Q: How Clean Is Clean In Vacuum Coating, Or How Long Is A Piece Of String? Part 1 Of 3

By Dr. Charles A. Bishop  
A: It is common for any film-substrate specification to include a phrase about the substrate being clean that can be quite general and without any definition of what is meant by clean. There may be reasons behind the vagueness such as the same film may be used for several different applications, each of which may have a different requirement for cleanliness. On other occasions, the inclusion of the phrase requiring the film to be clean was pointless as the incoming film was never tested. Only if the vacuum-coated film failed to meet the end specification would there be any attempt to find out the possible cause of the problem. 
Testing the incoming film can be time-consuming and costly. There are the problems of the ends of every roll being more contaminated with particles than the rest of the roll, and so how much of the roll needs to be unwound to reach a point where the film is typical? 
It may be more convenient to put the roll onto a winding system, wind the whole roll through, and either take samples periodically or measure the whole film in-line. It is possible the action of winding the film through will increase surface contamination. Airborne particles may fall onto the winding film or may be attracted to the surface by the triboelectric charge on the surface that is generated by the mismatch in materials between the substrate and the rollers. 

Plenty of particles just waiting to contaminate
The number of particles per cubic foot for a range of particle sizes (in microns) taken from a typical industrial atmosphere range from 200 million at 0.01 microns to 900 million at 0.10 microns to hundreds at 10 microns. This shows that there are plenty of particles that could end up on the film surface if it is exposed to the atmosphere either during sample-taking or winding. 
Although the above description shows that the peak number of particles is less than 1 micron in size, there are still a significant number of particles larger than 1 micron and, for many electronic circuits or barrier applications, it takes only one particle to result in a defect that can lead to product failure. 
If the quality of the incoming film can be guaranteed, then it is good policy notto sample every roll as every handling process can potentially reduce surface cleanliness. If the quality of the incoming film cannot be guaranteed, then the decision becomes more difficult. Evaluating every film before coating will be costly and, to prevent reducing the surface cleanliness, a controlled clean environment would need to be established.