High alloy bainite: why not?

[Rsq]

Background:
I just bought a lovely little k390 puukko from Todor Hristov, and while asking him about his heat treat and perusing the datasheet, something struck me. You never see high alloy bainite, presumably because it takes so long to grow, but there's no reason you COULDN'T force the transition.

I'm a mechanical engineer, so while I'm not a metallurgist or a materials scientist, I am reasonably comfortable with cct and ttt phase diagrams.

[It would surprise me if anyone on this board needed the background, but in case anyone reading this is wondering why I would wonder this, you can read all about bainite and mixed phase bainite here
https://knifesteelnerds.com/2018/07/09/bainite-vs-martensite/
and why it's rarely used in high alloy steels (it's often that it just takes too long, but both carbon and alloys that improve martensite (unlike Boron) can also inhibit bainite growth).]

I'm not sure how k390 compares to other high alloy steels in this regard, but If you look at the K390 datasheet, there's a CCT diagram on page 12. It looks to me like you can continuously cool according to selected cooling rate to attain a wide and controllable mix of martensite and bainite, with a consistent 10-12% RA, independent of the bainite/martensite ratio.

https://www.acerosbohler.com/media/productdb/downloads/K390DE.pdf

K390 seems like an excellent candidate for this because of both its ability to get very tough for its carbide fraction, even as full martensite, as well as the fact that no matter what you do to it, it will precipitate enough very hard carbides that it will exhibit ridiculous wear characteristics at any hardness. You can read about a really interesting (although not useful for cutlery, since we know we need high hardness to get a good edge) approach to K390 here

https://iopscience.iop.org/article/10.1088/1757-899X/118/1/012022/pdf

It looks to me like you could austenitize (in BU's datasheet's CCT plot at 1180 C, so some experimentation might be in order to investigate bainite growth from lower austenitizing temperatures) then marquench to an upper bainite growth holding temperature, here it looks like ~250C to 350C, for a very long time. The longer you can slow the cooling from 350 to 250 here, the greater the fraction of bainite I would expect. It looks like bainite starts to form at 250C when continuously cooled for 15 minutes. The effect of marquenching on bainite growth is unknown to me. You might encounter unexpected behavior by deviating from the charts, but that's why this territory is called uncharted.

I would expect things to get interesting around the 30 min to 2 hours cooling times where you see 21-51% bainite with only ~10% RA, and the balance being martensite (maybe even 4 hours? it's still HV 728, or right about HRC60 at 4 hours, and 72% bainite, but RA starts climbing to 15% and higher after this). This is why things get really interesting. You could control this quench rate/soak time to optimize the balance of bainite/martensite, then attempt to temper to convert as much of the 10-12% RA to martensite as possible.

There are a lot of unknowns (to me) here. The things that would need to be investigated include:
the effect on bainite of processes to temper martensite (Cryo?)
The effect of marquenching vs continuous cooling on the bainite/martensite fraction and grain
Whether bainitic K390 even actually offers toughness improvements over properly tempered, fine grained martensite.

So yeah, if someone that had the equipment, time, and interest (of which I have 2 out of 3, but an apartment is no place for a forge, heat treat oven, or tempering salts) to investigate high carbide, high alloy bainite, I think it would have potential. Even if it performed like a different alloy, it might just open up new uses for it, and I'm sure people (like me) would buy it just for the novelty and interest.

P.S.
It would be fascinating to see a metallurgist develop a steel like K390, but alloyed with boron to enhance bainite growth

 

[Nathan the Machinist]

"Ridiculous wear characteristics" does not translate into good edge retention. Good edge stability is challenging in these high alloy steels. A rapid quench helps.

 

[Rsq]

Are you referring to controlling grain size with a rapid quench? Edge stability requires hardness, which is where bainite would typically fail. I'm reading the datasheet as saying you can get mostly bainite and an undesirably large fraction of RA, and still hit HRC60. I would expect (from my understanding of the datasheet) that you could quench a little faster than this and still end up with a reasonable (desirable?) combination of martensite and bainite and still come out HRC62+ (in the area of the CCT that looks most promising to me). Can bainite be formed after a marquench to upper bainite temperatures? One of the things I really don't yet understand about steels is how and whether you can transform one phase into another without austenitizing on the way. I would expect a marquench to result in a higher fraction of converted martensite, as well as better grain, but then would you have to be also converting some martensite to bainite, or is this transformation only observed with continuous cooling?

I am curious to know more. What are the mechanisms by which high alloy grain boundaries (established above critical temperatures) erode during bainite growth at much lower (than critical) temperatures? I understand the bainite transformation to be driven by diffusion, but I would expect a lot of carbides to precipitate in time to pin dislocations, especially with a low austenitizing temperature and a rapid marquench to bainite temperatures, if this would work at all.

 

[DeadboxHero]

We don't want banite in k390, we want strength with a tempered martensite matrix that supports the carbides best. The banite is better for impact toughness not edge stability. I feel it doesn't make sense to seek impact toughness in a steel that has 15-17% carbide volume anyways. That is a waste of k390s capabilities.

Background:
I just bought a lovely little k390 puukko from Todor Hristov, and while asking him about his heat treat and perusing the datasheet, something struck me. You never see high alloy bainite, presumably because it takes so long to grow, but there's no reason you COULDN'T force the transition.

I'm a mechanical engineer, so while I'm not a metallurgist or a materials scientist, I am reasonably comfortable with cct and ttt phase diagrams.

[It would surprise me if anyone on this board needed the background, but in case anyone reading this is wondering why I would wonder this, you can read all about bainite and mixed phase bainite here
https://knifesteelnerds.com/2018/07/09/bainite-vs-martensite/
and why it's rarely used in high alloy steels (it's often that it just takes too long, but both carbon and alloys that improve martensite (unlike Boron) can also inhibit bainite growth).]

I'm not sure how k390 compares to other high alloy steels in this regard, but If you look at the K390 datasheet, there's a CCT diagram on page 12. It looks to me like you can continuously cool according to selected cooling rate to attain a wide and controllable mix of martensite and bainite, with a consistent 10-12% RA, independent of the bainite/martensite ratio.

https://www.acerosbohler.com/media/productdb/downloads/K390DE.pdf

K390 seems like an excellent candidate for this because of both its ability to get very tough for its carbide fraction, even as full martensite, as well as the fact that no matter what you do to it, it will precipitate enough very hard carbides that it will exhibit ridiculous wear characteristics at any hardness. You can read about a really interesting (although not useful for cutlery, since we know we need high hardness to get a good edge) approach to K390 here

https://iopscience.iop.org/article/10.1088/1757-899X/118/1/012022/pdf

It looks to me like you could austenitize (in BU's datasheet's CCT plot at 1180 C, so some experimentation might be in order to investigate bainite growth from lower austenitizing temperatures) then marquench to an upper bainite growth holding temperature, here it looks like ~250C to 350C, for a very long time. The longer you can slow the cooling from 350 to 250 here, the greater the fraction of bainite I would expect. It looks like bainite starts to form at 250C when continuously cooled for 15 minutes. The effect of marquenching on bainite growth is unknown to me. You might encounter unexpected behavior by deviating from the charts, but that's why this territory is called uncharted.

I would expect things to get interesting around the 30 min to 2 hours cooling times where you see 21-51% bainite with only ~10% RA, and the balance being martensite (maybe even 4 hours? it's still HV 728, or right about HRC60 at 4 hours, and 72% bainite, but RA starts climbing to 15% and higher after this). This is why things get really interesting. You could control this quench rate/soak time to optimize the balance of bainite/martensite, then attempt to temper to convert as much of the 10-12% RA to martensite as possible.

There are a lot of unknowns (to me) here. The things that would need to be investigated include:
the effect on bainite of processes to temper martensite (Cryo?)
The effect of marquenching vs continuous cooling on the bainite/martensite fraction and grain
Whether bainitic K390 even actually offers toughness improvements over properly tempered, fine grained martensite.

So yeah, if someone that had the equipment, time, and interest (of which I have 2 out of 3, but an apartment is no place for a forge, heat treat oven, or tempering salts) to investigate high carbide, high alloy bainite, I think it would have potential. Even if it performed like a different alloy, it might just open up new uses for it, and I'm sure people (like me) would buy it just for the novelty and interest.

P.S.
It would be fascinating to see a metallurgist develop a steel like K390, but alloyed with boron to enhance bainite growth

 

[shqxk]

Bainite is not always tougher than tempered martensite at similar hardness.

I have a low temp salt baht and have done several experiment on different low alloy (52100, L6 and some Daido's special steel) the major benefit of austempering is it require much less process compare to normally quench+tempering. Just quench the blade in 450F salt and lets it sit for 4 hours and done! Less problem on warpage too.

But from my experience with those steel mentioned, lower bainite structure at the same hardness noticeably has lower edge holding than tsmpered martensite.

 

[Rsq]

We don't want banite in k390, we want strength with a tempered martensite matrix that supports the carbides best. The banite is better for impact toughness not edge stability.

That could well be true that tempered martensite is the best matrix for the carbides. I just wonder if anyone has tried it. The mechanism of failure for k390 is almost always chipping. At least in my experience, it outperforms almost any other steel in most applications, but when it does fail, it does so in the worst possible way that takes the longest to fix. I attribute this to brittle martensite not being the best matrix for the carbides. While the steel can get to hrc 65 or much harder, it starts getting brittle. If it performs best at HRC 61-63 with martensitic microstructure, the grain refinement and ductility of a bainite matrix could be expected to improve the failure mode and generally improve performance, so long as it does not change the final hardness.

I refer to the discussion on p. 293 here
https://www.phase-trans.msm.cam.ac.uk/2004/z/3750-012.pdf

high alloy bainite, at best, could be a strict improvement. While it may well be that no one has tried it with k390, someone has definitely tried to make a steel of this class bainitic.

I don't know if this would have a net positive effect, but if it did, I could think of a number of explanations. I would think that a bainitic K390 (or any other ~10+% carbide steel) would begin to behave the way people wish a cobalt alloy would, but with a much harder (more stable) matrix.

I feel it doesn't make sense to seek impact toughness in a steel that has 15-17% carbide volume anyways. That is a waste of k390s capabilities.

Of course martensitic K390 has a place. It is outstanding as we typically use it. Bainitic K390 might not be a direct substitute for martensitic K390, but different heat treats for different purposes is always a valuable tool. Replacing ~30% of the martensite with bainite might have very little effect on test performance, but only reduce the likelihood of catastrophic failure modes. 70% bainite might have a steel that cuts like s90v with much better toughness. Maybe there would just be no way to get rid of the RA, and tempering the martensite would degrade the bainite and the whole thing would be a bad idea with K390 (but could still work with a different alloy. Maybe there is a high alloy boron steel that is unknown to knife makers, but someone reading this thread gets curious to test)

I don't know if it's a good idea, but I assumed if anyone does (and reads forums in English), they will probably be here. So far, the answers I have found have fallen into 3 broad categories
1. "if it aint broke, don't fix it"
2. "I don't do it that way because it just isn't done that way"
3. $$$

Obviously this is a niche question about eking out the last 5% of performance, even though it may cost more than the first 95. I doubt bainitic k390 is a large scale viable product, so I understand number 3. I don't think of these high alloy steels as having great edge stability to begin with, since they chip and are prone to fatigue failure at even moderate apex angles. Intuitively, I would think bainite would be a solution to low edge stability.

Edit: S shqxk
True. But bainitic microstructure has a more pronounced beneficial effect the higher carbon content of the steel. My question is whether you might not be able to "grow" what you could think of as a mixed metal compound with a ductile but hard mixed martensite-bainite matrix around the large carbide volume and get a performance that is not directly comparable to any steel I know of, but maybe closer to cobalt/carbide mixtures, but with a harder matrix.

 

[Stacy E. Apelt - Bladesmith]

My experiments will be on 52100 to develope high hardness in a partial bainite structure.

The whole setup is ready to go as soon as the new shop is up. The testing and analysis will likely take a year.

My comment to the OP is that it isn't as simple as he describes.

 

[BluntCut MetalWorks]

IMO - ht high carbide volume steels

Perfect annealed = low edge stability(ES) due to low strength ferritic matrix (~20s RC)

Increase to 10+% Co = improve ES with tempered mart and Co matrix (~40s RC, esp interface with carbides)... e.g. maxamet

Bainitic% (cementite + ferrite) = cementite increased hard phrase particle%, probably reduce ES. Distributed ferrite tiny increase impact/shock load and insignificant ductility

Ultra fine grain = big improvement in ES due to enhanced strength and toughness

Nano scale subgrain (mart/ra grain not PAG) = Huge help with ES (UFG inside each PAG), plus lowering misalignment angle of matrix interface with particle. *bhadeshia's experty and current attempts are (hoping for good results) on ra (with some bainite) subgraining.

Monograin = ES grail (to be clear: local - relevant to specific steel, not global - across all steels - ES optima)

 

[DeadboxHero]

At the of the day, it's the Geometry that cuts.


You're chasing raw impact toughness seeking more durablity but neglect that Strength is important for stability of thin cutting Geometry for knife performance.

Bainite would promote lower yield Strength and more carbide tear out.
That means the edge angle would have to be thicker to support the lack of strength and ability to hold Carbides.


Also,the grind would need to be thicker to prevent deformation behind the edge.

All this would drop the cutting performance.

The main benifits for bainite would be for shock resistance, but Carbides are detrimental for that kind of toughness anyways.

I feel we are putting a square peg in a round hole. Just use a different steel.

Curious, have you experienced higher Hardness K390? 65-66rc?

That could well be true that tempered martensite is the best matrix for the carbides. I just wonder if anyone has tried it. The mechanism of failure for k390 is almost always chipping. At least in my experience, it outperforms almost any other steel in most applications, but when it does fail, it does so in the worst possible way that takes the longest to fix. I attribute this to brittle martensite not being the best matrix for the carbides. While the steel can get to hrc 65 or much harder, it starts getting brittle. If it performs best at HRC 61-63 with martensitic microstructure, the grain refinement and ductility of a bainite matrix could be expected to improve the failure mode and generally improve performance, so long as it does not change the final hardness.

I refer to the discussion on p. 293 here
https://www.phase-trans.msm.cam.ac.uk/2004/z/3750-012.pdf

high alloy bainite, at best, could be a strict improvement. While it may well be that no one has tried it with k390, someone has definitely tried to make a steel of this class bainitic.

I don't know if this would have a net positive effect, but if it did, I could think of a number of explanations. I would think that a bainitic K390 (or any other ~10+% carbide steel) would begin to behave the way people wish a cobalt alloy would, but with a much harder (more stable) matrix.



Of course martensitic K390 has a place. It is outstanding as we typically use it. Bainitic K390 might not be a direct substitute for martensitic K390, but different heat treats for different purposes is always a valuable tool. Replacing ~30% of the martensite with bainite might have very little effect on test performance, but only reduce the likelihood of catastrophic failure modes. 70% bainite might have a steel that cuts like s90v with much better toughness. Maybe there would just be no way to get rid of the RA, and tempering the martensite would degrade the bainite and the whole thing would be a bad idea with K390 (but could still work with a different alloy. Maybe there is a high alloy boron steel that is unknown to knife makers, but someone reading this thread gets curious to test)

I don't know if it's a good idea, but I assumed if anyone does (and reads forums in English), they will probably be here. So far, the answers I have found have fallen into 3 broad categories
1. "if it aint broke, don't fix it"
2. "I don't do it that way because it just isn't done that way"
3. $$$

Obviously this is a niche question about eking out the last 5% of performance, even though it may cost more than the first 95. I doubt bainitic k390 is a large scale viable product, so I understand number 3. I don't think of these high alloy steels as having great edge stability to begin with, since they chip and are prone to fatigue failure at even moderate apex angles. Intuitively, I would think bainite would be a solution to low edge stability.

Edit: S shqxk
True. But bainitic microstructure has a more pronounced beneficial effect the higher carbon content of the steel. My question is whether you might not be able to "grow" what you could think of as a mixed metal compound with a ductile but hard mixed martensite-bainite matrix around the large carbide volume and get a performance that is not directly comparable to any steel I know of, but maybe closer to cobalt/carbide mixtures, but with a harder matrix.

 

[Rsq]

Stacy E. Apelt - Bladesmith
I'm sorry if I made it sound like I thought this would be easy. It's exciting to hear that this is something you're already experimenting with and I'll be looking forward to seeing the results.

DeadboxHero I have not experienced K390 that high. I have a few spydies, but the best example of K390 (until the Hristov puukko arrives) that I have is a Jamall/ Adam Kornalski hunter from a few years back. I've long lost the link and he doesnt work with K390 anymore, but it's definitely harder to grind than the spyderco k390. Right now it has one of the largest chips on the edge I've ever induced by cutting wood, but it also hadn't needed anything but stropping after 3 years of kitchen duty.

Lower yield strength would, indeed, be a very viable explanation for not wanting bainite. It will be interesting to see the results of the 52100 experiments as they progress

 

[Storm W]

That could well be true that tempered martensite is the best matrix for the carbides. I just wonder if anyone has tried it. The mechanism of failure for k390 is almost always chipping. At least in my experience, it outperforms almost any other steel in most applications, but when it does fail, it does so in the worst possible way that takes the longest to fix. I attribute this to brittle martensite not being the best matrix for the carbides. While the steel can get to hrc 65 or much harder, it starts getting brittle. If it performs best at HRC 61-63 with martensitic microstructure, the grain refinement and ductility of a bainite matrix could be expected to improve the failure mode and generally improve performance, so long as it does not change the final hardness.

I refer to the discussion on p. 293 here
https://www.phase-trans.msm.cam.ac.uk/2004/z/3750-012.pdf

high alloy bainite, at best, could be a strict improvement. While it may well be that no one has tried it with k390, someone has definitely tried to make a steel of this class bainitic.

I don't know if this would have a net positive effect, but if it did, I could think of a number of explanations. I would think that a bainitic K390 (or any other ~10+% carbide steel) would begin to behave the way people wish a cobalt alloy would, but with a much harder (more stable) matrix.



Of course martensitic K390 has a place. It is outstanding as we typically use it. Bainitic K390 might not be a direct substitute for martensitic K390, but different heat treats for different purposes is always a valuable tool. Replacing ~30% of the martensite with bainite might have very little effect on test performance, but only reduce the likelihood of catastrophic failure modes. 70% bainite might have a steel that cuts like s90v with much better toughness. Maybe there would just be no way to get rid of the RA, and tempering the martensite would degrade the bainite and the whole thing would be a bad idea with K390 (but could still work with a different alloy. Maybe there is a high alloy boron steel that is unknown to knife makers, but someone reading this thread gets curious to test)

I don't know if it's a good idea, but I assumed if anyone does (and reads forums in English), they will probably be here. So far, the answers I have found have fallen into 3 broad categories
1. "if it aint broke, don't fix it"
2. "I don't do it that way because it just isn't done that way"
3. $$$

Obviously this is a niche question about eking out the last 5% of performance, even though it may cost more than the first 95. I doubt bainitic k390 is a large scale viable product, so I understand number 3. I don't think of these high alloy steels as having great edge stability to begin with, since they chip and are prone to fatigue failure at even moderate apex angles. Intuitively, I would think bainite would be a solution to low edge stability.

Edit: S shqxk
True. But bainitic microstructure has a more pronounced beneficial effect the higher carbon content of the steel. My question is whether you might not be able to "grow" what you could think of as a mixed metal compound with a ductile but hard mixed martensite-bainite matrix around the large carbide volume and get a performance that is not directly comparable to any steel I know of, but maybe closer to cobalt/carbide mixtures, but with a harder matrix.

You are asking a thought experiment but no one has done the real experiment. Banite isn't done often and when it is it being done on steels from the other end of the spectrum from the a11 type steels. So it's a combination of $$$ and not expecting to find anything useful there. If you get around to trying it yourself and there are good things to be had let us know. I'm not just saying that to be dismissive. I don't have a lot of interest in those steels because how I use my knives I have never found them to be impressive. If you could get edge performance like tool steels I would totally be interested in the investment.

 

[Stacy E. Apelt - Bladesmith]

Storm W has a good point that will likely be a big part of my testing summation. The time and cost of producing a bainite/martensite mix ( or high percentage bainite steel) for a knife will not likely be cost effective. For long blades, swords, and things that will undergo FIF type abuse, there may be enough advantage ... but that would be the exception, not the norm.

I have said many times that high carbon steel, properly hardened and tempered for fine grain, forming ACAP 100% martensite will likely be nearly impossible to beat.