What is AMC?
Airborne Molecular Contamination (AMC) is chemical contamination in the form of vapors or aerosols that has a detrimental effect on a product or a process. These chemicals may be organic or inorganic in nature and includes acids, bases, polymer additives, organometallic compounds and dopants. The main sources for AMC are building and cleanroom construction materials, general environment, process chemicals and operating personnel.
What are the effects of AMC?
AMC can cause a multitude of adverse effects such as:
· Corrosion of metal surfaces on the wafer
· Degradation of HEPA/ULPA filter media
· Haze on wafers
· Haze on optics
· T Topping of chemically amplified photoresist
· Changes in contact resistance and voltage shifts
Who is monitoring AMC in my industry?
Almost all leading edge semiconductor companies are doing some type of AMC monitoring on a real time basis. This monitoring has been traditionally focused around the photolithographic area, but the area of coverage is now extending into other process locations.
My lab already does testing for AMC. Why should I use a monitoring system for AMC?
In general the testing done by most labs is static. This means the data cannot show actual trends over time, it can only show an average concentration level. The most common form of testing done by a lab is impinger testing. An impinger is put out into the environment to be tested for a fixed length of time. The sample collected is then analyzed for chemical concentration levels. The data provided from this type of testing only shows an average concentration level. Online monitoring gives you the ability to see AMC levels in real time. It can show you whether concentration in AMC is at normal background levels or is a specific contamination event, where the low and high phases are in the daily cycle.
What compounds do you monitor?
We are able to monitor all types of compounds. The most common chemicals that are monitored are as follows:
- Ammonia
- NMP
- Total Amines
- Total Acids
- Total Sulfur
- H2S
- HF
- HCL
What are the minimum detection limits?
The minimum detection limits depend on the type of chemicals you want to sample and the technology you want to use. Some technologies can view chemical concentrations into the part per trillion level; others in the parts per million level. It is important to first understand what the requirements are for your process and then to determine what the appropriate detection limits should be.
How many points can I monitor?
The AMC Manifold can be configured to sample anywhere from only 1 location to as many as 64 locations. If more than 64 locations are needed, multiple AMC Manifolds can be combined into a single system with up to thousands of sampling locations.
How often can I get samples?
The frequency of sampling from each location is determined by 3 things: number of sample locations, purge time, and sample time. Together, these values determine the Total Cycle Time – which is the time it takes for the manifold to sample from all locations and return to the original location for another sample. The Total Cycle Time is determined as follows:
Number of Sample Locations x (Purge Time + Sample Time)
So, for example, if you have 12 sample locations, a Purge Time of 5 minutes, and a Sample Time of 1 minute, your Total Cycle Time will be:
12 * (5 min + 1 min) = 72 minutes
Number of Sample Locations:
Naturally, the more locations you sample, the longer it will take to cycle through all locations and take another sample at the original sample location.
However, the AMC Manifold also allows you to choose specific sample locations for higher priority sampling, allowing you to sample multiple times from specific locations during the course of a cycle.
For example, you could choose to sample from a few particularly sensitive locations 3 times for every 1 time you sample from all other locations. This allows you to expand your sampling system without sacrificing the speed at which you sample more sensitive locations.
Purge Time:
After the manifold changes to a new sampling location, it waits for some time to allow the air from the previous location to be replaced by the air from the new sample location. This is called the Purge Time.
The value of the Purge Time is dependent on the response time of the sensors used, not on the manifold. Sensors with slow response times require longer purge times – which in turn increases the Total Cycle Time, reducing the frequency with which you can sample each location.
Even with sensors that have fast response times, the minimum recommended Purge Time is 5 minutes; for sensors with slow response times, the Purge Time may need to be as long as 30 minutes. It’s for this reason that it is important to select sensors with fast response times.
Sample Time:
Lighthouse recommends that the Sample Time is set for 60 seconds. This gives the sensors enough time to get a valid sample, and gives the AMC Manifold enough time to accurately determine the stability of the sample (regardless of the sensor used).
What is required to maintain the system (calibration, gases, etc…)?
There are two parts to this question.
1. First for the sampling system, the unit requires very little maintenance. The vacuum pumps will need to be maintained on a quarterly basis.
2. The maintenance requirements for the analyzers will be dependent on the type of analyzer used and the gases being monitored. Different analyzers will have different calibration requirements. Calibration frequency is often dependent on the desired accuracy of the sensor. Some sensors come with on board calibration systems while others require external hardware and standard gases to calibrate them.
Where should we install the sample points?
Sample points should be installed as close to critical processes as possible without interfering with the processes. It is common to install points both inside and out side of a process so if an increase in AMC does occur, it can be determined whether the increase came from the ambient environment or from within the tool itself. It is also common to install sample points up and down stream from the chemical filters. This will help to understand the removal efficiency of the chemical filters.
How much does a system like this cost?
The cost is dependent on what chemicals you want to monitor and how many locations you want to sample from. Prices can range from a low at $3,000 for a single sensor to $400,000 for an entire system sampling multiple locations.
Why do a I need to monitor AMC?
Monitoring for any type of contamination is an important aspect of contamination control. Monitoring specifically for AMC is important in industries where AMC can directly affect the product or process. Even the newest state of the art facilities are not immune from AMC related incidents. Incidents such as spills or contamination episodes result from tool or equipment failures and associated maintenance. Chemical filtration is affected by the environment; changes in the environment may result in performance changes in chemical filtration. Only continuous AMC monitoring can provide assurance that the facility is performing properly and can alert personnel when an incident has occurred. This type of monitoring allows for rapid responses to incidents that can be carried out immediately instead of days or weeks after the facility has been contaminated.
What sensors do you recommend we use?
AMC monitoring is very specific for the application; therefore, the sensor used should be based on the application. When picking a sensor you should consider 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
- Operation cost
Can I send my data to our existing data management system?
Yes. The AMC manifold reads the data from multiple sensors, using different protocols and signals but provides all the data via a single interface, using the industry-standard Modbus protocol. Almost every commercial automation and control system on the market – including most legacy systems – can read data using the Modbus protocol.
Do your instruments have any radioactive materials in them?
Lighthouse does not make AMC analyzers. We integrate various analyzers from different instrument suppliers into the sampling manifold. This allows us to match different techniques to provide a broad range of AMC monitoring.
Some instruments use a radioactive source to ionize the sample. This is found most commonly in sensors utilizing Ion Mobility Spectroscopy as an analysis technique. We recommend that you ask each instrument supplier this question.
How long can the tubing runs be from manifold to sample point?
The runs can be as long as you like, however, the longer the distance, the greater the chance for contaminating or diluting the sample. Contamination of the sample can come from leaks in the sample tubing. Dilution will come from part of the sample being absorbed by the sample tubing material. We recommend not running sample tubing any longer then 80 meters. At longer distances, it is a good idea to use a booster pump to maintain adequate sample flow.
What materials are used for the tubing?
The most commonly used sample tubing is teflon. Stainless steel can also be used, but it tends to be more expensive and is not compatible with all chemicals.
How frequently should we calibrate the sensors?
Frequency of calibration will depend on three factors.
1. The Zero Drift of the sensor per day – this is the amount of drift the sensor will experience from zero in a set period of time, normally a day or week.
2. The Span Drift – this is the amount of drift the sensor will experience from a fixed concentration amount over a day or week period of time.
3. What is the target level of detection you are looking for? The lower the level, the more frequently you will need to calibrate to keep the zero and span drift from growing larger than the minimum detection limit you want to achieve.
Why doesn’t the data collected match our impinger sample?
There are two parts to this answer.
1. Different analysis techniques will tend to produce slightly different results. Some techniques are more prone to interference and thus can show drastically different results.
2. An impinger sample is taken over a set period of time, normally 2 – 24 hours. Thus impinger data will only show an average concentration over the period of time the sample was taken. Real-time instruments will give a continuous readout and are designed to show trends in AMC levels.
What do you recommend to monitor in my process?
This is dependent on what type of manufacturing process you have.
Semiconductor industry:
- Photo lithography
- Metals
- Etch
- Copper Process Areas
- Photomask Manufacturing
Hard Disk Drive manufacturing:
- Wafer Operations
- HGA/HSA Operations
- Final Drive Assembly
- Media
How many manifolds do I need?
A single AMC Manifold can monitor up to 64 sampling locations and can collect data from up to 10 different sensors, all at the same time, depending on the types of sensors used and their output signal. Please contact your Lighthouse sales representative for details of how to integrate your specific sensors into the Lighthouse AMC Manifold system.