Forging induced martensite in AISI 4140? 5

[DutchChen]

Dear all,

Recently I got an analysis case on a forged product made of 4140.
The sample was first induction heated to 950 degree C for about 3.5 minutes and then air cooled after forging.
Material flow, which looks like white bands, are clearly observed after etching.
However, the hardness of these bands are about HV 650, twice as hard as matrix.

Their microstructure also show lath-like characteristic.

Therefore, I suspect they are martensite.
Two OM images are attached for your reference.
It has also been confirmed that such hard bands did not exist before forging.


Questions are: If they are indeed martensite, any explanation on the formation mechanism?
Otherwise, what structure/phase they could be?


By the way, such bands are still observed after austempering, however, their hardness are at the same level as matrix.
My theory is, the formation of such martensite induced by forging would create some micro cracks (due to volume expansion)in the boundary between matrix.
Therefore, it may bring negative effect on fatigue performance.

Appreciate for spending your time reading this post.

 

[DutchChen]

The other attachment is as below.
Sorry that I am new guy here and not familiar with posting procedure.

 

[metengr]

DutchChen;
It looks like either chemical segregation bands or a problem with the inital induction heating prior to forging that somehow resulted in retained austenite and upon cooling formed fresh martensite. The austempering would indeed soften these bands.

 

[TVP]

Shear localization can result in adiabatic shear bands that subsequently transform to martensite. Total strain, strain rate and deformation temperature are three factors that influence the development of shear localization. Here are some links to review:

http://www.dtic.mil/dtic/tr/fulltext/u2/a230999.pdf

http://www.metal2012.com/files/proceedings/metal_08/Lists/Papers/032.pdf

http://www.smf.phy.cam.ac.uk/Publications/Strength papers/642StrWalleyMMTA38.pdf

 

[metengr]

Very good point made about adiabatic shear deformation bands. This would imply an inadequate austenitization treatment before forging. I suspect the induction heating was not controlled correctly to ensure uniform soaking before forging.

 

[DutchChen]

Thanks for the sharing of shear localization.
From literature, adiabatic shear bands could cause transformation from ferrite to austenite.
However, I still can not figure out what makes these bands cooling so quickly that result in martensite.
The raw material is 5 cm in diameter and about 10 cm in length, air cooling does not seem to provide enough cooling rate?

 

[metengr]

Adiabatic shear deformation is not well understood -it happens. I would be more concerned about why your austenitizing temperature was too low. Under normal forging conditions at proper hot working temperature (fully austenitic) you won't have this problem.

 

[TVP]

DC,

You are thinking about temperature on a macro level. Transformation to martensite within adiabatic shear bands is on a micro level. The local temperature within the band is very high, but the temperature in the surrounding area is much lower. The temperature gradient allows the surrounding bulk material to "suck" the heat out of the shear band. If you are interested in exploring this subject in greater detail, there are books by Wright and Dodd & Bai on the subject.

 

[redpicker]

The more I look at the microstructure the more I think this is exactly what I would expect for 4140 air cooled from forging temperatures in this section size; a mixture of transformation products (upper and lower bainite) with martensite forming in the rich bands. 5cm x 10 cm isn't very big and forming martensite on air cooling is not uncommon for this material. A bar mill rolling 5cm (2 inch) bar would have to either slow cool off the mill or, more likely, anneal the material prior to straightening because of this.

Honestly, I don't think induction heating and/or forging has anything to do with it. If you heated this parts in a radiant tube furnace at 950C for two hours and air cooled, I'd expect to see the same microstructure.

rp

 

[CoryPad]

rp,

I think you are neglecting the 687 HV hardness in the white areas. That should not occur under the conditions you describe.

 

[redpicker]

CoryPad, why not? Actually, that hardness is almost exactly what I would expect in the rich bands of 4140.

4140 can be heavily banded, particularly if the producing mill is running high in Carbon and Manganese to save on Cr and Mo. Of course, it is the high C and Mn in the rich bands that allow martensite to form at cooling rates that produce bainite in the lean bands, so the martensite that forms in those rich bands has a higher hardenss.

As far as the OP's theory that this would affect the fatigue life after a proper heat treatment, I would doubt it, provided the subsequent heat treatment is effective. Heavy banding can reduce the effective hardenability, but as the OP mentioned the final heat treatment was effective in producing a homogenous structure at the appropriate hardness, I am not sure there is anything out of the ordinary.

rp

 

[CoryPad]

rp,

Grade 4140 can be banded (heavily) for the reasons you stated. But air cooling should not produce much in the way of quench severity, which in turn should not produce high hardness martensite. I would not rule out that possibility, but call me sceptical.

 

[redpicker]

You really aren't looking at quench severity, but transformation kinetics. In other words, while we're still dealing with continuous cooling; we need to look at the TTT diagram in addition to the CCT diagram.

If you look at a TTT diagram for 4140 steel, (say

http://www.uvm.edu/~dhitt/me124/JominyNotes.pdf,

page 12) you will see that it can take over 5 hours (that's right, Hours) for full transformation to occur at 1000F (537C). Realize that this is isothermal transformation, so you have to hold the material at this temperature, not just cool through this temperature, but this does illustrate just how sluggish the transformation is at these temperatures. We are also just looking at maybe 5-10% of the microstructure; 90-95% has already transformed to higher temperature decomposition products. The fact that there is no free-ferrite present in the microstructure indicates that it had to have cooled to below 1100F (600C) pretty darn quickly. Combine all this with the fact that we are also dealing with higher carbon and alloy content (since we are in the rich bands) which will shift the transformation curves to the right, there isn't very much skepticism needed.

rp

 

[metengr]

Hardness of martensite is indeed a function of carbon content not quench severity. Once again we don't have enough background information on why type of forging process/method or temperature, etc.

 

[CoryPad]

I wasn't suggesting hardness as a function of quench severity. Rather, the probability to form martensite during cooling (with low quench severity) as low. The adiabatic shear band mechanism seems most likely.