remaining 10% were generally related to the load lock or FOUP.

The number of aerosol particles sampled near the wafer was correlated with the number of particles observed by surface scanners, which in turn was correlated with the number of die kills. In other words, these particles killed die.

Some types of tools were more problematic than others.

Numerous examples were observed of properly installed and operating minienvironments that later malfunctioned. Causes included: fan failures, gasket leaks, filter failures, misaligned robotic chucks (see Example), crosscontamination from other tools, vortices due to wafer movement, and bearing failures (see Figure 2).

Figure 2: (Download this paper for all tables and figures) (252.1 KB)

Example: For a wafer crash (due to a misaligned chuck), particle counts reached over 30,000/cu. ft., which directly correlated with an increase on the top surface of 363 particles 0.25 µm in diameter.

*Particle Distribution and Circulation8

Particles generated at a single point tended to travel like a smoke plume: the stronger the laminar flow, the more compressed the particle plume (see Figure 3). The typical laminar flow, combined with the minienvironment's lowered ceiling, often evacuated the particle plume before it could spread wide enough to be detected by a single-point particle counter.

Figure 3: (Download this paper for all tables and figures) (252.1 KB)

Particle Entrainment and Recirculation

Laminar flow could not solve all problems. For example, when wafers were moved perpendicular, disturbing the flow, vortices were sometimes formed below wafer level. In many cases these vortices entrained particles from below the wafer, carried them upwards despite the prevailing downward laminar flow and re-deposited them on the wafer surfaces, where they were held by electrostatic forces (see Figure 4). Thus, particles residing below wafer level were recirculated, causing die kills.

Figure 4: (Download this paper for all tables and figures) (252.1 KB)

Example: For particles generated 6 inches below the wafer, the overhead passage of a wafer brought an increase of 850 counts/cu. ft., with a corresponding increase of 30 killer particles detected by scanning the top surface.

Minienvironments Require New Particle Monitoring Tactics

As a part of this study, a new set of tactics was developed for monitoring minienvironments, tactics different from those traditionally used for ballrooms or bay-and-chase layouts.

New Tactic: Monitoring Multiple Points

As seen in Figure 3, a single-point particle counter often failed to detect particle events inside a minienvironment. In response, Particle Measuring Systems developed and patented the Ensemble Manifold, which was incorporated into the MiniNet® Process Monitor. This device (see Figure 5) allowed a single particle sensor to sample simultaneously from multiple locations within a minienvironment.

Using the Ensemble Manifold to monitor at multiple points resulted in significantly increased coverage for detecting particle events (see Figure 6).

Figure 6: (Download this paper for all tables and figures) (252.1 KB)

Ensemble sampling also allows coverage of multiple planes or areas, which is quite useful in complex tools (see Figure 7).

Figure 7: Download pdf for all figures and tables

New Tactic: Sampling Below Wafer Level

Traditionally, users have placed their sampling probes near the wafer level. In minienvironments, however, there were two primary reasons for placing the probes lower.

First, since the plume spread more widely as it fell, probes did not need to be directly under the particle source to detect the plume. As can be seen in Figure 8, as long as the particle density was adequate, lowering the probes increased the overall coverage area, protecting against particle plumes from a wider range of possible locations.

Second, as seen earlier in Figure 4, recirculation due to vortices was capable of lifting particles from below onto the wafer. A major finding of these studies was that you cannot assume the product is safe just because the particles reside at lower levels than the wafer. Based upon these studies, it is recommended to start with the sample probes about six inches below wafer level. They can then be adjusted as more data are available. Lowering the probes reduces the need to identify precisely where future particle events will originate.

New Tactic: Detecting Particle Events, Not Smallest Particles

As smaller and smaller particles impact yield, the common impulse has been to monitor with more and more sensitive particle counters. In minienvironments, however, the work space was so clean that one could detect individual particle events. Each particle event generated a distribution of different-sized particles. Thus, the area could be monitored cost-effectively with a particle counter that detected each event, without needing to detect the smallest of particles (see Figure 9).

**Effectiveness of Particle Monitoring at 0.3 µm**

In these studies, particle events in minienvironments generated enough particles to be detected with a 0.3 µm particle counter. This is consistent with our earlier experiences in 200 and 300 mm fabs. Example 1: In wafer crash tests (caused by misaligned robotic chucks), over 90% of the particles adhering to the wafer's top surface were at 0.25 µm.

Example 2: Figure 10 shows data from a 300 mm lot sorter, where vibration twice caused recirculation of the settled particles. These events were easy to detect at 0.3 microns, but not at 0.5 microns.

New Tactic: Continuous Particle Sampling

Intermittent monitoring was inappropriate for the critical areas of a 300 mm fab. Since particles were so quickly swept out of the minienvironment, continuous monitoring maximized the probability of detecting intermittent events.

Complementary Tactic: Monitoring Chamber Pressures

Once a particle event was observed, often the particle source could be found more quickly by reviewing data on the differential pressure (e.g., minienvironment pressure minus chase or cleanroom pressure). If a particle event was observed, and the differential pressure had changed simultaneously, the problem generally was associated with the opening of ports, failure of fans, major leaks, etc. Conversely, if the pressure had not changed, the particle event was more likely to be associated with the robotics or tool. Tools with multiple pressure zones benefited from monitoring in each zone.

Example: Figure 11 shows a particle event in a 300 mm ion implanter, where particle counts rose to a level that resulted in numerous die kills. At the same time, the drop in differential pressure reading gave a clear indication of the source of this particle event, the failure of the fan filter units.

Conclusions

The development of the 300 mm fab has necessitated the use of minienvironments. While minienvironments can get the critical areas extremely clean, continuous particle monitoring is required to keep many of these tools functioning properly. The significantly increased die density, the greater susceptibility to small particles, and the increased cost of slow response have all combined to make particle monitoring a necessary and cost-effective component of yield maximization.

In fact, particle monitoring in minienvironments has been so effective that studies are now underway to evaluate the degree to which such monitoring will allow the reduced use of test wafers, potentially resulting in substantial cost reductions.

(Download this paper for all tables and figures.) (252.1 KB)

Contact us if you need more information or have questions.

Biographies

Ray P. Lucero is responsible for all certification and sustaining of minienvironments at Intel's 300 mm facility in Rio Rancho, NM.

Steven D. Kochevar is Particle Measuring Systems' lead Applications Engineer for minienvironment monitoring.

Scott L. Jorgensen is a Microcontamination Engineer at Intel Rio Rancho.

MiniNet® is a registered trademark of Particle Measuring Systems, Inc.