Cracking,Liquid Metal Embrittlement

Date 06-05-2003 17:00

Chris,
We have had a lot of galvanizing done for us over the years. Cracking has not been a major problem for us. When we do have troubles it's usually something like a long, thin angle welded to the flange of a thick beam, or similar with a high restraint situation.
I can't recall ever having a beam crack in a cope area except where someone had nicked into the radius and then improperly repair welded it.

Some people believe liquid metal embrittlement (LME) is a problem with galvanizing but I am not totally convinced due to a high success rate. Lots of pieces are dipped in molten zinc everyday with no problems. Now if some says LME is the reason for cracking in a high restraint situation, I can't argue.

We had some beams crack severely through the flange and several feet of web (one long crack, on 10 or 12 different beams. I think they were
W21 x 50s or W18 x 40s, about 30 feet long and rolled the hard way (camber) to about a 300 foot radius. The metal had been stretched a lot in those.

Another time we had 3 feet long 12 x 12 x 1/4 square tube, blow open so they resembled a plugged up shotgun barrel that had been fired. The odd part is the tubes were plain material and completely open, no restrictions in the ends whatsoever! Replacement pieces had no problems.

My standard advice regarding galvanizing is to talk with your galvanizer and let them know what you are making. I'm sure they have seen more than most of us and have worked out a good list of solutions and precautions to follow.

Chet Guilford

By apiguy

Date 06-12-2003 00:44

Check out a posting I had on Eng-Tips.com. I had a similar problem with a large number of beams. There was some really good feedback listed. Search by my name APIGUY

By jwright650 (*****)

Date 06-12-2003 10:55

This is a subject that has a lot of different people giving different ideas as to why this happens. I have found that here at our company, in close to 40 years of fabricating, we haven't had any beams crack when galvanized at the copes. However, we had a plant that was in Abingdon, Virginia and that plant experienced complete tractor trailer loads of galvanized beams with cope cracks after galvanizing. I was called and I thought I knew for sure before I arrived, I knew what I would find. I just knew I would find copes with a notchy finish or improper radius. To my surprize, the copes looked great. Now, having said all this, we had a new galvanizer move into the Bristol area, so it made sense to use him rather than trucking steel several hundred miles to the place we were using. Now that we had made the switch, there haven't been any more such cracks. We looked at heat numbers and such, but the steel was from four different suppliers/mills. All the beams that had cracked were W24 x 104's and all of them were dipped at the same galvanizer. Our assumption is that the pickeling/dipping process is different somehow at that location as we have had no such problems with any other galvanizer, and we have used many. I have called the AISC and looked up articles that their engineers had referenced, but all of them point to incorrect coping techniques, which was not our case.

Anyone with more info, please post.
John Wright

By RBeldyk (**)

Date 06-12-2003 14:52

I have found cracking of structural steel that has been galvanized. The cracking is occurring in the galvanizing bath hence most of the cracks are not visible until either NDE (MT or Eddy Current) techniques are used or after Loading. Visually the beams and structures look fine. The cracking that I am find is linear non branching at 40 to 50 deg from the centerline of the fillet weld, starting at a start or stop.

Here is a link to a few photographs http://pntbeldyk.wirefire.com/JAN%2027%202003%20MT/index.html , though not the best photos of the problem, it does show a typical crack.

If you could MT around a welded gusset plate that has been galvanized. Let the list us know if you find anything.

Rich

By jwright650 (*****)

Date 06-12-2003 15:50

Rich,
We had one of those W24 x 104's split almost into at the web. The beam was 30 ft long and the crack ran for 24 of the 30 ft (no NTD required to see that one). Most were only an inch or less in length and were found with MT and were radiating out of the coped area like you said into the web. Heat applied to the area opened up cracks that were not visible as well.

What was your fix on these beams?

We had gouged out the crack and filled. Then back gouged the other side of the web and filled back flush and ground flat. After it all cooled, went back, per AISC recommendations, and ran a reinforcing bead around the cope on both sides of the web.

By RBeldyk (**)

Date 06-12-2003 16:22

This is the generic procedure we used if it as a welded repair. Some repairs were by bolting on reinforing plates.

Repair -
1. Gouge out the crack with Air arc or Grinding.
2. Dress up by grinding
3. Use 7018 weld side 1 ( I would have preferred a Rutile coated electrode for the root passes, but not acceptable by code.)
4. MT
5. Backgouge
6. MT
7. Weld side 2
8. MT
9. Grind faces flush with base material
10. MT
11. Repair Galvanizing

Galvanizing repair procedure.

Galvanize Repair Procedure Using the Alloy developed by ILZRO (International Lead Zinc Research Organization) under Project ZMN-377

Alloy Trade name and manufacturers:

Galva-Guard - Teckcominco
Johnson #911 - Johnson Manufacturing Co.
Zaclon Galvanizing Repair Alloy - Gardiner Metals

Procedure:

1. Surfaces to be coated must be free of grease and scale. Degrease and clean with grinder.

2. Warm surface with torch.

3. Apply paste flux with a brush.

4. Heat surface to about 250oC (480oF) or until repair alloy melts upon contact with heated surface.

5. "Tin" surface by rubbing the repair alloy stick over the entire repair area.

6. Apply heat to repair alloy stick and melt off enough Galva-Guard to cover repair area.

7. Spread repair alloy with an appropriate application tool which has been dipped in flux and mold to desired shape and thickness. (eg. wood, ceramic, glass)

8. While still warm, use a wet cloth or sponge to clean excess flux.

By CHGuilford (****)

Date 06-12-2003 16:27

Interesting photos, Rich.
I have a few questions. How deep did those cracks along the weld toes go? What kind of wire were these welded with? Do you know the silicon percentage of your filler metal? Was the mill scale removed from the weld zone prior to welding?

With the exception of the crack across the plate, these look like you might have had a small amount overlap on the welds before they were dipped; like maybe there was thick mill scale that wasn't removed.

Another thought, welds with a high percentage of silicon will react differently and seem to "swell" from galvanizing, and the surface will harden. This situation can make welds look overlapped, also. Those who have had handrails galvanized will recall the smoothed out welds come back appearing not to have been ground at all.

Chet Guilford

By RBeldyk (**)

Date 06-12-2003 16:54

Crack went completely through the fillet weld and base metal, i.e., on both sides of plates and through the flanges of the beams.

(On other cases the cracks extended from the edge of the flange into the web.)

Welded with E71T1 Si content 0.57% (I had it analyzed)
Base material was ASTM A572 Gr 50, Si content 0.23%.

All plates and beams were abrasive blasted prior to welding, no mill scale.

Before Galvanizing but after welding all welds were 100% MT.

The overlap was caused by the difference in Silicon content of the base and weld metals.

By jwright650 (*****)

Date 06-12-2003 17:25

Rich,
Could some of your cracks come from the crater where the weld is terminated not being full section? I had problems years ago with the welders not filling thier craters, just pulling out of the puddle, without filling and they had cracks that actually started in the crater and went on out into the base metal, similar to the pics you have.
John Wright

By jwright650 (*****)

Date 02-21-2004 11:54

Another question,
I've been MT'ing tubular frames this morning that have L's between them serving as diagonals similar to the lattice work in a truss. These frames were galvanized(by a galvanizer that we had problems with before) and they come back cracked. The cracks appear to have come from the L's pulling the walls of the tubing. Where the L's are welded to the wall of the tube, the walls of the tubes have split about 5 to 6" each way from the intersection parallel to the length of the tubes. Additionally some of the L's have cracked perpendicular to the length of the L's. These L's looked as though they have been sawed squarely into.
Anyone else have similar situations?
John Wright

By RBeldyk (**)

Date 06-12-2003 18:06

Here are some other photos. The crack goes through the web of the beam. Note the weld is a small seal weld to keep acid out during the acid cleaning phase.

http://pntbeldyk.wirefire.com/images/seal%20weld/96230005.jpg
http://pntbeldyk.wirefire.com/images/seal%20weld/96230004.jpg

I have been cutting and grinding these out looking for this type of defect and haven't found any.

I did find one cracker crack after galvanizing, yet it did not cause an extended crack.

By jwright650 (*****)

Date 06-12-2003 18:20

By jwright650 (*****)

Date 06-12-2003 18:27

Rich,
Were these last two pics, pics of a beam splice with the rat holes filled in? I couldn't tell exactly what I was looking at. If so, I've seen this happen too, cracks in the rat hole went from one to the other, that severed the entire beam web.
John Wright

By RBeldyk (**)

Date 06-12-2003 19:01

You are looking at is the web seal weld joint between two beams that butted up next to each other with a 3/32" seal weld cover the joint. No structural strength, just there to keep the acid out.

This is such a small weld that residual stresses should be minimal and no stresses due to loading, in fact if it wasn't to be galvanized I would have left it open.

On some of these the crack was completely across the web while on others it extenrded only a short distance.

By jwright650 (*****)

Date 06-12-2003 19:15

I like the pics and being able to "see" what someone is talking about. I hope AWS or someone will set aside some memory for pics and a place to upload them for viewing. Even if we had a place to load the pics and a link to view them would be good. This way it would keep the refresh rate fairly quick on this forum and would not bog it down with memory hogging pics. Gerald Austin (pipewelder_1999) had volunteered to do this for us here on the forum. But I didn't want to take advantage of his generosity and bog his site down with pics.

Rich,
Thanks for sharing your pics and your research, I learn something new most every time as I read through the posts on this forum.
John Wright

By Niekie3 (***)

Date 06-14-2003 12:59

Hi John

There have been a number of posts on this board regarding the failures of steel whilst galvanizing. While some people point out that the welds and metal finishing needs to be done adequately, it still does not explain the failures, with or without a suitable finish.

If we look at the galvanizing operation, I see only three possible "driving forces" for the cracks:

1) The heat of the galvanizing bath.
2) The galvanizing itself.
3) Handling of the steel during the galvanizing operation.

I believe that if you heated the steel itself to the temperatures found in the galvanizing bath, for the time periods involved, without actually adding the galvanizing, there should be no significant failures to speak of.

If it was the handling, then there would be other associated dammage to make this obvious. It would not be a "mystery".

This leaves us with the conclusion that it is something in the galvanizing itself that is causing the problem. Research has shown that zinc definatelly causes liquid metal embrittlement in steel, under the right conditions. It therefore only makes sense that this is our problem.

The argument that if this was the case, then all steel being galvinized should show these defects, is much too simplistic an argument. Liquid Metal Embrittlement (LME) is a type of "environmental cracking", such as stress corrosion cracking or hydrogen cracking. As such, the same rules apply:

1) A susceptible microstructure and / or stress state.
2) A susceptible metal / medium combination.
3) The correct temperature and / or concentration.
4) It is a time dependent initiation and propagation mechanism.

If all of these are not present, the cracking will not take place. This is why it is more common in highly restrained weldments. This is also why it would be present at one galvanizer and not another, because they would run their metal baths at different temperatures and would immerse the metal in the bath for different time durations. This is also why an incorrectly finished product would stand a better chance of failure than one which has been correctly finished etc...

The fact that it occurs only sometimes and not always, does not rule out LME, rather it supports the LME theory because LME is based on so many interacting variables.

Regards
Niekie Jooste
Fabristruct Solutions

By jwright650 (*****)

Date 06-16-2003 11:41

Niekie,
I had often wondered if the solutions (soup mix) in the galv. bath may cause some of these things that you have pointed out. After talking to some of the galvanizers about this, they have shared some info that leads me to believe that they don't all use the same solutions to clean or coat with. I remember asking one of the guys why thier finished product looked so shiny vs. the dull finish of others, he told me he adds alot of nickle that others don't. Little things like this make you wonder what other ingredients may be added or subtracted from the solution to have an adverse effect to the steel under the 800F - 900F temps. and then then the rinsing in a cooler bath. Like you stated there are so many variables because no two galvanizers do it the same way. I'm sure economics drive what ingredients are used at some locations, while others strive to deliver a quality finish for the customer.
Still curious,
John Wright

By RBeldyk (**)

Date 06-16-2003 13:37

Galvanizing is 98% min. zinc, 2% others.

The 2% others could include Nickel, Tin, Bismuth, Aluminum, Lead...

These all effect the wettability of the galvanizining mix, heat transfer, apperrance, et ceteria.

By decreasing the surface tension (id est wettability) either by alloying elemnts and or kettle temperature one is more susceptable to LME. This allows the galvanizing elements to enter into microcracks and grain boundaries.

The thermal stress in thick sections can be quite high during the dip and removal process. Thick sections also require a longer dwell time in the kettle, exposing the steel on the surface to liquid metals for longer durations. LME is a time, temperature dependent process, hence, thin sections general do not show LME. However, thin galvanized steel exposed to elevated temperature has been know to crack due to LME. http://www.affiliatedinc.net/lab/case.html#Liquid%20Metal%20Embrittlement

By richardpargeter Date 06-17-2003 09:40

I have only just discovered this discussion, and coincidentally I have just mounted a 'knowledge summary' on the subject on the TWI website (Join IT) http://www.twi.co.uk/j32k/protected/pdfs/ksrjp003.pdf (Don't worry about registration if you are not already a user - it's free, and the article of interest is available to all, free, once you have registered.)

Liquid Metal Assisted Cracking (LMAC) is a real phenomenon, and has been around for years - I have a reference dating back to the 1950s. It is not frequent, but can be devastating when it occurs. The usual reaction, as I have detected in some of the correspondence here, is "are you sure it's not fabrication cracking?" or at least "are you sure there weren't fabrication cracks there, which caused propagation during galvanising?". I'm sure that poor details do at least encourage cracking, and we mustn't suddenly start blaming everything on galvanising either. Nevertheless, the Industry must recognise this as a cracking mechanism in its own right, which doesn't always require a defective weld or cut to start it. I've seen it on cold bent bar, with no weld or stress concentration present in the cracked region, and from well drilled holes.

I wish I could give definitive advice on how to avoid LMAC. Currently, however, there is not a sufficient understanding of all the contributory factors. (TWI has examined several occurrences of LMAC at first hand, and has also carried out some limited research into the subject. If anyone would like to co-operate in some further research into this, please let me know, as we are planning to do some more work in this area.) Nevertheless, we do know some factors which apparently increase the risk, and please follow the link above for my current understanding in more detail. However, one thing I would strongly advise is to inspect after galvanising. Most of the serious cases I have been involved with have only been discovered accidentally, during some other operation, in at least three cases after erection of major steel work.

I hope this helps.

Richard Pargeter
Consultant, Ferritic Steels and Sour Service,
Metallurgy, Corrosion, Arcs & Surfacing Technology Group
TWI, Granta Park, Great Abington, Cambridge, CB1 6AL, UK
richard.pargeter@twi.co.uk

+44(0)1223 891162 Ext.2358(phone) 894717fax)
TWI website: http://www.twi.co.uk/j32k/index.xtp