There are several different situations where "exciting" events can occur. Where there are reactive gases then it is very likely that some form of hazardous operation study will have been done and some procedures put in place to avoid any events occurring. This would include using oxygen in the vacuum system either for plasma cleaning or reactive deposition. To prevent too much oxygen being pumped through any oil lubricated backing pump where a thin film of oil with an excess of oxygen and heat from the pump running can become an explosive mixture. In this case the simple method of adding nitrogen to the backing pump inlet to dilute the oxygen to the same ratio (or below) as air would make the pump safe.
In the case of aluminium metallization, we have other things that can occur. Aluminium is very reactive and it has a very high affinity for oxygen and this includes the oxygen available from water. When aluminium is evaporated at very high rates in a metallization process there is a very large surface area of aluminium continuously produced. This is not just the coating on the web but also is the coating on the deposition shields. All of this surface area of aluminium is wants to oxidise, in fact it is impossible to prevent it from oxidizing. Papers have shown that both the surface of the aluminium is already oxidized before it leaves the vacuum system and the aluminium at the interface between the web and aluminium is also oxidised to a similar thickness. As the vacuum system is pumped down to low pressure there is very limited oxygen available for this huge active surface area of aluminium. There is however a large amount of water vapor present both from the outgassing of the surfaces in the vessel but also that is brought in with the substrate roll either in the substrate or interleaved between the layers of substrate as part of the air trapped between layers.
If we take it that the largest amount of oxygen is available as part of the water vapor then the aluminium is stripping the oxygen out of the water. What this means is that there is a large amount of hydrogen that becomes available and will be pumped through the system.
This hydrogen can burn in the presence of heat and oxygen. This may only occur at the roughing pump exhaust where the mixture has access to more oxygen and it may be heard as a popping sound that occurs occasionally. In talking to people this has been heard but as it has never appeared to interfere with production it seems that most often it was usually ignored.
As with pumping other reactive gases one way of reducing the risk is to dilute the gas just before it enters the backing pump. In some processes where it is expected that there will be some flammable gas in the exhaust they put the exhaust gas through a flame to make sure they burn it off and the remaining gas is benign.
If any of you have monitored or investigated the pumping exhaust gas and have details of fires, explosions, or simply information on the quantity of hydrogen passing through the pumps, I and other readers I am sure would be interested to read it, if you care to post a response.