How to Anneal Stainless Steel After Forging

Another rather large heat treating study! This one took quite a bit of time, effort, and money. If you want to support further research visit Patreon.com/KnifeSteelNerds and become a Patreon supporter. All of the money I receive that way goes to knife steel research. And you get some perks like seeing articles and videos early, and at a high enough tier you get a free Knife Steel Nerds mug!

Video 

Here is the video version of the following information:

Should Stainless Steel Be Forged?

Stainless steels are generally more difficult to move under the hammer than simple carbon and low alloy steels. They are typically more expensive. The forging range is usually narrower; you have to stop forging at a higher temperature or it may fracture. But of course stainless steel has the major advantage that it is corrosion resistant. There are many myths around forging stainless and carbon steels that are used to justify the use of only simple steels. These myths scare some knifemakers away from using stainless steels that might otherwise try them.

Is Stainless Steel Improved by Forging?

There is an old tradition in the knife world that says that only carbon steels should be forged. Some have gone so far to say that stainless steels do not “benefit” from forging in the same way that carbon steels do. This is somewhat difficult to refute, as the benefits of forging are often overblown to begin with. I have an older article on forging vs stock removal you can read here. However, I argue that high alloy tool steels and stainless steels have more potential benefits to forging than low alloy steels. The reason is because with simple steels all of the carbides are dissolved at forging temperatures and re-precipitated later; this makes the carbide structure easier to control with thermal cycling alone. High alloy steels have carbides that do not dissolve without melting the steel itself so there is more possibility of improving that structure through further working. Of course all steel purchased by knifemakers has already been forged and/or rolled from an ingot, so we are often talking about a relatively small amount of further forging.

D3 tool steel forged to different degrees thickness starting from a 10″ round ingot [1]

And the carbide structure has directionality to it, leading to different properties in the longitudinal and transverse directions (along the rolling direction and perpendicular to it). In my forged vs stock removal article I gave reasons for why forging blades to shape rarely leads to superior toughness, but if such a benefit was to be gained it would be more pronounced in high alloy and stainless steels.

M7 high speed steel with different degrees of reduction [2]. You can see that the carbide bands are elongated along the rolling direction.

What is a High Alloy Tool Steel?

Most of the information in this article will relate to not only stainless steels but also high alloy tool steels like A2, D2, CPM 3V, etc. Stainless steels used in knives are simply a subcategory of high alloy tool steels. Low alloy tool steels and simple carbon steels behave somewhat differently like 1095, O1, 52100, 80CrV2, and others. The line between low and high alloy can be somewhat fuzzy, some give it as 5% total alloy content. So if the chromium, tungsten, molybdenum, etc. add up to more than 5% it is high alloy. For our purposes a high alloy tool steel is any with at least 3% chromium. These steels are air hardening and their carbides dissolve at higher temperatures. This does not mean all of these steels will be annealed with exactly the same temperatures, hold times, cooling rates, etc. but similar ideas will apply to them.

How Hot to Forge Stainless and High Alloy Tool Steel

A common mistake with forging of stainless and high alloy tool steels is heating the steel too hot. It is a common misconception that because the steel is more difficult to move under the hammer that more temperature is necessary. When steel is overheated, the grain boundaries melt first, leading to steel breaking apart when it is forged. Some knifemakers mistakenly assume they must not have been hot enough, and try even hotter! The temperature at which grain boundaries melt is roughly the same with stainless steel as it is with simple carbon steels, and sometimes lower. Datasheets typically recommend 2100°F (1150°C), though some will recommend lower like 1900-2000°F (1035-1100°C). 2100°F/1150°C seems to work for most knifemakers I speak to. However, many knifemakers are used to lower carbon steels like 1084 or 80CrV2 which can handle higher temperatures and they are not used to dialing the forge down for higher carbon steels. This is why “cast iron” has very high carbon content (>2%); higher carbon means lower melting temperature so it is easier to melt the cast iron before casting. High carbon steels that are more commonly used in forging like 26C3, White #1, Blue Super, and ApexUltra are also more sensitive to overheating since they have relatively high carbon contents (>1.2%).

How to Normalize Stainless Steel

Stainless and high alloy steels are not normalized. The goal of normalization is not grain refinement but rather dissolving all of the carbides before air cooling. As I noted above, with most stainless steels the carbides do not dissolve until melting. Perhaps we could come up with some kind of creative treatment that could improve/change the microstructure prior to annealing but this is not common in the steel industry. Normalizing is not necessary and we will be skipping it with stainless and high alloy tool steels.

What Are We Trying to Accomplish During Annealing?

Annealing is the step we perform between forging and the final austenitize and quench. The steel is annealed by the manufacturer before you get it. For stock removal makers, annealing is not necessary except in rare circumstances. The purposes of annealing are multi-fold:

1. Soften the steel so it is ready for machining, drilling, bandsaw cutting, etc.
2. Prepare the steel for good response to austenitizing so that we don’t need excessive hold time or temperature.
3. Maximize the final properties after heat treatment including hardness, toughness, etc.

Problems with Annealing Stainless Steel – Long Times, Scale, and Decarburization

One issue with annealing stainless steel is that the recommended annealing procedures often require cooling from 1600F+ down to 1000F at 25 degrees Fahrenheit per hour (~15°C/hr). That takes over 24 hours! Not only does this take a long time but most knifemakers do not have furnaces equipped with inert gas or vacuum so this means that scale and decarburization are major concerns. Without any protection you could end up removing a significant amount of carbon from the steel. To mitigate this it is best to wrap the steel in foil during annealing. I found double wrapping to help some as well. This is not a surefire way to prevent any scale or decarb as the times are very long. Therefore it is best to leave some material to remove after heat treating to ensure all of the scale and decarb is removed. I have not experimented with coatings for annealing but they may also work.

Why AEB-L is a Good Stainless to Start With

The first steel I experimented with was AEB-L. It checks a lot of boxes for forging bladesmiths as it is a very fine carbide steel with properties that can be similar to low alloy and simple carbon steels. Some have called it “stainless 52100” for its fine carbide size and excellent toughness. It is also relatively low cost, as some bladesmiths have sticker shock buying expensive stainless steels when they are used to buying 1084 or 80CrV2 for less than $5 per pound. AEB-L’s relatively low carbide content also means it is somewhat easier to forge than other stainless steels; I have an article on which steels are most difficult to forge here. Having only chromium carbides means it is easier to grind and finish than some of the more exotic stainless and high alloy tool steels with very hard vanadium carbides. One potential downside to AEB-L is it is not typically available thicker than about 1/4″ (6 mm), so bladesmiths that like to forge from heavy stock, round bar, etc. will be out of luck.

My New Experiment with AEB-L

We started with 1/4″ AEB-L which my father, Devin Thomas, hot rolled down to 0.130″ (3.3 mm), which is a bit less than a 50% reduction. The temperature used for rolling was around 2100°F. I then annealed it in different ways and measured the annealed hardness, hardness after quench and temper, and toughness after quench and temper. The different annealing procedures I tried will be given in the sections below. The final austenitize I used was 1925°F (1050°C) for 15 minutes, plate quench, cryo in liquid nitrogen, then double temper at 350°F (175°C).

Traditional Slow Cool Annealing

I have an earlier article that discusses the mechanisms within steel that occur during annealing. The typical method is for heating the steel to a temperature where the steel is austenitic (the high temperature, nonmagnetic phase of steel), but typically lower than austenitizing before quenching. For a simple carbon or low alloy steel this is somewhere in the range of 1350-1450°F (730-790°C) while austenitizing before quenching is typically 1475-1550°F (800-845°C). This lower temperature means more carbide is present, then during slow cooling the soft ferrite forms while feeding carbon to those carbides and growing them, resulting in a “spheroidized” structure (round carbides).

Carbides increasing in size during slow cooling as the ferrite grows into the austenite

There are a couple differences with high alloy and stainless steels. For one, the temperatures for annealing are typically higher, with 1600-1650°F (870-900°C) being most typical. The steels do not transform to austenite until higher temperatures, necessitating the higher temperatures. Typically the hold time is longer at this temperature as well, such as two hours. Following that hold time, the cooling rates are also slower. These are “air hardening” steels so they are more prone to hardening if the cooling rates are not sufficiently slow. This is partially why low cooling rates like 25°F/hr are recommend in datasheets, though that is also true in many datasheets for low alloy steels. So for AEB-L I wanted to try 50°F/hr and 100°F/hr and see how the resulting properties compared with the as-received steel from the manufacturer. The temperature you must cool to for ensuring full transformation depends on the cooling rate and the steel. However, most annealing procedures will recommend a temperature below which no more transformation is likely to occur, such as 1000°F. I used 1100°F for the 50°F/hr anneal and 1000°F for the 100°F/hr anneal. If the steel is fully transformed it doesn’t particularly matter what cooling rate is used below that temperature.

The hardness after annealing was at first a bit surprising because even with the fast 100°F/hr the hardness was significantly lower than what is delivered by Uddeholm (As-received). Perhaps Uddeholm uses an even more rapid form of annealing.

AEB-L As-Received by Uddeholm

50°F/hr anneal

100°F/hr anneal

This hardness difference appears to be confirmed by the metallogaphy, as the carbides in the two anneals I performed have somewhat larger carbides. I next compared the resulting hardness and toughness with prior toughness testing done with the as-received steel.

The same 1925-350 heat treatment with the as-received material (labeled “Stock Removal”) is the 60 Rc point. So the 50°F/hr anneal resulted in slightly lower hardness with equivalent toughness, and the 100°F/hr anneal resulted in slightly better hardness and toughness. This is mildly surprising based on the annealed hardness and somewhat larger carbides in the annealed condition. But overall this is a good result, with the 100°F/hr anneal being relatively fast, having very low annealed hardness for ease in working, and then excellent hardness-toughness after final heat treatment.

Isothermal Hold Annealing

Another similar way to anneal is to do a hold at a lower temperature rather than a slow cool. After heating to 1600°F like with the prior annealing, I held at either 1300°F (700°C) or 1200°F (650°C). Holding at a higher temperature is roughly equivalent to a slow cooling rate, while holding at a lower temperature is equivalent to a faster cooling rate. I held at 1300°F for 4 hours and 1200°F for 6 hours, as I was concerned it would take longer at the lower temperature (based on published transformation curves for stainless steels).

I also decided to try a technique that is common with low alloy steels which is to do “grain refining” cycles prior to the 1300°F anneal. I heated to 1600°F for 30 minutes and air cooled, which I did twice before the same anneal. I did not find grain refining cycles to improve properties in 1084 in a prior experiment, but I thought it wouldn’t hurt to try it again. I labeled this condition “cycled.”

These were somewhat closer in hardness to the as-received condition. However, oddly the hardness of the 1200°F condition  was lower than the 1300°F. Yes I held it longer at that temperature but the transformation was completed in both cases (as will be shown in the metallography), so that should not have been a factor.

Isothermal Anneal 1300°F

1200°F Isothermal Anneal

The 1200°F anneal looks like it resulted in somewhat finer carbides thought they don’t look that different. And both look relatively similar to our slow cool anneals. There are probably subtle differences if we did a full statistical analysis, but the real differences are probably more apparent through the annealed hardness and the final heat treated hardness and toughness.

The properties look pretty similar between the three and I’m not sure there are any “real” differences between the three. They are all around 60.5 Rc and ~35 ft-lbs. Maybe the grain refining cycles or the lower 1200°F led to a slight improvement but those small improvements may disappear if we did the experiment multiple times and averaged the results.

Temper Annealing

Another type of annealing is quite different than the others which is temper annealing, which I have written about before. In fact this is the general type of annealing that I recommended for high alloy steels in Knife Engineering. With normal tempering of martensitic steel, the higher you temper the softer the steel gets. If you temper hot enough you get annealed, soft steel. However, normally we do an austenitizing and quench step from a lower than normal temperature prior to a very hot temper. I recommended this annealing treatment because of older studies that showed a superior grain size and toughness with high speed steels after a temper anneal rather than a more conventional anneal. But the tempering times for optimal properties were very long, 12-24 hours depending on steel and tempering temperature. Also I began to be worried that perhaps the anneal was better only for high speed steels, since they use very high austenitizing temperatures where grain growth is more common. Perhaps the temper anneal resulted in more stable carbides that don’t dissolve as readily at the high temps, therefore they still pinned the grain boundaries. For most steels where grain size is not as difficult to control, this benefit would be lost.

So based on those prior reported experiments I austenitized all of them at 1600°F for two hours prior to plate quenching. This also has the benefit of maintaining the same austenitize as the prior annealing experiments. I then tried three different tempering conditions: 1300°F 4 hours (Q13), 1400°F 4 hours (Q14), and 1400°F 24 hours (Q24). Long anneals were optimal for the high speed steels but I wanted to have shorter anneals which is why I also tried the 4 hour tempers.

The 1300°F anneal (Q13) is definitely harder than we would desire. The 1400°F for 4 hours (Q14) is better, and the 24 hour anneal (Q24) is definitely soft enough. Maybe we could do something in between 4 and 24 hours with 1400°F instead.

Q13 – 1300°F 4 hour temper anneal

Q14 – 1400°F 4 hour temper anneal

Q24 – 1400°F 24 hour temper anneal

Only the Q24 looks properly annealed with medium size spheroidized carbides throughout. The Q13 condition specifically has relatively large regions which still look like martensite, or only very small spheroidized carbides.

Comparing the toughness, Q24 and Q14 are very similar, while the Q13 was significantly worse. The average hardness I measured with the Q24 was around 60 but I got several readings that were lower; overall it seemed to test less consistently. Maybe that means the carbides were too coarse, again pointing to an optimal hold time being somewhere between the 4 and 24 hours I tested with 1400°F. The Q13 ended up being slightly lower in hardness which is just randomness. I did another set of small hardness coupons and it was more similar to the Q14. My hypothesis for why the Q13 had worse toughness is related to the inconsistent microstructure. Maybe those martensite regions led to large grains because there weren’t carbides available to pin the grain boundaries.

AEB-L Annealing Summarized

When looking at all of the conditions as a whole, the overall finding was that if we have properly annealed the steel the properties are relatively similar. We got good (low) annealed hardness with most of them, and the resulting hardness and toughness were also pretty similar. My favorite of the conditions was the 100°F/hr anneal. The overall time to anneal is around 8 hours, meaning you can anneal it overnight after forging the steel. And it had the best combination of hardness and toughness after the final heat treatment of 1925°F austenitize and 350°F temper.

Annealing MagnaCut

I also wanted to look at a steel that may require a slower cooling rate than AEB-L so I also tested MagnaCut. Of course I developed MagnaCut so I was more interested in it. But the reason why MagnaCut was a good steel to look at is because of its 2% Mo. Molybdenum increases “hardenability” of steel so that larger sizes can be air cooled and still fully harden. This also affects annealing. Below I have Time-Temperature-Transformation (TTT) curves for two steels after they were austenitized at a low temperature for annealing:

D2 Annealing TTT after austenitizing at 1600°F

M2 annealing TTT after austenitizing at 1625°F

It can take a minute to understand the charts if you have never looked at them before. The curve to the left shows how long it took for the steel to start transforming, and the curve to the right shows how long it took for the transformation to end. For D2 with 0.74% Mo, at the fastest transformation temperature (~1400°F) it took less than 30 minutes to fully transform. For M2 with 5% Mo (and 6.5% W), it took over 3 hours to fully transform at the same temperature. This is largely cause by the Mo difference. So I was concerned that a 100°F/hr anneal may not work with a steel like MagnaCut with 2% Mo. This should then translate to the many stainless steels with less than 2% Mo (440C, Elmax, M390, S90V, etc.), and to the several stainless steels that also have around 2% Mo (S30V, S35VN, S45VN, S110V). The most common steel with even more Mo is 154CM/CPM-154 with 4% Mo though its datasheet says it can fully transform at 1300°F after 4 hours so there is a decent chance it can also use an intermediate cooling rate.

We did the same experiment of hot rolling MagnaCut from 1/4″ down to 0.130″, I performed a few different annealing cycles from 50°F/hr or 100°F/hr and measured the annealed hardness. However, I also wanted to experiment with the annealing temperature prior to cooling so we looked at that for this steel. We will get to that after we take a small detour to discuss how those annealing temperatures are chosen.

How to Select an Annealing Temperature

Most steels have an available datasheet that lists a recommended annealing procedure. For example, the MagnaCut datasheet recommends the following for annealing: Heat to 1650°F (900°C), hold 2 hours, slow cool no faster than 25°F (15°C) per hour to 1100°F (595°C), then furnace cool or cool in still air to room temperature. So if you follow my recommendation and use the faster cooling rate of 100°F/hr you can simply use that recommended temperature (1650°F/900°C), and then cool at the faster rate.

But what if the steel does not have a datasheet, or the datasheet doesn’t give an annealing temperature? At that point we need to find the temperature at which the steel has transformed to austenite. With a simple carbon or low alloy steel that can be done with a magnet because austenite is non-magnetic. However, there is another point where steel becomes nonmagnetic called the “Curie point,” and high alloy steels reach that point before they transform to austenite. So another way we can try to determine it is to heat the steel to different temperatures and quench and see the point where hardness increases. I did this with both AEB-L and MagnaCut:

You can see that AEB-L shows a rapid increase in hardness between 1400 and 1475°F which is the region over which it transforms from ferrite to austenite. After it transforms to austenite, carbon goes in solution, then you can quench and the hardness is much higher after quenching. However, if you quench after heating to 1400°F, you didn’t transform to austenite and the hardness stayed low. It looks like we could anneal AEB-L from as low as around 1500°F, and the 1600°F I chose for my study should be pretty safe.

With MagnaCut, however, the transition is more difficult to see because carbon doesn’t go into solution until a higher temperature. It looks like the steel may have started transforming to austenite around 1550°F but this actually led to the steel slightly decreasing in hardness. The steel started increasing again around 1650°F and I think this means it is likely where austenite finished transforming and is probably a reasonable temperature to anneal from. But I wanted to see how the annealing temperature would affect the annealed hardness and the final heat treated properties.

There were several surprises in this experiment to me. One is that the temperature mattered much more than the cooling rate, at least for those two cooling rates. The other surprise is that the change in annealed hardness was quite linear. There was no big jump in hardness from annealing at too low of a temperature, even at 1550°F. The “as-received” hardness for annealed MagnaCut is around 22 Rc which is similar to the hardness measured for the 1600°F anneal.

The next experiment I wanted to do was ensure that we were getting good properties from annealing at 1650°F since the annealed hardness continued lower up to 1750°F annealing temperature. And the as-quenched hardness we measured from MagnaCut before didn’t really go up significantly until 1700°F. So I heat treated coupons from the 1650, 1700, and 1750°F anneals with both cooling rates. This also allowed me to compare properties with the two different cooling rates:

The hardness after heat treating was slightly higher with the faster annealing rate, which we would expect to see from having somewhat finer carbides. There wasn’t much difference between 1650 and 1700°F though there was a small drop by increasing to 1750°F. So I would probably recommend sticking with the 1650-1700°F range for annealing MagnaCut.

The toughness was slightly higher for the 50°F/hr anneal in line with its slightly lower hardness. Both the 50°F/hr and 100°F/hr anneals resulted in somewhat higher hardness than the as-received condition. You could austenitize somewhat lower or temper somewhat hotter to reduce the hardness if desired. The forged and annealed MagnaCut had a slightly higher hardness-toughness balance than the as-received material. However, toughness testing can be somewhat variable and I have managed to get ~16 ft-lbs with 62 Rc as-received before. I think the toughness may be slightly better or slightly more consistent with the forging and annealing. Probably not enough to justify forging blades instead of doing stock removal, but I have also heard some bladesmiths claim that you will “ruin” stainless steel by forging it yourself and that is certainly not the case if you do it right. I also don’t think that toughness results of forged low alloy and simple carbon steels are enough to justify forging. Knifemakers should choose forging or stock removal based on other factors.

Summary – General Stainless and High Alloy Tool Steel Recommendations

Don’t forge too hot – 2100°F is a good target temperature. A common mistake is thinking that stainless steels need higher temperatures to forge. They do not.

Don’t forge too cold. Stainless and high alloy tool steels need higher minimum forging temperatures. They become “hot short” (brittle) at higher temperatures than simple carbon steels. Depending on the steel and the datasheet, this minimum is given in the range of 1650-1750°F (900-950°C).

Protect the steel from atmosphere during annealing – Vacuum or inert gas is best but for many knifemakers this means double wrapping in foil (not double folding but double wrapping), and leave some extra steel to grind away.

Choose the right annealing temperature – Typically in the range of 1600-1650°F for most stainless and high alloy steels but can sometimes vary. Check for a datasheet if a temperature recommendation is given and that temperature should work fine. A 2 hour hold at the temperature is typical. If no datasheet temperature is available you may be able to use the method shown above with AEB-L and MagnaCut where I found the temperature at which the hardness goes up, indicating that austenite formed.

Cool at 100°F/hr – This is significantly faster than recommendations given in most datasheets but gives a better balance of speed (~8 hours instead of 24+, can anneal overnight) and final properties (higher final hardness and toughness). There may be some steels that cannot handle this faster cooling rate if they have very high Mo contents. If the steel ends up higher than 25 Rc at the end it may need slower. As a side note, 100°F/hr was also found to be optimal with ApexUltra, a low alloy steel, so maybe we are on to something here.

Cool to 1000-1100°F, after that the cooling rate can be faster. You can experiment with whether the final cooling point can be higher than 1000 or 1100°F by annealing both ways and seeing if the hardness is the same. Once the transformation is done, further cooling doesn’t matter. I probably wouldn’t go any higher than 1200°F in any case. This really only matters if that extra hour or two will significantly impact your workflow.

So to summarize, anneal at 1600-1650°F for two hours, slow cool at 100°F/hr to 1000°F. Before that don’t forge too hot or too cold, and protect it from atmosphere during annealing.

---

[1] Roberts, G A, and Robert A. Cary. Tool Steels. Beachwood, Ohio: American Society for Metals, 1980.

[2] Roberts, George Adam, Richard Kennedy, and George Krauss. Tool steels. ASM international, 1998.