Few issues haunt plasma processing quite like vacuum chamber problems. Plasma etching and plasma deposition tools require vacuum to be able to maintain the plasma and control the gas composition in the process. Other non-plasma deposition processes such as physical vapor deposition require a vacuum for the deposition species to reach the substrate without colliding with gas in the chamber. Creating and maintaining the vacuum can be complicated but is crucial to the success of the process.
The vacuum level required varies by the process. Most advanced plasma processes require a high vacuum, which is defined as pressures of 10-3 – 10-7 mbar [1]. This usually requires a two stage pumping process. A mechanical roughing pump is used to pump the chamber from atmosphere (~1000 mbar) down to around 10-3 mbar. Then, a high vacuum pump such as a turbomolecular pump or a cryogenic pump removes the sparse remaining gases to hold the vacuum at the desired level. Two kinds of pumps are needed because neither can function across the whole pressure range. Roughing pumps can move a lot of gas volume quickly but are ineffective below 10-3 mbar. High vacuum pumps can get to much lower pressures (higher vacuum) but do not function at atmospheric pressure.
To run a highly-controlled plasma process, the vacuum level needs to be steady and on target throughout. Every vacuum chamber will have some level of leakage into the chamber from outside. There is also often process gas being intentionally added to the chamber to provide chemical reactivity. The vacuum pump operates continuously to maintain the target pressure.
However, even if the pump can maintain the target pressure, leaks are still an issue because they are contaminating the process with unwanted gas species. Atmospheric oxygen, for example, is highly reactive and could be disastrous if present in a normally oxygen free process.
The first sign of a vacuum issue is usually that the chamber either cannot be pumped down to the target pressure or that it takes an excessively long time to reach the target pressure. The cause can be quite elusive as there are many possible causes requiring varying levels of effort to resolve. In the rest of this article, we will go over how to identify and resolve a wide variety of vacuum issues in an etch or deposition system.
One of the most straightforward reasons for poor vacuum level is that a vacuum pump is failing. Vacuum pumps require regular maintenance, but even with good care they can have unexpected problems. In a high-volume manufacturing environment, it is a good idea to have spare vacuum pumps on hand to switch in when one is suspected of needing repair. By switching out the pump, you can quickly tell if it was the cause of the issue. You can also continue to run production while the bad pump is repaired.
If swapping out the vacuum pump does not resolve the issue, there is probably a leak of some kind. Finding a leak can be an arduous process, but there are some tricks to make it much easier. The low-hanging fruit of looking for a leak is replacing or cleaning o-rings and gaskets. The chamber door o-ring is an especially likely culprit as it gets a lot of wear from the chamber door opening and closing for each wafer processed. It is a quick and cheap task to swap out the o-ring, so this is often done before bothering with advanced techniques. If the vacuum issue started right after maintenance and cleaning was done on the chamber, special attention should be paid to the o-rings or gaskets in joints that were separated in the maintenance process.
If the leak cannot be found through basic intuition, there are some tools to help locate exactly where the leak is coming from. The best way to pinpoint small leaks is to use the helium detection method. Helium gas is non-toxic, inert, and the atoms are very small. Because of this, it can be used to check for microscopic leaks and will not compromise the safety or the cleanliness of the chamber.
First, the chamber is filled with helium gas to a pressure slightly above atmosphere. This will cause helium to start flowing from the chamber to the air outside through whatever areas are leaking. Then, a handheld helium detector is used to probe all possible leak locations. The helium detector beeps when it finds a large concentration of helium, and the leak location can be investigated.
An alternative to the helium method can be used if helium or a helium detector is not available. The chamber is pressurized slightly with any inert gas (such as N2), and a foaming liquid such as soapy water is applied to all mechanical joints. Where the gas leaks out, bubbles will form in the liquid, visually marking where the leak is located. This method is not quite as effective for tiny leaks as the helium method is, because nitrogen molecules are larger than helium atoms.
The trickiest leaks to detect are “virtual leaks.” If you perform the whole helium leak checking process and do not detect any leaks, it is possible that you have a virtual leak. This means there is gas trapped somewhere internally that slowly leaks into the main chamber.
A virtual leak could be a crack or seam that fills with air when the chamber is not under vacuum. It could also be a pocket of air in the end of a screw hole that slowly leaks around the screw into the chamber when it is pumped down. In either of these cases, the leak will eventually stop when the pocket of air is finally pumped out. However, when the chamber is vented to atmosphere, these pockets can refill with air, and the issue will present itself again.
A virtual leak may not be problem if the air pocket is emptied fast enough. Otherwise, a hole may need to be tapped to relieve the pressure in the pocket if it can be found. The time lost from these leaks will add up over the course of the equipment’s lifetime. If a serious virtual leak is present, it is worthwhile to bring a tool expert on-site to resolve the problem.
Another source of a vacuum issues is a liquid such as moisture or grease that has a vapor pressure near the target vacuum level. If this is the problem, the chamber will pump down fine until the pressure is near the vapor pressure of the contaminant. Then, the vacuum level will remain steady until the contaminant is fully vaporized and removed. A contaminant like this is usually caused when the chamber is opened to atmosphere for maintenance. Water moisture from the air or skin oil (if a technician’s glove contacts their face and then touches the chamber) could cause this.
If this is the problem, it will eventually resolve itself when the contaminant is pumped out, but it could cause the loss of precious equipment time. A good solution is to heat up the chamber to raise the contaminant’s vapor pressure so that it is removed much more quickly.
There are many locations on a vacuum chamber that can leak both externally and internally (virtually). Having deep knowledge of how the specific system works is critical. When a vacuum chamber is not reaching pressure, it is always worth trying the quick fixes such as o-ring replacement or a chamber contaminant bake-out. But if the easier fixes are not successful, it can be valuable to have system experts available from the equipment vendor to work on your tool if you cannot resolve the issue quickly yourself.
Not only is it a loss of plasma tool time and engineer time to try to tackle more complex issues, but a non-expert could actually cause more damage or leakage if they don’t know the ins and outs of the tool. That is why it is essential to choose an etching or deposition tool vendor with the right service plan. Learn more by downloading our eBook “How to Choose a Plasma Etching and Deposition Partner.”
Citations
[1] Vacuum Technique, part of the A. C. G. (2021). The Fundamentals of High, Ultra & Extreme High Vacuum. Vacuum Science World. https://www.vacuumscienceworld.com/ultra-and-extreme-high-vacuum#high_ultra__extreme_high_vacuum_measurement.