Airborne Molecular Contamination Monitoring and Control
|
These chemical vapors can result from a wide range of sources, including cleanroom materials, make-up air, process chemicals, operating personnel, cooling systems and even the outgassing of certain aging materials.
Evaluating Gases
Ammonia contamination can be a serious threat. One of the most sensitive steps in the semiconductor manufacturing is photolithography. The technology has indicated increased sensitivity to basic contamination, with ammonia being identified as the dominant basic substance among AMC. The target control levels for ammonia and other AMC contaminations inside the process tool environment are under ~1 ppb. And as device geometries shrink in scale, the need for accurate ongoing monitoring of AMC levels in semiconductor manufacturing environments intensifies. AMC monitoring and control is equally as important as particle contamination, but even harder to identify. In fact, the level of basic impurities in the air can be critical at very low concentrations down to ppb levels and lower. For example, wafers exposed to ammonia can lead to structural defects, such as incorrect imprinted line width, short-circuits, metal corrosion and hazing on the wafers and optics—all of which could cause component failure. Ammonia contamination—which is difficult and expensive to fight—has many sources within cleanrooms, including personnel and several production processes. My lab already tests for AMC, so why should I use a monitoring system? Most labs conduct static testing, so the data generally can only show an average concentration level. However, online monitoring shows AMC levels in real time, and can indicated whether concentration is at a normal background level or is a specific contamination event. It can also show where the low and high phases are in the daily cycle. What compounds are monitored? Monitoring can be done for many types of compounds, including ammonia, NMP, and total amines acids and sulfur. What are the minimum detection limits? The minimum limits will depend on the type of chemicals and technology involved. You must determine the appropriate minimum detection limits based on the requirements for your specific process.
What is required for system maintenance? The sampling system requires very little maintenance, as the vacuum pumps need only to be maintained quarterly. However, maintenance requirements for the analyzers will depend on the type of analyzer and gases being monitored. Additionally, different analyzers will have different calibration requirements, which, in turn, depend on your desired accuracy of the sensor. Where should the sample points be installed? Installation should be as close to critical processes as possible without interfering with the processes. Typically, points are installed inside and outside of a process, as well as up and down stream from the chemical filters. How much does a system like this cost? The cost depends on the chemicals being monitored and the number of locations being sampled. Prices range from $3,000 for a single sensor to $400,000 for an entire system sampling multiple locations. Why should I continuously monitor AMC? Ongoing monitoring is essential to ensuring that a facility is performing properly and can alert personnel when an incident happens. As a result, responses to contamination incidents can be carried out immediately instead of days or weeks. What sensors are recommended? Sensors should be used according to the application. When choosing a sensor, consideration should be given to the following: target chemical, detection limits, dynamic range, response time, zero and span drift, potential interference, portable or online use, heat-up times, calibration method and frequency and operation cost.
Do the instruments contain radioactive materials? Some instruments do use a radioactive source to ionize the sample. Be sure to address this question with each instrument supplier. How long can the tubing runs be from manifold to sample point? The runs can be as long as you wish. However, the chance of contaminating or diluting the sample increases with longer distances. We recommend not running sample tubing any longer than 80 meters or at least using a booster pump for longer distances. What materials are used for tubing? Teflon is the most commonly used material. Stainless steel can also be used, but it is generally more costly and less compatible with a variety of chemicals. How frequently should sensors be calibrated? The frequency should be based on zero drift of the sensor per day (the drift the sensor will experience from zero in a set time, normally a day or week); the span drift (drift experienced from a fixed concentration amount over a day or week); and the level of detection desired. (The lower the level, the more often you will need to calibrate to keep the zero and span drift from exceeding your minimum detection limit.)
If you have additional questions or feedback on this topic, please click on the comments button below or call us at 1-877-688-2703 |
Airborne molecular contamination
There are primarily two types of AMC: internal and external. Internal sources of AMC involve process chemicals, accidental chemical spills, operating personnel and 
When addressing AMC, it’s essential to consider what type of gasses affect the process and at what level. It’s also important to keep in mind that there is no “universal” AMC sensor for monitoring. AMC monitoring is not like 
Can I send data to our existing data management system? Yes. The manifold reads the data from multiple sensors, but provides all data via a single interface using industry-standard 
What processes do you recommend monitoring? This will be determined by the type of manufacturing process you have. For example, the semiconductor industry generally involves photolithography, metals, etch, copper process areas and 
Permalink Comments (4) Posted By Dr. D
I recently spoke with a woman named Becky from 
"While looking at the technical specs, I notice that for counting efficiency, it says 50% per 
The fact is that it is the way particle counters are designed and calibrated. The lowest particle size is always 50% counting efficiency. Lets say we are calibrating the 0.3 
bell curve. As you need to have a cut-off point, you select the middle of the "curve" (or the 50% area), and call that 0.3 microns. In reality, the counter is counting all of the 0.3 microns partilces. It’s more of a term than actually describing the efficiency of the counter. This confuses MANY people!! 