That Day, After PM, The CPK Report Came Out, And The Whole Room Was Silent For Three Seconds
I still remember several years ago, we had a critical machine where the number of particles on the wafer surface suddenly surged. When the measurement results came out, the CPK value for that batch of products plummeted directly to 1.08, and DPMO soared to 6210! The production line was completely stunned, and the PM supervisor's face turned green. Everyone was at a loss; we had clearly followed the equipment PM procedures according to SOPs, so how could this happen? After a long investigation, we finally discovered the problem lay in a small, seemingly insignificant detail during the PM process, a detail intrinsically linked to the "cleanroom" environment.
Where The Problem Lay
Frankly speaking, many times we treat equipment PM as a simple process of disassembly, cleaning, and parts replacement. However, in a cleanroom, this logic can prove very costly. A cleanroom earns its name because its environmental control is extremely stringent. a single oversight from your end could introduce "contaminants" from the outside world. These contaminants are what we refer to as "particles." These invisible particles are deadly killers for wafers; at best, they affect yield, and at worst, they lead to direct product scrapping.
So, the key is that when performing equipment PM in a cleanroom, you must not only think about "fixing the equipment" but also "how not to contaminate the environment." In other words, if you fix a piece of equipment but cause a million particles to fall onto the wafers, then this PM was entirely in vain, or even detrimental.
What To Actually Do
Frankly, this tests your "cleanliness awareness." For instance, we often need tools during PM. Would you casually place tools on the equipment casing? Wrong! Tools themselves accumulate dust. The moment you place them down and pick them up again, those dust particles could fall into the open interior of the machine.
The correct practice is that all tools entering the cleanroom must be wiped clean and placed on a cleanroom wipe. To be even more meticulous, for PM of some precision components, we even require it to be performed under the machine's built-in FFU (Fan Filter Unit) to ensure the local environment achieves ultra-cleanliness below Class 100. Furthermore, the final cleaning step after PM cannot merely involve wiping with a cleanroom wipe; you must use a Particle Counter to measure the number of residual particles in critical areas to ensure it meets standards, for example: the particle count per square foot in a certain area must be less than 0.1. This is genuine validation.
The Most Common Pitfalls
The most common pitfalls I've seen are "cutting corners" or "habit becoming second nature." For instance, once, a new engineer, while performing PM on a CVD machine, for convenience, directly unwrapped the new spare part's packaging on a table next to the machine. What was the result? The plastic packaging of that spare part, at the moment of unwrapping, generated a large number of tiny plastic debris. These invisible debris drifted directly into the machine's interior with the airflow.
Subsequently, the yield of that batch of wafers plummeted. After tracing it back, a look at the SEM images revealed a large number of irregularly shaped organic particles on the wafer surface. It was then discovered that this was due to the "convenience" taken during that PM. To be honest, many people might think, "What difference does one particle make?" But when you accumulate dozens or hundreds of them, the yield is ruined.
One Thing You Can Do Today
During PM, imagine you are a particle. Where would you most want to hide? Avoid it.