15n20

[Jon Cook]

Last week, I found myself stuck at work into the wee hours, and some unexpected cash to burn. So, I decided to buy a chunk of .09" 15n20. I had it in my head that since it's usually paired with 1080 for patterning, I was basically getting a shinier 1080.

In the sweltering light of day, research seemed like a good idea. But, outside of @Aiden CC's stockman project, and a handful of others, I haven't found much on heat treating this steel by itself. What I have found suggests optimal is closer to O1. Can anyone confirm the need for a long soak at critical, or can I get good results with a simpler treatment?

I'm planning on making some small, stock removal EDCs, aiming at 60-62 RC if that makes a difference. 

[Joël Mercier]

I found this, not sure it's 100% accurate.

 

[Jon Cook]

That's pretty much what I'm seeing. Short of sending it out, or convincing the makerspace to let me repurpose their pottery kiln, I can't pull off a soak that long right now. 

I saw a thread suggesting normalizing O1 starting a good bit above critical can minimize the soak time. Would something like that work, or should I sit on it until I can build a proper kiln?

[Jon Cook]

Mods: it looks like I should have put this in metallurgy and other enigmas. Can we move it? My apologies. 

[Joël Mercier]

Since it is 0.090" thick, I am assuming you're doing stock removal? If that is the case, you do not need to normalize unless the stock was pre-hardened. And since it's thin, a short soak at austenizing temp is enough. There is no extra carbon to form carbides like in O1. 

[Jon Cook]

Yes to stock removal. Looks like most of it comes hot rolled at about 40-45, and I've read it likes to work harden. So, I think the plan is to normalize, then cut my shape. Normalize again if it work hardens. Slow ramp up to critical, hold it as long as I can without going over temp (based on decalescense/color), and quench in warm canola. This is probably going to leave more decarb to deal with after tempering as I can just manage a slight reducing atmosphere at full blast. Decarb loss shouldn't be anywhere near 80crv2, though. 

Edit: getting the nickel into solution seems to be the issue with the soak.

Edited August 25, 2018 by Jon Cook

[Joshua States]

There's this: https://www.bladesmithsforum.com/index.php?/topic/1654-heat-treat-for-15n20/

There's this: https://bladeforums.com/threads/heat-treat-15n20-procedure.1537672/

And because 15N20 and L6 are very similar in composition (or so I remember it) you can try the info in the attached file.

 

HT for L6-NF.pdf

[Alan Longmire]

I have made a few small stock removal knives from 15N20.  Treated it exactly like 1084, triple-normalized then quenched from decalescence.  No decarb, hardened fine.  I don't know the final Rc after tempering, but they passed the brass rod test.

You don't need to soak it because it's sold prehardened and tempered for bandsaw use.  The nickel is already in solution.

[D. Giagni]

My understanding is that the reason 15n20 is used with 1080/84 for pattern welding is because they both respond well to the same heat treatment.

[Jon Cook]

4 hours ago, Alan Longmire said:

You don't need to soak it because it's sold prehardened and tempered for bandsaw use.  The nickel is already in solution.

So the soak is assuming you anneal it per the data sheet. Good to know. I'm aiming to take advantage of the extra toughness and get a little more edge retention at the higher hardness.

[Jon Cook]

56 minutes ago, D. Giagni said:

My understanding is that the reason 15n20 is used with 1080/84 for pattern welding is because they both respond well to the same heat treatment.

That's a big part of why I bought it. Then I started reading what I could find on heat treat, and that didn't sound right. 

[Jerrod Miller]

For the record, Ni is larger than Fe, therefore it is substitutional in the matrix, not interstitial.  It is also not a ready carbide former.  Meaning it is always in solution.  

When we talk about getting carbon into solution, we are talking about getting it broken apart from carbides and into the interstitial sites.  Which in turn allows for FCC (austenite) to transform to BCT (martensite).  

[Alan Longmire]

D'oh!  I knew that, just trying to figure out why that description calls for a soak time on this relatively simple steel.  It's already close to eutectic, it's sold pre-hardened and tempered, so I really don't see any reason to soak it.

[Jerrod Miller]

Also for the record, I see no reason to soak this alloy.  

[Joël Mercier]

Probably to make sure the piece is heated evenly.

[Jerrod Miller]

1 hour ago, Joël Mercier said:

Probably to make sure the piece is heated evenly.

That takes seconds, not 5-10 minutes (or you're doing it pretty wrong).  And a "soak" is supposed to be after fully up to temp.  

[Joël Mercier]

Yeah, again it's probably for 1" or thicker cross sections

[JJ Simon]

Jerrod.
Does 1550 Austemp seem high on this.
Generally I heat treat my mix of 1075 and 15n20 at around 1500.
It has always worked well.
Figure being near 1550 we're at the edge of growing grain?

[Jerrod Miller]

1550 is definitely higher than needed, as critical is probably more like 1400 (possibly even less).  You should always heat to a bit above critical (like 50-75 degrees, but at LEAST 25, depending on alloy), because it takes a bit of temp loss between heat source and quench tank.  It is also helpful to cross the austenitization temp at a quick rate ("hit the ground running").  Grain growth technically starts below critical, but starts getting appreciable just after critical.  If you are getting up to 1550 fairly quickly (as blades in forges tend to do) then quenching right away you should be just fine.  If you are soaking at all (for some reason) then 1500 is probably a better idea.  Because of the alloy content in 15N20, I would think 1475 should be doable, but I'd aim for 1500.  Actually, as I've said before, I'd go be eye. 

I like to do my normalizations by going quite a bit over, watch it cool while studying the shadows (which shows the evenness of the heating, relative to section thickness).  Then do that multiple times getting cooler each time, judging by how soon the shadows form and how long it was back in the forge.  Each normalization cycle calibrates heat distribution (via positioning and moving of the blade in the forge) as well as the actual temperature above critical.  And of course it is grain refining every time.  

[Joël Mercier]

Jerrod, does that explain why a sub critical grain refinement cycle actually helps refine the grain? Austenite starts forming below the actual quenching temperature?

[Jerrod Miller]

Grain growth is a function of diffusion, which happens at a wide variety of temperatures, but being heavily temperature dependent, it happens faster at higher temperatures.  Grain refinement happens as a result of stress followed by diffusion (grain growth).  To be clear, part of grain refinement is grain growth.  I know I have covered this somewhere here.  This is the one I was thinking about, but I see there is an older one, too (and other smaller bits).