My Heat-Treating Process

In my last post, I described some of the theory behind the heat-treatment process. To show how the theory looks in practice, I thought I'd describe how I go about heat-treating a typical blade. This will be a long post, so I am attaching a PDF that you can print out if you'd like to. :)

I work with 1084 steel, which is pretty similar to what we've been discussing but isn't widely available. There's one guy who sells it to bladesmith types. It has (you guessed it) .84 percent carbon.

My Tools
My heat-treating equipment is as follows:

• Small forge made out of two refractory bricks, with a JTH7 propane plumber's torch serving as a heat source
• High-temperature thermometer and temperature probe
• A 6" piece of 2" iron pipe
• 1 gallon of specially formulated quenching oil (Parks #50)
• A hot plate from Wal-Mart
• An meat thermometer (the kind with a steel probe)
• A thrift store toaster oven
• Tongs, gloves, miscellaneous tools

You can get by without some of this stuff, but each tool plays an important role. Heat-treating is all about controlling variables. Knowledge is your biggest asset, but better equipment will allow you to put that knowledge to good use.

The Process
After I've forged a blade blank to shape and ground it close to its final dimensions on my belt grinder, I'm ready to heat-treat it. Heat-treating is the most important step in making a blade. The grinding and polishing won't matter much if the steel is poorly heat-treated.

I get everything set up before I start messing with hot steel. First, I put the iron pipe inside my forge. (It acts as a diffuser for the intense heat of the fire, protecting the blade and allowing it to heat more evenly.) Then I light the torch and crank open the propane a bit. It'll take a few minutes to get up to temperature. I get out my thermometer and high-temp probe and put the tip of the probe into the pipe. Now I can control the temperature in the pipe by manipulating the propane flow, and the thermometer will tell me how I'm doing.

I then set up the hot plate on the floor near my forge, with my container of Parks #50 quenching oil on the burner. I put the meat thermometer in the oil and turn on the hot plate. This preheats the oil to 130 degrees F, which works well for the steel I'm using.

I set up my toaster oven on the work bench and set it for 400 F. I also like to put some newspapers down on the shop floor to catch oil drips.

Once everything is set up and the forge has reached about 1500 F, the fun begins.

Normalizing
I start the heat-treatment process with a step that I didn't mention in my previous post. All the heating and cooling during the forging process, plus the heat and strain of the grinding, leave the blade blank with some internal stresses. To get rid of those stresses and set up the internal structure of the steel to respond to the heat-treatment, I do a step called "normalizing."

Normalizing involves heating the steel up past its critical temperature, then allowing it to cool slowly in still air. It's not annealing, which involves slowly cooling the blade over a period of hours. Normalizing takes less than 20 minutes, start to finish.

I stand the blade on its spine in the forge, inside the pipe. The temperature probe sits near the edge, allowing me to keep close tabs on how hot the steel is getting. When the steel first goes into the forge, the readout on the thermometer drops quickly—it’s like putting an ice cube in water. When the temperature reaches its previous level and stabilizes, the steel is the same temperature as the inside of the pipe.

This is where I run into a problem. My forge doesn’t heat very evenly right now. Even with the pipe, there is still a hot spot in the middle. I have to compensate for this by keeping it a little hotter than I really want to and moving the blade into and out of the hot spot. Overheating slightly gives me a little leeway to play with. If the tip gets 50 degrees hotter than I want, it’ll cool a bit when I shift the blade to heat the heel. I don’t like having to do this, and my next forge will be much more even.

Because of that, don’t take the temperatures I’m listing here too seriously. They’re working pretty well for me right now in my setup, but they may be off base with better equipment.

To normalize, I heat the steel to 1450 F, hold it there for a couple of minutes, then take it out of the forge and allow it to cool until it’s black again. This takes about two or three minutes. The color is all gone by the time the steel reaches about 800 F or so, and then it’s back into the forge again. I normalize three times.

By heating above critical and then cooling, I’m allowing the steel to recrystalize with each cycle. This refines the internal structure of the steel and relieves any residual stresses from the hammering and grinding. Normalizing helps prevent warping and cracking, and it allows the steel to harden more fully when it’s quenched.

Hardening
After the steel has cooled to black for the third time, it’s time to harden it. Normalizing usually takes about 15 minutes, so my quenching oil has reached my target temperature of 130 F. The forge is holding a steady temperature of about 1450 F. (I usually have to keep tweaking it, actually. I turn the gas up and down a bit, and move the torch closer to and farther away from the fire hole on the side. But I can hold it within 20 degrees if I baby sit it.)

I pop the blade in the pipe one more time and allow it to come slowly up to temperature. Heating slowly is desirable because it allows the interior of the steel to heat up, not just the outside.

I use my thermocouple to judge the hardening temperature, but there are a couple of other clues that can be useful. Iron looses its magnetism at 1414 F, and on 1084 steel this is close to the correct hardening temperature. The magnet approach is widely recommended for beginners, but it won’t let you get the maximum performance out of the steel in many cases. The reason is that most steels require some “soaking time” at the optimal hardening temperature. The maximum amount of carbon that can be dissolved into iron is right around .8 percent. With steels that have more than or less than .8 percent carbon, you need to apply more heat and more time to get all of the carbon into solution. For example, 5160 requires some soaking because the amount of carbon, which is .6 percent, is on the low side for a good blade. You need to get all of that carbon into solution if you want the steel to reach full hardness. In a steel like O1, which has around 1 percent carbon plus some alloying elements like chromium, the soak is required to allow the carbon to free itself from clusters of carbon and chromium atoms.

The steel we’re considering, 1084, has just about the perfect amount of carbon in it. So it requires the lowest heats and shortest soak times of any steel. I recommend beginners start with 1080 (more readily available), as it will give you good results with primitive equipment.

Another clue you can use to find the proper hardening temperature is decalescense. If you watch carefully as the blade nears the correct temperature, you can see the steel transforming to austenite. Black shadows flicker and pulse rapidly over the surface of the steel. This takes some practice and a dark environment to see, but it’s amazing to watch. When the flickering stops the transformation to austenite is complete.

As the blade comes up to temperature, I mentally rehearse the quenching process to make sure that I’m not going to drop the blade or trip over something while I’m transferring it to the quenching tank. After it reaches 1450 F, I allow the blade to soak for about three or four minutes. This is longer than is probably necessary, but it doesn’t hurt anything.

The biggest pitfall you want to avoid during the hardening heat is overheating the steel. As the steel overshoots the hardening temperature, the austenite grains start to merge, creating bigger grains. Bigger grains make the steel more brittle—bad bad bad. Overheating by 100 degrees is enough to have an effect.

Once the soak is complete, it’s time for the quench. This is the fun part. I grip the tang firmly with a pair of tongs, and in one smooth motion, I pull the blade out of the forge and immerse it, point first, into the oil. With my preheated Parks #50, there’s very little fuss—a few puffs of smoke, some bubbling on the surface of the oil. It’s important to agitate the blade in the oil, which accelerates the cooling. I’ve tried both agitating up and down and back and forth, and neither causes any warping. Agitation is has been recommended by industrial heat-treaters for decades, and those guys know their steel.

What happens during the quench? The goal of the quench is to cool the steel fast enough to transform all of the austenite into a structure called martensite. If steel cools slowly, the dissolved carbon will collect together, leaving pockets of plain iron. But if it cools quickly, the carbon gets trapped in the iron, forming martensite. Martensite is the hardest, most brittle form of steel.

The quench, therefore, needs to cool the steel quickly and evenly, but not too quickly. (If you use a quenchant that’s too fast for a particular type of steel, it’ll stress the steel and cause it to crack.) You’ll hear bladesmiths talking about missing the “nose.” This is a reference to a type of graph called a isothermal transformation diagram, which shows the different structures steel forms as it cools. To simplify, you need to get the steel down below a specific temperature quickly, and then the cooling speed becomes less critical. On 1084, for example, the steel needs to go from 1450 F to around 800 F in just under 1 second. After that, you have as much as 30 to 60 seconds to get it down to room temperature.

That’s the reason that I preheat the oil; warm oil is thinner and moves more freely, pulling the heat away from the blade more quickly.

If you do all these steps correctly, the chances of the blade cracking or warping are minimal. My personal opinion is that cracked blades are usually a result of using poor quality steel, failing to normalize, or overheating during the hardening heat.

Once the blade has cooled enough that I can touch it with my bare hands, I take it out of the oil and set it on the newspaper and allow it to cool to room temperature. At this point the blade should be 100 percent martensite, and it is loaded with internal stresses caused by the rapid cooling. One good visual clue is to look at the surface of the steel. As the blade heats up in the forge, black scale forms on the steel. If you get the hardening temperatures and quench just right, the scale gets blasted off the blade, leaving an otherworldly, ghostly gray-white finish on the blade.

The blade is probably between 65 and 67 on the Rockwell C hardness scale—crazy hard. If you dropped it on concrete, it would shatter like glass.

Tempering
This is where tempering comes in. By heating the blade again, you allow the martensite to soften slightly and you relieve some of the internal stresses. You’re sacrificing some hardness to gain resilience. This is a good tradeoff for a razor, because 65 HRC is virtually impossible to hone anyway.

After the blade has cooled and I’ve wiped off any excess oil, I take it over to my grinder and do a couple of passes on the contact wheel, cleaning the forge scale off the surface and leaving the hollows bright and clean.

It’s a good idea to test the hardness of the steel before you go on to tempering. The easiest—and least accurate—way to do this is to try to file the blade with a good file. Files are hardened to around 65 Rockwell C, and by observing how the file interacts with the steel, you can gauge whether it hardened or not. A file won’t dig into a hardened blade—it will slide off the surface as if the blade were glass. The sound it makes is a good indication, too. A hardened blade makes a high-pitched, brittle sound. (The file may take off some scale, which is soft, but once it gets down to clean metal is should barely even scratch it.)

My tempering oven right now is a small toaster oven. These things vary in accuracy, so I got an oven thermometer to confirm the temperature. You could also use a kitchen oven.

To temper the blade, you heat it to a prescribed temperature and hold it there for two or three hours. The tempering range of steel starts around 350 degrees F and goes up to 600 or 700 degrees. Around 350 F, the martensite starts to transform into tempered martensite, which is slightly softer and much tougher than untempered martensite. The higher the tempering temperature, the softer the steel will get.

For razors, I usually temper at 400 to 425 degrees F for two hours. I’ve experimented with tempering blades at 350 F, and the steel is usually too hard to hone easily. On one O1 blade that I made and tempered at 350 F I spent more than five hours trying to get it honed. I eventually gave up and gave it another tempering cycle at a higher heat, and after that it honed right up.

Antique razors seem to be between 58 and 62 Rockwell C, and several of the modern custom makers shoot for a hardness of 62 to 64. I’ve actually made a couple of blades that were pretty soft, maybe 52 to 54, and they still shaved well. The actual number doesn’t matter much, in my opinion, unless you’re looking to produce and sell blades and want to be able to make specific hardness claims. There is a wide range of “acceptable” hardness levels—the blade will still shave nicely, although edge holding and ease of honing will vary.

As the blade tempers in the oven, the bright, shiny parts that I cleaned up will start to discolor again. This is oxidation forming on the surface of the steel. These tempering colors are a good indication of what’s going on inside the steel. The first color to appear is a faint bronze tint. The bronze darkens to brownish, then it starts to go iridescent reddish purple, followed by blue, then finally gray. For blades, you want to temper in the medium bronze range. The exact colors vary by the type of steel, so you’ll need to experiment. I usually go by the temperature, and the tempering colors just provide a nice secondary reference.

After the temper, the blade is ready for finish grinding. It’s important to keep the steel from getting too hot—anything over 300 or 350 degrees will start to soften the blade again. By letting the blade get too hot, you’re doing further tempering. That’s why the blade will go blue if you let it get too hot. Those are temper colors. The best way to prevent overheating the steel is to grind without gloves on and dip the blade in a bucket of water every couple of passes. You will cook your skin long before the steel gets hot enough to affect the hardness.

So that’s it. You can do a good heat-treatment with simple tools and some basic knowledge. Improving your knowledge will allow you to get better results with the tools you have. Better tools give you more control over the process, but you have to have the knowledge to put them to use.

I welcome discussion and corrections as necessary. :)

Josh

That's a great explaination of a process that would otherwise seem very foreign to me... Thanks for the insight and I'm sure some guys thinking about diving into this just might be pushed over the edge with this. Very cool,,thanks for sharing.:bow

Mike, I'm looking over that edge now, and now it doesnt seem so deep. thanks Josh, gotta feeling I am going to take the plunge soon.:eek: Of course I am going to see what the experts say before I commit.

Now to video it .... ;-)

Thanks Josh!

Quote:

---

Originally Posted by poona

Now to video it .... ;-)

Thanks Josh!

---

Always leave 'em wanting more. :D

Josh

Great post. I appreciate you taking the time to spell this out for us.

Thank you for posting this Josh. A very interesting plethora of information.

Josh,

Thanks for for the walk-thru on the hardening and tempering processes.:bow Now all I have to do is get grandpa's cast iron forge with hand cranked fan going and it's martensite city.:roflmao:roflmao

Long ago when I fancied myself as boy blacksmith I did use that forge when it was still on the farm to make a new burner grate for grandma's stove. No heat treating, grinding, or honing necessary, thank heavens.

Very good post! Thanks for putting this together.

The only observation I have is the "soaking" time. Most of the recommended soak times are for a piece of steel that is one inch thick. Our blades are much thinner so I do not think that we need to soak them that long at critical temp. I am trying to catch mine maybe 30 seconds after the magnetism is gone. I am using 1095 and old files.

I agree on using 400 as a temp for tempering. 350 is not enough.
I temper for one hr, let cool to room temp, then repeat. I have finished one blade using that routine and it is hard but it can be honed.

Randy,

You're right about the recommended soak times being for steel 1" thick. But from what I understand it's still good to soak any piece of steel for at least a couple of minutes, just to be sure you have full solution. The spine on my blades is usually 1/4" thick, and I like to make sure it's evenly heated all the way through.

1086 and similar steels come into full solution pretty quickly. Steels like O1 need five to 10 minutes in blade cross sections. I think the soak time for O1 is 30 minutes per 1" of thickness, so you can extrapolate downward from that. :)

The soaking thing is assuming that you're able to control the temperature to the point where you aren't going to overheat the blade. If you have some control, soaking won't hurt anything except maybe cause a little extra decarb, which you'll grind off anyway.

If you can't control the temperatures fairly well, soaking is risky. Overheating is just about the worst thing you can do to a blade...

Josh

Also, on the tempering cycle... A lot of guys recommend doing two or three one-hour cycles, but from what I understand there's no difference between doing two one hour tempers and one two hour temper. The reason guys like Kevin Cashen let the blade air cool every hour is to do hardness testing. Kevin likes to zero in on an exact RC by tempering in stages.

Maybe Mike will jump in and clarify a few things... :)

Josh

Josh,

Great post, I just read, reread then read it again. I really appreciate you taking the time to document and post this.

Charlie

Not much to clarify Josh. This is a good summary of the whole process, and a very good piece to defuse the need to have all sorts of high test equipment. If a simple system does the job, so much the better. I notice that you substitute the human computer for the digital controller in this arrangement. That saved you about 400 bucks right there, but then you don't have to feed or entertain a digital controller....LOL.

I'm lucky. I have the good equipment, but I still baby sit my salt pots. Even for the thin section blades I make, I soak them for at least 3-4 minutes. I can see the limitations of doing that in a more aggressive atmosphere like an open forge because you have to make the blades thicker to account for the scaling. But you can design a simple, controllable forge like Josh's that will "hold" at a temperature and that's the secret to beating the grain growth problem.

It's all good stuff. Being aware of the variables is the key. Controlling them with the tools you have and being able to explain how you approach them is really helpful, especially when educating customers.

Quote:

---

Originally Posted by JoshEarl

Also, on the tempering cycle... A lot of guys recommend doing two or three one-hour cycles, but from what I understand there's no difference between doing two one hour tempers and one two hour temper.

Josh

---

There is a small difference between using one temper cycle versus two or three: any Retained Au,stenite is converted to Martensite during the cooling after the blade is heated to tempering temperatures, so the blade has some fresh Martensite along with the (more desirable) tempered Martensite after the first cycle. A second tempering cycle would temper the new Martensite and not alter the old stuff if you use the same temperature, but would make for a slightly better/more uniform microstructure. But the extent of this is based on the type of steel and effectiveness of the heat treating process.

There are even a few guys that recommend 2 or 3 two hour tempering cycles to be absolutely positive that the R.A. is converted and that the Martensite is tempered properly, but they mostly work with swords etc. So, for a razor, 2 or 3 one hour cycles wouldn't hurt and would be a slightly better procedure.

Otherwise, Great job!

Thank you Josh, Mike and Russel. This area of knowledge is vast and it takes a whole lot of people sharing their technical knowledge, methods and experience to make it all work. The one thing I am sure of is that the handmade razors we are making have a better heat treatment than the mass produced razors.

Hey I thought I would include a good tutorial for engineering students on the process. It has all of the the isothermal transformation graphs and explanations on how they work. In fact I learned back at Virginia Tech of a tutorial almost exactly like this. Ahhh how I miss it!

http://www.engr.ku.edu/~rhale/ae510/heattreat.pdf

Russel,

I've heard this same thing, actually. I asked Mike about it once, and he said it didn't make a difference. Maybe he can chime in and clarify...

I can see both sides. On one hand, it makes sense that the steel would need to cool back down below Mf (the point where austenite ceases to convert to martensite), which if memory serves is around 200-250 degrees F. Cooling to room temperature would allow that to happen. The next cycle would temper the new martensite.

On the other hand, there is such a thing as "auto-tempered" martensite, which is an atypical heat treatment done by quenching the blade to 400 degrees F and holding it there for the amount of time you'd usually use in a tempering cycle. This prevents some of the stresses that normally occur during quenching, and I believe it results in a slightly tougher structure. It does require some more advanced equipment, otherwise it might be more widely practiced.

This might be one of those things that varies a bit based on the particular steel alloy under consideration. Mike's advice to me earlier might have been based on the simple steel I was using.

Mike? :)

Josh

Quote:

---

Originally Posted by Russel Baldridge

There is a small difference between using one temper cycle versus two or three: any Retained Au,stenite is converted to Martensite during the cooling after the blade is heated to tempering temperatures, so the blade has some fresh Martensite along with the (more desirable) tempered Martensite after the first cycle. A second tempering cycle would temper the new Martensite and not alter the old stuff if you use the same temperature, but would make for a slightly better/more uniform microstructure. But the extent of this is based on the type of steel and effectiveness of the heat treating process.

There are even a few guys that recommend 2 or 3 two hour tempering cycles to be absolutely positive that the R.A. is converted and that the Martensite is tempered properly, but they mostly work with swords etc. So, for a razor, 2 or 3 one hour cycles wouldn't hurt and would be a slightly better procedure.

Otherwise, Great job!

---

Yeah, I meant to clarify that it mainly applies (as far as I know) to a steel that has a tendency to have retained Austenite after heat treatment. I just wanted to make the point that it could be easily integrated into any H.T./tempering procedure as a type of safe guard just in case you're left with a little soft stuff after the quench.

(BTW, since we last traded info on smithing procedures I've picked up a digital pyrometer/controller from Auber Instruments. Definitely a helpful tool, much thanks for the suggestion of getting one.)

I meant to get to this earlier but forgot.

At about eutectoid (which is a carbon content around 0.77-0.85% depending on which text you use for reference) all the available carbon will go into solution at the austenitizing temperature.

More carbon will essentially result in a supersaturated solution of austenite that when quenched, will have retained austenite because there is a finite amount of space between crystals to form martensite. The more alloying content, the more likely there will be retained austenite. One of my favorite exotic steels has, despite the best heat treatment possible, by the book etc., about 25% retained austenite and it's still hard and tough as the hubs of Hell. Retained austenite is more brittle/unstable that martensite so the fracture risk is higher/less tough.

For the most part, simple steels heat treated correctly will have little to no dangerous retention. More complex steels, it is something that will have to be accounted for. For the high alloy stuff, one method to reduce it is the liquid nitrogen cycling. But, even then you only reduce the amount of RA and may not necessarily eliminate it. Repeat tempering is definitely necessary afterward.

The phenomena of long term conversion of RA to martensite, as in, over years, has led to the belief that swords get more brittle over time. Likely because some austenite converted to untempered martensite and left a hard brittle spot to be found generations later.

As to the tempering cycle question. I have mostly always used a two hour temper. I think I experimented with three one hour cycles just so I'd know for myself what happened and frankly couldn't see any difference. But then I'm using simple steels and don't have much of a retained austenite problem to have to convert. Plus generally, the Ms, or martensite start temperature, is higher than the tempering temperature, so most austenite is well on its way to martensite before dropping down further to Mf or the temper bath.

Also, given that the conversion of austenite to martensite occurs at or near the speed of sound, the likely amount of retained austenite to convert is going to be really very small and far less time than the available hour or two of tempering.