Thanks to Thomas Busch for becoming a Knife Steel Nerds Patreon supporter!
Misc. update: I have added a set of supporting micrographs to the introduction to Austenitizing steel.
To begin describing what bainite is it makes sense to start with martensite first. To form martensite we heat up the steel to high temperature to transform to a phase called austenite where we dissolve carbon in between the iron atoms (see Austenitizing Part 1), then quench the steel to lock in the carbon and form a hard phase called martensite (see What Makes Quenched Steel so Hard?). Following that we temper the martensite to allow some of the carbon out and increase the ductility of the martensite; the carbon comes out as very small carbides, a compound of iron and carbon (see What Happens During Tempering?). In the article on martensite formation I shared the following YouTube video to see the formation of the martensite laths:
You can see that the laths grow almost instantaneously once they start forming (nucleation). It is a rapid transition once a sufficiently low temperature is reached to drive the martensite formation. You can see this quenching process in a Time Temperature Transformation (TTT) diagram when the line indicates a rapid quench to avoid all other types of transformations such as pearlite, and is cooled through the martensite start (Ms) and martensite finish (Mf) lines instead:
Then the steel is reheated to a low temperature for tempering where small carbides come out, and are replaced by larger carbide “plates” at higher tempering temperatures :
Why spend so much time on martensite formation if this article is supposed to be about bainite? The two share similar features. Bainite is almost like directly forming tempered martensite. The TTT diagram shows that to form bainite we quench to an intermediate temperature and hold rather than directly to room temperature:
This quenching process is often done with molten salts that can hold the intermediate temperatures required for bainite formation. The process of holding at the intermediate temperature to form bainite is called “austempering.” Whereas martensite laths forms rapidly without diffusion, bainite also forms laths but diffusion is required to form carbides as the laths grow; the carbon is allowed to diffuse out as carbides while the bainite grows :
You’ll notice that the image is labeled as “lower bainite” formation. This is as opposed to “upper bainite” which forms at somewhat higher temperatures but is probably less interesting in a knives context because it is lower in hardness than lower bainite. You can see in the simple schematic above how the bainite laths form and that the carbides are forming within it as the laths nucleate and grow. You can also see in the schematic that the carbides are all aligned in the same direction within each individual laths. Each lath has its own orientation that the carbides follow. However, with martensite the carbides are more randomly oriented within individual laths :
On a broader scale you can see the formation of lower bainite in a 1080 steel with increasing time at 300°C, from 70s (a), 200s (b), 800s (c), and 2000s (d) :
Sometimes it is so difficult to differentiate between lower bainite and tempered martensite that high resolution electron microscopy is required to see the orientation of the fine carbides to differentiate the two. So if they are both lath-like phases with small carbides inside and can be difficult to tell apart, why would we expect there to be a difference in properties?
Properties of Bainite and Martensite
Verhoeven in his book  originally written for bladesmiths reported that his review of the literature revealed that bainite shows greater toughness than tempered martensite at hardnesses greater than or equal to 50 Rc. He did not provide any explanation as to why. Sometimes extreme differences in toughness between bainite and martensite have been published :
Tempered Martensite Embrittlement
To compare the toughness of bainite and martensite they need to be the same hardness first, because higher hardness almost always means lower toughness, putting martensite at a disadvantage if it has higher hardness. Martensite has a higher maximum hardness for a given carbon content than bainite does :
Therefore martensite has to be tempered down to the same hardness level as the bainite to realistically compare properties. Unfortunately, the degree of tempering required often puts the steel into the “tempered martensite embrittlement” (TME) region around 450-650°F where toughness is reduced :
With TME the hardness continues to decrease with higher tempering temperature but the toughness is either flat or decreases slightly. This leads to a lower hardness-toughness balance; it would have been better to temper lower and thus have both higher toughness and hardness. The toughness is reduced in that temperature range in part due to the loss of retained austeniteand also because of the carbide “plates” that are formed with higher tempering temperatures (see What Happens During Tempering?). A more recent study on bainite and tempered martensite in a 0.78%C steel found that tempered martensite had lower toughness than bainite at comparable hardness due to tempered martensite embrittlement . Bainite is not immune to large carbide particles, however, particularly at higher austempering temperatures. Therefore, in certain cases the formation of lower bainite at sufficiently low temperatures to avoid large carbide particles has superior toughness to tempered martensite that is tempered in the TME temperature range. This is reversed with high austempering temperatures where bainite forms large carbides. Fortunately, low temperatures are generally where knifemakers want to form bainite because the strength of the bainite is higher with lower austempering temperatures :
Another major source of embrittlement with high carbon steels is “plate,” rather than lath, martensite. There is a transition above about 0.6%C where these plates of martensite form instead which are much more brittle than lath martensite, in part because of microcracks that form:
However, even the very highest carbon bainite still forms laths, not plates, and so this embrittlement does not occur for bainitic microstructures. The study which Verhoeven cited for improved toughness of bainite looked at steels with sufficient carbon to form plate martensite, up to 0.85%C. And the superiority of bainite increased with increasing carbon content. This increasing gulf in properties with increasing carbon content likely indicates that it is plate martensite that is reducing the toughness of martensitic steel. Studies with low carbon steels as show tempered martensite has better toughness than bainite, and it is steels above about 0.5%C that show superior behavior of bainite rather than the cutoff being hardness as Verhoeven described :
Studies that specifically looked at the effect of plate martensite in high carbon steels when comparing with lower bainite found that the plate martensite was a major reason for lower toughness of tempered martensite . A study of 52100 hardened at sufficiently high temperature (for high carbon in austenite) to induce plate martensite formation found that the superior toughness of bainite was because of low toughness of plate martensite, though even then the difference at comparable hardness was relatively small, showing an improvement with charpy v-notch test but only a small difference with a modified fracture toughness test :
One advantage of austempering is that quenching is less severe and the steel is therefore less prone to cracking. Pre-existing cracks lead to poor toughness or even failure during tempering. Martensite is a larger phase than austenite, and rapid quenching leads to an uneven temperature distribution of the steel during quenching. The steel also changes size based on temperature; it grows slightly with increasing temperature based on the thermal coefficient of expansion. Therefore, the combination of size changes with uneven temperature distributions (the interior is hotter and larger than the surface) along with martensite formation (forms first at the colder surface) means the steel can see deleterious stress distributions which can lead to cracks. Bainite formation is much slower than martensite because it requires diffusion, and with the time required to begin bainite formation the steel is much more likely to have a uniform temperature distribution prior to transformation than with martensite formation. Sometimes a process called martempering is used which is not for the formation of bainite but to quench to above the martensite start temperature and allow the steel to reach a uniform temperature prior to cooling to form martensite. Martempering is performed specifically to avoid quench cracking.
High Carbon Steels and Martensite Toughness
Will all high carbon steels have relatively poor toughness with a tempered martensite microstructure? You may be looking at the composition of some tool steels with >1.5% carbon and wonder why anyone is even using them. Well, with alloy additions the carbon content of the austenite, and therefore the final martensite, is reduced. For example, here is a diagram showing how the “eutectoid” carbon concentration is reduced with increasing chromium additions:
A2 with its 5% Cr or D2 with its 12% Cr are designed to have just under 0.6% carbon in solution to avoid plate martensite and excess retained austenite. So even though these steels will not form bainite in a reasonable time with their high alloy content, it probably doesn’t matter because they are not usually forming plate martensite anyway. Simple carbon steels such as 1080 and 1095, however, are in danger of forming plate martensite. The likelihood of forming plate martensite can be reduced through the use of lower hardening temperatures , though that is easier to accomplish with a digitally controlled furnace rather than the commonly used forges with simple carbon steels.
Steels Suitable for Austempering and Bainitic Microstructures
As shown in the plot of carbon vs hardness, only the highest carbon steels are suitable for achieving the high hardness that we want in knives. This is probably a good thing since it is the high carbon steels that show the biggest improvement from a bainitic microstructure. However, if very high hardness is required then bainite is obviously not an option because its hardness is lower than martensite. Higher alloy additions slow down the formation of bainite, and with sufficient alloy content bainite formation is so slow as to be untenable, as shown for high speed steel T1 below (labeled 18-0-1), where at 200°C it takes over 4 hours just to start bainite formation . This limits the application of bainite microstructures to relatively simple steels such as 1095, 52100, and O1. Even these simple steels require relatively long hold times to ensure complete transformation to bainite. A temper is probably recommended after austempering in case there is an incomplete transformation to bainite, as any remaining austenite will transform to martensite upon final cooling. TTT diagrams for a few common low alloy steels that may be suitable for austempering are also shown at the end of this article .
Mixed Microstructure of Bainite and Martensite
In some cases superior properties have been found by having a combination of both lower bainite and martensite. Here is an example of 52100 austempered at 270°C (518°F) for different times, where there is a peak in both strength and toughness at about 30 minutes which gives some bainite (formed during the hold) and some untempered martensite (formed after cooling to room temperature following the hold) :
The introduction of about 20% softer bainite has been found to lead to higher strength than 100% martensite , because as the bainite forms within austenite carbon diffuses out of the bainite in the remaining austenite as it forms which increases the final strength of the martensite. This led to the peak in hardness at about 30 minutes, as this would correspond to about 20% bainite, with the remaining microstructure being untempered martensite. With increasing time at austempering temperature more bainite was formed for lower hardness. However, this peak in hardness and toughness has not always been reproduced, as is shown with another study on 52100 with a similar austempering temperature of 275°C :
Mixed Microstructures with Martensite Formed First
At the end of this article TTT diagrams for several low alloy steels can be seen . The far left line shows the time required for the transformation to start, and the far right line indicates the time required for the transformation to be complete. There are also final hardness values which gives an indication of the hardness that is achieved with complete transformation for each temperature. Note that in some cases the transformation finish time extends below martensite start, as shown in the schematic diagram above. Below that temperature some martensite forms first (during quench to the hold temperature) and then during the hold bainite forms. It is important to realize, however, that often the time required for complete transformation below Ms is an extrapolation. The time to complete the transformation can be reduced by the presence of martensite that is formed during the quench to below Ms, but the diffusion rate is also slow at low temperatures, so the behavior can be complex. The mechanical properties of steel with below-Ms austempering have received little study .
If the steel is first quenched slightly under martensite start to form some fraction of martensite followed by reheating above martensite start, it has been found that the time required for bainite formation has been reduced in part because less austenite is present to transform and in part because the existing martensite increases the nucleation rate for new bainite laths . The martensite laths that form first act as nucleation sites for the bainite that forms. This “pre-quench” process can be seen in the schematic above. This process can allow the use of low austempering temperatures for high hardness but still complete the reaction in a reasonable time. A study on 52100 found that using this process showed an improved combination of impact toughness and hardness by forming 33% martensite with a quench to 200°C prior to reheating to 240°C for 18 minutes to form bainite, with a temper at 200°C afterward . However, since there was only one comparison with a quench and temper (martensite) condition which had a higher hardness, it is unclear if the balance of toughness and wear resistance was truly higher. It has been proposed that the formation of the initial martensite leads to a refinement of the bainite that forms afterward for an improvement in properties; however, the properties of the steel that was not given a pre-quench and thus only formed martensite following bainite formation had similar properties, so it may not matter what order the phases formed but simply the fraction of each. Based on these studies it is possible that a mixed martensite-bainite microstructure has a higher hardness-toughness balance but it is not a certainty.This process requires two salt pots or perhaps a high temperature marquench oil for the first step, but would be an interesting area for further study.
Some studies have shown bainite having a small advantage over tempered martensite in wear resistance at a similar hardness , generally due to either a higher fraction of small carbides vs the tempered martensite or due to retained austenite. The number of studies looking at abrasion resistance rather than the less useful sliding wear resistance is small, however, and the carbides and retained austenite are both affected by the type of austempering treatment and composition of the steel. Under sufficiently high pressure the retained austenite transforms to fresh martensite and this transformation increases wear resistance. Retained austenite is often undesirable in knives, so an improvement due to retained austenite may not be beneficial overall for a knife. However, there may be an effect of the carbides or lower bainite microstructure itself. The figures below have some variation in how they represent wear resistance; however, in all cases higher hardness means better wear resistance, so that is the key to interpreting them.
Conclusion and Summary
Tempered martensite and lower bainite are very similar in that they are both lath-like microstructures with small carbides within. Martensite has the potential to be stronger (higher hardness) which can mean better resistance to edge rolling or permanent bends in knives. Previous studies indicating that lower bainite has superior toughness to tempered martensite appear to be due to comparisons with embrittled martensite from plate martensite formation or tempered martensite embrittlement. Therefore if those embrittlement mechanisms of martensite are avoided the two microstructures have similar toughness at the same hardness, so either has similar resistance to edge chipping or breakage in knives. Lower bainite may have a slight advantage to tempered martensite in terms of wear resistance, but that may be reversed if the martensite is at a higher hardness. Forming a completely bainitic microstructure can take 2 hours or more with low alloy steels, but this can be accelerated by quenching below martensite start prior to reheating to form bainite. Short austempering treatments with incomplete transformation can also be used as long as the resulting martensite is tempered. High alloy steels are unsuitable for austempering due to the slow rate of transformation. The small amount of research that has been completed on mixed martensite-bainite microstructures is promising and potentially shows a small increase in the hardness-toughness balance.
Caron, R. N., and G. Krauss. “The tempering of Fe-C lath martensite.”Metallurgical Transactions3, no. 9 (1972): 2381-2389.
 Bhadeshia, H. K. D. H., and D. V. Edmonds. “The mechanism of bainite formation in steels.”Acta Metallurgica28, no. 9 (1980): 1265-1273.
Samuels, Leonard Ernest.Light microscopy of carbon steels. Asm International, 1999.
 Verhoeven, John D.Steel metallurgy for the non-metallurgist. ASM International, 2007.
Olund, P., S. Larsson, and T. Lund. “Properties of bainite hardened SAE 52100 steel.” In18 th ASM Heat Treating Society Conference and Exposition including the Liu Dai Memorial Symposium, pp. 305-309. 1998.
Horn, R. M., and Robert O. Ritchie. “Mechanisms of tempered martensite embrittlement in low alloy steels.”Metallurgical Transactions A9, no. 8 (1978): 1039-1053.
Tu, Meng-Yin, Cheng-An Hsu, Wen-Hsiung Wang, and Yung-Fu Hsu. “Comparison of microstructure and mechanical behavior of lower bainite and tempered martensite in JIS SK5 steel.”Materials Chemistry and Physics107, no. 2-3 (2008): 418-425.
 Chandler, Harry. “Heat Treater’s Guide.”ASM International, Geauga County(1995): 661-663.
 Niccols, Edwin H.Literature Review: Impact Toughness of Bainite vs. Martensite. No. WVT-TR-76012. WATERVLIET ARSENAL NY BENET WEAPONS LAB, 1976.
 Bowen, P., and J. F. Knott. “Size effects on the microscopic cleavage fracture stress, σ F*, in martensitic microstructures.”Metallurgical Transactions A17, no. 2 (1986): 231-241.
 Das, Santosh Kumar. “Structure and mechanical properties of Fe-Ni-Co-C steels.” (1968).
Kar, Rameshchandra J. “Optimization of Strength and Toughness in a High Carbon Steel.” (1976).
Kevin Cashen’s Guide to 1080 & 1084
Kaleicheva, J. A. “Structure and properties of high-speed steels after austempering.”International Journal of Microstructure and Materials Properties2, no. 1 (2007): 16-23.
Chakraborty, J., D. Bhattacharjee, and I. Manna. “Austempering of bearing steel for improved mechanical properties.”Scripta Materialia59, no. 2 (2008): 247-250.
Young, C. H., and H. K. D. H. Bhadeshia. “Strength of mixtures of bainite and martensite.”Materials Science and Technology10, no. 3 (1994): 209-214.
Kilicli, Volkan, and Mucahit Kaplan. “Effect of Austempering Temperatures on Microstructure and Mechanical Properties of a Bearing Steel.”Scripta Materialia59, no. 2 (2008): 131-254.
Tian, Junyu, Guang Xu, Mingxing Zhou, and Haijiang Hu. “Refined Bainite Microstructure and Mechanical Properties of a High‐Strength Low‐Carbon Bainitic Steel Treated by Austempering Below and Above Ms.”steel research international89, no. 4 (2018): 1700469.
Dong, J., H. Vetters, and H-W. Zoch. “Shortening the duration of heat treatment in the lower bainitic range.” International Heat Treatment & Surface Engineering, no. 5 (2004): 555-560.
Li, C., and J. L. Wang. “Effect of pre-quenching on martensite-bainitic microstructure and mechanical properties of GCr15 bearing steel.”Journal of Materials Science28, no. 8 (1993): 2112-2118.
Misra, Ambrish, and Iain Finnie. “A review of the abrasive wear of metals.”Journal of Engineering Materials and Technology104, no. 2 (1982): 94-101.
Xu, Liqun. “Abrasive wear of ferrous alloys.” (1991).
Lefevre, Justin, and Kathy L. Hayrynen. “Austempered materials for powertrain applications.”Journal of materials engineering and performance22, no. 7 (2013): 1914-1922.
Moore, M. A. “The relationship between the abrasive wear resistance, hardness and microstructure of ferritic materials.”Wear28, no. 1 (1974): 59-68.