Brass fitting corrosion

[Mkuchta]

I have an corrosion mechanism in a machineable brass Swagelok fitting (copper,30%zinc,2.5% lead). The fitting has was in use 20+ years in a PH 10 steam environment with <10ppm of hydrazine (which decomposes to ammonia). The part failed at the first thread of the fitting. The fitting appears to be yellow brass colored on the outside with a pink, copper color on the inside. I initially thought it was dezincification however EDX analysis showed almost identical composition in both areas. The only difference was a small increase in oxygen in the copper colored area. The yellow brass area shows a normal microstructure after etching however the pink copper area shows a finely porous surface with no grain structure. If this is not dezincification, what else can it be?

 

[EdStainless]

The Cu has been dissolved and then re-plated out on the surface. High pH and ammonia are a good reason to not use a Cu containing alloy. I hope that no one got hurt.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[metengr]

From what you described, it sounds like a dezincification mechanism.

 

[Mkuchta]

Yes it does sound like dezincification but there are no areas of only copper- the zinc is still there. After EDX analysis showed these results I took the sample to another lab for verification. I think what I am looking for is some other corrosion mechanism.

No one got hurt and all fittings are being replaced with stainless.

 

[swall]

I would suspect SCC due to the amines.

 

[ApC2Kp]

Mkuchta,

I would tend to agree with the SCC as suggested by swall.

Some people might have heard that SCC was first observed in brass cartridges. During the civil war, some army units had stored brass cartridge ammunition in chicken coops. The ammonia from the chicken manure caused SCC failure of the brass. Modern cartridges now require annealed brass.

The pink copper color inside the fitting could be the remainder of thread sealant compound. Difficult to say what thread compound might have been used 20 years ago. If Teflon tape was used for thread sealant, then the steam temperatures might have been hot enough to decompose Teflon. Chlorine and fluorine from decomposed Teflon would be very aggressive on brass.

 

[rconner]

Was just curious was the environment the same (as described) inside and outside the fitting involved? What exactly do you mean by "last thread" location, and what was the specific environment to which the brass speciifcally at the fracture location exposed? Lastly, to what specifically was the threaded fitting attached?

 

[Mkuchta]

The fitting is exposed to the steam/ammonia/amine environment on the inside only. The copper colored discoloration is in the areas exposed to the condenser environment. The outside environment is room temperature air and no discoloration was seen in these areas.

The way I understand it,SCC is a cracking mechanism. This copper colored material on the ID of the fitting is not SCC. It appears to be oxygen related corrosion. This copper colored area is a woody, slightly porous material with no grain structure. EDX dot maps show copper, zinc, and lead separated out in what appears to be flakes (seen at 8k magnification in the SEM). There is slightly less copper in the porous area but the major difference is that this area has more oxygen. This is unusual for a condenser where oxygen is kept low however this may be an area that has not been purged. This fitting has been in use 20+ years and currently all brass fittings are being replaced by stainless.

Teflon tape is not used as a sealant. EDX analysis did not show any fluorine or other unexpected contaminants. In any case, the tape would be on the outside of the threads. This copper colored corrosion is on the inside where it is exposed to the environment. The best I can come up with is that it's an oxygen corrosion mechanism.

 

[btrueblood]

Ed pretty much nailed the problem in his post on March 16th. But to try and elaborate a bit on his explanation, and assuming this fitting is a stagnation region, I would propose the following theory. The oxygen probably appears in a transient, each time the system is drained for cleaning or other maintenance. Thus, the corrosion of the fitting i.d. surface proceeds during this downtime. Subsequently, the hydrazine-rich fluid re-wets the surface, and the oxides are reduced back to base metal. But, the base metal is a coarse, porous coating now, since the oxide took up more room on the surface. Subsequent corrosion cycles, therefore, can allow penetration of the surface layer with each aeration/deraeration cycle, producing the porosity you see (until the fitting is penetrated sufficiently to fail due to the pressure load).

Cu + O2 -> Copper oxides

then some time later

Cu oxide + N2H4 -> Cu(metal) + H2O + N2 (gas)

I'm neglecting the role of ammonia, but it pretty much is a side player here. Ammonia (well ammonia+water) will dissolve the oxides, and tend to wash the surface, and you would see something more like pitting corrosion; probably both reactions occur (otherwise you'd see growth of the surface towards the i.d. as the base metal becomes more porous or less dense).

Clear as mud? I probably am missing something too.

 

[btrueblood]

Sorry, I forgot the other part of this, which is that it took 20 years for the fitting to fail. This is also best explained by transient exposure to oxygen, which limits the oxidation to those times when the system was drained (no ammonia to carry away the oxides then). The brass surface then is really only being "attacked" during re-fills of the system, when the surface oxides are being reduced back to base metal. Once the oxygen is scavenged by the hydrazine, further attack stops. Until the next drain/refill cycle.

...figure one exposure every year, 20 years, perhaps .005 inch of corrosion progress per cycle...sound plausible?

 

[metengr]

btrueblood;
Normally, I could not agree more with your posts. However, in this case not sure that I would agree. The zinc is still present as mentioned above. If you had copper remaining with no zinc, this would be dezincification. In this case, the OP has a layering effect with copper and zinc.

 

[metengr]

Here is a quote from ASM handbook on Corrosion regarding dealloying

Dealloying
Dealloying is a corrosion process in which the more active constituent metal is selectively removed from an alloy, leaving behind a weak deposit of the more noble metal. Copper-zinc alloys containing more than 15% Zn are susceptible to a dealloying process called dezincification. Selective removal of zinc leaves a relatively porous and weak layer of copper and copper oxide in brass. Corrosion of a similar nature continues beneath the primary corrosion layer, resulting in gradual replacement of sound brass by weak, porous copper that would eventually penetrate the brass, weakening it structurally and allowing liquids or gases to leak through the porous mass.
Plug-type dealloying refers to the dealloying that occurs in local areas; surrounding areas are usually unaffected or only slightly corroded. In uniform-layer dealloying, the active component of the alloy is leached out over a broad area of the surface. Dezincification is the usual form of corrosion for uninhibited brasses in prolonged contact with waters high in oxygen and carbon dioxide (CO2). It is frequently encountered with quiescent or slowly moving solutions. Slightly acidic water, low in salt content and at room temperature, is likely to produce uniform attack, but neutral or alkaline water, high in salt content and above room temperature, often produces plug-type attack.
Brasses with copper contents of 85% or more resist dezincification. Dezincification of brasses with two-phase structures is generally more severe, particularly if the second phase is continuous; it usually occurs in two stages: the high-zinc ? phase, followed by the lower-zinc ? phase.

 

[btrueblood]

metengr,

Sorry, I knew I wasn't being clear. De-zincification implies the zinc goes away entirely (due to its oxide's high solubility in water). Here, the zinc (and other alloying metals) is getting left behind in the porous deposit, although we don't know how much relative to the original alloy content. I'm just trying to point out how the zinc and other metals could get reduced and re-deposited by the hydrazine. I used copper in the illustrating reaction by mistake. The zinc (and lead, tin, etc.) can get oxidized and reduced in the same manner as the copper, only at different rates depending on pH, nobility of the base metals, etc.

It is possible, the copper is not oxidizing at all, just the base metals, but then the re-flooding of the area with hydrazine solution is reducing and redeposting some of the zinc/lead/tin ions.

Dunno, not a materials expert, just know a little bit about hydrazine, and wanted to throw out the possibility of a cyclic oxygen exposure to account for the corrosion being seen here.

 

[Mkuchta]

FYI, EDX quant analysis showed no difference between the normal and corroded areas. Dot maps showed slightly less copper in the corroded areas. Thanks everyone for your ideas. Btrueblood's idea seems reasonable. I've never seen this before and haven't been able to find it documented anywhere so I may never know.