Grain Refinement

[stephanfowler]

okay, thinking caps on:

given a proper bladesteel of unknown exact composition such as a patternwelded billet

During the process of welding the grain will grow beyond acceptable levels for a proper blade (i think we can all agree on this)

in order to accurately harden the material I must assume that the origin grain size is both important and also somewhat defined.

heres my issue, during the welding process i blew up the grain, in order to have the hardening results i want i need to undergo the heat treat starting with grain size "X"
if i don't know with any reasonable precision how bad the grain got during welding how can i ensure that i refined the grain back to X through "proper" normalization technique.

is this simply a trial and error situation, i must assume that the precise answer varies with the different alloying elements, so theoretically i should expect to trash several pounds of a given alloy combination before nailing down the correct normalization cycle.


my main point here is challenging the good old "heat it till a magnet don't stick and let it cool in vermiculite" deal. I understand that it works "okay" but how can i do it BETTER

i will be looking at investing in salt pots and PID's in the future and want to do a little theoretical work if you get my meaning

 

[Kevin R. Cashen]

Grain refinement doesn’t take that much when one considers the mechanisms it works with. It is now funny to look back on all the bladesmiths who for many years sold their product on “grain size”, or how skilled you had to be to avoid grain growth. To heck with grain growth, it is easy enough to fix grain size that as long as you are aware of it, it is a rather minor issue. The problem is that bladesmiths tend to work without tight temperature controls, judging by color and eye, so something as elementary as controlling grain size became an issue. I tell folks to carefully consider bladesmiths that taught their control of grain size since it is kind of like listing being able to write your name as a job qualification on your resume… not exactly confidence inspiring.

I deal with the grain growth in damascus issues all the time with more concern than most due to the alloys I prefer. O1/L6 will self destruct if you don’t carefully take control of the grains size as you bring it down from welding temperatures. With simpler steels walking away from a welding session will still leave the grains in a horrible condition but you may be able to come out all right the next time you fire the forge but not all alloys are that easy.

First of all, you need to realize that you cannot shrink or “break up” the existing grains, that is just bad terminology and understanding by bladesmiths. What you can do is make new grains that more closely fit your tastes. There are two ways to make new grains, you can transform the austenite into another phase and then back into new austenite grains, or you can induce spontaneous recrystalization through deformation. The latter could be handled by heavy forging as the billet drops to below recalescence (critical).

Simply cooling the billet in the air until the magnet sticks again and the reheating to a reasonable temperature, say 1500F will completely remake the grains within the steel, repeating this process without exceeding the grain coarsening temperature will continue to reduce grain size. The process can be accelerated by reducing the temperature that you go to and even more so by how fast you cool the steel. I am not an all out enemy of multiple quenching; I just feel that the focus gets taken away from proper normalizing by the hype. I keep a large tube of old oil near the forge to quench into after forging in order to speed up the grain refinement. The driving force behind recrystalization is stored energy within the crystalline system, thus hitting the steel with the hammer or rapidly cooling and reheating the steel will drive it along quicker.

As grain size decreases the rate of diffusive based reactions will increase so you will from other structures much more readily upon cooling and the grains will begin to grow at lower temperatures as you continue, so one can reach a point of diminishing returns on the cycling and grain size if you are not careful to chase that recrystalization temperature down the curve, but this should keep us awake at night since an ASTM grain size of 9 or 10 should have most of us giddy with the equipment we work with and will make a fantastic knife. Also remember if you are working with simple steels that as grain size decreases the rate of pearlite or bainite (upper) formation drastically increases so there may be a grain size that you definitely want to stop at if you want full hardness.

If you carefully refine your grain in this step you are entirely justified in having doubts about the old heat it up and put it in vermiculite trick, or even worse stuff it in the forge and walk away. These are recipes for lamellar anneals which involve coarse pearlite, and thus recrystalization. It will require heating the steel to a temperature that may be above the ones you ended with in grain refinement, thus undoing much of your work, and of course the slow cool will not do you any favors either. This is why I work with spheroidal annealing, it is well below any temperatures that can affect grain size and it does not put carbide in the grain boundaries.

I would not be too concerned about the initial grain size of “X” since with proper technique it should be no problem getting back to it and most likely a little finer.

I have a very old thread here dealing with the quirks of different steels and how they recrystalize that is worth digging up in order to emphasize the point that one cannot treat all steel the same even in normalizing, but I am unable to find it, I thought I had posted it on this forum, hmmm… I will find a cop on another forum that I did put it on and post it after this reply, just be aware that the discussion is over 3 years old so some details may be lost to me.

 

[Kevin R. Cashen]

A tale of two steels - 12-01-2004, 09:01 PM
________________________________________
I had the forge running today and after banging out a couple of pattern welded blades I decided to use up some scrap metal to satisfy some curiosity.

From this I have an interesting little experiment to share with you all, in hopes that it will spark some great conversation. I had a real good idea of what my results would be with this exercise, be fore I did it but I had no idea how pronounced it would be, so it was a good learning experience for me as well.

One thing I am really excited about is that this is the first images that were prepared, photographed and uploaded to the internet, all from my new lab! The wireless networking and the digital cameras work better than I expected!

The first series of images are of 1084 that has been heated to approx. 2100F. and then normalized at different temperatures without cooling in between. The bar was heated to 2100F. and then a specimen was sliced off, quenched and broke to reveal the grain. This was all done before the parent bar cooled below critical.

Image #1 is the steel grain structure as it was at 2100F. Image #2 is a specimen that was sliced off after the bar had cooled to the point that it just started to regain magnetism and was then reheated to approximately 1450F. Image #3 is a specimen from the same bar that was allowed to cool to around 1000F and was then reheated to approx. 1450F. and sliced, quenched and broke.

The results are obvious to see. #1 shows marked grain growth, #2 shows a very rapid reduction in grain size simply by heating from a stage at which some pearlite could begin to form. #3 shows a very nice grain from yet another reheat after complete pearlite was allowed to form. The grain is fine but the break is not clean due to the formation of fine pearlite before I could get the slices quenched. I could do this better, the next time, if I were to put a bucket of water directly below the cutoff saw and have the hot pieces drop straight in.

No forging was done on these pieces; all the grain refinement was the direct result of recrystalization from a microstructure that produced increased nucleation of the fresh grains. With 1084 this can be done by cycling at a relatively high temp without letting things go cold; the bar never did drop below 800F.

Now we have the next series of images:

This is L6 steel. #1 is heated to approximately 2100F. then sliced and quenched. #2 is from the same bar after it has cooled to approx 1100F. and then reheated to approximately 1500F and then sliced. #3 is from the same bar that has been allowed to cool to about 950F. and then reheated to 1500F then sliced, quenched and broke. #4 is the same bar allowed to cool to around 700F and then reheated to 1500, sliced quenched and broke.

As I said, I predicted this outcome; I just didn’t think that I could get such good examples of it under my camera. To spur some good conversation, lets discuss what all this means and how it can help us in getting the most out of our selected steel.

 

[Karl B. Andersen]

Fantastic stuff.

 

[stephanfowler]

marvelous -

that answers my base question very well but leads into a whole new realm......

basically this confirms the assumption that the forging process does not necessarily effect the outcome of HT in relation to grain "dynamics".


effectively at any point that the steel goes above a given temperature (different depending on alloy??) the grain is returned to a blank slate, and much greater effect depends on how the steel returns to a lower temperature (quickly, slowly, etc)

if this is the case then given a patternwelded example, how would you determine which steel process to follow.... i.e.

i have a 256 layer billet of W2 and O1 (odd combination i realize)

if i have determined that i get excellent results for plain W2 by process A, but horrible results for O1 by same process

And i get excellent results from O1 by process B, but horrible results for W2 by same process

should I ----

test an entirely new process
go with process A and the O1 be damned
go with process B and the W2 be damned
forget about mixing these two steels entirely......

hypothetically

 

[Kevin R. Cashen]

My vote would go for-

forget about mixing these two steels entirely......

Because with such a mismatch grain refinement and normalizing procedures will be the least of your worries, and as you have already pointed out they alone will be no small headaches.

It is kind of like going into a doctors office and asking him what to do because it really hurts when you do "this”, and then demonstrate by driving a red hot spike through your hand. I doesn't take a genius or a psychic to guess what his immediate response will be.

 

[wiggins]

Kevin, this is good stuff. Thanks for posting it. I understand that each of the pictured samples were cycled from lowering heats. What I don't understand is if each sample was cycled one time, or if each sample was a progression from the previous sample. Bill Wiggins

 

[stephanfowler]

My vote would go for-

As it very well should, however, consider a situation where each steel has some respective property that is highly desirable to your chosen end use which is not found in the other steel,

is it just a judgement call as to whether to deal with the headaches associated in such a terrible mismatch of alloy possibly resulting in a suboptimal end product,

or go with a suboptimal blend of steels in the first place


(i realize that i'm dealing in the realm of hyperbole to the point of tomfoolery, im just strutting down the primrose path of my thought process, learning new stuff)

 

[Greg Obach]

very nice pictures..

thanks for showing

Greg :thumbup:

 

[mete]

Since we're talking about grain refinement let me add grain refinement by alloying. We see additions of small amounts of V or Cb [~.020 V ] to 'refine 'grain. What actually happens is that for a grain to grow it absorbs an adjacent grain. To do this the grain boundary must move. V goes to the grain boundary and slows down the movement of the boundary thus 'refining' grain. So grain refining alloying elements don't eliminate the problem , only reduce it.

 

[Tai Goo]

"Grain refinement doesn’t take that much when one considers the mechanisms it works with. It is now funny to look back on all the bladesmiths who for many years sold their product on “grain size”, or how skilled you had to be to avoid grain growth. To heck with grain growth, it is easy enough to fix grain size that as long as you are aware of it, it is a rather minor issue." Kevin

I agree with Kevin.

Grain refinement isn't the end all of end alls. It's nothing new that a finer grain is tougher than a course grain. Proper processing will produce a refined structure. However, the stress involved, if taken too far, could actually become a problem. Each heat is a thermal stress cycle, etc...

If there were a holy grail or end all of end alls, it would be predicting exactly how many stress cycles, (thermal and/or mechanical), and to what degree (or under what circumstances), "over a length of time", any given piece of steel or metal could take before fatigue (or failure) occurs.

... and how to fix it, if possible and/or practical, under the circumstances,... from a theoretical standpoint.

 

[Tai Goo]

... Hope that helps.
LOL

 

[Tai Goo]

Along with fatigue and failure, I forgot to mention burning it out or rendering it useless in one way or another, in general terms...

 

[stephanfowler]

Tai - understood, the pretext of my question was more along the lines of "if i'm really good with this steel, and really good with that steel, can i combine them and still be good even if they don't naturally seem to compliment each other" in a nutshell

Also, just to clarify, i believe kevin's first response and subsequent repost of an old thread more than answered the information i was looking for, i'm just interested in continuing the discussion....

 

[Tai Goo]

Tai - understood, the pretext of my question was more along the lines of "if i'm really good with this steel, and really good with that steel, can i combine them and still be good even if they don't naturally seem to compliment each other" in a nutshell

Also, just to clarify, i believe kevin's first response and subsequent repost of an old thread more than answered the information i was looking for, i'm just interested in continuing the discussion....

I didn't see a single question mark in your original post,... so I'm still not sure what the question is. However, yes it is possible depending on the variables, as long as you haven't "overworked" it, there is enough carbon in the steel, and you haven't rendered it useless in one way or another,... theoretically.

It is basically trial and error... and you can always do it better.

 

[stephanfowler]

I didn't see a single question mark in your original post,....

Tai - that would be because there was not a single definable question, just a rambling juxtaposition of incoherent thought that had been naggin at me in regards to screwing up perfectly good steel by overheating in the patternwelding process.......

 

[Tai Goo]

Tai - that would be because there was not a single definable question, just a rambling juxtaposition of incoherent thought that had been naggin at me in regards to screwing up perfectly good steel by overheating in the patternwelding process.......

Me too, and the rest of us it seems...

Overheating in the early stages of pattern welding isn't usually as critical as overheating in the finished billet or blade. Forge welding is overheating, but you can get away with it and even use it to your advantage, if it and the rest of the processing accounts for it...

You have to find your margins for error and try to stay within them.

 

[Mike Krall]

What HT and/or physical characteristics of O1 and W2 make them incompatible? And on the other side of that coin, what is the allowable range in the "important characteristics" that constitutes "compatible"?

Mike Krall

 

[stephanfowler]

which brings us full circle to where i was kind of going originally, how to define the process for establishing those margins.


(to clarify - I'm not actually asking anyone to define the margin FOR me, im looking for input on how best to go about defining that margin for myself.)

 

[stephanfowler]

What HT and/or physical characteristics of O1 and W2 make them incompatible? And on the other side of that coin, what range of important characteristics constitutes "compatible"?

Mike Krall

in very basic terms, O1 does best with a slow quench, W2 does best with a fast quench