So You Think Your Nitrogen Stored Parts Are Not Corroding.
Maybe They Are!

John Franey
            Lucent Technologies Bell Labs

When people think of corrosion they typically have visions of rusted out cars that have seen too many miles of salted pavement in a Midwest environment. The sheet metal of car panels made of steel alloys that are 31750000 angstroms (0.125”) thick can be converted to dust in a matter of 3 years or so. While this gross corrosion is true in the outside environment, in the clean room/micro electronics world catastrophic corrosion occurs on an atomic scale of tens of angstroms.

To make the jump from car corrosion to corrosion in the clean room we are looking at six orders of magnitude less corrosion being critical. This is a very important number to consider as this discussion evolves.

For many decades materials sensitive to degradation in normal atmospheric environments have been stored in protective environments comprised of inert gases. There are many gases on the list. These are Helium (He), Neon (Ne), Argon (Ar), Krypton (Kr), and Xenon (Xe). The problem with the idea of using these gases in very large quantity in industry to protect high volume electronics lines is the cost per cubic foot.

Along came Nitrogen. Nitrogen comprises 78% of the earth’s atmospheric gases. It is plentiful and in great supply. Now we know that nitrogen is not an inert gas and does react with other gases, but it needs a driving force to do that. For instance to make ammonia  (NH4) you must heat Iron to a high temperature and pump Nitrogen and Hydrogen over the hot iron to cause the reaction to take place. Even with that only about 15% of the gases react at a time. Given the forces needed to cause a reaction to take place with Nitrogen it was found that if the nitrogen was dried out (water removed) the resulting gas was very inert. Most factories use storage tanks of liquid nitrogen refilled by suppliers to charge common supply lines in the plant with the boil off gaseous nitrogen. When manufactured as a gas it is relatively pure and dry. The process Liquid nitrogen is the same nitrogen that makes up 78 percent of our atmosphere, just much colder. Nitrogen boils at minus 196 degrees Celsius.  So, to make it into a liquid we have to cool it down to at least minus 196 degrees Celsius.  At normal temperatures, of course, nitrogen is a gas.  Like any gas, its molecules are spaced relatively far apart, and are moving very rapidly.  Because of this, the nitrogen molecules don't interact with or attract each other.  To turn nitrogen gas into a liquid, we need to slow the molecules down by cooling them.  We start with air.  Air is compressed in a compressor, pushing the air molecules close together.  Then the air is allowed to flow down a tube through a chamber.  For the gas to expand in the chamber, the molecules need energy.  They get this energy by absorbing heat from the chamber, cooling it.  The process is repeated until nitrogen begins to condense into liquid.

As you can see the refining process provides very clean gas. This is where the problem is. Very clean may not be good enough for a clean room environment.

Normally very clean gases have a stated purity of 99.9%. That is an impressive number. However, if you remember the levels of corrosion that we stated as being a problem in clean rooms being six orders of magnitude less than outdoor corrosion on cars we see that 0.0001% is a significant number. The 99.9% purity can actually have impurities on the order of 1,000 parts per million (PPM). The significance of that is that is that corrosive gases with concentrations of 0.01 PPM (10 Parts Per Billion (ppb)) are known to cause catastrophic corrosion on pure reactive surfaces such as exist in clean room environments.

To carry the scenario to a actual situation, the Nitrogen gas boil off from in-house liquid nitrogen supply line was analyzed for Sulfur Dioxide (SO2) over a 24 hour period. The result is shown in the graph below. 

As you can see in the graph above the corrosive SO2 can show up in your inert gas with concentrations ranging from 30 to 70 ppb. That is a considerable amount of corrosive gas.

One major factor that will save you is that SO2 and many other gases nee water (H2O) or some other catalyst to start reacting with copper et al. There is very little water in the N2 supply.

However one must consider that all gases will absorb onto and into all materials. This means that a certain amount of residual corrosive gas will be on your product, and sooner or later the product must come out of the ‘safe’ environment you have supplied for it and be exposed to the earth’s normal atmosphere which contains plenty of water to drive the sulfur to react with any molecule that has the smallest propensity to be coerced into joining atoms. This is when you will find the product has a certain amount of corrosive gas reactions and yield hits take place. Usually in the form of impure surfaces that do not attach well whether the attachment is an epoxy bond or a weld.

So the moral of this story is : Your clean Nitrogen may not be clean enough.

There are many materials out there that getter unwanted gases, many of them are used in molecular sieves and just plain absorbers. Look into them. It is cheap insurance to have for the sake of loosing a wafer, of say, GAs amplifiers.

Also consider protecting you product during shipping and storage from the corrosive atmospheres residing in storage houses, trains, planes, and ships.

Yours truly,
John P. Franey