A Few Heat Treating Questions
Posted 25 November 2012 - 08:42 AM
I'm preparing to place the first order of steel for my new workshop, and while I have primarily used 1084 and 1095 in my knife making career to date, I figured now would be as good a time as any to re-evaluate. (Additionally, since I'll be learning two new forges and a new heat treating oven, I figured now would be as good a time as any to learn some new steels.)
I have encountered a couple of general heat treatment questions that I thought I'd throw out to the community. (Especially since Kevin Cashen is a member of said community.)
A) What is it about certain steels (1095 and 5160, for instance) that causes one to quench better in water or brine, while the other does better in oil?
--I have to assume that the answer has something to do with heat transfer rates (i.e. induction), vapor formation, and/or general viscosity, but beyond that basic observation, it's a bit of a mystery to me.
Similarly, why is it that some steels (like 1095) are best quenched in water/brine but can be quenched quite successfully in oil if the work piece does not exceed a certain dimension (such as 1/4" thickness, in the case of 1095).
--Again, I have to assume that it is on account of heat transfer rate; perhaps the oil will fail to draw enough heat out of the center of the piece, leaving it so hot that it will counteract (or at least critically delay) sufficient cooling of the outer surfaces?
C) My guess above raises another question: if a piece of 1095 thicker than 1/4" is quenched in oil, what are the resulting non-optimum characteristics of the steel?
--If I'm correct, then I would guess that a thin enough piece (still over 1/4") will sort of give itself a rough "temper" after the initial case hardening, while a thick enough piece will come out of the quench completely not hardened (or perhaps "unhardened" would be a better term).
D) Switching gears slightly, what are the merits of multiple quenches on a given piece of steel. (Note: I realize that each steel will respond differently to multiple quenches. Please assume, for the purpose of this question, that I'm asking about a steel wherein multiple quench is an optimal procedure. And then please tell me which steel that is.)
--In this case, my very general understanding is that each subsequent successful quench will convert more of the austenized steel to martensite, resulting in a blade that is both harder and more uniform. But what I don't understand at all is why that works. For each quench, the steel must be heated to its critical, austenizing temperature. If my general understanding is correct, then that has to mean one of two things (I think): a) that each unit of post-martensitic austenite is more likely to change back into martensite on a subsequent quench, OR that something in the process renders ALL of the steel more likely to experience a martensitic transition the more it is successfully quenched. (Note: by a "successful" quench, I mean that all necessary parameters for effective hardening have been met: steel temperature, quench medium, quench temperature, and the IT/TTT requirement.) But again, no clue why that happens.
Anyway, thus far, I'm considering ordering batches of the following steels:
I'm interested in 52100, as I've heard that you can get quite a lot of performance out of it, but I am admittedly a bit intimidated by the "fiddly" working and heat-treating protocol it demands.
Posted 25 November 2012 - 11:58 AM
A. The answer is as simple as one word- “alloying”, and complex enough to fill a whole book (I recommend “The Hardenability of Steels” by Siebert/Doane/Breen-ASM, and “Elements of Hardenability” by M.A. Grossman). Carbon content will determine how hard the steel can get, but added elements in the alloying will determine how deeply the steel can harden. For example- a steel with only .4% carbon will only reach 50 HRC (Hardness Rockwell “C” scale) whether the steel is ¼” thick or 1/16” thick since it has limited carbon to contribute to the hardening process. If you increase the carbon, the maximum hardness possible will increase but the cooling effect may not be enough to cool the ¼” piece, while the 1/16” piece is no problem. It is as simple as thicker steel is harder to cool than thin steel. So by increasing your carbon content the 1/16” steel can reach 60HRC+ but the ¼” steel will still only reach the low 50’s simply due to its mass. In other words, the depth of hardness is limited.
We overcome this with alloying. Adding chemical elements that have large atoms which can run interference in the iron slows the carbons ability to come out of solution on cooling and so it increases the time we have to cool it and still get full hardness. This levels the playing field so that the ¼” thick steel can reach 60HRC+ as easily as the 1/16” steel. In other words, the alloying increases the depth of hardness, not the maximum hardness level, which we get from the carbon. This I why I refer to some steels as “deep hardening” (O-1, L6 etc…) and some as “shallow hardening” (W2, 1095 etc...). The cheapest and most ubiquitous alloying element that results in deeper hardening is Manganese (Mn). It is in all steels for other purposes, but all of the 10XX series steels have it to some level. It is variances in the Mn levels which causes 1095 to require a bit faster quench than 1084, not the carbon levels. In fact it is best not to use the extra carbon in the 1095 for hardness levels, if you don’t have to.
B. Pretty much covered under “A”
C. What all of this depth of hardening is all about is avoiding the carbon coming out of solution at around 1000F when it is cooled. Steel exists in various phases depending on the temperature and rate of cooling. The hardened form of steel is called “martensite”, and is the highly strained result of carbon atoms trapped between the iron atoms where they normally wouldn’t be at room temperature. If we allowed the steel to cool slowly, from the high temperature where the carbon is naturally in solution, the carbon would normally come out of solution at around 1000F to form a segregated carbon/iron structured phase known as “pearlite”. Because there is no strain distortion from trapped carbon pearlitic is soft. So the secret to hardening steel comes down to cooling it fast enough through this range to avoid any pearlite formation. If the steel does not cool fast enough you will get a mixed structure of the two phases of pearlite and martensite. This is particularly problematic since the problem will not always be readily apparent. For example- a file will skate off from such an edge because it has enough martensite to allow it, but mixed in will be little localized pearlite colonies which will lower your overall hardness and effect things like edge retention.
D. Is a tougher question because the methods you mention are in a gray area and, to be honest, a bit controversial. I will try my best to stick with the areas where known metallurgical facts can be applied, and apply them to give detailed explanations; I would strongly urge you to ask for the same whenever getting input on the topic as the subject is rife with assumption and hyperbole. There is no steel where “multiple quench” is the optimal procedure, just as there is no steel where austempering, or marquenching is the optimal procedure, and I am still looking for an instance where industry recommends the procedure as it is practiced in bladesmithing. Each and every treatment used is based on specific properties you are looking to increase in the material, and this is why most steel will have a few methods listed and a range in temperatures used.
To avoid the sensational and controversial aspects of other terms allow me to switch to the concept of multiple heat cycles rather than quenching. There are two main areas where multiple heat cycles can affect the steel- grain size and carbide condition/distribution. In normalizing, the high heat with air cooling results in the dissolution of lunky carbide concentrations and will help create a uniform grain size. Faster rates of cooling will avoid the carbon pooling back up and, while it doesn’t affect the current grain size, it will create conditions that will increase the number of grains on the next heat. So by using a sequence of cycles we can refine both grain and carbide. This could be used in any part of knife creation after the forging if we are not using the traditional bladesmith practice of lamellar annealing (nonmagnetic blades insulated in wood ash, vermiculite etc. overnight), which undoes it all. In fact there are a lot of differences cited from multiple quenches that are “accentuated” by the methods and tools used by bladesmiths. Just one example would be- in the absence of a controlled method of heating the blade, precise holds at specific temperatures, necessary for optimum solutions, may not be possible. But one can play it safe and err on the low side, putting a fraction of carbon into solution and then trap it there with a quench, the next heat would then be able to take another bite out the needed carbon, and so on.
The complications are actually the opposite of what you may suspect. Really fine carbides are great, but really fine grain can have its downside. Finer grain does the opposite of what the alloying that I mentioned in “A” does, it increases the rate of pearlite formation, so the finer the grain, the faster you need to cool the steel to fully harden it, and we are once again back to looking for added alloying to help us out. This is just one reason why the results many claim to get from multiple quench cycles tend to be steel specific (e.g. chromium can be your friend).
On your steel selection, I would go for the simplest stuff to work with and develop a good proven basic heat treating schedule for each and then start you experimentation from that solid foundation. I would advise against pursuing claims of greater performance based merely on the claims themselves with no details as to the mechanisms you are trying to trigger. An entire career could be spent chasing your own tail if you don’t plan a course to follow. 52100, is a very good steel. Could it possibly ever live up to all the tales that have been spun about it in our business? No, I don’t even think unobtainium could do that. What it can do is throw all kinds of complications your way if you are not aware of its quirks and have the controls to deal with them, some of these complications may be overcome by getting creative with heat treatments using standard bladesmithing tools… well you probably know what I’m getting at here.
52100, has around .2% more carbon than is needed for fully hardening the steel, in fact if this carbon is put into solution it can drastically reduce overall hardness. That extra carbon can also bunch up in places that create some nasty problems. Just as 1095 is a bit tougher to deal with than 1084, 52100 is even more touchy than 1095. It was carefully crafted to give optimal properties when heat treated by the bearing making industry, so high compressive strength with abrasion resistance from plenty of residual carbides is its ideal. Industry uses advanced procedures to put the extra carbon into fine carbide packages and just enough free carbon to fully harden the steel. Needless to say, they don’t do this with hand held torches or bladesmith forges. This is not to say that the steel can’t make an excellent knife, but so can many other steels without the learning curve for a guy starting out.
I think all knifemakers begin their career hoping there is a special steel or special heat treatment that will allow them to make the best knives possible, but experience and time invariably reveals that just becoming familiar enough with a steel to master the basics that we can build our skills on is the true path to excellence. I use all kinds of high tech gadgets, but it is knowledge of the basics that allowed me to use them in the making of good knives, rather than just making it easier and faster for me to make lousy ones.
Posted 25 November 2012 - 12:57 PM
Zack as Kevin states, I would pick one or two steels and concentrate on getting the heat treat down for these steels. My personal preference would be 1084 and 5160, if I could only have two steels. I also use a lot of W2, but the heat treating can be a little more complicated. I do not use a lot of 52100, but give Ray Kirk an e-mail as he uses a lot of it.
You can satisfactorily harden 1095 and W2 with oil, BUT, it has to be an oil that is extremely fast. Think parks50. Brine and water can be a lot of fun, in more ways than one, just get ready for frustrations like cracking.
Anvil Top Custom Knives
Posted 25 November 2012 - 01:14 PM
I sent an email to Ray Kirk as soon as Zack posted the Topic this morning and asked him to reply with his techniques for heat treating 52100 steel. Ray is known for his exceptional work forging 52100 integral knives.
ABS Journeyman Smith
Send an email to Dan
Posted 25 November 2012 - 02:02 PM
Kevin, as usual your comments were clear and easy to understand. Thanks very much--that clears up all of the questions I asked, and a number of smaller "peripheral" issues I'd never really sorted out.
Dan, thanks for Emailing Ray. Hopefully we get a response!
Posted 25 November 2012 - 05:37 PM
Kevin pretty well nailed it on the steels. I started out on 52100 and learned most of what I know from Al Pendray and Charlie Ochs. Every batch of steel is a little different. Kinda like each other but not quite. When you want to work a steel, get as large a quantity as you can afford. Get the heat treat right for that batch and it will pretty well work with the rest of it.
That being said, if you decide to venture into working with 52100, either buy a large amount or get it from some one that has one batch that they sell. I have had one batch of 1 3/4" dia 52100 that I got in 1995. I still have a lot of it left. It is a little lower in carbon than the limits of 52100 specifications but was high enough in carbon to do what I needed done. I also use other round bars and try to buy as much as I can when I make a purchase. This eliminates the time getting the heat treat right.
If you use common practices used in forging knives, that is, as you get closer to your final shape, you reduce the amount of heat you put into the blade, You will end up with some fine grain. After forging, I thermocycle my blades twice and then the last time I bring it up to a dull red and put it in the vermiculite. 52100 will air harden and make it nearly impossible to drill if you don't. I use both an acetylene torch and a heat treat oven to heat treat the 52100 blades. Depending on the service and blade size. The oven is set for 1550 for 20 min. This is what I quench at and the torch, I use a magnet and go a bit higher in a low light area.
For making a knife from 52100. The rules must be followed. Thermo cycle when finished forging, cool slow enough to drill, drill with sharp bits and cutting fluid, use a slow speed and don't let the bit slip in the bottom of the hole. It also work hardens readily.
When you learn how to get the most from a steel, any steel, then I would say you could introduce another steel into your arsnal of materials. When you learn to heat treat one steel and get it right, there are very few people in the world that can tell if it is optimum for sharpness and edge holding qualities. A lot of that depends on edge geometry and blade thickness.
I have cussed the 52100 at times and usually it was my fault because of the problem. No short cuts and it will treat you right.
If you have any questions about how I do the forging and steps I take to insure it performs the best I can, You might email me a link to the question here as I don't get on too much. That way, every one can see the answer I have for it. Not saying it will be completely correct, but as correct as I can get it.
Enjoy the new forges, they always take a little time to get used to.
Posted 25 November 2012 - 05:52 PM
Posted 30 November 2012 - 10:14 AM
Since it does take much greater concentrations of Mn to increase depth of hardening there are other elements that are much more effective, although not as cheap as Mn. By, far the most common element added to steel to increase depth of hardening is chromium. In relatively small amounts Cr can drastically increase depth of hardening to accommodate oil or even air quenches in larger concentrations. 1060 is often impossible to quench to a totally pearlite free condition, but add around 0.7% Chrome to it and you have 5160 which can through harden in almost any oil.
Other elements can have effects to greater or lesser extents but are not as economical as Mn or Cr. Tungsten, vanadium, molybdenum and others can have profound effects in either direction depending on concentration and interactions with other elements. And others, such as nickel, we can hope contribute very little since their contributions would be negative (e.g. retained austenite).
But all of these elements still need the carbon to work with. It is the carbon that determines the maximum obtainable hardness, while these elements effect how fast you have to cool things to get that hardness. One test that was used early on to determine hardenability (or, more accurately, depth of hardenability) is the Jominy quench test. In this procedure a standard diameter round bar is quenched in a water spray and is then cross sectioned for hardness testing. On the outside surfaces, nearest to the cooling effect of the spray, the maximum hardness is recorded and then subsequent testing is done moving toward the center of the sample. “Shallow hardening” steels will have a very large deviation from the surface to the center, with the center often being dead soft, but “deeper hardening” steels will have much less of a discrepancy from the skin to the center.