Category: Toughness
4 thoughts on “Toughness”
Leave a Reply
Why Cold Steel Is Brittle
Thanks to Tracy Mickley from becoming a Knife Steel Nerds Patreon supporter!
Effect of Temperature on Strength
Steels become stronger at lower temperatures. This can be measured with the “yield strength” which is the load to permanently deform the steel. This deformation is in the form of a shape change, ie. if you are bending steel it stays bent, when hammering hot steel it dents, when flexing an edge it rolls. This is perhaps easier to visualize for those that have forged steel because hot steel is easier to forge, and is progressively more difficult to forge as it cools down. This increase in strength at lower temperatures continues below room temperature, so steel at cryogenic temperatures is stronger than at room temperature. Here are values for yield stress for 410 stainless steel heat treated to 39 Rc [1]:
Cryogenic Processing of Steel Part 2 – Toughness and Strength
Thanks to Gator, Russell Dodd, and Matt de Clercq for becoming Knife Steel Nerds Patreon supporters!
Introduction
Part 1 of the Cryogenic Processing series covered the transformation of retained austenite to martensite and the increase in hardness that occurs. That is the least controversial aspect of cryogenic processing of steel. The other two primary properties of steel affected by cryo processing are toughness and wear resistance. Both of these properties can be difficult to pin down as they have high variability. Tool steels are known for their relatively poor toughness which means we are often comparing small numbers.
Detour – Tempering
One important interrelation to keep in mind with subzero and cryo studies is the transformation of retained austenite in tempering. With sufficiently high tempering temperatures all/most of the retained austenite is transformed without any cold treatment. This depends on the alloy content, as low-alloy 52100 will have lost its retained austenite with a 500-600°F temper while high alloy steels need over 900°F. You can read more in the article on tempering. With high alloy steels the loss of retained austenite also coincides with “secondary hardening” which is a high temperature tempering treatment that increases hardness [1]:
Above is a tempering chart for Caldie steel (0.7C-5.0Cr-2.3Mo) which shows both hardness vs hardening temperature and also retained austenite. You can see that at low tempering temperatures (<400°C) the retained austenite is basically constant. You can also see that the hardness decreases with higher tempering temperatures up to about 350°C and then it increases to a peak at around 520°C (950°F). Therefore tempering in the secondary hardening region above 400°C can lead to both high hardness and also the elimination of retained austenite.
Subzero or cryo processing prior to tempering also shifts the tempering-hardness curve to lower temperatures when using the secondary hardening range of tempering [2]:
This means that in general, a lower tempering temperature is required to achieve the same hardness level with secondary hardening. Using the same tempering temperature as without a subzero treatment will lead to a greater degree of tempering. More tempering can be good or bad depending on the situation. Excessive tempering can lead to coarsening of tempering carbides which can reduce toughness. However, if the tempering was insufficient without subzero, the use of subzero processing may increase toughness due to shifting the “optimal toughness” range.
Toughness
In an earlier article where we tested the effects of heat treatment on Z-Wear toughness
The Sharpest Youtube Channel in the World
Thanks to Rusty Craig, Dylan Curtis-Reeve, Brendan Porter, and Ian Cox for becoming Knife Steel Nerds Patreon supporters!
Kiwami Japan
A popular Youtube channel called “kiwami japan” includes several videos of making knives out of unusual materials such as jello, pasta, chocolate, etc. The video on making a knife out of cardboard has over 20 million views which means that these videos have reached a broader audience than just knife makers or enthusiasts. As a materials engineer I find the videos interesting from a materials perspective, but they are entertaining in other ways as well. The videos are a bit quirky so I decided to take a dive into these videos and try to figure out what is going on. I also e-mailed the person who makes the videos and he answered a few of my questions. I will refer to him as “Kiwami” for the rest of this article though I know that is not his name. Kiwami means extreme in Japanese.
The person making these YouTube videos hides his identity. I think for the true fans there might be some small reflections to use to piece together what he looks like (I don’t have that much time), but in general he either doesn’t show his face, or covers it up. He also rarely speaks and many of the videos include only the sounds of whatever he is working on:
That GIF of him holding a sickle is perhaps more menacing than the channel appears, in general. He often uses subtle humor in the videos such as being expressive with his hands:
Or by adding anthropomorphic elements to materials he is working on:
Or by showing how much time has lapsed in a project which is taking a particularly long time:
His videos started out with making various things with cheap materials, like “I made a karambit knife with a sickle” or “I made a Nunchaku with 4 dollars.” The first knife video was the karambit, and he followed it up with a video of making a butterfly knife. The making of knife videos started to ramp up with “
Manually repair very rusty Japan’s $500 kitchen knife
Super Steels vs Regular Knife Steels
Thanks to Daniel Jackson for becoming a Knife Steel Nerds Patreon supporter!
Super Steel
I see frequent references to “super steel” online, and I was curious about how long that terminology has been around. I did searches on bladeforums as it is one of the oldest knife forums. The number of references to “super steel” has increased over time, but so have the number of posts on bladeforums. I saw how many references to “super steel” there were in each year, and then as a proxy to how many posts there were on bladeforums I did a search for “154” and saw how many references there were each year. Google tops out at 200 results but at that point the dataset was big enough to get an idea:
So referring to steels as “super steel” or the category of “super steels” is at least as old as still-existent knife forums on the internet. Reading through the descriptions of “super steel” now and for as long as bladeforums has existed, they are typically defined as one or more of the following [1][2][3][4]:
- New (relatively)
- Excellent edge retention
- Difficult to sharpen
- Stainless
Not all of those qualities are universally used. Sometimes non-stainless steels such as 3V or Infi have been called “super steels” [4]. I am not sure if a steel must have high edge retention to be called super but in general the “new” steels that come out have high wear resistance and edge retention. Super steels are often described as having high edge retention but greater difficulty in sharpening, however. Different steels slowly lose the title of “super” over time. In the early bladeforums era, VG-10 was sometimes called a “super steel” [5] but I don’t see it called super much anymore [6]. This confirms the “new” part of the definition. I’m not sure why edge retention or wear resistance became synonymous with super rather than other properties like toughness, but this is where we have ended up.
Pre-Internet History
Unfortunately, searching through magazines and books that predate the internet is not as easy as searching through bladeforums. However, I did find one reference from Outdoor Oklahoma 1978, where a very modern sounding description of “super steel” is found:
“Some hunters are a bit reluctant to opt for super steels because these have a reputation for being hard to sharpen. It’s true, good edge holding qualities go hand in glove with hard steels and hard sharpening. Some steels, especially stainless…”
And that’s where my free view through Google Books ends. Reading through the descriptions of super steels on bladeforums I find similar descriptions to this day. In 1978 basically none of the current “super steels” were even in existence, meaning that the article was likely referring to 154CM, 440C, or both as super steels. Those steels are not called “super” any more which again confirms that steels tend to lose their super title over time.
Edit 6/27/2022: I found an even earlier reference to the term “super steel” though the above quotes are still great because they reflect a similar mentality to today. This new reference I found is from the first issue of The American Blade Magazine (now Blade Magazine) from 1973 in an article by John Wootters called “Blades for Game”: “In this day of super-steels, there is no reason why a hunting knife shouldn’t have a hollow-ground blade. Such a grind offers less drag in meat-slicing and is easy to keep razor-sharp. the higher the hollow-grind bevel lies on the blade, however, the less “spine” or strength the blade will have, and the less abuse the knife can be expected to stand. If the steel is not absolutely top quality, however, a flat bevel offers more resistance to edge-chipping.”
The second issue of American Blade Magazine in an interview with Ted Dowell mentions that “He is still field testing the new ‘super stainless,’ 154-CM, and remains unconvinced although he offers it as an option to those who want it.”
Is “Super” a Positive or Negative?
Even in that 1978 article it was stated that some don’t want super steels because of difficulty in sharpening. The sometimes negative connotation of super steels as being nothing more than a “flavor of the month” or being too difficult to sharpen continues to this day. Therefore, it is not clear to me if the term super steel was originally coined as a negative or positive description. Many discussions on bladeforums about super steels continue to be about whether we need the so-called super steels or whether the old classics are good enough or even superior [7][8][9].
Current Views of Super Steels
While many decry the super steels as being unnecessary, the conflation of “high wear resistance” and super, or premium steel, continues. For example, in the Knife Informer article rating knife steels [10], the steels are categorized from “Super Premium” down to “Low End” with the differentiating property being edge retention. See this article for more information on articles that rate and rank steels. Because these steels are viewed as being superior, they are often perceived as also having high toughness despite their high wear resistance. In the linked article on steel ratings I pointed out that M390 is often given high scores for toughness despite Bohler not providing any toughness data on the steel. Toughness testing here at Knife Steel Nerds has also found unspectacular toughness values for M390, though it has only been tested at relatively high hardness:
The reason why almost any steel will have relatively low toughness that is designed for very high wear resistance is the large amount of carbide that is present in the microstructure. You can read more about the effect of carbides on toughness in the article I wrote on microchipping and in t
he summary of edge stability theory
Tests of Knife Edge Toughness
Thanks to Jaethan Reichel for becoming a Knife Steel Nerds Patreon supporter!
In an earlier article I wrote about the microscopic mechanisms by which chipping and micro-chipping occurs in edges. However, that article did not cover specific tests of edge toughness. Correlating conventional toughness tests with edge toughness is difficult for many reasons:
- Fracture toughness tests do not incorporate the effect of crack initiation which is the primary mechanism by which micro-chipping occurs, and is part of the energy required for macro chipping. Unnotched and to some extent c-notch or u-notch impact tests do include crack initiation, however.
- Toughness test specimens are generally much larger than the tip of a cutting edge. Therefore, the statistical occurrence of large carbides or inclusions that act as crack initiation sites is very different [1]. Other small microstructure features such as retained austenite may have different effects in such a small volume.
- The small cross section of a thin edge means it is likely to deform rather than chip even at relatively high hardness (see the article on bending of steel for more information).
- The loss of sharpness through micro-chipping and gross fracture through chipping likely occur through somewhat different mechanisms (see my article on chipping for more information). For micro-chipping it may be useful to perform actual sharpness tests to measure the degradation of the edge.
- Measuring the effect of edge thickness, edge angle, shape, etc. is obviously not possible without a test of actual edges.
Factors that Affect Toughness
Higher hardness, impurities, retained austenite, larger grain size, greater carbide volume, larger carbides, and smaller spacing between carbides all reduce toughness. Thicker edges and more obtuse edges require more energy for deformation or fracture. However, determining the extent of influence of each of these different variables requires edge toughness testing.
Tests of Edges
Tool Steel Simplified
In Tool Steel Simplified they describe a process for impact testing of edges [2]. They used a 3/8″ piece of M2 with a 45° single bevel edge:
They then impacted the edge with a pendulum at different heights. As shown below the knife was held at an angle relative to the impact. At lower impact energies the edge would deform, they increased the energy until the edge would deform sufficiently to fracture. Greater height meant higher impact measured in inch-lbs:
They then produced a plot comparing edge toughness against tempering tempreature for the M2 steel austenitized at 2225°F:
Highest toughness correlated with the lowest hardness, and showed a great decrease in the initial secondary hardening range (800-1200°F) due to increase in hardness and the transformation of retained austenite (see my article on tempering). They compared the knife edge toughness against several other measures of mechanical properties:
The comparisons were made with torsion tests which are simply twisting a bar of steel until it fractures. The strength required to break is the torsional strength. The strength required to yield, or begin plastic deformation, is the elastic limit. Torsion impact toughness is a rapid twist of the steel so that toughness is measured. Poor correlation was found between the edge toughness test and torsional strength, torsional elastic limit, and torsional impact toughness, with a strong correlation with torsional ductility. Torsional ductility is the amount of twist a bar undergoes prior to fracture, and is similar to strain in a tensile test. However, there are other tests that they did not compare with, such as pendulum impact tests or fracture toughness. Fracture toughness, for example, shows a peak in toughness at a similar tempering temperature [3]:
I don’t know if they ever continued the use of this toughness test. It was removed from the 4th edition of Tool Steel Simplified, probably in part because all references to torsion testing were removed from the 4th edition. No other steels, edge angles, etc. are reported in the book.
Agricultural Blades
Impact tests have also been reported for agricultural blades [4][5] which are single bevel as well:
A test was performed with a pendulum impact tester using simple single bevel blades to mimic the cutting edges of the agriculture blades. The test overall is very similar to that described in Tool Steel Simplified. The hammer was dropped from an unspecified height and the degree of deformation was measured. I emailed the primary author of the study for more information such as the energy used of the impact tester, what type of impact hammer was used, orientation of the sample relative to the hammer, etc. but he wouldn’t tell me anything. However, it appears that the amount of deformation and/or chipping of the edges was relatively significant:
When comparing different steels it is apparent that hardness is an important parameter, presumably because higher strength would reduce the amount of deformation that occurs in the edge. However, there are still differences between steels independent of hardness:
Almost all of the steels that were tested are medium carbon steels meaning they are unlikely to have much in terms of carbides while 1.3243 is a high speed steel which does have some primary carbide making fracture initiation easier. Perhaps that is why XAR650 showed superior behavior to 1.3243 despite its lower hardness; the high speed steel fractures more easily due to carbides. However, none of that is confirmed or described in the referenced paper.
The researchers also tested a range of different edge angles where, as expected, a more acute edge led to more deformation during the impact testing.
Edge Stability
Edge stability tests were designed by Roman Landes and reported in different publications but most prominently in his book [6]. The test was conducted by pressing a 2 mm Titanium Nitride coated rod into the knife edge with a load of 1 kg for 10 seconds. The initial tests were performed with 20° single bevel edges:
A series of 10 indentations were performed on each edge and the extent of deformation or chipping was measured. The tests were performed on a range of steels and heat treatments to test the effect of steel properties on “edge stability,” or the resistance to deformation and chipping. In the future I am going to write a couple articles on edge stability so I’m not going to go in depth on the edge stability test in this article. However, I think it was important to mention it because it would be strange not to bring it up.
Future Research
I think the impact tests of edges show promise for quantitatively characterizing edge toughness. I think doing the impacts in combination with a sharpness test would help to characterize the effect of impacts on the micro-scale. Perhaps tests could be separated into different regimes such as sharpness loss, deformation, chipping, etc. Tests would be necessary to determine where those dividing lines might be. With a well designed test it would be possible to explore what steel property and edge geometry parameters are most significant for edge toughness. Some have asked what is the point of toughness testing if the knives are designed simply for cutting and not for any heavy chopping. However, edges also lose sharpness due to micro-chipping or rolling, particularly in thin edges which are desirable for superior cutting ability. The combination of an impact test with a sharpness tester is appealing for measuring the effect of small impacts on sharpness. Also multiple low energy impacts could be used to simulate sustained use. I’m currently exploring my options for purchasing a small impact tester to start on edge toughness tests. With such tests we can determine what correlation there is, if any, between charpy impact tests and edge toughness testing. The number of knifemakers I have convinced to make charpy toughness specimens for me to test is still relatively small but maybe it would be easier to convince them to make simple knives since that is what they are used to producing. To see some of the charpy impact testing we have done you can read the article on Cruwear toughness and
the article on CruForgeV toughness
Toughness testing – Cru-Wear, Z-Wear, Upper vs Lower temper, Cryo vs No Cryo
Thanks to Paul Hart and James Covington for becoming Knife Steel Nerds Patreon supporters!
I recently completed some toughness tests on samples that were heat treated by knifemaker Warren Krywko. The steel was donated by Chuck Bybee of Alpha Knife Supply. The samples are subsize unnotched charpy specimens with dimensions as specified on the bottom of this page: http://knifesteelnerds.com/how-you-can-help/ If we can get more people to make toughness specimens we can have more comparisons between steels, hardness points, heat treatment parameters, etc. Patreon dollars are for the purpose of paying for machining, shipping, testing, etc. for tests like toughness and CATRA edge retention, so if you are able to contribute that way please visit the Knife Steel Nerds Patreon page.
Warren likes to use CruWear and PM Z-wear in his knives because he likes its combination of good toughness and wear resistance along with high hardness. Both are copies of the older VascoWear developed by Vasco, the company that James Gill worked for who I wrote about here: The Development of High Vanadium Steels, M4 and the First Tool Steels Book. James Gill didn’t develop VascoWear but I won’t be writing about the history of VascoWear in this article. Maybe another time. It is also sold by Carpenter as PD-1 and probably other companies under other names.
Warren and I discussed comparing low vs upper tempers (400 vs 1000°F), with and without cryo, and the ingot version vs the powder metallurgy (PM) version of the steel. Crucible shows a significant increase in toughness for the PM version [1][2] so we wanted to see if our toughness testing shows similar behavior. We also saw a big difference between them:
This makes sense because of the great refinement of the carbide structure from the powder metallurgy process. You can see a comparison between micrographs of ingot and PM 154CM in this article:
Micrographs of Niolox, CPM-154, and AEB-L
How Chipping of Edges Happens at a Microscopic Level
Thanks to Warren Krywko, Joseph Cannell, and Timothy Thomas for becoming Knife Steel Nerds Patreon supporters! Your contributions will help fund more research on knife steels.
To discuss chipping we have to start with fracture mechanics of materials, and in this case steel. Chipping itself is just fracture, so by definition resistance to chipping is controlled by toughness. Unfortunately there are many definitions of toughness. I covered one definition of toughness in the article on spider silk, which is the area underneath the stress-strain curve:
There is a brief intro to tensile testing in the article on flexing and bending. In the tensile test a material is pulled until it fractures. The stress (σ) is the load divided by the cross section, and strain is the change in length divided by the original length. Therefore, stress and strain are normalizing terms that are affected by the size of a specimen. A load applied to a small sample results in a larger stress than the same load on a large sample. With the definition of toughness given by the area underneath the stress-strain curve, greater values of toughness are given by stronger materials (high stress) with high ductility (high strain).
Fracture Toughness
An important measure of toughness is the resistance to fracture in the presence of a pre-existing crack. This is important because most parts in service have some level of imperfection to them. For example, a knife edge sharpened to rougher finishes will have larger scratches than a finely sharpened edge, and these scratches can be thought of as pre-existing flaws [1]:
Fracture is measured in terms of K, or “stress-intensity factor” which is the stress at the crack tip. K is proportional to the stress and the square root of the size of the flaw. Therefore a larger flaw means that less stress is required to cause fracture:
Using specimens designed for testing fracture toughness you then get a test that looks something like this where the stress required for crack propagation can be measured by pulling apart the specimen:
Another method for measuring crack propagation is with a three-point or four-point test such as this one where they measure the fracture toughness of watermelon, because science is important:
Fracture toughness is measured in dimensions of MPa*√m (stress*√length). An important reported toughness value is KIC, or plane strain fracture toughness, which is a material property for pulling apart fractures (no twisting or shearing) using specimens that are sufficiently thick that the value is no longer affected by thickness:
Crack Initiation and Fatigue
If the flaw or crack size is sufficiently small then the material will yield (deform) first rather than fail rapidly by fracture. Therefore, a sufficiently large crack must form first before crack propagation can occur. Crack initiation occurs due to different mechanisms than crack propagation. Because fracture toughness testing is performed with pre-cracked specimens the effect of crack initiation is not included.
Charpy Impact Testing
Impact testing is performed by dropping a heavy weight on a pendulum through a sample and measuring the degree that the sample resisted fracturing. If there was no sample present the weight would reach the same height on the other side. The more resistant the sample is to fracture the lower the height is that the weight reaches on the other side. Charpy impact samples are often “notched” so that fracture occurs in a consistent region and has lower variability [2]:
Fracture Toughness vs Charpy Toughness
The two major differences between charpy toughness and fracture toughness are 1) charpy testing is dynamic (falling pendulum breaks sample) rather than static (specimens slowly pulled apart) and 2) fracture toughness is performed on pre-cracked samples while charpy samples are typically not pre-cracked. Charpy toughness testing cannot generate a “material property” in the same way that fracture toughness can because the values are always affected by sample size. Therefore, “serious” fracture analysis is typically performed with fracture toughness testing. Fracture toughness is usually more expensive than charpy impact testing, however, and even with these differences charpy v-notch testing and fracture toughness results are often highly correlated [3][4]:
One of the reasons that charpy v-notch and fracture toughness are highly correlated is because the presence of the notch in the charpy specimen greatly reduces the energy required for crack initiation. However, in the case of unnotched samples the situation is somewhat different because there is no stress riser for lowering the energy required for crack initiation. This makes unnotched testing somewhat more variable because the crack can initiate anywhere over a certain distance because there is no notch. However, this also means that the effect of crack initiation on fracture is better captured by unnotched specimens, or gentle notches like c-notch or u-notch. Because unnotched or c-notch charpy testing incorporates both crack initiation and propagation, it correlates much more strongly with the area under a tensile curve [5]:
Fatigue Initiated Fracture
Chipping and micro-chipping does not have to occur with a single impact, but can happen by the material being stressed multiple times:
The damage to a material due to repeated loading is called fatigue. A three-point bend fatigue test can be seen in the following video:
The higher the stress applied, the smaller the number of cycles a sample can withstand prior to fracture. With sufficiently low applied stress the material will never fracture (at least with martensitic and ferritic steel), and this stress level is known as the “fatigue limit” or “endurance limit” [6]:
Below a certain number of cycles, the material reaches a region of “low-cycle” fatigue (high stress) rather than “high-cycle” fatigue (low stress), usually the cutoff point is given as 1000 or 10000 cycles, so failures in knives are almost always in the low-cycle fatigue regime. Typically low-cycle fatigue is at sufficiently high stress levels where small amounts of yielding (permanent deformation) occurs. Micro-chipping and chipping of edges occurs in the low-cycle fatigue region [7]. Low-cycle fatigue is more greatly controlled by toughness and less by the slow crack growth dominated by high-cycle fatigue [8]. This can be understood through analysis of the effect of ΔK on crack growth [7]:
This plot shows the crack growth rate on the y-axis and ΔK on the x-axis. ΔK is the difference in stress intensity factor, K, during loading and unloading (leading to fatigue), ΔK = Kmax – Kmin. So if a sample is loaded to a certain stress and then unloaded it would be ΔK = Kmax – 0 = Kmax. As described earlier, higher stress would mean a higher K value and a larger crack would mean a higher K value. At sufficiently small stress there is no growth of cracks, but above ΔKTH the cracks grow. Region I is slow crack growth, but the rate of growth increases with higher ΔK. Region II shows the region with moderate crack growth rates that are consistent over a relatively wide range of ΔK values. Region III is characterized by more rapid crack growth, and with sufficiently high ΔK then the piece breaks in one cycle and is therefore equal to or greater than the fracture toughness, KC. With sufficiently high stress applied a sample can jump into Region III or even to KC, but K is also controlled by crack size. Therefore, even with a fixed stress applied the ΔK can grow because the crack is growing, meaning that a sample can move through all three regions shown in the plot. Region III can describe low-cycle fatigue as it has a high stress-intensity factor on a crack tip. The higher the stress the closer the behavior is to fracture toughness, KC. Therefore because chipping and micro-chipping is controlled by low-cycle fatigue, we would expect toughness to have a strong effect on chipping of knife edges.
Fatiguing of materials in practice
While testing in the laboratory usually uses a constant stress or strain, use of knives is by humans, and humans do not apply constant stresses, so the pattern might look more like this:
In this case the damage to the material is cumulative; the application of a low number of large stresses reduces the number of smaller stress cycles that the material can withstand.
Crack Initiation
While crack initiation can certainly occur at previous flaws, in the absence of sufficient flaws they occur at microstructural features instead. With steels used in knives there is usually some amount of hard, brittle particles called carbides that contribute to wear resistance. Often large carbides are present where cracks form, either at the interface between steel and carbide or cracking of the carbides themselves [9]:
Larger carbides require less applied stress to crack than smaller carbides. In a comparison between conventionally cast 1.2379 (D2) steel and a powder metallurgy steel HWS it was found that the stress required for crack initiation was vastly different (note difference in magnification) [7]:
The stress required for cracking of carbides in the D2 was found to be only 700-900 MPa, while steels with smaller carbides required higher stress levels to crack the carbides. The cracking of these carbides is particularly bad because the cracks can easily grow along segregated carbide bands [7]:
However, cracking of carbides was not seen in the HWS powder metallurgy steel, but instead the cracks initiated at impurities because they were larger than the carbides, and therefore the stress required for cracking at impurities was lower than for carbides. The stress required for crack initiation was still much higher than that required for D2, however, on the order of 3400-3600 MPa [7]:
Crack Initiation vs Fracture Toughness
One interesting thing found in the study comparing D2 and HWS was that the fracture toughness of HWS was actually lower than D2. HWS had fracture toughness of 20 MPa*√m while for D2 it ranged from 22-28 MPa*√m depending on orientation. The researchers reported that because the carbides are so small and evenly distributed in the powder metallurgy steel that they are, on average, closer together than for D2. Therefore a pre-existing crack can more easily grow throughout the steel by jumping from carbide to carbide [9]:
In testing of shearing dies operated by trimming of high strength sheet steel, it was found that the HWS PM steel did not have superior fracture resistance to D2 steel when in the presence of large initiated cracks, and they attributed this to the ease in crack propagation of the PM steel despite its resistance to crack initiation. Another conventionally cast steel with a lower carbide volume, however, showed superior resistance to fracture due to its higher fracture toughness from a small carbide volume combined with a relatively large average distance between carbides. In the presence of small cracks which lead to micro-chipping, superior behavior was found in the PM steel. Therefore, the type of loading and use of a knife edge is important in determining which steel will show superior behavior. They concluded that the PM steel was better for short cracks and high stresses while the D2 was better at resisting crack propagation of large cracks. Similar results were found in a comparison of four different PM steels with similar carbide volume but different carbide size distributions. The PM steel with the largest carbides and highest hardness was found to have the best fracture toughness because the distance between carbides was greatest [10]:
Surface wear and machining grooves
In the study comparing HWS and D2 in shearing dies it was found that both wear and fatigue occurred in the steel. They found that the wear of the dies led to lower stress required for crack initiation which is part of what increased the chance for fracture in the PM steel they tested. It was also discovered that grooves left over from machining led to much lower stresses required for crack initiation. They recommended that PM steels be used either when smooth surfaces can be guaranteed or with coatings applied to limit crack initiation. This was confirmed in another study on Vanadis 6 where a milled surface was found to lead to fracture without any plastic deformation [11]. In the case of poor finish quality the superior resistance to crack initiation doesn’t matter because the fracture toughness of PM and conventional steels is similar.
Effect of carbide volume on low-cycle fatigue
In a study comparing D2 (SKD11) with a modified steel with lower carbon and chromium for lower carbide volume and smaller carbides (M-SKD11), they found that with low-cycle fatigue the stresses allowed were much higher with the lower carbide volume steel:
The D2 steel had 13.5% carbide volume, and a maximum carbide diameter of 21.5 microns, while the modified steel had only 4.5% carbide volume with a maximum carbide diameter of 14.4 microns. There is also a comparison point with 4340 steel, which would be expected to have significantly lower carbide volume than either, but because of its lower strength (correlated with hardness) it cannot withstand the same degree of stress. At high stress levels the researchers found that cracks initiated at the surface with large carbides at the origin of the fracture. The stress level required for surface initiation in D2 was 1100 MPa, but for the modified steel was 1800 MPa. This is why when loaded to a similar stress level the modified steel with less carbide and smaller carbides lasts for 1-2 orders of magnitude more cycles than D2. The difference is greatest for a small number of cycles and high stresses which is the region of interest for knives when it comes to chipping resistance.
Effect of carbide volume and Hardness on Fracture Toughness and Impact Toughness
Higher carbide volume has also been found to lead to lower fracture toughness and lower impact toughness as measured by charpy or izod. Therefore a lower carbide volume improves resistance to both crack initiation and crack propagation. Higher hardness also reduces both fracture and impact toughness, and the effect of carbide volume becomes smaller at high hardness [12][13]:
Effect of Grain Size
I wrote an article about the effect of grain size on toughness: How Does Grain Refinement Lead to Improved Properties?
To summarize that article, smaller grain size means that a growing crack must re-initiate and change direction as it meets grain boundaries. Therefore a smaller grain size means higher toughness:
I also showed some examples of toughness numbers with different grain size in articles on austenitizing, particularly Part 2 and Part 3. Here is a chart I shared in Part 2:
Effect of Retained Austenite
It is very common for custom knifemakers to use cryogenic processing of their steels which eliminates most of the retained austenite in the steel. I wrote a forum post about what retained austenite is and what cryo treatments do. Toughness testing usually shows an increase in toughness with greater retained austenite, as can be seen in this figure for 440C [5]:
However the concern sometimes is that through stressing of the steel with retained austenite present that the austenite will transform to (untempered) martensite and therefore the brittle phase will act to reduce toughness. However, fatigue testing of material paints a different picture, where the retained austenite containing steel, which converts to martensite during cyclic loading, has higher resistance to fracture [5]:
From that figure it can be seen that the steel with retained austenite had both a higher ΔKTH (stress intensity required for crack growth), and at high ΔK approaching KC (low-cycle fatigue and fracture toughness) the resistance to fatigue is better. The behavior in the steady crack growth region is similar between the two. During crack growth there is stress on the surrounding austenite and it transforms to martensite. Martensite is a larger phase (less dense) than austenite and therefore during the transformation the expansion from the transformation leads to “plasticity-induced crack closure.” A stress applied to the crack closes it and therefore the stress required for crack propagation is greater. A micrograph showing a crack growing within a ferrite (red)-austenite (green) microstructure is shown below, where transformation of the austenite to martensite can be seen around the crack (green austenite becomes black martensite) [14]:
Importance of Orientation
Steel is typically rolled to the final thickness, and carbides and impurities are oriented along the rolling direction. This leads to segregation of carbides in “stringers” that can be preferential bands for crack formation [15]:
“RD” is the rolling direction, “TD” is the transverse direction (perpendicular to the rolling direction), and ThD is the thickness direction. With the aligned segregated carbides, it matters which orientation a toughness test is performed:
With the transverse toughness specimen the notch and the breaking direction are aligned with the segregated carbide bands. Therefore the crack can grow along those bands of carbides and the toughness with transverse specimens is lower when compared with longitudinal specimens. Powder metallurgy helps to reduce carbide segregation and therefore improve transverse toughness but has a smaller effect on longitudinal toughness [16]:
The direction of loading of edges has been found to have an effect on their fracture. Generally knives are oriented along the rolling direction so that the carbide bands are parallel to the edge:
Therefore side loading requires less stress to fracture the edges and chipping can occur along the carbide bands [15]:
If the knives are oriented along the transverse direction instead, then the tips of the blades would fail more easily in a similar fashion:
Micro-Chipping
Micro-chipping of edges does not require the stress to reach the level of catastrophic growth of cracks, but can occur through the initiation and growth of small cracks that link up and lead to fracture. Micrographs illustrating this can be seen here [7]:
This is exacerbated by the segregated carbide bands that are aligned along the rolling direction, which is what helps drive the crack propagation and linking parallel to the edge. While powder metallurgy steels have a lower degree of banding and segregated carbides, the short propagation of cracks can still occur, linking up the cracks to form a micro-chip [7]:
Micro-chipping is on a “micro” scale meaning that the chips are often not visible to the naked eye but require magnification. Therefore, micro-chipping is usually perceived as a loss of sharpness rather than obviously visible chipping. The short growth and linking up of carbides occurs through crack initiation at carbides and then growth through low-cycle fatigue.
Chipping
Chipping, as differentiated from micro-chipping, is on a more macroscopic scale and requires high stresses that exceed the fracture toughness, KC, of the steel to allow rapid propagation of large cracks. Therefore those types of chips occur either in a single high application of stress or a very small number of cyclic stresses. Avoiding these types of chips requires either a change in use, higher toughness by changing material or heat treatment, or change in edge geometry [17]:
Conclusions and Micro-chipped Knife Edge
In 2007 Sandvik metallurgists sent me micrographs of several steels along with images of various damaged edges to demonstrate failure mechanisms of knives. They have low-res versions on their site [18]. I think it’s a good idea to analyze these to see if the mechanisms discussed in the rest of this article line up with our observations:
First of all I think it’s clear that there are scratches in the edge from sharpening which as described earlier can act as preferential sites for crack initiation. In the second image as marked with arrows you can see the cracks that formed parallel to the edge as shown in this article. And the chips themselves are longer than they are tall perhaps indicating that the cracks formed by cracks linking up along the rolling direction. Therefore I would say that mechanisms of micro-chipping and chipping are all aligning along similar lines from multiple sources. Chipping is controlled by steel toughness though there are different mechanisms in play. Micro-chipping is more controlled by low-cycle fatigue and linking up of cracks that form along carbides. Chipping is on a more macro scale and the stress must reach the fracture toughness of the steel. Larger carbides, a greater volume of carbides, and segregation of carbides all increase susceptibility to chipping. Hardness, grain size, and retained austenite all affect toughness and therefore the crack formation, propagation, and chipping and micro-chipping in addition to the carbides.
[1] https://scienceofsharp.wordpress.com/2015/10/30/burr-removal-part-1/
[2] https://www.twi-global.com/technical-knowledge/faqs/faq-what-is-charpy-testing/
[3] Novotny, Paul M. “Toughness index for alloy comparisons.” Advanced materials & processes
You should make a Facebook page
Just out of curiosity, did you ever test ztuff with a high temper. At one time I did some testing, and found it to be much tougher if tempered at around 1000 degrees. This was compared to tempering at 400 degrees!
I haven’t tested the high temper on Z-Tuff.
I think it would score quite a bit higher on you’re charts! Maybe in the future I could heat treat a few samples and send them to you. I’m interested personally!