The below HTML version provides a quick overview.

Abstract

Process tools used in 300 mm semiconductor production rely on minienvironments to reduce the exposure of wafers to particles. Real-time particle monitoring of minienvironments gives tool owners the ability to characterize normal operating behavior and to detect potentially harmful particle events. By also monitoring changes in air pressure, temperature, relative humidity, airflow velocity, electrostatic attraction, etc., the user can more rapidly determine the causes of particle excursions.

The MiniNet® 310 Particle Monitor provides this enhanced insight into tool behavior, including providing alarms and supporting data as to when a process should be shut down.

This application note provides guidelines on installing the MiniNet in a minienvironment. It includes sample probe placement, integrating analog environmental sensors, and configuring monitoring software to collect particle and analog data. Finally, frequently asked questions are discussed.

Introduction to the MiniNet Particle Monitor

The MiniNet 310 particle monitor provides real-time monitoring of particles and differential air pressure in minienvironments. It uses a diode laser based sensor to size and count sampled particles in 0.3, 0.5, 1.0 and 5.0 µm classifications, based upon a sample flow rate of 1.0 CFM. It also offers the ability to integrate data from 1-3 additional 4-20 mA environmental sensors. This information often can be used to more rapidly determine the source of the particle event.

Example: In one study microcontamination engineers determined that particle events occurred simultaneously with increases in the minienvironment pressure. Tool owners knew that these pressure increases were associated with FOUP dockings. This led them to an investigation of how FOUP dockings were related to increases in particle counts.

One or more MiniNets are controlled by Facility Net software, which is loaded in a dedicated PC. This software not only controls the monitors' sampling patterns, it also collects, displays and stores the resulting data, generates reports, alarms upon excursions, and automatically sends reports, e-mail or alarm pages to the appropriate microcontamination and process engineers.

Ensemble Sampling: The underlying principle used in the MiniNet particle counter is ensemble sampling, i.e., sampling simultaneously from multiple locations. This is effected by using an "ensemble manifold" to draw aerosol samples simultaneously from 5-7 ports, at a combined rate of 1 ft3 per minute (CFM). This combined, "mixed air" input is then sampled by a particle counter; the particle counts represent the total particles from all sites - rather than from one site at a time.

As can be seen in Figure 1, the ensemble manifold is placed on top of the MiniNet's sample inlet, located on the top surface of the MiniNet. Tubing is connected to each of its seven inlet ports, a sample probe is attached to each tube, then the sample probes are placed at appropriate locations in the minienvironment. The built-in differential air pressure sensor is installed so that the high pressure tubing samples from a positive pressure zone of the minienvironment, while the low pressure tubing samples from the cleanroom.

Other MiniNet 310 particle counter features include a cooling fan, HEPA output filter, power source, dual indicator LEDs, and Ethernet data communications.

Installing the Sample Probes

As the semiconductor industry gains more experience with the use of minienvironments, microcontamination professionals have begun to recognize that new tactics are required to detect particle events. Compared to a ballroom-type cleanroom, the minienvironment's enclosure has a greatly reduced volume, tool-specific filtering, carefully designed laminar flows, and much more complex airflows. All of these factors combine to make particle events much more localized, hence much more difficult to detect. Further, with the internal air replenished several times per minute, events can be much shorter in duration.

The single most important factor in detecting particle events in minienvironments is the careful placement of the sample probes. Sample probe location is a complicated decision, as it requires a comprehensive analysis of the possible particle generation and transport mechanisms of both tool and minienvironment. This section describes:

² How particles move and are detected

² Identifying particle sources and movement in the minienvironment

² Determining where to place probes

² General tips for probe installation

² Examples of where sample probes were placed in evaluations of specific tools

Particle Movement and Detection

Particles move similarly to a plume of smoke. Instead of being carried up by hot air, the particles are carried down by laminar flow (see Figure 2)*. If the air current is strong, the particle plume remains tightly concentrated; otherwise, it spreads out more widely.

For a particle event to be detected, the particle plume must fall upon a sample probe. Particle counts will be highest when the sample probe is both in the plume and close to the particle source; they will drop when the probe is located further away from the source. Counts will drop to zero (or baseline) when the probe is located outside of the particle plume.

Defining the Particle Envelope

There are numerous potential particle sources in all but the simplest minienvironments. Every port, every filter, every potential leak point can become a particle source over time. Likewise, the robotics can generate friction at virtually every position they can reach.

Before placing the sample probes, it is essential to work with the tool owner to complete a particle analysis of the tool and its enclosure. Potential particle ingress or generation spots must be identified. The wafer path must be specified.

Identifying the potential sources is only the first half of the job. Next, one must consider the particle envelope that could be generated from likely sources. Defining this envelope is essential in identifying the best locations for detecting particle events. Thus, airflows must be identified, including lateral airflows caused by ports, openings or multiple pressure zones within the minienvironment. Based upon these data, the spread of projected particle plumes can be visualized.

Probe Placement

To detect a particle event, a sample probe must be placed where the particle plume will fall upon it. The likelihood of detecting particle events is maximized by placing the probes where the odds are the greatest:

² A probe can be dedicated to a single, highly-likely particle source by placing it inside the projected particle plume from that source.

² Alternatively, a probe can be placed inside the intersection of the projected particle plumes from multiple sources, each of whose likelihood is low to medium.

Considering the many potential particle sources, hence the great expanse of the particle envelope, one quickly realizes that in a minienvironment a single sampling point is likely to miss many particle events. The solution is to sample simultaneously and continuously at a number of different locations. While you may not be able to sample at enough places to detect every particle event, you should at least cover the most likely ones.

When the cost of particle sensors is considered, the situation clearly calls for ensemble sampling in minienvironments. For example, the MiniNet's ensemble manifold allows one particle counter sensor to simultaneously and continuously monitor 7 points (at approximately 1/7 CFM each), providing a much wider detection of particle events than monitoring only one point at 1 CFM.

The MiniNet particle monitor provides seven different opportunities for detecting sample plumes. Accordingly, the sample probes should be placed to detect the most critical, most likely particle events.

MiniNet Particle Counter Installation Tips

The following rules of thumb are based upon Particle Measuring Systems' studies of minienvironments:

² In general, the most critical places to monitor will be along the wafer's path through the minienvironment, plus at the critical locations and potential particle generators (such as load ports, wafer slots, filters, and robotic chucks). See Section 2.5 for examples of actual sampling locations for specific tools.

² Install the sample probes near or just beneath (e.g., 4" to 6" below) the working level of the process.

By installing beneath the point where you want to detect the particles, you increase the odds that the particle plume will fall upon the probe, hence increasing the odds of detecting the particle event. (See Figure 3.)

If, however, you install too far below, the plume could spread so widely that the counts will fall too low to be detected.

² Avoid installing sample probes at heights well above the process level (e.g., near the ceiling fan filter unit) as these locations will generally result in low or zero counts, sometimes missing significant particle events that impact the wafer.

² Avoid installing sample probes just above a floor or table, as counts then may reflect particle re-circulation instead of original particle plumes.

² Try to install sample probes where they can detect any particle excursions that pass over the wafer or other critical areas.

² Always install the sample probes to point directly into the airflow.

² After initial probe installation, review the baseline data to see if you want to make minor adjustments in sample point locations.

Contact us if you need more information or have questions.

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