Is 1084-15N20 The Best? Pattern-Welded Damascus

Patreon

Several new experiments were done to support this article. I am able to do these experiments thanks to the contributions of Patreon supporters. If you would like to support knife steel research visit Patreon.com/KnifeSteelNerds

YouTube Video

Here is the video version of the following information:

Stainless and San-Mai

This article got too long so I will be discussing stainless steels, high alloy steels, and San-Mai laminates in a future article.

Historical Steel Combinations

I have a book on the history of steel in knives in “modern” times called The Story of Knife Steel: Innovators Behind Modern Damascus and Super Steels. I have a very short summary below of some of the content in that book but if you want to learn more about steel combinations and patterning in Damascus, you should read the book.

Early Bill Moran pattern-welded Damascus used O1 and mild steel. Moran and other 1970s and 1980s-era Damascus smiths believed that a large carbon difference would lead to the difference in etching behavior to provide a contrast. Dr. John Verhoeven and bladesmith Howard Clark in 1998 did a study on pattern-welded Damascus [1] where they found that carbon would diffuse evenly between two steels and thus have the same carbon content throughout. Instead they found that other alloying elements led to contrast after etching. Carbon is a very small element which is “interstitial,” meaning it sits between iron atoms. Larger elements like manganese, chromium, and nickel are “substitutional” atoms that replace iron atoms and thus diffusion is much slower.

Before the study by Verhoeven and Clark, Damascus smiths had already moved on to other combinations. Instead of mild steel a high nickel, low carbon steel A203E became popular for the “bright” layer in Damascus. This recommendation was made in 1977 [2] by the “Damascus Steel Research Team” of Daryl Meier, Jim Wallace, and Robert Griffith. They also recommended pure nickel metal as a bright layer. Anciently, nickel steel was used in some blades from metallic meteorites, which typically have 5-10% nickel. Early “dark” layer choices in the 1970s and 1980s were typically simple high carbon steels such as W1, W2, and 1095. Along with low alloy high carbon tool steels like O1 and O2.

A203E has only 0.1% carbon and there was an increasing desire by bladesmiths to use two high carbon steels for better hardness and wear resistance. Tim Zowada began using O2 and L6 high carbon steels from Carpenter around 1990-1991. It became increasingly common to use bandsaw blades which were typically made with high nickel steels such as 15N20. However, it was more common at that time to call this material “L6,” despite the fact that L6 has never been used in bandsaw blades as far as I can find. L6 is somewhat similar because it has 1.5% Ni but it is different than bandsaw blade steels.

Popularization of 1084 and 15N20

The steel combination of 1084 and 15N20 has become ubiquitous, though it did not become popular until the mid-to-late 1990s. For example, the Jim Hrisoulas book The Pattern Welded Blade from 1994 does not even mention 15N20, and he mostly wrote in that book about combining high carbon and low carbon steels. The 1084/15N20 combination was popularized in part by the “Montana Mafia,” made up of Damascus smiths such as Rick Dunkerley and Shane Taylor. Dunkerley reported that he began using this mix in 1995 from a suggestion from Devin Thomas [3]. I asked my father Devin why he suggested these steels and he said that a lot of bladesmiths would ask him for help with forge welding and etching issues but a lot of people were using difficult steel combinations. He felt that 1084 and 15N20 were easy to forge weld and would give a good contrast.

15N20 has 0.75% carbon along with 2% nickel, so like other nickel alloyed steels it can act as the “bright” layer in Damascus. 1084 has only slightly higher carbon at 0.84% and has elevated manganese when compared with 1095 (0.75 vs 0.4% Mn). 1075 has the same 0.75% carbon of 15N20 so I also asked my father why it was 1084 and not 1075. He said that 1075 was typically available as wide sheet which would require shearing down to narrow pieces for the billet. 1084 was available as narrower “bar” which could be bandsaw cut for the stack with 15N20. For some knifemakers it was appealing that 1084 had more carbon than 1075. Regardless, 1075, 1080, and 1084 are roughly equivalent.

Rick Dunkerley folder in 1084 and 15N20

Etching Response

Notes on Comparisons

Comparing etching response of different combinations can be somewhat tricky because of a few factors:

1. Different etchants may lead to different levels of contrast
2. Etching and polishing technique affect contrast
3. Photography has a big effect on contrast. You may have experienced times when you saw a Damascus knife in an image with very clear contrast but once in person you see it is much more subtle.

So despite those difficulties I am going to make some generalizations about level of contrast with different combinations. If you disagree with my assessment feel free to say so in the comments.

Nickel

It has long been understood that the nickel content in steels like 15N20 allows them to resist etching and thus provide the “bright” layer in Damascus. However, when you etch 15N20 on its own it etches dark just fine so I think it may be some kind of anode/cathode effect at work where the 1084 is etched preferentially rather than the 15N20 resisting etching on its own. Higher nickel seems to lead to greater contrast than lower nickel, so 15N20 with its 2% nickel does pretty well. There are some rarer steels with 4% nickel, such as 4800 steel powder or Bohler K600 (1.2767). L6 has somewhat lower nickel at 1.5% and 8670 even lower with ~0.9%. Unfortunately there aren’t any low alloy nickel steels with significant vanadium or tungsten additions so wear resistance typically needs to come from the other steel used in the combination.

Nickel knife steels. Note: I found conflicting information on 4600 and 4800 compositions so I left the Mn blank. Some list there being 0.5% Mo. The most common versions have little or no carbon, and the “KC” versions have carbon added to them, usually graphite. I have also heard that some distributors take 1080 powder and add nickel powder.

Manganese (and Carbon)

Higher manganese leads to darker etching. For example, in the Verhoeven/Clark study they found that with 1018 (0.75% Mn) and 1086 steel (0.4% Mn) the 1018 was the “dark” layer despite its lower carbon content. This demonstrates that the old view that carbon differences led to etching contrast was incorrect.

Image from [1]. Notice the higher carbon steel is the “light” layer.

The higher Mn in 1084 relative to 1095 and 1075 leads to it being somewhat darker when in combination with 15N20. Some steels with even higher Mn etch even darker such as O1 (1.2%) and O2 (1.6%). The European version of O2 called 1.2842 has 2% Mn and etches darker than perhaps any other common steel.

Chromium

Chromium is a bit more mysterious as to its effects. In the Verhoeven/Clark study they found that 52100/L6 had poor contrast. 52100 has high chromium (1.5%) but also low manganese (0.35%). I have seen knives here or there with 52100 and 15N20 and the contrast can be “ok” in some of them, such as this feather pattern bowie by Aaron Wilburn:

Aaron Wilburn feather Damascus in 52100 and 15N20

1095 steel has similar carbon content to 52100 and slightly more Mn (0.45%) but without chromium. It typically etches darker than 52100 when in combination with 15N20, as seen in this knife:

Blake Nichols knife with Greg Shahan Damascus in 1095 and 15N20

O1 still etches quite dark despite having 0.5% Cr and so does 1.2842 despite having 0.35% Cr. 80CrV2 is also generally considered sufficiently dark in combination with 15N20 even though it has 0.5% Cr and only 0.4% Mn. 5160 has relatively high Mn (~0.85%) and Cr (0.8%) and it can also etch relatively dark despite the Cr addition.

So I think overall it appears that low amounts of Cr do not have a large effect on etching response, but higher amounts like the 1.5% in 52100 may affect how dark steel etches.

Vanadium, Tungsten, Molybdenum

As far as I can tell these elements do not effect etching. However, high vanadium and/or tungsten additions are good for higher wear resistance and edge retention in Damascus so I have included the composition of several of them below:

“3-color Damascus”

Sometimes steel combinations can be used that give multiple shades of grey/white. A common example would be a dark etching steel in combination with 15N20 and pure nickel. While 15N20 does resist etching it doesn’t do so nearly to the extent of pure nickel so three different shades are achieved.

Sometimes an intermediate shade steel can be found between 15N20 and a dark etching steel. One example I found was 1.2842/80CrV2/15N20 Damascus [4]. There can also be some contrast between 2% Ni 4600 and 4% Ni 4800 as seen comparing the eyes (4800) and beak (4600) in the roadrunner by Cliff Parker below:

Cliff Parker Damascus roadrunner. Bright lines are nickel sheet, eyes are 4800, beak is 4600, and the rest is 1084/1018 mixed.

Forge Welding Temperatures – Crumbling, Melting

It is quite common for Damascus smiths to use very high forge welding temperatures such as 2300, 2350°F (1260-1290°C) or sometimes even hotter. This typically works for steels like 1084 and 15N20. However, with higher carbon the melting temperature of the steel is reduced. The grain boundaries melt first. After the grain boundaries melt the steel will crumble during forging. This is especially important to keep in mind when steels have higher than 1.2% carbon like 26C3, ApexUltra, 1.2562, Blue Super, White #1, etc. However, even 1% carbon steels like W1 and 52100 can cause problems for Damascus smiths that are used to pushing the temperature as much as possible with lower carbon mixes like 1084 and 15N20. With these very high carbon steels, a reduced temperature is likely necessary such as 2150-2200°F (1175-1205°C). A longer soak time can make up for reduced forge welding temperatures.

Ease in Forging

I have a separate article on difficulty in forging of different steels. I have a loose ranking of some typical steels in terms of difficulty in forging in that article. Typically, higher carbon and higher alloy steels are more difficult to forge. One reason is because those steels are more likely to have undissolved carbides at the forging temperature. Those carbides don’t deform as well as steel which increases the difficulty.

Steels below the eutectoid (~0.77% carbon) are very easy to dissolve the carbides. The carbides would likely be dissolved well below the forge welding temperature. Steels with very high carbon like 26C3 (1.25% C) may still have some carbide near the forging temperature. 52100, even though it only has 1% carbon, also has 1.5% chromium which raises the temperature at which carbides dissolve. However, even 26C3 and 52100 will have very little carbide when compared to something like D2, and the remaining carbides would be much smaller. Some low alloy steels have vanadium and tungsten additions for greater wear resistance like CruForgeV and ApexUltra. Those vanadium and tungsten carbides are still present at forging temperatures and can affect forgeability.

How Well Do They “Stick”?

When forge welding in an oxygen atmosphere, oxides form on the surface which prevents diffusion bonding. Flux dissolves the oxides, allowing the two pieces of steel to bond. Some alloying elements will change the type of oxide that will form. The most common example is chromium in stainless steels, which famously cannot be forge welded with the conventional flux method. The chromium oxides are not dissolved by flux and therefore forge welding is not typically possible. Manganese and nickel do not seem to have a large effect which is why 1084 and 15N20 are easily forge welded using flux. The higher the chromium content, the more difficult the steel is to be forge welded. However, if the steel is in an oxygen-free environment then difficult steels can still be forge welded.

Cracking After Forging

High hardenability steels can partially harden after forging and this can lead to cracking, sometimes several hours later. Oil hardening steels like O1, O2, and L6 are more sensitive to this behavior. It helps to slow cool from below a certain temperature (like 1400°F/760°C or so). Slow cooling from very high temperatures can lead to carbides forming on the grain boundaries that are difficult to eliminate. Low alloy steels can be normalized to dissolve the carbides though air hardening steels cannot be normalized. After cooling to room temperature (whether slowly on purpose or unintentionally quickly), the steel can by cycled or annealed to remove the stresses in the material so that cracking doesn’t happen.

Austenitizing

Hardness

Semi-frequently I get questions from knife enthusiasts (and even some Damascus smiths) about combining non-typical steels together. Sometimes bizarre combinations like MagnaCut and 15N20. Often times the reason these steels are not compatible is because of the austenitizing (hardening) temperature range. 15N20 is optimally austenitized from about 1400-1475°F (760-800°C), while MagnaCut is austenitized from 1950-2200°F (1065-1205°C). As you can see, these two ranges do not overlap. If the MagnaCut was austenitized from 1475°F it would be very soft (and corrosion resistance would be lower than 1950+ as well). If the 15N20 was austenitized at 1950°F or higher it would have very poor toughness from grain growth. Some steels can be pushed a bit higher or lower than typically recommended in a datasheet but this is obviously not an example of where that can happen.

MagnaCut and 15N20 is of course an extreme example but austenitizing ranges can also be relevant even with low alloy non-stainless steels. For example, Mn and Ni push down the temperature at which steel transforms to austenite, and Cr increases the temperature where steel transforms to austenite. You can see this in the chart below where I austenitized steel in 25°F increments and quenched in oil (from a normalized microstructure).

You can see that 15N20 with the 2% Ni addition is fully austenitized (from a normalized, pearlite microstructure) from 1325°F (720°C) while 1084 was still dead soft after quenching from that temperature. The 1084 instead showed full hardness at 1350°F (732°C), only 25°F hotter. 80CrV2 has 0.5% Cr, however, which increases the temperature at which carbide dissolves, and full as-quenched hardness was not achieved until about 1400°F (760°C). This is from a normalized microstructure, not an annealed microstructure. The annealing procedure can have a large effect on heat treating behavior. Below is a chart of 80CrV2 after austenitizing from different temperatures:

As you can see, there is a significant difference between a pearlitic starting microstructure (normalized) and annealed. And in this case even a sizeable difference in the annealed microstructure of the two different suppliers. The Cr addition makes the starting microstructure have a particularly large effect on the heat treating response. This is why Cr alloyed steels often have a somewhat higher recommended austenitizing temperature than Cr-free steels such as 1500-1550°F (815-845°C).

15N20, only alloyed with Mn and Ni, shows a much smaller effect of starting microstructure on heat treating response:

The “Fast DET” anneal was my own annealing procedure with a faster cooling rate. The “Steel company anneal” is the steel as it was received from the manufacturer. You can see that with the steel as-received the peak hardness is not achieved until about 1400°F which is ~75°F higher than a normalized microstructure. But 1400°F is a much lower temperature than was necessary for factory annealed 80CrV2. This helps explain why Cr-free steels are usually austenitized from a somewhat lower temperature such as 1475°F (800°C).

Toughness – Grain Growth

However, there is also an upper end of acceptable temperature, which is where grain growth occurs. For chromium-alloyed steel, toughness typically peaks around 1500-1550°F (815-845°C) and drops from there. This was found with 5160, 52100, and 8670, for example. Carbides can “pin” grain boundaries and prevent grain growth. Because chromium stabilizes the carbides to higher temperatures, grain growth does not occur until somewhat higher temperatures than steels alloyed only with Mn and Ni. Vanadium can extend this temperature range further, as the vanadium carbides do not dissolve until higher temperatures. This is why small vanadium additions (~0.2%) are made to certain steels like W2 and 80CrV2. In my tests of 80CrV2, toughness had not dropped even with an austenitizing temperature of 1580°F (860°C), though I did not try any higher. But 15N20 does not have any chromium or vanadium so we did a new series of tests with 15N20. We also added two other steels, L6, a 1.5% Ni steel with a chromium addition; and 1.2519, a steel with chromium, tungsten, and vanadium (all three of those elements can help prevent grain growth). Each of them were heat treated from the annealed condition as-received from the steel manufacturer.

In terms of hardness, the steels behave in ways I have been describing. The L6 and 1.2519 increased in hardness with higher austenitizing temperature. This is because the steels were alloyed with chromium, delaying the dissolution of carbides. 15N20, lacking chromium, was flat in hardness between 1400 and 1600°F.

The L6 peaked in toughness at 1500°F which is similar to other chromium-alloyed steels. The toughness actually increased between 1400 (760°C) and 1500°F (815°C), which is because we are dissolving some carbide (carbides are brittle),  but not getting too much carbon in solution or grain growth. So the optimal austenitizing temperature for L6 would be around 1500°F, as that gives us a good combination of hardness and toughness. 1.2519 had flat toughness between 1400 and 1500°F though hardness also increased significantly over that range. Therefore, the hardness-toughness balance peaked at 1500°F even though the toughness did not change between 1400 and 1500°F. Toughness dropped slightly at 1550-1600°F (845-870°C), which is either due to grain growth or perhaps too much carbon in solution (I will explain that next).

With 15N20, the toughness dropped significantly even with 1500°F. In an earlier test with 15N20 I also found good toughness from 1475°F (800°C). So grain growth and a drop in toughness was found even with 1500°F, the same temperature that gave the peak properties for L6! I was surprised that toughness dropped off so rapidly for 15N20 at 1500°F, which I did not think sufficiently high to cause significant problems, even in a steel with little carbide like 15N20. This may have affected the toughness results of two 15N20 combinations I tested in my prior study, one with 1.2419 and the other with CruForgeV. 1084 is also very sensitive to overheating, grain growth, and a reduction in toughness. So despite how easy it is for 1084 and 15N20 to be hardened, I recommend using a furnace to avoid grain growth and a toughness drop.

80CrV2 and 15N20

So 15N20 needs to be austenitized at 1475°F (800°C) or below to avoid poor toughness, but 80CrV2 is more optimal from 1500-1525°F (815-830°C). Perhaps L6 is the more optimal steel to combine with 80CrV2, and 15N20 is more optimal to be used with 1084. However, the 80CrV2 can be processed to be more compatible with the 15N20. If starting from a normalized microstructure, the steel can be austenitized from a lower temperature to match the 15N20. Normalized steel is somewhat harder, and therefore more difficult to drill and bandsaw cut. With normalized 80CrV2 I found it to be about 25 Rc. This value can vary some based on the cooling rate (greatly affected by the size of the piece). If another grain refining step was added by heating to 1400°F (760°C) and air cooling, the hardness is dropped to about 19 Rc without significantly affecting the hardening response. Generally I prefer to use a “Fast DET” anneal before austenitizing instead, but this is one of the exceptions to that general recommendation.

Toughness – Carbon in Solution and Plate Martensite

Grain growth is not the only source of reduced toughness with higher austenitizing temperatures. The higher you austenitize, the more carbide is dissolved, putting more carbon “in solution,” which is why hardness is increased with temperature. However, as carbon in solution increases past 0.6% you get more “plate martensite” which is brittle and reduces toughness. You can read more about plate martensite in this article. With a steel like 15N20, the carbon content is only 0.75%, so this is not as much of an issue. However, when steels have over 0.85% carbon, they are more in danger of low toughness from plate martensite. Below is the example of O1 steel, austenitized from 1425 (775°C), 1475 (800°C), and 1550°F (845°C). Despite 1475°F not being hot enough for grain growth in O1, there was a significant reduction in toughness by austenitizing at 1475 instead of 1425°F. This was further reduced by austenitizing at 1550°F. So this is yet another reason to pay attention to optimal austenitizing ranges for different steels before combining them in pattern-welded Damascus.

Normalizing and Annealing

I know it seems slightly out of order to talk normalizing and annealing after austenitizing (to quench), but we are typically austenitizing the steel to perform these two steps as well. Compensating for normalizing is not too difficult with various Damascus combinations. It is generally best to take the higher required normalizing temperature of the two. For example, if you had combined 26C3, which needs something like 1700°F (925°C) to normalize, and combined it with 15N20 which only needs 1550°F (845°C) or so, you would normalize at 1700°F. The temperature needs to be sufficiently high to dissolve all the carbides, and the 1550°F for 15N20 would not be sufficiently high for the 26C3.

For annealing, similar rules apply. You generally want to use the higher required annealing temperature to ensure that both are annealed. If too much carbide is dissolved in one of the steels, that can mean you get pearlite at the end instead of spheroidized carbide and thus a little higher annealed hardness. But that is usually an acceptable tradeoff. As I mentioned in the 80CrV2/15N20 section, for some combinations a pearlitic, normalized structure may make the steels more compatible in heat treatment. In this case you would normalize like normal at the required temperature for 80CrV2 (1550-1650°F). Then heat to the annealing temperature for 80CrV2 (1400°F), but instead of slow cooling do air cooling. This is a grain refinement cycle instead of annealing. Normally grain refinement cycles aren’t necessary when annealing because the low temperature anneal also refines the grain. Generally I prefer annealed steel outside of specific cases such as mixing chromium and non-chromium steels. Or steel that will be heat treated in an uncontrolled forge.

I have a separate article on normalizing and annealing with recommended procedures and temperatures.

Quenching and Hardenability

Steels also have different levels of “hardenability,” or how fast they need to be quenched to be fully hardened. For example, there are “water hardening,” “oil hardening,” and “air hardening” steels. Even within low alloy non-stainless steels there can be significant differences. For example, L6 is a high hardenability oil hardening steel, while W2 is a very low hardenability water hardening steel. The two can be used together, but there is a significant gap between the two steels for hardenability. This means that to fully harden the W2/L6 Damascus you need to quench very rapidly for the W2 to ensure it fully hardens. If a slow oil was used you would likely end up with hardened L6 but soft W2. Even 1084 and 15N20 have a difference in hardenability because the 2% Ni addition in the 15N20 means its hardenability is higher despite the lower Mn. I have a loose ranking of hardenability of different steels in this article.

Tempering

Tempering leads to darker etching than untempered steel [1]. Some people using Damascus for decorative purposes (thus not needing high hardness) will skip the heat treatment but this leads to poorer contrast. Martensite etches better than ferrite, and tempered martensite etches even better.

In terms of compatibility, tempering is usually not a big deal. Most steels can be tempered at ~400°F and achieve good properties. Some steels have a slightly different range than others. 15N20 and 8670 needed to be tempered at least at 350°F for achieving high toughness and 5160 needed 375°F. The steels were not necessarily brittle below those temperatures, but did not have the high toughness the steels are known for.

You can also see that the toughness of 15N20 dropped when tempering at 450°F, which is called “Tempered Martensite Embrittlement.” The temperature at which TME is first seen changes some by steel, and some do not see the phenomena until 500°F or even higher. However, 400°F is safe for any steel I have so far tested.

Toughness and Edge Retention

I already have a whole article on the testing of toughness and edge retention of pattern-welded Damascus so I won’t rehash all of that here. But there are a few key points to mention here when it comes to compatibility:

1. Combining a low toughness and high toughness steel generally seems to lead to toughness similar to the low toughness steel. It fractures at the “weakest link” and there isn’t much benefit to adding a high toughness steel in terms of overall toughness.
2. High wear resistance steels with high slicing edge retention generally still cut very well even when combined with a steel with lower wear resistance. So combining CruForgeV and 15N20 or ApexUltra and L6 led to edge retention similar to CruForgeV or ApexUltra. However, this was due in part to the improvement in edge retention seen in ladder patterning.
3. While ladder patterning benefited edge retention it also slightly reduced toughness.
4. A “Damascus Cutting Effect” was seen in a 1095/Nickel Damascus. Most steel combinations do not have hard and soft layers but using nickel does provide this potential benefit.

When not attempting to have the “Damascus Cutting Effect” it can be beneficial to have two steels with good wear resistance. For low alloy non-stainless steels it makes sense to look at steels with significant vanadium and/or tungsten additions like ApexUltra, CruForgeV, 1.2519, 1.2419, 1.2562, Blue Super, V-Toku1, etc. However, for low alloy steels the main choices for a “bright” layer are 15N20 and L6 and neither has much wear resistance.

If making a Damascus combination for high toughness there must be two high toughness steels. 1084 and 15N20 are a decent combination for this. 1084 is somewhat lower in toughness but the two together still tested at ~34 ft-lbs in a random pattern (as opposed to 45-50 ft-lbs for 15N20).

Summary and Conclusions

It is best to choose steels that have similar temperatures for heat treating. High manganese is best for a dark layer after etching while Nickel is best for a bright layer. Other alloying elements can make forge welding difficult when using conventional flux methods, especially high levels of chromium. It is important to adjust heat treatment based on the steels that are combined together. For example, 15N20 sees a large toughness drop when austenitized at 1500°F. So when 1084 and 15N20 are used together it is best to heat treat with a furnace so that the steel isn’t overheated. The optimal temperature range for austenitizing 80CrV2 is too high for 15N20. If 80CrV2 and 15N20 are used together it is better to use a normalized starting structure so that a lower austenitizing temperature of 1475°F (800°C) can be used with the 80CrV2 and still achieve high hardness. L6 may be more optimal than 15N20 to use along with other chromium-alloyed steels though it has lower nickel than 15N20 (thus contrast is not as good), and is also more difficult to work with because of its alloy additions. High carbon steels are more sensitive to overheating because the grain boundaries melt, leading to crumbling during forging. High vanadium and tungsten steels are best for high wear resistance combinations, though they typically need to be combined with 15N20 or L6 which unfortunately do not have much wear resistance.

---

[1] Verhoeven, John D., and Howard F. Clark. “Carbon diffusion between the layers in modern pattern-welded Damascus blades.” Materials characterization 41, no. 5 (1998): 183-191.

[2] Meilach, Dona Z., George Martin, E. A. Chase, and Theodore Davidson. “Decorative and sculptural ironwork: tools, techniques, inspiration.” 1977.

[3] Dunkerley, Rick. “His Forge Burns Hot for Mosaic Damascus.” Kertzman, Joe, ed. Blade’s Guide to Making Knives. Krause Publications, 2005.

[4] https://www.kovares.com/product-page/165mm-damascus-petty