Purified (< 0,01 microsiemens/cm) water corrosive? 4

[Skogsgurra]

Hello,

I have seen some corrosion in silicon carbide seals that are used in process water pumps in medicine fabrication. Some people think that the corrosion is electrical and that could very well be so - I have measured high frequency currents (from frequency inverters) in the 100 mA order of magnitude.

But some guys tell me that this kind of water is very aggressive and corrodes stainless steel, and probably SiC seals as well. I know nothing about these chemical things and ask you if that is true or not. Any examples? Remedies?

 

[Metalguy]

Generally speaking, the lower the cond. of the water the better. Corrosion is electrochemical, and reducing the cond. favorably affects the electo part-usually. But it isn't so simple because many ions affect the cond. Some are "more corrosive" than others.

Oxygen is also a big factor, and doesn't affect the cond. In some cases the water becomes less corrosive if some metal ions are present-so called water hungriness. SS needs small amounts of O to remain in the passive state, which can greatly lower corr. rates. But more than that amount can/will increase them.

What temps. are involved? How are you measuring the current? How close is the corroding part located to something that could be cathodic to it? If very close (~1" or less) there may be a galvanic effect, even tho the cond. is low.

I'm not familiar with the corr. if SiC, but someone here probably is.

 

[moltenmetal]

Interesting problem. To my knowledge, SiC is not attacked by water, steam or condensate, to quite high temperatures. I know of no reason that removing the last few ppm of impurities from the water should matter. I can't comment on your induced current theory without knowing the particulars, though it's clear that you can dissolve many things if you provide enough electrochemical driving force.

Ultrapure water is also, to my knowledge, NOT corrosive toward stainless steels. There may be issues with minor quantities of metal ions released from stainless steels which may result in contamination of the ultrapure water, particularly soon after the material is brought into service, but that's not the same as saying that ultrapure water will aggressively corrode the stainless steel. It's this common misconception that your colleagues are probably repeating. Hot, oxygenated pure water IS aggressive toward copper and brasses (particularly in areas of high local velocity), but in closed deoxygenated systems both of these materials also stand up quite well- regardless of how pure the "pure" water is.

Are you certain that the damage you're seeing is corrosion? Could it not be disintegration caused by poor-quality silicon carbide? Or perhaps it's just wear? Water is very non-lubricious.

 

[Skogsgurra]

Metalguy,

Thanks. I do not know about the O2 contents. Will ask the specialists. The temperatures are same as environment - i.e. room temperatures. I measured the current with a rogowski coil around the shaft going from motor to impeller, the seal is between this shaft and house.

The SiC is inserted into the SS. It is a ring, so it is close to the SS. Do not know if SS is cathodic w/ respect to SiC. Do you?


Molten

Your statements support my thinking. Thanks. Interesting part that the water gets contaminated in the beginning and that this can be mistaken for corrosion. I think that it can be the explanation I need, but not quite sure yet.

We had a specialised institute (Norsk institutt for energiforsking) investigate the damages and they think that it is corrosion.

Thanks guys. A small step for mankind. A quantum leap for me ;-)

 

[IRstuff]

DI water can be corrosive in the sense that some materials such as glass will be attacked or dissolved into the water.

There was a famous incident about 30 yrs ago where a university lab announced the discovery of "polymerized water," which had properties unlike any other form of water. After about six months or so, they, much to their chagrin, announced that it was simply plain water and that the beaker glass had dissolved into the water and thus had contaminated their tests.

However, the concentrations were extremely low, so it's unlikely that you would actually see physical corrosion effects. Since you are measuring currents, then it's certainly more likely that there is some sort of electrolytic reaction going on.

TTFN

 

[moltenmetal]

To my knowledge, conventional glasses are not soluble in water. Glasses do dissolve in strong base, but not in water, regardless of the water's purity. Sodium silicate ("water glass") is soluble, but again is strongly alkaline. At pH neutral or lower, silicic acid (silica gel) precipitates.

 

[unclesyd]

Water is corrosive to all common glasses. About 10 mils a year on your glassed lined water heater, though there has been some improvement in the last few years. That is also why you analyze sodium in plastic equipment.

Extremely Pure water has been known to attack silicon carbide. I'm trying to find the reference at the present, as I think there were other parameters involved.

We tried SiC/SiC seals in pure water total disaster. We went to C/Si seals evidently they are performing well.

moltenmetal,
Sodium silicate is an excellant solvent for glass, especially fiberglass.

 

[Skogsgurra]

Very interesting, uncle Syd!

That could easily be the problem we are having. I have done some searching myself but not found anything substantial. I would apreciate a reference or two.

 

[IRstuff]

http://www.newton.dep.anl.gov/askasci/chem00/chem00528.htm

http://www.corrosion-doctors.org/Household/Glass.htm

http://www.osti.gov/bridge/product.biblio.jsp?osti_id=787189

http://www.isis.rl.ac.uk/isis2000/highlights/GlassCorrosionH10.htm

not quite on subject:

http://www.corrosionlab.com/Failure-Analysis-Studies/28059.corrosion.zinc-housing.htm

http://corrosion.363.net/glass/accessory/resistance.htm

http://www-atalante2004.cea.fr/home/liblocal/docs/atalante2000/O3-04.pdf

http://nvl.nist.gov/pub/nistpubs/sp958-lide/181-183.pdf

TTFN

 

[Skogsgurra]

Well, I think I had better leave this to the chemists. It is obvious that something outside of my box is happening to the silicon carbide. BTW, is silicon carbide similar to glass? Thanks for the references, IRstuff.

 

[moltenmetal]

IRStuff & Unclesyd- this is absolutely fascinating! Now that I've been through the references, 10 mil/yr corrosion on a glass liner in pure hot water service (i.e. even scrupulously free of fluoride and base) is totally believable! It seems that even traces of silica in the water are inhibitory toward corrosion, so ultrapure water is particularly nasty in this rare instance.

I can certainly understand how sodium silicate would definitely be nasty toward glass. That doesn't surprise me in the slightest.

I've personally seen SiC used successfully in pump seals and valve trims in all sorts of water services and even some highly acidic water/steam services. If you have any references on SiC corrosion in ultrapure water I'd be very interested to see them!

Thanks again, guys, this is extremely helpful and useful stuff!

 

[Skogsgurra]

Moltenmetal wrote: "If you have any references on SiC corrosion in ultrapure water I'd be very interested to see them!"

Comment: Me too!

 

[sreid]

I would not totally discount the electrical theory. The ball bearings in inverter driven motors have been destroyed due to ball to race electrical arcs. Are you getting pitting on both the shaft and the seal?

 

[unclesyd]

Don't put me down yet, I'm still looking for the reference that I mentioned concerning the pure water and SiC. I haven't found it as yet since when I retired I never had expected to need a reference on SiC.
I did find a reference to SiC vs SiC running dry, not very good.
Also I made a note that one of the silicon carbides absorbs water (2%-5%) but made no notes of which one sintered or reaction bonded.
I am still looking, as I am looking also for some old steel spec's.