Dark Patches on Bores of SS321 Tubes 7

[deics]

Problem - After a tube leaked in a new exchanger under a hydro test all tubes were checked with ECT and some were found to give indications above reference level.

Tubes were checked by fiberscope. Spiral marks resembling roll straightener marks were seen in the bores. The tube that leaked had a fissure in the tube associated with a spiral roll mark.

The tubes were pulled, hydrotested at 1000psig re-checked by ECT and UT and even at the Mill. RT was performed at location of indications. No defects were found.

However careful rechecking by fiberscope showed dark, discolored patches in the bores. Micro-examination showed there to have been an aggravated intergranular attack by the pickle after solution anneal in these locations compared with the usual light coloured bore surface to a depth of about 0.030mm.

Also there was a reduction in thickness in the dark areas of about 0.020mm.

The dark patches are caused by traces of lubricating "soap" used for the cold drawing of the tubes that appears to have remained in the bores during the degrease before annealing.

Carburisation and grain boundary carbides are postulated as the cause of the aggravated attack by the pickle in the dark areas.

The tubes have been characterized as "corroded" by the Client.

There remain in the bundle tubes with indications below the reference level. Fiberscope shows that they have similar dark patches. ECT in Phase Analysis Mode shows reduction of thickness of <0.020mm.

Question 1: Are such a variations in the wall thickness as a result of of a variable pickle at the Mill usual?

Question 2: Both Light and Dark Areas show intergranular attack although the dark area is worse the depth of the attack is about the same. SEM MPA shows identical and adequate compositions for this material. Based on this information:

a)- Could the corrosion resistance of the two conditions be expected to be significantly different or about the same?

b)- Is there some way that the tubes can be checked to demonstrate that they are not sensitized in the dark areas?

The solution anneal will have dissolved the grain boundary carbides and the water quench will have kept the carbon in solution plus the Chromium would diffuse back to the depleted zones but maybe some degree of hysteresis would result in retained sensitized areas, and that is what the pickle attacked....

Question 3: A piece of this tube was re-pickled to remove the dark areas. When checked by ECT in Phase Analysis Mode it showed 0.035mm reduction in thickness the same as the thickness reduction on the piece examined by optical microscopy which had only one pickle. Can any valid deductions be made as to the corrosion resistance of the tube as a result of the second pickle. (6 hr 10%HF+HNO3 at room temperature)

Question 4: Does anyone know the typical removal rates by pickling of scaled and Chromium depleted bore surface material after an anneal of for example 1150 degC for 2 min on this or similar grades of Stainless? Does 50 microns
sound reasonable? Is 100 microns too excessive?

I apologize for the many questions but any help or advice would be greatly valued.

Sincere thanks - deics

 

[unclesyd]

First where did you get that pickling solution and times? It is chewing up you tube.
If you had a requirement for these tubes to be annealed it should have been a bright annealed.
To carry drawing lubricant into an annealing furnace is an absolute no no.
This material should not be senitized if the QA was done properly.
All my tests and measurements over the years is that the metal loss from pickling is in the micron range. When this work was done our best physical measurements were only to <.0001&quot;. Very seldom was any IGC discernable by optical microscope. We kept the HNO3/HF ratio high to avoid IGC.

These tubes are corroded as stated by the owner. He will have to determine if the extent of corrosion is detrimental to his process.

It sounds like you have the tools so pull the worst tube/tubes and examine a cross section of the worst area and see what you have.

 

[mcguire]

If the tube was carburized on the ID by this contamination, and this is quite possible, then the carbon penetration into the grain boundaries would overwhelm the titanium's ability to combine with it, and because carbon diffuses so much faster than chromium, you would never fix the grain boundary depletion at possible annealing times and temperatures.
So you will have grain drop-out when you pickle. This can cause this degree of wall loss. The first pickle dissolved the grain boundaries, grains fell out, the surface appears black ( diffraction effect)reaining material analyzes OK, further pickling does no more damage.
I no longer have the data to answer question 4, but it will be less than 10 microns.

 

[kenvlach]

unclesyd and mcquire gave excellent responses. I will add that any free carbon or organic on the surface, even if not reacted as carbide, combines with nitric acid to have an etching effect (note that a nitric - alcohol etch is used in metallography of SS).

Re the necessity of cleaning prior to heat treating and avoiding oxidation, see
How to protect metallic surface from oxidation when heating
thread330-74100

MIL-H-6875H HEAT TREATMENT OF STEEL gives protective atmosphere requirements. A free download at

http://assist1.daps.dla.mil/quicksearch/

ASTM A380-99e1 Standard Practice for Cleaning, Descaling, and Passivation of Stainless Steel Parts, Equipment, and Systems. Pretty self-explanatory, and definitely no 6 hr pickle in 10% HF-nitric acid. Even if you don’t purchase, the abstract has some useful information.

http://www.astm.org/cgi-bin/SoftCart.exe/DATABASE.CART/REDLINE_PAGES/A380.htm?L+mystore+wapq3924

 

[deics]

Unclesyd

The pickle was reported from the Mill.I am not sure what the % of Nitric was. Something might have been lost in the translation.Japanese! I'll check and get back to you. It's a cold pickle. Room temperature. I was surprised at the duration but that was double checked. Maybe the tubes will be ordered Bright Annealed next time but for the last 7 times they haven't been and no problems in service so.... and the last lot from the same Mill.

I note your answer on cold drawing lubricant not being carried into the furnace under any circumstances.The Mill tried a couple of tests, by the way. On one test they caked the &quot;soap&quot; on 3mm thick and were able to duplicate the effect. Then they smeared it on thinly...still pretty severe contamination....nothing! It just burnt away. The Mill, of course, degrease before the anneal and reportedly they come out clean. There is no evidence at all of the sort of caked on condition. Got us all a bit puzzled that has.

The condition is probably not detrimental to service as it is dry, hot CXHX and nitrogen.

We already pulled the worst tubes that gave ECT indications over reference level and looked at them.The results came from those tubes. Covers are welded on now so, it's accept as is. This means some tubes have small areas where wall is thinned by something less than 0.020mm...something under 7 thou....

Anyway thanks very much for your assessment....most helpful,

deics

 

[deics]

mcguire

The first pickle was the Mill pickle the second was our attempt to get rid of the dark stuff before we had seen the optical micros.

This was not a diffraction effect.There was definitely a residue remaining on the tube. It's one of the reasons why we re-pickled a tube sample that showed the dark patches. Even after the second pickle we could still see some signs of it but this turned out to be like black dust. We wanted to check that sample with Probe ECT and it just wiped the stuff off. At that point we just wanted to see what the effect was of removing the stuff was on the ECT indication.

Anyway the less than 10 microns is helpful. Thanks

deics

 

[deics]

Kenvlach

In this case they used a 10% Oxalic Electrolytic Etch. One thing I wondered about is if the lab had over-etched. The unetched condition did not look so bad at the surface.

The Mill is very old experienced firm and their cleaning is performed correctly as far as we can ascertain. Degrease before anneal seems to get the tubes clean. See my mail to unclesyd. They did tests. This &quot;soap&quot; they use had to be really caked on to duplicate the effect. Thin smears just disappeared. Something very odd there.

Anyway, many thanks for your input. Interesting point.

deics

 

[mcguire]

When you anneal stainless, any contaminant will disrupt the oxide and give a discoloration after annealing and pickling. Usually it's superficial, because it burns off in anneal, even with carbonaceous materials.
If you exclude oxygen, as in bright anneal or within a tube, then carbon, which doesn't oxidize, can diffuse in. This is more pernicious because you don't know it's there until you pickle and the grains boundaries are attacked.
The pickle was fortunate for you because it alerted you to a condition that would certainly have causes a later field failure. With a bright anneal you would not have known there was a problem.

 

[deics]

mcguire

That is also a useful point of view on the BA vs Pickled Condition. I really appreciate that.

The reason we got where we are on this is because the internally applied ECT found the problem in the first instance and visual checking latched on afterwards as to what was going on plus a lot of other investigation work really demonstrated. I think most folks would have shrugged their shoulders on seeing the ECT results and accepted that they were seeing some sort of surface anomaly causing indications.

By the way the pickle was not exactly as I stated before....hope unclesyd reads this...apologies for wrong info previously. Happens all the time round here.

We finally got through to the man on the shop floor instead of the Sales Rep and he said:

4%HF 15% HNO3 Balance H20. Time 50-180minutes depending on the pH variation with depletion of the pickle.
Temperature is 0-40deg C.(Room temperature range for this neck of the woods).

Degrease before anneal is a special alkali wash with 45-50% NaOH, Chelating Agent 5-10%, another Chelating Agent 1-5%,Poly Car-BON acid 1-5%(That's the way the man said it on the phone your guess is probably better than mine!), balance is water.

Again many thanks for your input. Most valuable

deics

 

[unclesyd]

You are getting closer on the pickling solution. We found that 2%HF/20%HNO3 worked the best for routine pickling. We would raise the HF on occasion to 3%.
We run our bath base on the Iron concentration and the expertise of the operator, who determined the acid concentration by its activity. I can't tell you how he did his testing, you would laugh to hard.
Sooty carbon should be long gone prior to achieving annealing temperatures. We have a cleaning process that produces sooty carbon (carbon black)that operates at 1000°F. Prior to the process being commissioned we ran tens of tests with tubes full polymer-carbon up 1800°F with no diffusion of carbon into the stainless, 304L, 321.

One trick to using a pickling solution to clean S/S is to condition the oxide of the tube, contaminates, etc. prior to pickling If it possible treat the part in a solution of 4%KmnO4 + 10 NaOH at 200°F for 3 Hrs, rinse, and pickle with the part. You might have to do it several times heat induced scale.

Using 20% HNO3 + HF the lower the ratio of HF to HNO3 the brighter the part will be. Use the lowest concentration possible.

We use the above two treatment routinely to clean long (200' to 300') serpentine S/S tube bundles after they have been thermally treated to remove polymer.

 

[deics]

unclesyd

That's useful information about the cleaning solutions which I will pass on and make use of in our investigations.

I never laugh at shop floor experience. They have to do the stuff I only ever read about in books. Eyeball and brain ar unbeatable combination.Operators have both of those commodities in abundance.

Mill tests show complete burn away of this &quot;soap&quot; with no sooty deposits except in nonsensical amounts. Our attention now turns to the degrease bath but that is a proprietary agent supplied by the specialist company that supplies the soap. I don't believe it's causing thes dark deposits on the tubes either.

One more possibility is the large amounts of Sodium in the 'soap&quot;. Another suggestion that I have from a contact in a mill in Europe is that the Sodium would develop a chromate during heat up during anneal and the subsequent agressive oxidation would be more agressive in such areas would equally cause chromium depletion and a more severe attack by the pickle in those areas.

Another thing I found out, which has me kind of foxed, is that the lab that did the micros, actually etch tested them with 10% oxalic at 1A/cm2 for 1.5 minutes instead of just etching them for microexamination. Would that be a routine procedure? I worked in a lab long ago and we never did that as I recall.

I thought that might explain the aggravated extent of intergranular corrosion seen on the normal and dark areas of the tube micros? I think I mentioned the micros didn't look so bad in unetched condition. The lab also failed to mention about this etch testing thing in their report until I went back and asked about the etch they had used.

Again thanks for the help.You replies are really opening up new possibilities for resolving this problem.

deics

 

[TVP]

Austenitic stainless steel alloys do not etch as easily as plain carbon steels. You need etchants that are more aggressive than nital. So, electrolytic etching using oxalic acid is a common method for revealing microstructure of these alloys.

 

[deics]

TVP

Thanks for your comment.

I guess my question was is it normal procedure to subject a microspecimen to an etch test at 1A/cm2 & 1.5 minutes....as opposed to usual electrolytic etching of SS using 10% Oxalic Acid and a lower current density and short time, say 10-15 seconds. ASTM A262(?) apparently refers to this Etch Test but I don't have a copy here in Japan. I'm hoping somebody does and can put me right.

deics

 

[TVP]

Etching time varies with this procedure. Revealing grain boundaries takes longer, usually 45-60 seconds. Perhaps mcguire can provide his thoughts on this...

 

[unclesyd]

decis,

321 S/S doesn't show it's grain boundries say as quick as say 304. As you noted we usually started with 10-15 seconds at .5-.8 A/Cm^2. We would repeat the etch until we got what we wanted. Over etching shouldn't have gone after the 321 S/S as you report. What does the sample look like unetched, mount and polished. IGC should be visible on the tube surface under a Stereo Microscope at 100X.
Keep us posted on the investigation.

Check this site out. There are some non-electrolytic etchants.

http://www.metallography.com/

 

[deics]

unclesyd

I got hold of the ASTM A 262. Standard Practice for Detecting Susceptability to Intergranular Attack. The oxalic acid etch test is a rapid method of identifying by simple etching those specimens of certain steel grades which are essentially free of susceptability to intergranular attack associated with chromium carbide precipitates. These specimens will have low corrosion rates in certain corrosion tests and therefore can be eliminated from testing as acceptable.

Apart from the oxidised/carburised surface condition which has been nibbled away by this Etch Test, there are no signs of ditched or stepped structures which show susceptability to intergranular attack.

I think we will have to do a further test from this Standard like ferric sulphate + sulphuric acid test.

Thanks again. Will let you know how it all comes out.

deics