Vacuum-deposition process variables
The vacuum process also will have variables, some of which are unlikely to be recorded and which may be of little significance, but this is an assumption unless the variable has been monitored for sometime and proven to have no effect on the product. If the effect is small, it may be hidden by other variables that have a much bigger effect. When logging data, it is preferable to keep logging these variables for a long time until they are conclusively proven to have no effect, rather than discard the data early and later find they do have effects, albeit small.
Process variables can include things such as time between venting and pump-down, atmospheric temperature and humidity, atmospheric particulate count, was the system cleaned after each vent cycle and if so was it a quick clean or a full shield-removal clean, and if there is a collection bucket for the melting frost from the cryopanels, what was the volume. All of these are part of the vacuum-deposition process and may affect the process or are an indication of something different during processing. The amount of water collected by the cryopanel gives an indication if the roll of film had the same water content in the polymer and wound in the roll between film layers, or if the quantity of water was significantly different it could indicate the roll was different or the system has a water leak. But if it is not monitored this difference could go unnoticed.
Cleaning of the system is often related to the difference in pump-down time but not necessarily directly to pinhole levels or barrier performance. The atmospheric conditions also may play a part as the longer the time the system is open the more moisture may be absorbed on the surfaces slowing down the pump-down time. In atmospheric winding systems, the particulate count has been shown to increase when the pollen count is high and this, in turn, affects the pinhole numbers and size. The effect on a vacuum roll-coating process is likely to be less as the roll is unwound only once the system is closed, but there will be extra particles in the system and, due to turbulence during pump-down, these can appear on the coated film.
I hope you can see from this why I may appear skeptical when someone replies that their process was identical, and the output just changed for no reason at all. It is easier to collect data and discard it at a later date, but it is impossible to go back in time for data that was never recorded. It also is common to find that nobody is really interested in adding more sensors and more data-logging, especially when the system is working well, and so the priority is low, and there is always something else that takes precedence. Until, of course, the product fails in which case this leaps up to the top of the list of priorities until the problem is solved, and it then drops like a stone back to the bottom again.
I leave with you the more difficult question: “What else should we know about our process?”