O-1 steel air cooled breaking

[Woody]

I saw that video, junk science at best, totally lacking in any scientific proceses. No one had documented how long the material is held at temperature, the exact temperature the material was heated to or anything else that could be duplicated by others. Once again is there any publication that documents this or is it just alchemy and blacksmith hear-say. On Page 69 of this document Metallurgy of Steel for Bladesmiths and Others who Heat Treat and Forge Steel by John D. Verhoeven Emeritus Professor Iowa State University it is documented that repetative heating and quenching in oil reduced the grain size in the steels tested.

In the above mentioned article where repetative quenching has reduced the size of the grain, he states " The toughness of steel is improved as grain size becomes smaller. For this reason it is important to do heat treatments at as low a temperature as possible to reduce grain growth of the austenite " And the last I knew, repeated quenching did not make the steel less hardnable.

I think that it is about time that this Forum either fully documented this normalizing theory by finding a published reference that proves it, or drives a stake it's heart and gives it a decent burial.

[MattBower]

It's not a video, Woody. If you actually "watched" it I don't know how you could think that it is.

Second, you're wrong that it's not repeatable. The write-up clearly states that the steel was heated just to decalescence, removed from the forge, and allowed to air cool to black. You're telling me you don't think you can manage that? The decalescence point is better known as Ac1. It will vary from steel to steel, and with heating rate, but it marks the transition to eutectoid austenite (although not dissolution of carbides in hypereutectoid steels). The neat thing about it is that, with a little practice, you can see decalescense happen, which is the technique they used in that experiment. (This is easier to see in some steels than others. The reason the phenomenon can be visible has to do with latent heat of phase transformation.) But if you happen to have a kiln or HT oven and would prefer to run the experiment in a way that doesn't require you to actually observe decalescence, I'd be very happy to help you set up the experiment.

Now, the thing that really confuses me is that you're citing Professor Verhoeven for the principle that repetitive heating and cooling cycles reduce grain size, but you're also rather aggressively challenging a claim that repetitive heating and cooling reduces grain size. Multiple quenches will accomplish the same thing as multiple normalizing, and faster [edit: when I say "faster" I mean that starting with two identical samples of a fairly large grain size, heated to the same appropriate temperature, if you air cool one and quench the other to martensite, it's my understanding that the martensite sample should have a slightly finer grain, although both will be smaller than the original sample] -- but of course normalizing doesn't carry with it the danger of quench cracking, which is why I, for one, prefer to normalize repeatedly rather than risking multiple quenches.But don't take my word for it; let's go back to Prof. Verhoeven, on the very same page you're relying on:

If one wants to reduce the grain size of a ferrite/pearlite steel, a simple technique is to heat into the austenite region holding the maximum temperature as low as possible and the time as short as possible and then rapidly cooling back to room temperature as fast as possible without forming bainite, assuming it is not desired. Since smaller initial grain size promotes smaller final [edit: was originally "initial" -- fixed that] grain size, repeating this cycle several times will enhance fine grain size.

Bear in mind that we're talking about forming fine pearlite grains here, so we have to read carefuly enough to realize that when he says, "cooling back to room temperature as fast as possible," he doesn't mean "so fast that we form martensite." Thus, at least with respect to steels with sufficient carbon to harden, he can't be referring to a full quench like we'd do if we were hardening. "As fast as possible (without forming martensite)" is a reasonable description of air cooling for most simple steels. (As you pointed out, multiple cycles of quenching to martensite would in fact work for grain refinement, too, but that's not what Prof. V is discussing in this pargraph. Here he's talking about forming fine-grained pearlite and ferrite, not fine-grained martensite.) In fact I suspect there's a typo in that paragraph; I think when he wrote, "without forming bainite," he meant martensite. The reason I say that is that rapid cooling isn't something we normally associate with bainite formation; to get much bainite we normally have to quench fast to below the pearlite nose, then hold a little above the martensite start temperature for a very long time. So it doesn't really make sense to me that he'd be worried about bainite formation in this context. Martensite, though, is a different matter. But I could be missing something there.

Please also note that, a few lines further up the page, Verhoeven explains that "new austenite grains begin to form on the old pearlite grain boundaries as shown schematically in Fig. 8.4."

If, as Prof. Verhoeven says, repeated heating and cooling into and out of the austenite region reduces the grain size of pearlite, and new austenite grains begin to form along the boundaries of exisiting pearlite grains, and if the austenite grain size determines the grain size of martensite after we quench (which it does -- he addresses that at the beginning of the final paragraph on p. 69), then it's no mystery that forming fine pearlite as described above, followed by heating to the appropriate transition temperature and quenching to martensite, will produce fine-grained martensite. That's exactly what we're doing with multiple normalizing cycles, followed by a quench to harden -- assuming we execute them properly, of course.

Finally, you say that "the last I knew, repeated quenching did not make the steel less hardenable." We have already established that repeated quenching makes grain size smaller. You seem to agree with that proposition, at least. So let me quote Prof. Verhoeven from p. 83 of the very same book you're relying on:

To improve hardenability one needs to force the various start curves that appear on the IT diagrams to longer times. Physically this means that one must make it more difficult for the 3 product constituents, ferrite, pearlite and bainite [note: again, I'm a little confused about why bainite is an issue here, but it doesn't matter for our purposes] to form in the austenite. There are 2 principal ways in which this can be effected:

(1) increasing grain size, and (2) adding alloying elements.

(1) Effect of grain size on Hardenabiltiy It was pointed out above that the product constituents virtually always form on the austenite grain boundaries. The amount of grain boundary area depends on grain size. A larger grain size will reduce the amount of grain boundary area per unit volume, which will shift the start curves to longer times and improve hardenability. Hence, the position of the curves on IT diagrams depend on the austenite grain size. For this reason, as shown on Fig. 9.3, published IT diagrams always specify the grain size.

So let me ask you: if increasing grain size moves the start curves on the IT diagrams to the right, indicating greater hardenability -- which is what Professor V just said -- what must we necessarily conclude would happen if we reduced grain size? Which way would the curves move, and what would be the effect on hardenability of the steel? Remember: we've already established that changing grain size moves the curves, and that moving the curves affects hardenability. So . . .

[Woody]

I know its a slide show or pictures with written text, the whole point is exactly what temperature was each piece of steel heated to? Answer unknown. How long was each piece soaked at temperature? Answer, Unknown. Was each piece of steel cooled at the same rate for the same time? Answer unknown. Was each piece of steel quenched from the same temperature? Answer, Unknown. If it isn't documented and repeatable no matter what the results it's junk science. Take that neat little slide show into any university and submit it for peer review and you would get laughed off the campus.

Verhoeven quenched his steel in rapidly stirred oil, quite a difference for air cooling wouldn't you say. He also states "rapidly cooling back to room temperature." Which don't seem to be the slower cooling that is the normalizing process but more like an air quench.

This brings up another question large grain = harder metal, why are all the knifemakers triple normalizing and many triple quenching to shirnk the grain size to get better edge holding properties? It would seem counter productive.

As far as I can tell the objective of normalizing is to relieve the stress in the metal produced by the forging process and prevent cracks and breaking when the metal is further processed. Once again, everything I have been able to find on normalizing calls for a soak time at a specific temperature and then air cooling, nothing I have been able to find calls for a repetition of this process, you would think that if there were some benefit to be had, it would be documented somewhere. I would like to find that documentation.

I really didn't mean to hijack this thread, in regard to the original question about the O-1 breaking. Obvious answer is you hit it when it was too cold and cracked it.

[MattBower]

Woody, I have other things to do today, so this will be my final post on this issue. Bottom line: I'm starting to suspect that you're arguing just for the sake of being contrarian. Here's why.

First, your questions in your previous post make it obvious that you don't know the difference between hardenabilty and hardness, or understand the basics of the interaction among grain size, hardenability and brittleness. That's Ferrous Metallurgy 101 stuff. I'm a huge proponent of getting rid of the "alchemy and blacksmith hearsay" in our craft, but the first step in that is at least a little self-education in metallurgy (for those of us who don't have the option of getting formal education in the subject). If you have read Dr. Verhoeven's work, you apparently haven't absorbed it, or you'd understand that "large grain = harder metal" isn't an accurate summary, and that large grain comes with a serious down side. (By the way, triple quenching and triple normalizing, all in a row, is vast overkill unless you're doing something funky with your steel, and I don't recommend it.) On one level, that's OK. I printed my copy of Verhoeven's book more than five years ago, and there's a lot in it that I still don't understand. Even as much as Dr. Verhoeven tried to dumb his subject down for guys like us (and he dumbed it down a lot), it's still far from a simple read. I don't blame anyone who doesn't "get" all of it. But until you have a handle on some basic principles, you ought to be careful about writing off as "alchemy" things that you don't understand.

Second, you've established an absurdly high standard for belief, which strikes me as nothing more than a tactic to justify your position. If your standard is that you won't believe anything written here unless it's suitable for publication in a peer-reviewed academic journal, that's fine, but that means you'll believe almost nothing written on IFI about any subject. And I am very skeptical that you really adhere to that standard. That's a reasonable standard when we're trying to prove some totally novel scientific theory, but that's not what we're doing here. Which leads me to my third point.

Third, you asked for a "documentable reference" that confirms the benefits of repeated normalizing. You already have it; you named it yourself, and I pointed out to you a couple of specific paragraphs that support the practice. You respond by misreading the text in order to avoid having to come to grips with what I gave you. For example, you say that "Verhoeven quenched in rapidly stirred oil." That comes froms Verhoeven's description of one experiment, in which the goal was to form fine-grained martensite. As I explained to you in my previous post, Verhoeven, on the very page you're reading, discusses two separate things: how to form fine-grained pearlite/ferrite, and how to form fine-grained martensite. You're quoting from the paragraph on martensite in order to try to craft an argument that quenching in rapidly stirred oil is necessary in order to form fine-grained pearlite/ferrite, which is what we're trying to do when we repeatedly normalize prior to hardening. I tried to preempt that sort of confusion in my previous post, when I wrote,

Bear in mind that we're talking about forming fine pearlite grains here, so we have to read carefuly enough to realize that when he says, "cooling back to room temperature as fast as possible," he doesn't mean "so fast that we form martensite." Thus, at least with respect to steels with sufficient carbon to harden, he can't be referring to a full quench like we'd do if we were hardening. "As fast as possible (without forming martensite)" is a reasonable description of air cooling for most simple steels. (As you pointed out, multiple cycles of quenching to martensite would in fact work for grain refinement, too, but that's not what Prof. V is discussing in this pargraph. Here he's talking about forming fine-grained pearlite and ferrite, not fine-grained martensite.)

To be fair, I probably should have added, "for most simple steels in blade-sized cross-sections." In really large sections, cooling in still air might not be fast enough. Regardless, though, all you've got to hang your hat on is the idea that air cooling isn't fast enough to achieve the pearlite/ferrite grain size reductions that Verhoeven is talking about in blade-sized cross-sections. But Verhoeven doesn't say that. He leaves it somewhat ambiguous (because he's teaching principles, not writing recipes -- more on that in a second). You just want to read it that way.

You see, I don't think you're actually looking for a metallurgical reference that supports the concept of repeated normalizing. I think you're looking for an end-user manual that gives you a recipe that involves repeated normalizing, and until you see that, you're going to write it off as witchcraft. But real textbooks don't give recipes; they teach principles, which it's up to the reader to apply to a specific situation, taking into account all the relevant variables. That's why Verhoeven wrote a roughly 200 page book that's heavy on theory, rather than simple heat treating instruction sheets for the 20 or 30 most common blade steels. End-user publications like the Heat Treater's Guide simplify principles into recipes that can be applied by folks with minimal knowledge of metallurgy. And that's great; I'm a big fan of the Heat Treater's Guide. It contains tons of useful info. But as has been discussed here many times, it's oriented toward a particular type of end-user. And in order to come up with its recipes, it necessarily relies on some assumptions about how those end users are going to operate. I am certain that it does not assume that end users will be treating their steel like we bladesmiths do during the forging process.

That's all I have to say on the subject. You believe what you want. But I am going to continue to promote repeated normalizing (within reason) to reduce grain size and improve toughness because I know first-hand that it works, and because I know that there is a solid metallurgical foundation for the practice.

[Woody]

I am well aware of the difference between hardenability and hardness, but why the personal attack. You are the one that pointed out that large grain is harder see your quote below.

"So let me ask you: if increasing grain size moves the start curves on the IT diagrams to the right, indicating greater hardenability -- which is what Professor V just said -- what must we necessarily conclude would happen if we reduced grain size? Which way would the curves move, and what would be the effect on hardenability of the steel? Remember: we've already established that changing grain size moves the curves, and that moving the curves affects hardenability. So . . . "

5.4 Normalizing
This is a very common form of annealing. The method is to austenitize, then air cool to room temperature. You get...? Yes, pearlite. The biggest advantage is to get a uniform microstructure and to soften up the metal for subsequent operations like machining. I'ts done after cold-working to re-crystallize the microstructure. The stresses imparted to the metal crystals cause them to "break up" and re-form when austenitized.

Soak periods for normalizing are typically one hour per inch of cross-sectional area but not less than two hours at temperature. It is important to remember that the mass of the part or the workload can have a significant influence on the cooling rate and thus on the resulting microstructure. Thin pieces cool faster and are harder after normalizing than thicker ones. By contrast, after furnace cooling in an annealing process, the hardness of the thin and thicker sections are about the same.

http://www.stainless-steel-tube.org/The-Heat-Treatment-of-Steel.htm here again the process calls for a soak time but no mention of repetitions of the process.

Dr. Dmitri Kopeliovich

Normalizing is a process in which a steel is heated to about 100°F (55°C) above the upper critical temperature, followed by soaking and cooling in still air at room temperature.
Normalizing treatment is similar to the full annealing treatment. The difference is in the cooling method and rate – full annealing involves slow controlled cooling if the furnace or in some medium providing slow cooling rate.
As normalizing requires less time, it is more economically efficient heat treatment method than full annealing.
Normalizing relieves internal stresses caused by cold work while grain growth is limited by the relatively high cooling rate therefore the mechanical properties (strength, hardness) of a normalized steel are better than in an annealed steel.
Since the cooling rate in the normalizing heat treatment is not controlled, the resulting structure is dependent on the thickness of the steel part, therefore the effect of increased mechanical properties is greater in thin parts.
Quality of surface after machining of a normalized part is also better than in an annealed part. This effect is caused by increased ductility of annealed steel favoring formation of tearing on the machined surface.

http://www.ehow.com/how_7831831_normalize-temperature-carbon-steels.html

You seem to make a lot of assumptions about what my motives are, they are plain and simple. Find documentation that states repeated normalizing is ncecssary. Everything I have listed above calls for a soak time none call for a repetiton of the process.


.

[Woody]

I need to add a bit more to my last reply. We originally started with me questioning the necessity of normalizing repetedly because I can find nothing that documents this as being benificial. You have used a process that bears similarity to normalizing only in that the metal is heated to the specific temperature, in normalizing the metal is slowly cooled back to room temperature in still air after a soak time at temperature. in the paragraph you quoted "New Grains formed by Phase Transformation" from Dr Vs paper, the metal is rapidly cooled in air quite a difference I would think. Einstein said that it is foolish to do the same thing and expect different results, I would submit that it is equally as foolish to do different things and expect the same results. One can not prove one process by using as evidence results obtained by a different procdure.

The objective of normalizing is to relieve stress in the metal, as I read it, this is accomplished by changing the grain structure to pearlite which is a very uniform grain structure. This is accomplished by heating the metal to a specific temperature, soaking at that temperature for a given period of time. The soak time is given as one hour per inch of thickness. in a knife blade of 1/8 inch thickness that would give us a soak time 7 1/2 minutes or 15 minutes for a quarter inch thick blade.

The shrinking of grain size in the paragraph titile "New Grains formed by Phase Transformation" calls for "a simple technique is to heat into the austenite region holding the maximum temperature as low as possible and the time as short as possible and then rapidly cooling back to room temperature as fast as possible without forming bainite assuming it is not desired." The soaking at temperature in the Normalizing Process would promote grain growth. " Grain Growth As austenite is heated to higher temperatures or held for longer times at temperature the average grain size is found to increase. This process of grain growth occurs by smaller grains shrinking in size until they disappear and larger grains growing in size; the net effect being an increase in the size of the average grain.

The confussion seems to be what we call normalizing and the results we expect to obtain from it. So what most recommend as "normalizing" is not actually normalizing because there is no soaking of the metal at temperature and what is expected of the process is both stress relief and grain reduction. It would seem that the latter would be difficult to achieve if the correct normalization were conducted. So what we have is more of a "Phase Transformation" that will reduce the grain size and hopefully relieve stress.

I would think that a better proceess would be to do the normalizing correctly with proper soak time once to relieve stress, then if grain size reduction is your goal to do the Phase Transformation process to achieve that end. I am not sure of the need for this in knifemaking however, since Dr V states the following was done to obtain an ultra fine grain structure. " A series of similar experiments was performed here on 3 steels to examine the effectiveness of thermal cycling alone, no cold working was employed. The steels were heated by immersion in a salt pot. Initially the steels were austenitized for 15min. at 1650 oF and oil quenched in rapidly stirred oil. Then the steels were given 3 thermal cycles consisting of a 4 minute austenitization in 1450 oF salt and a quench inrapidly stirred oil. It would seem that holding the steel at temperature as outlined here would result in grain growth that might negate any reduction in grain size that was accomplished by "Phase Transformation"

[MattBower]

As I already said at least twice, if you are doing multiple quenches there is no reason to repeatedly normalize; repeated quenching to martensite can accomplish the same thing, and faster. But the down side to repeated quenching is the risk of losing the blade to quench cracking. There are less risky ways to accomplish the goal.

Beyond this, some of your complaint seems to boil down to semantics. You think that what I'm doing isn't "normalizing" because it doesn't look exactly like some of the normalizing recipes you've seen. That's what I get from your second-to-last paragraph, above. And that's true: some of the recomended normalizing cycles I've seen, and that you're referring to, call for a soak at temp -- and at a higher temperature than I am using, too. (Frankly, I don't understand why they recommend that because, as you yourself suggest, the relatively high temps and long soaks would seem to invite grain growth in simple steels. I can understand something like that in hypereuctectoid steels to get the carbides into solution, but I would want to follow it with grain refining cycles.) But with that said, "air cooling is often called normalizing," (Verhoeven, 32) and "the data on normalized (air cooled) steels illustrates two general characteristics (ibid., 38). There are other examples in Verhoeven in which he uses the words "normalizing" and "normalized" as essentially synonymous with "air cooling" and "air cooled" (from austenite, which is implied but not always stated). The Heat Treater's Guide says, "in a thermal sense, normalizing is an austenitizing heating cycle, followed by cooling in still or agitated air." (Ibid., 27.) It also says that, "Normalizing is applied, for example, to improve the machinability of a part, or to refine its grain structure . . . " (Ibid.) So, putting it all together, an austenitizing heating cycle followed by cooling in still air (air cooling), to refine grain, meets these simple definitions of normalizing. And that's exactly what we're doing. So I continue to think it's completely legitimate to call this type of thermal cycling normalizing.

Last, you said (to me) that, "You are the one that pointed out that large grain is harder see your quote below." I'm not sure where you're getting that. The quote in question is below. I see the word "hardenability" three times. I don't see the word "hardness" at all.

"So let me ask you: if increasing grain size moves the start curves on the IT diagrams to the right, indicating greater hardenability -- which is what Professor V just said -- what must we necessarily conclude would happen if we reduced grain size? Which way would the curves move, and what would be the effect on hardenability of the steel? Remember: we've already established that changing grain size moves the curves, and that moving the curves affects hardenability. So . . . "

I did say in an earlier post that it is possible to refine grain so much that it becomes difficult to properly harden a blade. Hardenability can affect final hardness, but I never suggested they're the same thing.

Let me add one final thought here. Industry has developed a lot of heat treating methods that simply aren't reproducible with the extremely simple equipment most of us are using. If you have a kiln or proper heat treating furnace with good temperature measurement and control, you can do a lot of neat stuff, and work with a lot of steels that are out of reach for most of us. (This is the main reason I still avoid working with steels with more than about 85 points carbon.) But since most of us don't, folks have developed simplified techniques to get close to some of the industrial methods while accommodating the limits of our equipment. The fact that they're simplified doesn't mean they don't work, when executed properly and applied to the right steels.