By: Graham Fredeen
Isothermal Transformation Diagram or TTT (Time, Temperature, Transformation) Curve
One of the most important things that should be learned when heat treating steel is how to read a TTT curve and what the TTT curve illustrates.
Picture temporarily removed
A TTT curve illustrates the transformations of the microstructure that steel will undertake given various rates of cooling over a range of temperatures. Above is an example TTT curve. TTT curves will vary depending on the type of steel.
To read a TTT curve it is important to be able to identify the various features presented in the curve and understand what they represent and how the curve “flows.” The TTT curve does not just show the state of steel at a given temperature, but rather illustrates the various microstructure states a piece of steel will have when cooling from a temperature point above critical temperature (the temperature at which Austenite forms). Thus the chart only applies to steel that is in an Austenitic state and is being cooled, and thus the beginning temperature is always around 1500 F (a rough temperature that is close to most steel’s critical temperature, however individual TTT curves will illustrate exact critical temperatures for a specific steel). To read the chart, assume that the steel is at critical temperature, determine the temperature you are going to cool the steel to, in a given amount of time. Drawing a line from that temperature through the point of cooling at the given amount of passed time will yield a line, which ever states the line intersects will be the microstructure transformations the steel will undergo. The most useful thing the TTT chart can be used for is to determine the rate of cooling necessary for steel to harden.
Features of the TTT curve:
The line labeled A1 illustrates the temperature at which Austenite formation occurs, and divides the boundary between Austenite and Unstable Austenite.
The most prominent feature, and possibly the most important feature is the pearlite nose line, which shows when the formation of pearlite will occur (pearlite is composed of Ferrite and Cementite, hence the F+C). This is the large silver protrusion labeled pearlite F+C (Ferrite + Cementite) and includes the lighter A+F+C (Austenite + Ferrite + Cementite) section. This section is highly important because if the rate of cooling intersects the pearlite nose line, pearlite will form and the steel will not harden/will not harden completely. This is crucial to know when trying to harden steel.
You will note the very narrow corridor created by the pearlite nose line around 1000oF. In order to avoid pearlite formation, the steel must be cooled fast enough from critical temperature to pass through this narrow area without hitting the pearlite nose line. This is the reason why steels must be quenched when hardening, to ensure that they cool fast enough to miss hitting the pearlite nose line. In shallow hardening steels (things like lower carbon steels in general, and plain carbon steels near the medium and lower end of the high carbon range without alloying elements) the channel between the pearlite nose line is much narrower, since there is less carbon in solution it will diffuse at a faster rate so the steel must be cooled more quickly. With deeper hardening steels (steels with higher carbon content and alloying elements which affect the rate of carbon diffusion) the area between the pearlite nose line is much wider, meaning that the steel can be cooled more slowly while still hardening. This necessity for different rates of cooling is why different quench mediums are used. Shallow hardening steels require faster quench mediums like water and brine, and deeper hardening steels require a slower quench like oil, or air. By looking at the time needed to bring a given steel below the pearlite nose line the proper quench can be selected. A quench that is insufficient to miss the pearlite nose will result in varying amounts of fine pearlite formation and will keep the steel from reaching maximum hardness. Often steel manufactures already give suggested quench mediums for a given steel after analyzing the TTT chart.
The area labeled Unstable Austenite illustrates the area where Unstable Austenite exists. Unstable Austenite is Austenite which is “in between” transformation so to speak. This is Austenite that will undergo some transformation depending on the rate of cooling and is not a state that the steel can remain in, but is only a transition state, it will quickly transition into the more stable microstructures as soon as conditions will allow.
The next important feature on the chart is the Martensite Start (Ms) point (indicated by the green line around 400oF). This is the point at which Martensite begins to form and when the steel actually starts to harden. Before this point, Martensite will not form and the steel will not actually begin to harden. You will note that the most crucial cooling time is the time needed to cool the steel to miss the pearlite nose, afterwards cooling can be much slower and martensite will still form.
Hardening refers to the process used to (that’s right you guessed it) harden steel. The hardening process involves heating steel passed its critical temperature into an Austenitic state, holding the steel above this critical temperature for a period long enough to ensure that all of the carbon present in the steel has entered solution (and formed Austenite), and then quenching the steel to rapidly cool it in order to form Martensite and harden the steel. As discussed in the TTT curve explanation, the rate of cooling is crucial in order for Martensite formation to occur. If the Pearlite nose line is intersected during cooling the steel microstructure will not be completely Martensitic and will contain small amounts of pearlite, and will not be as hard. Additionally in many steels (especially those with high amounts of carbon and/or alloying elements) it is possible to get the steel “too hard” or to form too much Martensite. Martensite has a body centered tetragonal structure which takes up more space than the body centered cubic structure of pearlite. As a result when martensite forms the steel actually expands, and the crystalline structure becomes more rigid (hence harder), this more rigid structure can not take as much deformation as the more elastic pearlite structure, so if the martensite expansion is to much, the steel will crack. As a result, some steels require slower rates of cooling than others to help cut down on some of the Martensite formation and prevent cracking. The rate of cooling of a particular steel is determined by the quench medium used. Common quench mediums include brine, water, oil, and air. Brine being the fastest, air being the slowest. The speed of the quenchant depends greatly on its density and viscosity. The more dense the fluid, the greater thermal capacity it has (meaning it can absorb more heat), and the less viscous a fluid is, the easier it can flow and circulate, meaning the faster it can dissipate heat. Oil is less dense than water and more viscous, meaning it has less thermal capacity, and will not dissipate heat as quickly, therefore it is slower in cooling steel.
Picture temporarily removed
Additionally the hardness achieved within steel will depend on the amount of Martensite that forms, which is dependent on the amount of carbon which is retained in solution when quenched. This is of course a factor of firstly the rate of cooling which occurs in the steel, the faster the rate of cooling the more carbon will remain in solution and more Martensite will form (this is what the TTT curve illustrates, Pearlite being composed of iron and carbon which is not in solution). The amount of carbon present within the steel itself will also dictate hardness. The more carbon present in the steel, the more carbon which can enter solution when heated and with more carbon in solution longer time is needed for the carbon to diffuse back out of solution so greater amounts can be retained in solution when quenched and more Martensite will form. However, the greater the carbon a particular steel has does not necessarily mean that it will form more Martensite and that it will become harder. This depends on whether the steel is in or has surpassed its eutectic state. Eutectic refers to a condition in chemistry when two elements can be alloyed together on an atomic level, but only up to a specific percentage, at which point any additional secondary element will retain a distinct separate form. In other words, given a certain amount of iron (ferrite), there is only a certain amount of iron carbide (cementite) which can diffuse into solution with the ferrite to form Austenite, which can then be used to form Martensite (the eutectic point for austenite is roughly .8% carbon). Once the eutectic point has been reached additional carbon can not enter into solution and will remain as separate bands of cementite and will not affect the amount of Martensite which can form. Steel which has more carbon than can go into solution is known as Hypereutectic. Steel which has less carbon than the eutectic point is said to be Hypoeutectic. And steel containing the same amount of carbon as what can enter solution is known as a Eutectic steel. Hypoeutectic steels will not attain maximum hardness. Eutectic and Hypereutectic steels will have maximum Martensite formation and can attain greater hardness. The additional carbon in Hypereutectic steel can slow diffusion rates and keep more carbon in solution and the presence of separate bands of the excess carbide that form along side the martensite can add different performance characteristics to the steel as well.
Alloying elements also have an effect on hardness because they affect the rate of carbon diffusion within the steel (usually slowing the rate of diffusion), so more carbon is retained during the quench and more Martensite is formed.
The term temper is often misused and misunderstood. Tempering is often confused with hardening. This misconception is probably a result of old smiths referring to the blades final state of heat treatment as its temper. Most people do/did not know that there are more steps involved in the heat treatment of a blade besides hardening, therefore they assume/d that the term temper must apply to the hardening process. This is of course not the case as tempering actually performs a function much the opposite of hardening.
In simple and basic terms, one can think of tempering as slightly softening the steel after hardening to take away enough hardness to prevent brittleness while imparting toughness and strength.
Tempering serves a few main purposes. After the hardening quench it is possible for some Unstable Austenite to be retained within the Martensitic structure of the hardened steel. After the temperature of the steel has been lowered to the point where Martensite begins to form, the longer it is allowed to cool from this point the more Martensite will form. If there is not a long enough cooling time below the Martensite start point (which is often the case when quenching a blade) some of the unstable Austenite can be trapped and not allowed to transform. This unstable Austenite creates a steel structure which is not as hard as it could be and creates points of weakness and failure within the blade. A blade with large quantities of unstable Austenite after quenching may not appear to have hardened at all (I have come across this instance a few times with large sword blades). Tempering brings the steel close to the temperature at which Martensite begins to form (this is called the Martensite start or Ms point). Near this temperature unstable Austenite will convert into Martensite.
Secondly, and perhaps the main reason for tempering, is that fully hardened steel (Martensitic steel) forms a Body Centered Tetragonal crystalline structure (also known as Alpha Martensite)
Picture temporarily removed
This structure is harder due to the additional lengths and strains in bonds between the individual atoms. These stronger bonds are ultimately what makes the martensitic structure stronger and harder. A conceptual way to think about this is that this structure takes up more room than the Face Centered Cubic structure which creates a more dense and tight packed structure with high rigidity. Additionally due to the Tetragonal distortion the structure is not as deformable to shear or bending forces. The combination of the tightly packed structure and the lack of deformation results in “hardness” (hardness being really nothing more than a strong resistance to deformation). This type of structure has large quantities of abrasion resistance and “strength” (compressive and tensile). However due to this lack of deformation, the structure is very brittle especially in impact strength and under bending forces. This type of structure is therefore not suitable for a blade which must be able to undergo impact stress, as well as bending stresses. This is where tempering comes into play. When a blade is heated to tempering temperatures, it is heated just enough to free some of the trapped carbon from the BCT structure and allow it to diffuse out of solution. This allows the BCT structure of the Alpha Martensite to relax some and form Beta Martensite. This removes some of the rigidity of the structure and allows for greater deformation and higher elasticity, making a structure that is both hard and resistant to abrasion, and “tough,” being able to take large impact and bending stresses.
Finally tempering also serves to relieve stresses built up within the steel imparted during the hardening quench. The rapid cooling and contraction caused from the quench creates large stress build up within the steel, as does the distortion of the standard BCC structure of Pearilitc steel into the BCT structure of Martensitic steel. As Martensite forms, it takes up more room than the standard Pearlite and the blade actually expands. This expansion paired with rapid contraction from cooling and the rigidity of the martensite structure means stress build up is huge. This stress must be relieved or the blade will risk failure due to cracking or distortion. Tempering, in allowing small amounts of carbon diffusion, relieves this built up stress.
Here is a chart that shows the effects tempering has on steel hardness versus its strength.
Picture temporarily removed
Annealing, simply put is the process of softening steel. You don’t always want steel to be hard during the entire process of working it. Annealing steel allows it to be more easily worked by imparting qualities of malleability and machineability. In simple terms annealing softens the steel so you can more easily file and grind it to put the final shape on your blade. In a completely annealed state, the microstructure of steel exists as pearlite (which is nothing more than bands of ferrite and cementite [iron and carbon forming iron carbide]).
The Pearlite microstructure is achieved by heating the steel past critical temperature into an austenitic state, and then cooling the steel slow enough to allow the carbon to diffuse out of solution.
Picture temporarily removed
As seen in the TTT curve, the slower steel is cooled, more pearlite will form and hence the softer the steel will become. The slow rate of cooling necessary to anneal can be achieved in multiple ways. Firstly the piece of steel can be placed in an insulating medium (such as vermiculite) which will allow it to cool very slowly. The steel can also be placed in a controlled furnace where the temperature is slowly ramped down. Often leaving a piece of steel in the forge after the gas/fire has been cut off will anneal it enough for most bladesmithing applications.
Normalization is one of the most important steps in heat treatment, yet is often one of the most over looked steps.
As steel is worked into a blade it is undergoing different heating and different stress as it is forged and ground and shaped. Different portions of the blade are heated and cooled at different rates and different internal stresses are placed on different portions of the blade. This difference in heating creates different microstructure and grain size to exist within different portions of the blade and allows stress from working the blade to be retained within this structure. These differences in grain structure and retained stress will create un-uniform hardening within the blade, additional and unneeded stress within the blade, and deformation in the blade, all of which are unwanted when making a blade that must perform.
Grain growth in steel occurs at temperatures higher than critical. Steel is usually forged a good way above critical temperature. The longer time the steel spends at these higher temperatures the larger the grain will grow. Large grain structure within steel is not desired, it creates large planar faces and grain boundaries for fractures to occur over making the steel “weaker.” Instead a smaller and more refined grain structure is better for blade performance. The smaller grain structure not only makes the blade less likely to fracture but it aids in sharpening and edge retention.
In order to refine grain structure within steel and create uniformity within the grain structure, as well as relieve stress, normalization is necessary. The normalization process involves heating the steel to an austenitic state, holding this temperature for a short time, then allowing the steel to air cool uniformly. This process is often done multiple times and sometimes gradually decreasing the temperature with each cycle. By heating the steel into an austenitic state and allowing it to soak, all of the carbon in the steel is allowed to enter into solution, removing any differences which may exist within the microstructure of the steel imparted from the forging and shaping processes. Allowing the steel to cool in the air uniformly creates the same microstructure throughout the entire blade. Additionally by heating to critical (and not very far above) and allowing the blade to cool in the air helps reduce and refine the grain structure, and create a uniform grain structure throughout the steel.
Proper normalization will greatly increase blade performance by creating a desirable grain structure as well as a uniform structure prior to hardening. This prevents areas of differing hardness (soft and hard spots along the blade). By relieving the stress and creating a uniform microstructure, warping during quenching is also drastically reduced. And the refined grain structure reduces the chances for blades fracturing during the hardening process.
It is advisable to run multiple normalization cycles during heat treat. This helps to insure that there is as much uniformity within the steel structure as possible. Decreasing the temperature gradually (what is known as a “step down normalization cycle”) helps to refine the grain structure (lower temperature limits grain size). 3 step down normalization cycles are the recommended normalization procedure for heat treatment.
Austenite was the microstructure state of steel discovered by Sir W.C. Roberts Austen, hence the name. It exists as a solid solution of iron and carbon. Austenite is a transitional microstucture that then trans forms into the other microstructures (martensite, pearlite, bainite etc...) based on rates of cooling. Iron exists in a standard state as a body centered cubic structure, referred to as “alpha iron.”
Picture temporarily removed
Alpha Iron (BCC)
When heated, the standard ferrite (iron) structure opens up and expands, creating what is known as “gamma iron” and transitions from a body centered cubic structure to a face centered cubic structure.
Picture temporarily removed
Gamma Iron (FCC)
Due to the expansion in structure, the gamma iron is now capable of holding much more in the extra space and begins accepting carbon to enter the structure and go into solution.
Austenite is unstable at temperatures below critical temperature and will eventually transition into another microstructure. If cooled slowly, the carbon will diffuse out of solution as the FCC structure begins to contract and return to a BCC structure, forming separate bands of iron carbide and iron (cementite and ferrite) making pearlite. If the rate of cooling is more rapid, the carbon is not given enough time to diffuse out of solution as the structure contracts and it becomes trapped. This trapped carbon results in the formation of martensite (or bainite, depending on the rate of cooling), in which the contracting structure is distorted from a body centered cubic structure into a body centered tetragonal structure due to the additional presence of the carbon within the interstitials.
Bainite forms when steel in an austenitic state is cooled fast enough to miss the pearlite noseline, but slow enough to avoid entering the martensite start temperature. To achieve a bainite structure often a process known as marquenching is necessary, where the steel is quenched into a medium that is hotter than the Ms point and held at this temperature, preventing the steel from forming martensite and entering the bainite phase instead. Bainite has some characteristics of both pearlite and martensite. It is significantly stronger and harder than pearlite, but at the same time it does not have the brittleness of martensite. Bainite characteristics are similar to those of tempered martensite.
Picture temporarily removed
Cementite, another name for iron carbide (Fe3C), is the form that carbon takes in steel. Iron carbide is actually a ceramic and is extremely hard. Pure iron carbide is used heavily in machining tooling for its hardness and wear resistance properties.
Picture temporarily removed
Ferrite refers to the “iron” part of steel. Ferrite exists as a collection of iron atoms which arrange themselves in a body centered cubic structure at room temperature.
Picture temporarily removed
Martensite is the hardened structure of steel, and was discovered by Adolph Martens. Martensite is formed by cooling Austenite at a fast enough rate that the carbon that is diffused into the FCC ferrite structure is trapped by the contraction of the crystalline structure as it cools. As the structure contracts around the carbon atoms while trying to form a body centered cubic structure, the structure is distorted, forming a body centered tetragonal structure.
Picture temporarily removed
As the FCC structure of Austenite contracts, the carbon present within the structure is not allowed to diffuse out of solution and form pearlite. As a result the martensite which forms will have the same composition as the parent austenite, and will retain the same grain structure as well, making it a diffusionless process. The BCT structure of martensite takes up more space than the more tightly packed BCC structure of pearlite thus lengthening and straining the bonds between the atoms, which strengthens the bonds and creates rigidity in the structure. This rigidity in the structure resists deformation and is what creates “hardness.” The resistance to deformation is also the reason why martensite can be highly brittle in nature, as rather than flex and deform under stress it will fracture.
Picture temporarily removed
Pearlite is the microstructure of softened or annealed steel. Pearlite is composed of ferrite (alpha iron) and cementite (iron carbide). Pearlite is formed when austenitic steel is cooled slowly from critical temperature, which causes the steel to cross the pearlite nose line. During this slow rate of cooling, the FCC structure of austenite begins to slowly contract and the carbon atoms which were in solution, filling the excess space of the FCC structure, begin to diffuse and move out of the crystal matrix. These carbon atoms form cementite and align themselves along side the now BCC structured ferrite, forming layers of ferrite and cementite. This lamellar structure can undergo large deformation as the layers of ferrite and cementite do not produce a very rigid structure. This ability to deform makes pearlite very “soft” and gives the structure good malleability and machineability characteristics. The pearlite microstructure is often desired in bladesmithing to make grinding and shaping the steel into a blade much easier. Pearlite is attained through the annealing process described above.
Picture temporarily removed
Basic Heat Treatment Procedure
Heat treatment refers to the application of the metallurgical knowledge stated above to perform thermal cycling on a piece of steel to attain a desired microstructure for a desired performance characteristic/s. In simplest terms, heat treatment is the process of hardening steel, and then tempering the hardness back to a desired point to get the steel to perform the way you want it to (or at least that’s the idea, whether it does or not is a different story). The basic steps involved in heat treatment are normalization, hardening, and finally tempering, in that order (some include annealing the steel for the final shaping and grinding before heat treatment is done, but for this explanation, annealing will be considered a separate process from heat treatment).
After the blade is finished being shaped through forging, grinding, filing, and sanding it is then ready for heat treatment. It is important to note that when shaping the blade prior to heat treatment, the blade should be left slightly oversized to allow for descaling and cleanup. Additionally the edge of the blade should be left a bit thick (roughly a thickness between a dime and nickel is recommended). During heat treatment the blade is heated above critical temperature, which firstly frees some of the carbon in the blade and the increased temperature increases the oxidization reaction that the steel undergoes (this is why scale forms on blades when they are heated, it is essentially an accelerated “rusting” process). If the edge is left too thin during heat treatment, it is possible for decarburization to occur, which will leave the edge soft and prevent it from performing as intended. The thicker cross-section of leaving the edge thicker helps slow the effects of oxidization and prevents the edge from becoming decarbonized.
The first procedure in the heat treatment is the normalization of the blade. Normalization is done to create uniform grain size and uniform microstructure within the steel and to help refine the grain size and relieve stress within the steel. Normalization is a very important step in heat treatment which many tend to overlook. Inconsistent grain size and inconsistent microstructure can cause many problems during hardening, such as warping and cracking. Additionally lack of normalization will adversely affect blade performance. Inconsistent hardening can occur, forming hard and soft spots along the blade. Inconsistent grain structure can cause weak spots and higher risk of blade failure. Large grain that is not refined through normalization creates large planar faces for fractures to occur, and can hinder edge performance. Many of the problems blade makers have with heat treatment (warping and cracking) exist due to lack of normalization, and can easily be prevented or drastically reduced.
To normalize a blade, the blade is heated up past critical temperature, allowed to soak at this temperature for a short time, then removed from the forge/furnace and allowed to cool uniformly in the air. The critical temperatures for steels vary depending on their carbon content and presence of alloying elements (most critical temperatures for steel fall around the 1500F- 1600F range). It is important to look up the critical temperature for the particular grade of steel being used. It is very important during normalization to attain an even heat along the entire length of the blade. This can be difficult in both coal and gas forges, which can have hot and cold spots. The longer the blade the more difficult it is to attain even heating as well. One thing that can be done to help this problem, without the need of expensive heat treating equipment is to take a steel pipe or tube and place it in the forge and allow it to heat up directly in the forge. Place the blade inside the pipe and allow the blade to be heated indirectly by the pipe. This indirect heating allows a much more uniform temperature to be achieved along the length of a blade. It is also important to have good control over the blade temperature. The normalization process should be done only at critical temperature, if the blade is heated much beyond critical temperature excess grain growth can occur within the steel. A pyrometer is the ideal way of telling the temperature, though a magnet, consistent lighting, and a good eye will suffice with practice. Note, do not trust judging steel temperature solely by eye, subtle lighting differences can cause large differences in the appearance of the steel. It is much safer to test with a magnet to determine when you are close to critical temperature and then visually adjust by eye. Remember not to stick the magnet in the forge with the blade, the heat of the forge will quickly demagnetize the magnet and make it useless, pull the blade and quickly check with the magnet and then put the blade back into the fire.
Once the blade has been brought up to critical temperature, allow the blade to soak at this temperature for a short time. Usually 2 minutes or so will suffice for most small blades, though longer will not hurt (the thicker the cross-section of the blade the longer the soak time is needed). The soak time ensures that the entire blade has reached critical temperature and gives the carbon in the steel enough time to fully diffuse and go into solution, forming pure austenite.
After the blade has soaked, remove it from the forge and allow it to cool uniformly in the air. It is best to hang the blade by the tang, or clamp the tang in a vise with the blade not touching anything besides still air. If you set the blade down on something, uneven cooling will occur and the grain structure of the steel will not be the same.
After the blade has cooled to the point where you can comfortably handle it, place it back into the forge and bring it back up to critical and repeat the procedure. Usually by practice, multiple normalization cycles are performed to ensure uniformity in the steel structure and to help refine grain as much as possible. The total number of normalization cycles performed depends on the individual’s own methodology, though by standard practice 3 cycles is a good number. One cycle is not enough, two cycles might get you by, three will pretty well ensure the blade is normalized well, and anything beyond, while not harming the blade, does not really produce much, if any, noticeable benefit. Additionally during normalization blades may warp slightly due to the built up stress. After the blade is cooled between cycles, the blade can be straightened. After the next cycle the blade should not warp as much. You can repeat the normalization cycles until the blade goes in and comes out straight. Some blades may continue to warp slightly during heating process, this is due to inconsistency in the grind, causing one side of the blade to have more material, which causes greater thermal expansion than on the other side of the blade. If this occurs, you can slightly “pre-curve” the blade in the opposite direction of the warp to roughly the same magnitude and the blade should straighten for hardening.
Another variation that can be done with normalization cycles is to perform step down normalization. Step down normalization refers to gradually decreasing the temperature of the steel with each cycle performed. This gradual decrease in temperature helps to further refine grain structure within the steel over the standard normalization. While step down normalization is not “necessary” it does have some benefit. When heat treating a blade I usually perform 3 step down normalization cycles. The first cycle I heat the blade to critical temperature (or a hair above), with the second I decrease the temperature slightly so the blade is at critical, and with the third I bring the blade slightly below critical. Some choose to normalize directly after forging but I prefer to wait until after grinding is done. Grinding produces heat which can slightly alter the grain of the steel and it will impart some small amounts of stress in the blade, and it is my belief that it is best to normalize after all of the grinding and shaping has been done. Plus, annealing the blade after forging acts as a rough normalization and a big stress reliever, so I don’t see the need to normalize directly after forging is complete.
After normalization it is time to harden the blade. The blade can go into the hardening heat directly after the last normalization cycle, or it can be done at a later time. To harden the blade, it is brought up to critical temperature and again allowed to soak at this temperature to ensure complete austenite transformation. Again, it is best to get as even a temperature along the blade as possible. When the blade has soaked at critical long enough, quickly (and carefully) remove the blade from the forge and move it to the quench tank. In a swift motion, quench the blade. When the blade is in the quenchant, try not to move it around, moving it can create inconsistent cooling and cause warping. When the blade is finished cooling it can be removed from the quenchant. It is important to complete the quenching procedure very quickly before the blade can cool very far below critical temperature, as you have seen on the sample TTT chart, there is only a very short time window (in many cases under a second) to get the blade below critical in order to miss the pearlite nose line. Allowing the blade to cool too much before quenching will cause incomplete hardening. As discussed before, the quench medium will depend on the carbon content and alloying elements a particular steel has. In general, the lower the carbon and the less alloying elements a grade of steel has, the more shallow hardening it is and the faster the quench medium needed to fully harden it. The higher the carbon and if there are alloying elements present, the slower the quench. Determine the correct quench medium based on the TTT chart and manufacturer recommendation for the particular grade of steel, as well as the desired performance characteristics for the blade. Also note from the discussion on the TTT curve, that after the blade has been cooled past the pearlite nose line, the rate of cooling can be much slower and martensite will still form. This knowledge can be useful when hardening long blades which tend to develop warps. If you quench the blade and pull the blade just before the martensite start point, you can correct any warps while the steel is still in an austenitic state (meaning it’s pretty soft and easy to bend) and as the blade cools, it will continue to harden.
After hardening, the blade is in a fully martensitic state. In this state the steel is very hard, yet due to this excessive hardness it is also very brittle and would not make a good blade. In order to make the blade tough, some of the hardness must be removed. To do this the blade is tempered. Wait until the blade has cooled long enough to reach the Mf point (martensite finish) and then immediately place the blade in the tempering oven to temper. It is always recommended to temper a blade immediately after hardening. The hardening process creates massive amounts of stress within the steel, as the blade cools it contracts, but as martensite forms it expands, these two counteracting stresses can create all kinds of problems if they are not relieved. Many times a blade can come out of the quench without issue, but if you set it down for a few hours, it can crack after the quench if not tempered. Additionally, not tempering a blade directly after quenching means it is in a very fragile state. If you try to work on the blade, or set the blade down and it falls of the work bench, or you drop it, etc, you run the high risk of accidentally breaking it. You can quickly run a file along the edge to confirm that the blade has hardened (though sometimes this can be deceptive if there is decarb along the edge or if there is retained austenite, the blade can appear as somewhat soft still), and you can sand away a bit of the scale to see the tempering colors, but do this relatively quickly and do not try to do any thing else like start on the final sanding or polishing. Most blades are small enough to fit in a toaster oven or conventional oven and this is what I recommend for tempering. If the lady of the house has a problem with you using her oven to temper your blades in, toaster ovens can be found at thrift stores for pretty cheap, and they are certainly worth every penny to maintain the peace. That or wait until she is out of the house for a good three hours or more…
I do not recommend pure torch tempering. For a proper tempering, it is best to let the blade soak at tempering temperatures for extended periods of time. Additionally the torch is difficult to control the temperature and does not create very consistent temperature along the length of the blade. If this is all you have, it can produce a serviceable blade, but it is not ideal and you will get better performance out of your blade with a cheap toaster oven for tempering. If you must torch temper, sand the blade and clean it so you can see the temper colors run, heat the blade with the torch, staring at the spine (do not heat from the flat of the blade as warping will occur), let the temper colors run from the spine to the edge, when the edge has reached a straw color, stop heating. You must be very careful as its very easy to over heat the edge, you will actually have to stop heating slightly before the edge turns straw or else you will over heat. Allow the blade to cool on its own, do not quench it. You may need to quench the edge to keep it from over heating, but it is best to allow things to cool on their own. Sand off the temper colors and repeat this at least 2 more times.
Tempering is one of the highly debated procedures where there is no one “right” way to go about it. Some people will only temper for one 1 hour cycle, some will do four 2 hour cycles. The important thing is to realize what tempering does (please refer to the tempering section for a more in-depth explanation). I would recommend running a bare minimum of two, 1hour temper cycles. This will help to ensure that all or most of any retained austenite is converted to martensite and will allow enough alpha martensite to convert to beta martensite (allow the blade to soften enough) to make it tough and remove the brittleness. 3, 1 hour temper cycles are better (I usually run 3 temper cycles, at a minimum of an hour a piece, sometimes up to 2 hours a piece, depending on the steel and the desired performance characteristics of the blade). The exact tempering temperatures and times will depend firstly on the steel itself and secondly on the performance characteristics needed for the blade itself. For shorter blades that do not need to flex as much or will not be put under as much impact stress, you can temper at lower temperature. For longer blades that will flex more and be put under more stress, higher tempering temperatures are necessary. Tempering is not an exact science and each blade produced will have its own tempering method. The best method is to firstly determine what the performance characteristics you need in the blade are, and then run an initial tempering cycle at 375 F (for smaller knifes). After the initial temper cycle, remove the blade from the oven and allow it to cool, once cool, grind a rough edge on it. Clamp a brass rod in a vise and perform an edge flex test. To do this, lay the edge on the brass rod and slightly angle the blade up, apply moderate pressure downward and draw the side of the edge across the rod and visually observe the edge as you do this. The edge should flex over the rod and return back to its original shape. If the edge chips out, you know that the blade needs to be tempered at a higher temperature. Increase the tempering temp by about 25F-50 F (depending on how the edge performed), run another temper cycle and repeat the edge flex test. If the edge returns to its original state, you have reached the correct tempering temperature, leave the temperature where it is and finish the remaining tempering cycles. If the edge folds over you have over tempered the blade, you can either leave the blade in the current state, or (unfortunately) the only way to correct this is to re-normalize, re-harden, and then re-temper the blade. Once the edge is at its proper state, the rest of the blade can be tempered more, depending on its intended use and necessary performance. To do this, you can get a shallow bath of water to immerse the edge in to keep it cool. Heat the spine of the blade with a torch to draw the temper back on the spine to impart additional toughness. The standard tempering procedure is sufficient for most small blades, and this additional torch temper is really only needed on longer blade or big choppers that will see a lot of abuse.
This is just a basic over view of metallurgy and heat treatment procedures for bladesmithing. While this is about 20 some odd pages of writing and diagrams and pictures, it really only begins to scratch the surface of heat treatment and metallurgy. There are many many other variables and conditions in heat treatment, depending on the steel, blade, differing techniques, and different performance characteristics which this simply does not cover. This is designed to give good foundational knowledge and understanding of what is going on in the steel when you heat treat a blade which can then be built upon with future research. I would be happy to answer any questions which may arise to the best of my ability. If I can not answer the question I will try to point you in the right direction or try to find someone who can.
I will also update this with other useful links and specific steel heat treat information as I come across it. I have a bunch of it spread out across my bookmarked sites to sift through and try to find. If anyone else has any useful links, charts, or diagrams, please send them to me so I can update this with the new information.
Some useful charts, links, etc:
Verhoven link: More in-depth paper on metallurgy for bladesmithing, very good information.