Welcome! Log In Create A New Profile

Advanced

Research on ultra fine grain steels

Posted by CliffStamp 
Research on ultra fine grain steels
December 21, 2014 04:54PM
Steel is made out of grains, just like there are grains in wood. A fine grained steel will have an ASTM grain size of 8-9, beyond that they are considered super fine or ultra-fine or some other adjective. A grain size of 8-10 means the austenite is 10 to 20 microns in grain size.

In some detail (who doesn't love charts) :

-grain size : [drive.google.com]

Recently there have been development in nano-structured steels which have grain sizes of 25+ . They can also be carbide-free bainite and they are not only extremely tough (they are used for armor plating) but the wear resistance in standard silica tests is likely to be surprising :

1006 (mild steel) : 83
4340 at 53 HRC : 40
1060 at 63 HRC : 17
1084 (nano-bainite) at 56 HRC : 8.1
D2 at 58 HRC : 7.8

Full paper : [www.msm.cam.ac.uk]

A useful conversion tool : [www.carbidedepot.com]

Just to clarify, the alloy content of that steel is 1084 with some light alloying to ensure bainite and prevent carbide formation and allow some retained austenite. The full details are in the paper.

--

As a side note, being able to say you are make a knife out of nano-structured bainite which is tough enough to make armor plating and has wear resistance equal to D2 - well that is a pretty decent promotional claim, especially when it is actually true.



Edited 1 time(s). Last edit at 12/21/2014 04:55PM by CliffStamp.
Re: Research on ultra fine grain steels
December 22, 2014 05:49AM
Cool article, thanks Cliff. However, if I understand correctly, the way the wear resistance is measured here may have little relevance to cutlery. Basically, if my understanding is corect, because of the high friction the outer layer of the bainitic steel reaustenitises and work hardens. But yeah, work hardening bainite would attract a lot of fans, starting with the Clark indestructible katana and finishing with the "H1 68HRC by sharpening" crowd.
Re: Research on ultra fine grain steels
December 22, 2014 07:54AM
I would not argue that it is ideal, but note that is the standard test for three body abrasive wear which is what almost all of the references to abrasive wear refer to. There are other methods such as pin-on-disk or ball cratering, if they are used they would be cited, ball cratering is getting more popular because it focuses wear on isolated spots. The main CATRA machine is essentially a abrasive wear test (2-body) for knife edges and in general it will tend to correlate well to the standard rubber wheel abrasive test, however I am not sure the results would be identical for the exact points you raised.

In general though, I think an extremely critical question is always to ask is any statistic/measurement being cited actually relevant to the application being considered.
Re: Research on ultra fine grain steels
December 22, 2014 08:12PM
For a while now, I look up to Prof Bhadeshia's works & perspectives.

Now, my rambling...

Mn 2.28% + Si 1.9% + Co 1.55% pinned & nucleated grain and at the same time preserve excess aust (from carbon excess of 0.4%) from cementite precipitation. To me this is a local potential minima in an energy crater (bounded by not very high rim). It might be good for breadth & depth surfaces where shock obsorption and high ductility are intended target/use.

It probably won't be good for edge tools because of work-hardening RA -> martensite and cementite. High kinetic (impulse forces) impacts can cascade this conversion through the interconnecting RA network sandwich by bainite sheaves. Some brittle converted mart & cementite will fall out, other will propagate/cascade impulse force to further conversion a way from point of interaction. in the field of converted, Mn+Si+Co are becoming more/less nano inclusions because spacing & favorable bond to RA are now lacking.
Re: Research on ultra fine grain steels
December 22, 2014 08:35PM
The austenite to martensite is clear in the paper as is the very high hardness of the transformed martensite which matches the hardness of untempered martensite as expected. It has to be clear in any of these tests what they are measuring, in this case the abrasion is reduced both by the austenite transforming which absorbs energy and further by the then full hard martensite resisting it. I would be curious how a CATRA run would hold. But I agree in general, especially if a knife was going to be used for impacts, untempered martensite is likely not a great idea for blades.
Re: Research on ultra fine grain steels
December 22, 2014 08:43PM
Untempered martensite bands won't be as bad as cementite bands. bainite sheaves sandwiched these bands...

Abrasion/friction end up as thermal, which probably favor mart conv than fe3c precip. but impulse/impact/chop energy would induce fe3c conv.
Re: Research on ultra fine grain steels
December 22, 2014 08:51PM
Ref : [ojs.cvut.cz]

"Superior Properties of Ultra-fine-grained Steels" : J. I. Leinonen

This paper is more about ferrite and how to produce steels which have extreme strengths and toughness values, even at ultra-low temperatures by producing extremely fine grained austenite through hot rolling just below the crystalization temperature. The grain size is ASTM 14-15. There is some interesting information on the dramatic increase in yield strength and toughness as a function of grain size though the charts are extrapolated and they are specifically talking about ferrite. They are working in the 1 GPa range which is only about 30 HRC normal martensite, but they are getting it in steels which are much weaker/more brittle in traditional senses. Plus they are constrained by wanting levels of retained austenite for working hardening.
Re: Research on ultra fine grain steels
December 22, 2014 08:56PM
Quote
bluntcut

Abrasion/friction end up as thermal, which probably favor mart conv than fe3c precip. but impulse/impact/chop energy would induce fe3c conv.

What is this based on exactly? Are you talking about at the apex or above it?
Re: Research on ultra fine grain steels
December 22, 2014 09:13PM
Apex & affected depth behind apex.

Based on energy configuration. Where impulse energy has high energy peak, could translate into of equivalent of thermal in 1000+F, so cementite could precip on the way up or down (from extra high).

Quote
CliffStamp
Quote
bluntcut

Abrasion/friction end up as thermal, which probably favor mart conv than fe3c precip. but impulse/impact/chop energy would induce fe3c conv.

What is this based on exactly? Are you talking about at the apex or above it?
Re: Research on ultra fine grain steels
December 22, 2014 09:14PM
"Ultrafine-grain Effect on Martensitic Transformation in a Hypereutectoid Steel - Fuliang Lian, Yongning Liu, Hongji Liu, Junjie Sun, Xuejiao Sun"

"Martensitic transformation in an ultrafine-grained (UFG) steel was investigated. The UFG steel was a plain carbon steel containing 1.0 wt.% C and prepared by thermomechanical treatment. The grain size was approximately 1 μm. The sample was reheated at 1023 K for 5 min and then cooled in water to conduct a martensitic transformation. In addition to martensite, lamellar pearlite was observed in microstructure. When the heating time was prolonged to 7min, after quenching process, matensite was only observed. The results indicate that when heating time is shorter, the grain size is small (2.5 μm) and lamellar pearlite can be formed as quickly as martensitic transformation in quenching process. When the heating time is prolonged, grains grow (4 μm) and no lamellar pearlite is observed. This phenomenon is greatly related the increasing resistance for martensitic transformation and accelerated diffusion of carbon atoms due to the refinement of grains in the steel."

In summary, a 1% C steel (1095) had ultra-fine grain (15 ASTM) and when it was water quenched it formed a significant amount of pearlite due to the very fine grain. In order to get full martensite with a water quench, the grain had to be coarsened to increase hardenability. Beyond 12-13 ASTM then water isn't a sufficient quenchant for 1% carbon steels.

In short, if you want to get UF grain in martensite, it looks like you will need more than water as a quenchant to enable full martensite - which is pretty much exactly what bluntcut noted in his quench comparisons awhile back.
Re: Research on ultra fine grain steels
December 22, 2014 09:20PM
While I have seen impacts be noted to cause transformation of retained austenite, it isn't due to an actual increase in temperature as much as it is just the localized strain due to deformation. Are you talking about impacts which don't grossly damage the blade and are just part of the slow wear process?
Re: Research on ultra fine grain steels
December 22, 2014 09:36PM
Those pesky RA are great for body-armor but poke them with high energy impact with equiv of thermal 600+F, those will convert into cementite. Problem is these RA in a form like strings, hence cementite interconnected strings/bands outcome -> crack propagation...

Pardon my dances around ND (non-disclosure) in my posts.

Quote
CliffStamp
While I have seen impacts be noted to cause transformation of retained austenite, it isn't due to an actual increase in temperature as much as it is just the localized strain due to deformation. Are you talking about impacts which don't grossly damage the blade and are just part of the slow wear process?
Re: Research on ultra fine grain steels
December 24, 2014 10:01AM
"Lath martensite in 1.4%C ultra-high carbon steel and its grain size effect"

Abstract :

"1.4%C ultra high carbon steel (UHCS) was prepared in order to study the substructure of martensite transformation. Because of ultra-fine spherical carbide, the growth of austenite grain, whose average size was 2.5 micron, was prohibited. After quenching, there was a great deal of lath martensite. The sub-structure was composed of a large quantity of dislocations and twins. Through calculation, it was determined that twin shear stress increases faster than that of slip due to the reduction of austenite grain sizes. A model based on twin and slip shear stresses has been proposed, which yields critical grain size ranges from 1–4 micron. The result is in agreement with measured results."

In short :

-at ultra fine grain sizes, ~15 ASTM, twinning is suppressed and martensite forms by slip and thus lathe martensite is produced

-the ultra fine grain and lathe martensite produces extreme strength (4.7 GPa) and ductility (26% compression stress)
me2
Re: Research on ultra fine grain steels
December 24, 2014 02:32PM
I haven't looked into many of the other papers posted lately, but the stuff on UHC steels is pretty interesting.
Re: Research on ultra fine grain steels
December 24, 2014 03:36PM
There are people using UHC steels (Roselli) but no information on how they are hardening them. If they are not generating the UF grain then the compression strength and ductility won't be similar to what is being discussed. Given how it can be achieved through quench cycling, it isn't difficult for a maker to do. The only problem is making sure you are really getting the UF grain which will take a little work. Ideally you would want to correspond with someone doing the research. In many cases these people, if you hit the right one (like Verhoeven), will collaborate.
me2
Re: Research on ultra fine grain steels
December 24, 2014 05:31PM
From the original paper, the processing from the mill required both a thermal and mechanical component. If that structure from the mill is altered by the final processing, I'm not sure I'd describe it as not being difficult to regain the ultrafine grain size. The research by an author on BFC on damascus showed simply cycling in a knife making kiln type set up resulted in a grain size coarser than the as forged damascus.

Also, while perusing the Verhoeven classic (for knifemakers anyway), I saw what appears to be a conflict in data. The hardenability experiment provided shows a fully hardened sample of 1084 with an ultrafine grain size (ASTM 15) with a highly agitate oil quench.

It also brings up another point to further expandd the discussion, which is the use of alloying to increase hardenability. For example, take an UHC steel alloyed with nickel, chrome, & molybdenum subjected to grain refining processes and I would hazard a guess that an oil quench would be perfectly adequate to avoid any higher temperature, diffusion based, non-martensite phases (pearlite, upper bainite, or perhaps even lower bainite).

It was brought up in the Spyderco thread, but I wanted to discuss it here, as it relates to the UF steels. Ed Fowler always comes up in discussions of ultrafine grain sizes. I've always wondered what the as quenched hardness of his blade edges is. With a grain size of 14ish, and the use of a relatively slow oil (Texaco Type A, 18 second oil from discussions on his site) I'd again hazard a guess that there is something other than martensite.
Re: Research on ultra fine grain steels
December 24, 2014 08:02PM
Quote
me2

Also, while perusing the Verhoeven classic (for knifemakers anyway), I saw what appears to be a conflict in data. The hardenability experiment provided shows a fully hardened sample of 1084 with an ultrafine grain size (ASTM 15) with a highly agitate oil quench.

The rate of agitation and even quench volume would have effects as would the relative size of the same piece as would alloying composition as you note further. At the sizes we are talking about here even rounding would be a difference as one is claiming 12-13 is ok (4 micron) and another 15 (2 micron).

Quote

It also brings up another point to further expandd the discussion, which is the use of alloying to increase hardenability. For example, take an UHC steel alloyed with nickel, chrome, & molybdenum subjected to grain refining processes and I would hazard a guess that an oil quench would be perfectly adequate to avoid any higher temperature, diffusion based, non-martensite phases (pearlite, upper bainite, or perhaps even lower bainite).

I would think so but I don't think it is trivial to claim the other properties would not be effected.

Quote

It was brought up in the Spyderco thread, but I wanted to discuss it here, as it relates to the UF steels. Ed Fowler always comes up in discussions of ultrafine grain sizes. I've always wondered what the as quenched hardness of his blade edges is. With a grain size of 14ish, and the use of a relatively slow oil (Texaco Type A, 18 second oil from discussions on his site) I'd again hazard a guess that there is something other than martensite.

The one I had was very soft, however I don't think they are all like that. The curious thing about Ed's data is that at times you get a lot of references to grain sizes and such but no micro-hardness tests which only cost ~$7 a shot. The work he does get done to measure grain sizes and such is much more involved.
Re: Research on ultra fine grain steels
December 25, 2014 11:04PM
This thread and the UHC one and other recent threads here really bring home that there are still, at a minimum, thousands of researchers in hundreds of institutions working towards post-grad degrees and post-doctoral work and senior researchers getting sizable grants to investigate the reasons behind and approaches for improving the strength/toughness balance of fairly basic steel mixtures in the high strength areas. Some otherwise sound folk seem to have decided that all that ever can be known is already in the sacred texts or elsewise covered in their pronouncements and presumably that all these grant committees need sacking and this ongoing activity is equivalent, or less, than the continued playing of the band on the foundering Titanic.

While there is a wealth of existing knowledge that goes woefully underexploited, there is still plenty of active areas of work uncontroversially ongoing in the real literature relating to mechanical processing (thermomechanical or otherwise), intensive quenching (using high pressure/velocity water jets like a pressure cleaner to rapidly but briefly quench), induced superplasticity and other areas which very well may or even may not relate to useful properties in this particular odd little subject of simple bladed cutting tools.

Cliff’s recent posting from the archeologic literature also reminds us that we may not be the first to pass this way. While the mechanisms may have not been known in the past, and potentially are still not, it doesn’t mean that real effects were not stumbled over by dedicated craftsmen of yore working with particular raw materials in conducive environments.

The insistence that this or that is the one, only and final Truth, and if by some chance it isn’t then we don’t care because it doesn’t matter anyway because we are happy here and don’t need anything better, is very reminiscent of the observation (unfortunately I know not of whom) that a conclusion marks that place where one grew tired of thinking. Especially when employed by folk whose designs, unlike Kyley say, appear very far from functionally optimized.

To take that attitude, we may as well all just settle for an expenditure somewhere around $10 to $50 and get better at sharpening and cutting and let the rest go. Admittedly, this is actually where I find myself in real life nearly all the time these days, but I still wonder what could be better and like to leave that door open.

I appreciate that there are more ways to stuff things up than to make them better, essentially the second law of thermodynamics – entropy rules - and that good enough is good enough and is rarely approached as it is, but that doesn’t mean that the show is over. Not by a long way.

Or, if it is, no-one told these guys:

[www.mpie.de]

[www.intensivequench.com]

[www.jim.or.jp]

[www.diva-portal.org]

[cdn.intechopen.com]
Re: Research on ultra fine grain steels
December 27, 2014 06:24PM
So on that first link on sandgrouper's post above they are talking about austenitizing, deforming mechanically to reduce grain size, and quenching? That seems workable as plate... I mean to a knifemaker... heat in furnace, hydraulic squeeze or air hammer in special dies, quench.

The second link really spoke to me as well. I have long pondered if distortion can be reduced by cooling the entire part so fast it has less opportunity to distort... apparently so

Still more reading to do... want to get into this 4.7GPa stuff
SQ 52100 grain?
March 07, 2015 11:35AM
I am a newb at optical micrographing & interpretation, please don't hesitate to correct & dish out pointers. At 1Kx, it should able to resolves ~0.25um. However I don't know whether my scale is anywhere near actual.

SQ 52100 nital 3% etched 3-5 minutes (perhaps, too long to yielded sponge like topo)

edit: 64rc

Your thoughts?



Edited 1 time(s). Last edit at 03/07/2015 11:40AM by bluntcut.
Re: SQ 52100 grain?
March 07, 2015 12:43PM
Is this quenched and tempered? In order to see the austenite grain ideally you wanted quenched only, and if tempered, below 200 C to ensure you are in the first stage of tempering only. Once you pass 200C then the retained austenite decomposes, the transition carbide coarsens to cementite, RA turns to bainite, and it gets very unclear.

The primary carbide is there at < 2 microns, the austenite grain size is unclear and in general is very hard to read if tempered.
Re: SQ 52100 grain?
March 07, 2015 01:05PM
Thanks Cliff.

This blade was SQ+cryo-dip and 1st tempered at 275F mentioned in pic (135C). I don't see any carbides. I believe those white egg shapes are partially acid dissolved grain at boundaries. 52100 doesn't has enough Cr to aggregate to large clumps like that. heheh, newb speaking anyway.

With more xref about scale, I think, the scale is right for 1Kx resolution but this newb willing to learn ...

Quote
CliffStamp
Is this quenched and tempered? In order to see the austenite grain ideally you wanted quenched only, and if tempered, below 200 C to ensure you are in the first stage of tempering only. Once you pass 200C then the retained austenite decomposes, the transition carbide coarsens to cementite, RA turns to bainite, and it gets very unclear.

The primary carbide is there at < 2 microns, the austenite grain size is unclear and in general is very hard to read if tempered.
Re: SQ 52100 grain?
March 07, 2015 01:51PM
Those are tiny, 1-2 microns and the volume fraction is very low, on the order of 1%. See for example, "Effect of Austempering and Martempering on the Properties of AISI 52100 Steel, P. Vamsi Krishna" which shows the same size/distribution of carbides in 52100 as hardened, volume fraction is of course higher when annealed. Note that iron carbide will also form in/with the chromium carbide as will any other trace carbide former. Chromium is the most friendly out of all carbides, it invites everyone else inside.
Re: SQ 52100 grain?
March 07, 2015 03:53PM
Interesting - my micrograph looks like the 'austempered 10 minutes soaking time' pic. Here more info on my ht

Lengthy prep
aust 1425F (not a typo, I have a 1420F with 65rc after temper) [www.bladeforums.com]
SQ to room temp
wash with soap
dip in LN2 (5 minutes)
temper 275F for less than 1hr.

grind -> polish -> etch -> micrograph

Now, I need to learn how to read my micrograph


Quote
CliffStamp
Those are tiny, 1-2 microns and the volume fraction is very low, on the order of 1%. See for example, "Effect of Austempering and Martempering on the Properties of AISI 52100 Steel, P. Vamsi Krishna" which shows the same size/distribution of carbides in 52100 as hardened, volume fraction is of course higher when annealed. Note that iron carbide will also form in/with the chromium carbide as will any other trace carbide former. Chromium is the most friendly out of all carbides, it invites everyone else inside.
Re: SQ 52100 grain?
March 08, 2015 09:12AM
I spoke to Roman, he noted that they could be chromium carbides, they are the right size/volume, but they could also be inclusions as you said, they would show up exactly the same, hence why if you really want to know you have to do a more layered analysis, ideally you use a SEM and you can do a sub-structure analysis and you can also use x-ray diffraction. However, back to practical inference :

-try to do a lighter etch, it looks like you are getting topological features due to random etching depth
-work on resolving/focus

Now to be clear, the second part is something i don't have a lot of patience for. In high end microscopy it is almost like being a Sniper (from what I read, no direct experience). Movement has to be kept to a absolute minimum, the sample has to be locked in place and pictures have to be taken with no movement of the sample or instrument. This is why the small images of edge are always blurry because they are always in motion. Clay Allison (from Wicked Edge) takes this as seriously and has build custom clamps/mounts to keep everything rigid and he gets really nice results.

Of course with light microscopes you are also going to be at the resolution point at some level.
Re: SQ 52100 grain?
March 08, 2015 11:44AM
Cliff - thanks very much your time & brain. Please extend my thanks to Roman.

I calibrated my new scope, so the scale is confirmed as accurate.

Here is a diff img from same blade with 30 seconds nital 3% etched at 1000x mag. There is a high volume % of round shape objects, which 5-10x higher than Cr alloyed particles volume%.


btw - my D2 carbide refinement only partially worked. I spotted a large continental size carbide - a whopping 30um across. I need to aust at a higher temp to bust these large ingot carbide. Didn't see grain size well but I think it's around 9-10 astm.

Quote
CliffStamp
I spoke to Roman, he noted that they could be chromium carbides, they are the right size/volume, but they could also be inclusions as you said, they would show up exactly the same, hence why if you really want to know you have to do a more layered analysis, ideally you use a SEM and you can do a sub-structure analysis and you can also use x-ray diffraction. However, back to practical inference :

-try to do a lighter etch, it looks like you are getting topological features due to random etching depth
-work on resolving/focus

Now to be clear, the second part is something i don't have a lot of patience for. In high end microscopy it is almost like being a Sniper (from what I read, no direct experience). Movement has to be kept to a absolute minimum, the sample has to be locked in place and pictures have to be taken with no movement of the sample or instrument. This is why the small images of edge are always blurry because they are always in motion. Clay Allison (from Wicked Edge) takes this as seriously and has build custom clamps/mounts to keep everything rigid and he gets really nice results.

Of course with light microscopes you are also going to be at the resolution point at some level.
Re: SQ 52100 grain?
March 09, 2015 05:09PM
FWIW - here is a reduced from full size of gray scale negative on the same spot as pic-ed above



Also infered from figure 4.23 SEM pic in Verhoeven book page33. Round shape objects in prev post's pic are cementite. I've mentioned before how I was changing carbide configuration. Well, I over inflated carbide size, where my aim is around 0.5um. At any rate, will worry about reduce cementite size backdown to submicron once grain size issue resolve.
me2
Re: Research on ultra fine grain steels
March 10, 2015 03:38AM
How were these polished?
Re: Research on ultra fine grain steels
March 10, 2015 08:32AM
Grinded a flat piece to 800 grit belt. Waterstone progression from 240 to 12K. then 0.5um CrO stropped. Mirror finish but there are still some low grit scratches. I washed the surface with dish washing soap and clean with alcohol prior to 30 seconds etch (nital 3% methanol). I did tried 15 seconds but surface didn't bloom much, hence not much to see but a smooth surface. Next time, I will try 20 and 25 seconds. Image stitching could be done to gain depth - I save this time consuming for later when I know what I am doing at a simple level.

Suggestions/advises/interpretation are appreciated.
Quote
me2
How were these polished?
Re: Research on ultra fine grain steels
March 10, 2015 09:41AM
The gain boundries look to be visible in the second picture, 4-5 microns in size.
Sorry, only registered users may post in this forum.

Click here to login