Common Pumping Problems.

Water has to be the biggest cause of pumping problems.  The source of the water can vary from, a wetter than normal substrate, more ambient moisture absorbed when the machine is at atmospheric pressure or even a water leak directly into the system.

Other problems can relate to a lack of routine maintenance or too infrequent maintenance.  This can be everything from not cleaning the deposition shields frequently enough to not checking the oil levels in the pumps or ballasting the roughing pumps to remove any excess water vapour.

Below I will highlight each of these problems and some of the methods for identifying the possible source of the problem. 

Let us look at the sources of water in vacuum in a little more detail.  The rolls of substrate material can vary in moisture content. There are two sources of water within each roll, the water that is absorbed into the substrate and secondly the water that is contained in the air trapped between the layers as the roll was wound up.  The water within the substrate is dependent upon the type of material as well as the humidity during both manufacture & subsequent storage.  The substrate material can vary from those with almost no water absorbed to paper that probably has the greatest amount of water absorbed.

Once the system has been pumped down to the desired base pressure if the winding system is started it is common to see a rise in pressure as more of the trapped gas is released from the roll & also as more of the substrate surface area is exposed to the vacuum.  After the surface water has been stripped out the rate of outgassing from the substrate surface will be limited by the rate of water diffusion from the bulk of the substrate to the surfaces.

All the rest of the vacuum system will also be giving up water vapour from the surfaces.  The amount of water absorbed depends upon the surface structure.  The deposition shields get coated by material with low energy & often at angles other than normal to the surfaces which results in very porous structures.  These porous coatings have a huge surface area and can act like a sponge for water vapour that gets absorbed as the system is brought up to atmospheric pressure.

What often gets overlooked is that beyond the deposition shields there will still be some scattered material that gets past the deposition shields and coats the rest of the chamber or other surfaces.   Whereas the deposition shields do tend to get cleaned periodically the rest of the chamber may not & so over a long period of time the coating will build up and will also present a highly porous very high surface area coating that will also absorb water vapour like a sponge.

These surfaces will tend to give up the water vapour at a faster rate when subjected to heating.  Thus when the deposition sources are heated up a pressure rise can also be expected as all these shields are heated up.  These shields often have a large thermal mass & so they will continue increasing in temperature throughout the process cycle encouraging ever more water vapour to be released. 

Some operators use dry air & others pure nitrogen when bringing the system to atmospheric pressure to try to minimise the amount of absorbed water vapour.  Another simple method I have seen used was to have the normal air pass through a tube of silica gel as is drawn into the vacuum chamber when vented.  As the air passes over the silica gel it is dried and if the tube is transparent the change of colour of the silica gel can be monitored and as the colour changes with increasing water content it can be dried periodically & refreshed.

In general most vacuum systems benefit from the use of cryogenic panels or coils that specifically are used to pump the water at very high speed.  This is particularly so for paper metallizers where the paper may have approaching ~10% water content.

Another different way of working that has been used to good effect is to change the normal pumpdown procedure.  Often the pumpdown cycle involves roughing the chamber to a fixed crossover pressure & then to use the diffusion pump and pump down to a low base pressure.  However pumping to a low base pressure can give rise to some oil backstreaming and it does not necessarily help get rid of the water vapour very quickly.  Some have used a different pumping protocol to improve the speed of desorbing the water from the chamber surfaces.   The system is pumped down using the roughing pumps and an inert or dry gas is introduced for a period of time as the pressure approaches the cross over pressure. The benefit of this is to increase the time the surfaces are bombarded by the gas whist under some vacuum. As the surface is bombarded the energy can help remove some of the surface water vapour.  If the pump down were continued to use the diffusion pump the pressure would fall & the surface bombardment would decline with the reducing gas density thus the speed of which the water is released would decline.   By holding the system at a high pressure the increased bombardment makes the water removal more efficient & so when the change is made to the diffusion pump the pressure fall will be more rapid & in some cases the base pressure is reached more quickly than by the more conventional pumping process.

With the high levels of moisture present and the repeated high load from pump-downs it is important to manage the pumping system well.  It is possible to lose diffusion pump oil through to the roughing pumps and so their oil level needs to be monitored regularly and topped up as required.  Some of the roughing pumps can accumulate the water vapour & this can turn the oil into an emulsion. Often this just requires the pumps to be ballasted which is simple passing some extra air through the system causing the pumps to run hotter & the water vapour becomes entrained in the air and is exhausted.   I have seen a number of systems suffering from these basic problems.  Whilst the diffusion pump can still work with the oil level below the marked minimum the pumping performance is reduced & the risk of cracking the oil increases which can then mean a major cleanup & extended downtime hence it is easier to maintain the correct oil level.

It is only by routinely monitoring the pumpdown performance by plotting the pressure change with time & marking the pump changeover points that it is possible to compare the current performance to previous ones.  Ideally there will be a pumpdown graph for the system when it was new or when it was clean & dry. This gives a benchmark for what can be achieved.  Typically operators use the time to the desired base pressure as a trigger for action. If the time takes too long then the system will get cleaned out next time around. This will be expected to improve the time because the huge water vapour absorbing high surface area coating will have been removed.

With small water leaks there can be a problem in identifying them until the system gets cleaned out as the absorbed water can be large enough to hide small leaks & it can be they remain unidentified until they increase in size.  Sometimes they can be identified during the pumpdown because the pressure does a few cycles of pressure rising & falling as the water from the small leak freezes as the water evaporates & so the evaporation rate falls & so to does the pressure. This is followed by the ice melting & a sudden surge of gas causing the pressure to rise again & the fast evaporation causing the water to freeze again & hence the pressure to decline again, as before. This cycling sometimes stabilises and the crossover pressure is never achieved but other times the crossover is eventually reached but a poor base pressure is reached.

The main recommendation from all of this is that routing cleaning & maintenance is not only good practice but also can be both cost effective and it will also help in the production of high quality coatings because of less contamination.

Comments

[Rick Spencer]

Hi, this is going to sound like a plug for my website, but there's a paper there that might be useful. The paper is on the best cryo-coil placement to optimize your vacuum level.

The paper was presented by Charles at last years SVC, "Coating improvements through better control of water vapour in vacuum deposition". It's available for download at :-
http://www.alacritas-consulting.com/downloads.html

regards

Rick Spencer

[bryan]

When pumping a balzers 500a diffusion pump it gets to all the right settings and pressures but once i go into high vac and bleed in gas for an ion etch the chamber pressure starts to rise and the diffusion pump starts to stall could this be a cooling problem with the pump.