Retained austenite/ martensite refinement

L

Lazy H

Joined
Aug 19, 2014
Messages
12
Hello, from reading a bunch on different forums, even if room temperature is lower than the martensite finish point the blade will still benefit from cryotreating the steel, raising the hardness and improving the other properties. Does this mean that the martensite finish point is more of a "close enough" point without specialized equipment or does further cooling do something else to the martensite that already exists, and since tempering cycles can transform retained austenite to martensite, then does that mean that cryotreating will have any effect in between temper cycles? I know from chemistry that it would be nearly impossible to have a full 100% transformation, and that there will always be some minute (possibly immeasurably small amount) of retained austenite in the steel, but is the transformation from austenite to martensite only undergo a single change once martensite start point is reached or can the "fracturing" of the steel grains that martensite is known for possibly increase in amount, even if the total amount of martensite stays the same?

I have a lot of different theories on metallurgy like this but it can be hard to find unbiased facts about this kind of thing.
 
Martensite finish is interesting research, for sure. I've been spending way too much time (someone started a thread about this very thing!) studying martensite start and finish temps. I can say with pretty good certainty that figuring out a very accurate martensite finish temperature is VERY difficult. Martensite Start temp is a bit easier, but even then, it is difficult to determine accurately.

There are essentially two things going on when taking a steel, such as stainless, down to sub zero temps. First thing is very direct....full martensite conversion. Obtaining absolute 100% martensite may indeed be impossible, always approaching the limit but never getting there! When a steel is quenched, it has a certain amount of retained austenite. Usually (disregarding the experimentation of leaving some RA for added toughness/strength) we try to get as much martensite as possible. Let's say a steel like AEB-L has a martensite finish temperature of -90F. I'm not saying that IS what it is, but as an example (if anyone knows actual Mf for AEB-L I would love to hear from you!). If we did NOT employ a sub zero treatment that can reach -100F, then the AEB-L would likely have retained austenite. You may reach 60C, but not 63C. So the sub zero quench will help to convert the RA over, adding to the overall hardness level.

The second thing that we may see during the use of proper "cryo" (as in liquid nitrogen) will be, not only full martensite conversion, but the precipitation of super small greek letter carbides. I am not sure if the sub zero (-100F) precipitates carbide formation or not. But by FAR the most important thing we are after by employing either sub zero or cryo is full martensite conversion. Eta carbide precipitation is a bonus!

You mentioned that if Mf was higher than room temp, that sub zero treatment will improve a blade. OK....that is highly debatable. I have an opinion. OPINION. And to clarify, we are going to be talking about carbon steel, NOT stainless, since we are interested in steel with Mf above room temp. Stainless has Mf well below 0F. Let's take our beloved 52100 steel as an example. And if I am way off base here, I would be thrilled to be corrected!!!

52100 has an Ms of around 400F, maybe 350F-400F. It's Mf temp is maybe 150F-200F, above room temp. If I were to take 52100 (assuming prior heat treat has set it up correctly) and harden at around 1475F with a solid soak, and proper quench speed, I would expect to obtain 66 easily, possibly 68, C scale, once the steel has cooled down to room temp (below Mf). You really can't achieve a hardness level much more than that. So I would not expect to gain any more hardness by employing a sub zero quench. What would I gain? Possible ETA carbides.....possibly. How much to ETA carbides aid a knife? I have not a clue. There would be some retained austenite in this sample, but by choosing the austenitizing temp to just put enough carbon into solution to achieve max hardness, and leaving the rest as carbides, we have very minimal RA. Can sub zero convert that over? You bet. But there is very little RA, so the sub zero would likely not manifest a noticeable improvement.

Now let's take 52100 that is set up right, but this time let's harden it at 1550F or 1600F. After the proper quench and cool down to room temp, if you were to Rockwell that piece it may not be 67 or 68, likely NOT. Probably around 63 or 64? Just guessing, please don't take the numbers literally, but as an example. How would I get to the max RC level of 67-68? You would then employ a sub zero or cryo treatment to convert that RA over to martensite. Because you hardened at such an elevated temp, RA can be a relatively high percentage, and you'll likely want to convert that over.

I don't know much about the employment of sub zero/cryo after tempering has been done. I know that it is done on some of the higher alloy stainless steels out there, but I can't comment too much about that.
 
The simple answer is - it depends. It mainly depends on the steel type.

Ms and Mf are reached at the normally expected places for the steel type used. The exact spot is rarely all that important. For simple carbon steels, the transition points are at about 400F and 200F. Add a little alloy and the Mf drops to around room temp. Some of the higher alloy carbon steels have a Mf down below -10F. Stainless steels are more complex, but for standard calculations use 400F for Ms and -100F for Mf.

Sub-zero treatment with a dry ice slurry reaches the Mf on most any high alloy or stainless steel.
Cryo, at -350F or lower, reaches the Mf and then gets down to where carbides change. There is some disagreement on whether these changes stick around through the two or three tempers.

Suffice to say that following the standard procedures for the steel in use will get good results. Going beyond these parameters most likely will not gain anything. Doing a procedure not needed or recommended will not gain any thing, and may even do some harm. If stainless steel "A" calls for cryo, ands the results give a 2-3 Rockwell point increase and an increase in toughness, that is great, but don't expect that same increase for 1095 done with those procedures.

My personal take on improving ones HT is not to think outside the box, but to tighten the box. What I mean is getting the temperatures "verifiably" accurate, and getting the procedures as standardized as possible. This will give consistent and repeatable results. With this, small changes can be tracked and the gains or losses of those changes determined.
 
Stacy gives good advice and is right on the money here.

There are a couple things I might add.

First, as we all know, Mf is an antiquated concept. It's an asymptote, and a non zero one at that, so even a steel quenched to absolute zero will retain some RA.

Second, the as quenched hardness number isn't important because of anything special about that value, it's important because that value tells you something about the homogeneity of the structures you formed. Nearly 100% martensite, evenly tempered to some usable hardness (lets say HRC 61) will behave much differently than some messy mixed structure tempered back to that same value. A high as quenched hardness simply means the steel was properly hardened and is ready to be brought to some appropriate working hardness. To put it another way, 52100 quenched at HRC 66 and tempered back to HRC 61 is going to hold an edge much better than that same steel quenched to HRC 63 and tempered back to HRC 61.

Soft zones such as RA and carbon lean martensite from inadequate soak, and brittle spots from large carbides etc are like the perforations in a postage stamp. While less significant in a big block or chunk (like most steels are utilized) they play hell in a thin knife edge. I can think of no other application for steel more critical of a heat treat than a cutting edge. And most industry standard heat treats are aimed at hitting a particular measurable RC number with minimal distortion and risk of cracking for the widest possible range of shapes and material. Knives are niche.

RA does not decompose into just martensite. That's a misconception we need to work on here.

For the purpose of cutlery, generally speaking and particularly with the more complex steels, it is better to form as much martensite as possible in the primary quench. Converting RA later is better than never, but there will also be ferrite and bainite and other structures and it will never be as cohesive nor as complete as a good solid quench to start with.

-said Nathan the not-a-metallurgist Machinist...
 
Mete -the metallurgist is here.
Anyway ,going backwards .Be very gentil with final grind as it's very easy to damage the microstructure you worked so hard to do !
To help clarify some confusion . Quench converts austenite to martensite .To make that conversion more complete we use sub-zero treatment [normally about 0 F] The martensitic transformation is a shear type of reaction and is very fast.Higher carbon and alloying elements are more likely to benefit by sub-zero treatment !
Cryogenic cooling is different.It is not a martensitic transformation and requires some time .At this point how much time is needed is AFAIK still in question. The best estimate I can give [ unless someone has more info ] is 4-6 hours. What happens is that cryo causes the lattice to be tweeked forming areas that have room for the small " eta" carbides to precipitate in these areas ! A 1-2 point increase in hardness can be expected.Again higher carbon and alloying content will gain more than others. Those who have done it right are happy with the improvement in properties !
Basic rule ++ The higher the carbon and alloying content the more critical the time and temperature for HT. ++ Sub-zero and cryo are beneficial for those same steels !
 
WOW! Now I know a lot more than I did, but actually less than I thought! Really though, thanks for all the great info, you guys really got me to thinking now. It's nearly ten pm and all I can think about is going out to the shop and heat treating something!
 
http://www.nitrotechnics.com/texten/fundchang.htm

This is some nicely explained things about cryo. There interest here is wear resistance rather than hardness.
There are other interesting links if you look for ''eta carbides in steel ". One found that best time was 36 hours at 77 Kelvin .This of course was a lab study. At one point you get into the practical things like $$ !
How much improvement vs how much extra cost ?

Back to basics austenite to martensite is shear type , fast type transformation. Formation of eta carbides , pearlite , bainite , precipitates are all diffusion type slow transformations .

Sleep on that guys !!
 
For we knifemakers it is important knowing that alloy elements in solution into the austenite tend to stabilize it... it is also true for carbon in excess of 0.8%.
Talking about simple hypereutectoid carbon steels, we see the graphs having Mf below zero, but those charts assume full solution at Acm.
For our purposes as was pointed out by our friends in this thread, we usually need to put in solution just the amount of carbon allowing full hardness, let's say austenitizing at 1475 °F.....we put in solution just less than 0.8% carbon, and that will be converted into martensite just at room temperature. The rest of carbon will stay out of solution in the form of undissolved carbides that didn't contribute to austenite stabilization.
 
"How much improvement vs how much extra cost ?" That is how I see using sub zero or cryo on carbon steels. There may be some minute conversion of RA, but at what gain and at what cost? Certainly not like we see with the higher alloy steels, which some basically NEED the cryo bath.

If anyone has the time, can you comment as to why cryo is often employed after a tempering cycle? Not a snap temper, but a proper temper. Thanks. I think it may have something to do with the carbon slipping out? BTW, the Ms of AEB-L is about 350F, according to the chart provided by the document "Secondary carbide dissolution and coarsening in 13% Cr martensitic stainless steel during austenitizing." That would indicate Mf higher than I would have guessed.

Stezann hinted at something that is very important. Many of the "charts" or even "formulas" given that show Ms of a carbon steel lower than the 300-400F average are indicating that EVERY alloy went into FULL solution. Usually when we heat treat a carbon steel that is not the case. An alloy like 1.5% carbon white steel would have Ms very low. IF you put all of that carbon in solution. If you did that, the sub zero would be your friend.
 
Last edited:
Oops, for some reason I'm not receiving updates on this thread, I'm amazed by just how much info there is to absorb now!!!

samuraistuart said:
You mentioned that if Mf was higher than room temp, that sub zero treatment will improve a blade. OK....that is highly debatable. I have an opinion. OPINION. And to clarify, we are going to be talking about carbon steel, NOT stainless, since we are interested in steel with Mf above room temp. Stainless has Mf well below 0F. Let's take our beloved 52100 steel as an example. And if I am way off base here, I would be thrilled to be corrected!!!
Click to expand...

I had read about subzero being better even below Mf, though I was guessing it was because of how you could never COMPLETELY eliminate retained austenite, now I'm curious about the structure and mechanisms of these precipitated carbides that are being mentioned and how the grain structure affects their growth and is affected by their presence.

Nathan the Machinist said:
RA does not decompose into just martensite. That's a misconception we need to work on here.
Click to expand...

I was just reading a paper earlier today called "Tempering bainitic steels" that had me asking that same question: how exactly does retained austenite change and why? Possibly the answer depends on the surrounding structures or the method (mechanical, thermal) of removing the retained austenite?

mete said:
Cryogenic cooling is different.It is not a martensitic transformation and requires some time .At this point how much time is needed is AFAIK still in question. The best estimate I can give [ unless someone has more info ] is 4-6 hours. What happens is that cryo causes the lattice to be tweeked forming areas that have room for the small " eta" carbides to precipitate in these areas ! A 1-2 point increase in hardness can be expected.Again higher carbon and alloying content will gain more than others.
Click to expand...

Are these "tweaked" areas similar to the dislocations found in bainitic steel? I was reading that forming a bainitic structure before final quench would form nucleation points for small carbides to form when heated slowly to critical temperature, and since the carbides would often at least partially survive austenization that they would act as nucleation points for martensite grains, the overall effect being that the martensite grains would be smaller because they had more locations to form. Is this possible and if so is it a similar mechanism to what you are referring to?

Thanks for all the responses, sorry if my questions get a little annoying. My curiosity is insatiable.
 
NOT the metallurgist here. :D I have heard and read that us knife makers should be striving to keep the amount of RA below 5% for all of our sharp cutting thingy applications. I have also been told that some of the "industry standard' HT recipes for some highly alloyed steels can leave as much as 20% RA without cryo and that for tooling purposes, some don't consider that to be a big deal just like huge honkin' carbides.
As related to the eta carbides and the higher hardness, I have also heard that you can get higher as quenched hardness perhaps by as much as 1 or 2 points from steels that we commonly use from faster and more even and efficient austenizing methods like molten salt, lead, etc. Are these two phenomenon unrelated?
 
Stuart, for our purposes, should we even be TRYING to get all of the carbon into solution for out applications? The current opinion on some hyper-eutectic carbon steels like 52100 seems to be that we shouldn't if we want that steel to perform as well as it can for fine cutting tasks as opposed to having some brute abrasion resistance more suitable for bearings.
samuraistuart said:
"How much improvement vs how much extra cost ?" That is how I see using sub zero or cryo on carbon steels. There may be some minute conversion of RA, but at what gain and at what cost? Certainly not like we see with the higher alloy steels, which some basically NEED the cryo bath.

If anyone has the time, can you comment as to why cryo is often employed after a tempering cycle? Not a snap temper, but a proper temper. Thanks. I think it may have something to do with the carbon slipping out? BTW, the Ms of AEB-L is about 350F, according to the chart provided by the document "Secondary carbide dissolution and coarsening in 13% Cr martensitic stainless steel during austenitizing." That would indicate Mf higher than I would have guessed.

Stezann hinted at something that is very important. Many of the "charts" or even "formulas" given that show Ms of a carbon steel lower than the 300-400F average are indicating that EVERY alloy went into FULL solution. Usually when we heat treat a carbon steel that is not the case. An alloy like 1.5% carbon white steel would have Ms very low. IF you put all of that carbon in solution. If you did that, the sub zero would be your friend.
Click to expand...
 
I had a discussion with the rep from B-U at Blade in 2012 or 2013 and I commented that it looked like their HT chart said that the temperature of the typical dry ice and acetone or kerosene Slurpee didn't quite get down the the temps that they recommended for cold treating Elmax. He told me that I was correct. Dry ice does the trick for simpler stuff like AEB-L obviously, but as some folks have already stated, that Mf point (apply named in the case of some of the super steels. LOL) can be a moving target.
 
as was pointed out one thing is to reach Mf (dry CO2 for short time is enough) to ge rid of stubborn RA, another story is to promote the precipitation of eta carbides, which requires hours in liquid nitrogen and then tempering...and it is unrelated to RA, at least those temperatures are way overkill for this purpose as far as i know.
 
JDM...my current understanding (NOT the metallurgist either!) is putting full carbon in solution, let's say ALL of the carbon in a high carbon white steel (1.5%), is not ideal, as the RA would be pretty massive. The ideal situation being as you mentioned.....put .8% or so in solution (achieving max HRC level), leave the rest of the carbon as carbide. Thumbs UP!

I have always wanted to do an experiment:
1. Take a hypereutectoid, let's say 52100, and heat treat like we normally do. Normalize/thermal cycle/single quench at 1475 soak. Temper. Test it out.
2. Take the same 52100, normalize/thermal cycle/single quench at 1550F soak. sub zero bath. Temper. Test it out.

Concerning the full on cryo in LN....I had heard three possible benefits. Full RA conversion. Eta carbides to precipitate during temper. A "re-arrangement" of the martensitic matrix. (that last one really intrigued me). But according to a tech paper I read (I can't remember which one it was), they said there was no noticeable change in the tetragonal structure at all. I was disappointed!

Finding the time tho.....?
 
Stuart, after I posted the last question is occurred to me that my reading tells me that the salt pot, etc method of austenizing may be acting to decrease any minor ausgrain growth that you might get from higher heat even if you are VERY careful just because you don't have to hold the piece at austenizing temp (soak) for quite as long. As we know, time and temperature are both our friend and enemy. ;) On the other hand, it sounds like there is some strange structural hoodoo stuff going on with the eta carbide formation....maybe.
 
The temps that the B-U chart had for Elmax were below -120 (can't remember if it was C or F) which is colder an your typical dry ice "boiling" point at 79C/109F, but still significantly warmer than the -196C boiling point of nitrogen. (C02 has no liquid phase to speak of at ambient pressure) I recall being told by someone that unless you are shooting for the eta carbides, you don't need to actually immerse the blade in LN and perhaps you don't want to. The ambient temp in the air space above the surface is likely sufficient to do the RA conversion trick.
stezann said:
as was pointed out one thing is to reach Mf (dry CO2 for short time is enough) to ge rid of stubborn RA, another story is to promote the precipitation of eta carbides, which requires hours in liquid nitrogen and then tempering...and it is unrelated to RA, at least those temperatures are way overkill for this purpose as far as i know.
Click to expand...
 
The LN cold does transform more RA But it ALSO tweeks the lattice to permit the formation of eta carbides on tempering. They are two different processes ,RA transformation is fast shear type, and tweeking the lattice is a slow process as is the formation of eta carbides following that.
Don't worry about getting 100 % martensite .There are a number of things happening to prevent that. Your job is to grt the bestHT for that steel and that use.
 
You must log in or register to reply here.
Back
Top