Bearing Steel Material Analysis : Non-metallic inclusion? 2

[metalfate]

Dear eng-tips material engineering master,

I would want to ask about inclusion in bearing steel. For two months we conduct investigation regarding material defect on bearing steel. From SEM/EDAX data, we believed that the "foreign material" was grinding stone, which contain Aluminum, Silicone, and Oxygen in very high ammount. But, i am not sure that the "foreign particle" was grinding stone. Is there any chance that the foreign material was non-metallic inclusion? If there is, with only using SEM/EDAX analysis, what kind of test method appropriate to conduct the test? Or, am i missing something here?

Thank you for your reply. =D

 

[metengr]

NSKLab;
Over the last several years I have been involved with wind turbine gear box high speed and intermediate speed pinion gear tooth failures that were considered infant mortality (failure after 6 to 12 months)

Most of these failures were caused by subsurface inclusions introduced during steel making. The inclusions were typically aluminum and silicon oxides.

As a side note, we worked with various wind turbine OEM's to eliminate those suppliers that did not provide higher quality (cleaner - no inclusions) steel for pinions and gears.

 

[TVP]

As metengr stated, there is a very high probability that the non-metallic inclusion is a remnant from the steelmaking process. It is possible to determine the source of the inclusions if you have the following:

- high quality images of the inclusions (Scanning Electron Microscopy)
- composition analysis of the inclusions (Energy Dispersive X-ray Spectrometry)
- knowledge of the steelmaking process (BOF vs. EAF, ladle slag composition, refractory type used in ladle, tundish, etc.)

 

[metalfate]

@metengr
On several investigation, we too believed in subsurface inclusions causing the cracking defects on bearing steel after heat treating. Our experience conclude that the size must be higher than 10 mikrons with cluster of its within some area to give significant effects to the bearing steel. We also know that it is possible for Mg, Al, S, and Ca can form inclusions. But, after further studies, grinding stone also having those similar elements, minus Magnesium. Our hypothesis is, the grinding stone was stabbed inside the steel bearing parts from grinding process. We have searched all grinding stone in production line, but we cannot find them. I am afraid that our search was fail from the beginning.
Is it possible for Magnesium to form subsurface inclusions? If grinding stone element also having Al-Ca, how can we differ that particle are from Sub surface, or from grinding stone? Your opinion will be such a great help metengr =D

@TVP
Thanks, i will continue to review the process. If you and @metengr interested, i can sent both of you the attachment files. Email me at mahendra@nsk.com.

Have a nice day =D

 

[metengr]

NSKLab;
I would be interested in the follow-up information because I have been directly involved with SKF roller and inner race bearing failures, once again with wind turbine gear boxes.

I do seriously question your hypothesis though because some of the roller bearing failures I have recently seen from SKF were initiated by subsurface fatigue cracks (they appear to look like butterfly wings) caused by extremely high subsurface shear stresses within the carburized layer.

 

[metalfate]

@metengr
We will post if anything new found. Yes, we also found similar defect with butterfly-like sub surface inclusion. But i think this one is a lil bit different.
Oh, i want to ask also, is there any effect to the inclusion, if bearing steel undergo heat treatment? We suspect this problem arise from steel maker until this morning.

 

[Tmoose]

The "classic" ball bearing spalling failures should always initiate sub surface.

 

[TVP]

NSKLab,

The Mg is a significant clue as to the origin of the particle, because it is not an alloying element for steel, and is not typically part of the composition of the grinding media. MgO is part of the refractory lining of the ladle furnace, and interaction with the synthetic slag during ladle processing can remove MgO spinel particles (exogenous inclusions), which become trapped in the steel during casting and rolling. Here are some references that you can read for more information about NMI in steel:

http://www.swinburne.edu.au/lib//ir/onlineconferences/HTP2012/HTP_21.pdf

http://www.msm.cam.ac.uk/phase-trans/2011/Bearings/index.html#SECTION000137000000000000000

http://ccc.illinois.edu/PDF Files/Publications/06_MTB_Zhang_IngotInclusions_reprint_post.pdf

http://kth.diva-portal.org/smash/get/diva2:217063/FULLTEXT01

http://kth.diva-portal.org/smash/get/diva2:10404/FULLTEXT01

 

[metalfate]

Dear all,
Due to some circumstances, i change my handle to metalfate, nice to meet you all.

@Tmoose, yes it is.This research was conducted to better understand the origin of such phenomenon.

@TVP, great! Thanks for the information. This will help us to move our research to next step. From our knowledge, there are some types of inclusion which are not produced by heat treatment and abnormal composition in metals, which highly caused by external particle such as sand and lining refractory, but we were not realize that Mg-O was one of the lining material composition. My last question: are those lining particle have similar hardness with grinding stone or another abrasive materials?

 

[TVP]

MgO is not that hard (~ 500-775 HV), but the MgO·Al2O3 spinel inclusion is very hard (2100-2400 HV):

http://www.metaljournal.com.ua/assets/Uploads/attachments/Shufeng-87.pdf

 

[metalfate]

Dear TVP,
For weeks of our simulation, we often found Al-Mg-O particle with occurence particle size about 5 micron on the surface of the bearing steel. By this assumption, we believed that if the occurence on the surface is very high, then it is almost sure that we will found similar compund inside the bearings steel. So we cracked a lot of the rings using pressure machine, to see if there are inclusion inside the fractured area to proove it. But until now, we did not find it yet -___-
If the occurence in the surface was high, why we never see it on the fractures area? It seems, odd...

 

[Metalhead97]

metalfate,

As TVP said Mg-oxide inclusions usually come from the refractory during melting. If the metal was melted in a zirconia crucible Mg, Al, and Si oxide inclusions are typical. Another source of oxide inclusion can come from the refractory in the heat treat furnace. The other source of ceramic inclusions come from, as you pointed out, grinding processes.

Chemistry is important, but so is morphology of the defect. Is it round or stringy? Are these inclusions associated with scratches in the surface? If so that is a big clue. Can you attach a picture of the particle(s).

MH

 

[metalfate]

Here attached SEM images file for the suspected spot. It locates on the raceway, with a high probability. The size was ranged from 3~10 um.

Metalhead,
The data we collected was like this spot, round and having Al-Mg-O compound on them. We tried to understand the spot by cracking the bearing into half and SEM-analysis on the fracture area, hoping to find some Al-Mg-O compound inside them. And, as i suspected, we never found them. This method conducted based on this hypothesis : We often found the spot, even by randomly picking bearing steel. So, if the occurence is very high, so there is chance that we will find that spot too, in the middle of the ring. So, we cracked the ring, lot of them. I was going to propose to wire-cut the area near the spot, then grind-polis it so we can see the cross section of the spot.

Here are follow up question:
If the occurence is very high in the steel surface, why we cannot find in the middle of the steel by cracking them? I think the method was wrong. But im not sure. What do you think?
Why we always find rounded spot, if there was randomly attached, instead of plate like (it may look like a string, i suspect...)?

Thank you for the advice anyway, dear all =D

 

[Metalhead97]

Metalfate,

Based on the size of the scratches and the size of the particle that particular particle was not dragged on the surface and embedded. It would be very tough to cross-section that particular spot your metallographer will likely miss it. From a data-collection standpoint if you took 10 random 500x SEM images how many of those spots would you see? What is the mean spacing between particles? I'm trying to get a feel for the frequency of the inclusion. Based on what I've seen so far it appears you have a inclusion from the crucible. It's likely a spinel Al-Mg-O, which can form from the MgO and Al2O3 additives in Zirconia.

It would be harder to see the particles on a rough, fractured surface than the raceway. It really depends on how much ductility you have on the fracture surface; if there is a lot of ductility good luck trying to find the particles. However, I'm sure they are there.


MH