Brittle failure in 1095 hunters (Calling Mr Cashen!!!!)

NDallyn

NDallyn

Joined
Jun 22, 2006
Messages
814
So I decided to try out some new material (1095 - at least it's new to me). Before I let any blades out of the shop, I decided to do a little testing. I rough ground 5 blades out of 1/8" thick 1095 from admiral steel to about 1/16" thick at the edge (my usual pre-heat treat thickness), drilled and profiled the tang as I usually would. My heat treat method (now, after testing, quite obviously flawed) was as follows:

-edge heated with a torch to non-magnetic (2/3 of the distance from edge to spine)
-immediately water-quenched in a preheated tank (125 degrees F) to hand warm
-transferred immediately from the quench to a preheated temp. oven (400 deg. F)
-tempered for 3 cycles (1 hour per cycle at 400 F, allowed to air cool to room temp between cycles)

Keep in mind that none of these blades cracked in the quench (as far as I could see or hear). Rockwell hardness, according to the charts, should be around Rc 62.

Each of the blades was clamped in the vise approximately 1/3 of the distance from the tip to plunge and pulled by hand (no snipes or torque wrenches here) in the attempt to do a 90 degree bend. I say attempt because not one of the blades went past 45 degrees before snapping like a twig. Here's what I observed:

-3 of the 5 blades broke in multiple places with little or no permanent deformation
-2 of the blades had a brittle longitudinal fracture along the bounderies of the quench line (possible carbide segregation along this line?)
-as you can observe in the macro pic of the cross-section, there is a course structure near the spine (soft area), a finer structure along the edge, and a very fine structure in the center at the boundary of the heated area (quench line)

I must admit I had some apprehension about posting these results, I wouldn't want anyone to think I'm putting out crap - but, I figure that showing the results of my own testing would help a few of the beginners. I have done bend tests like this with O-1 and have not had any issues. This is why I test everything before I send a blade out - I would rather bust a few blades in the shop than ever send out a flawed product.

Obviously, my ht methods on these blades were severly flawed. If Mr. Cashen or anyone else can add anything to help me refine my methods and put out the best blades I possibly can, I would appreciate it.

Have a good one,
Nathan
 
NDallyn said:
So I decided to try out some new material (1095 - at least it's new to me). Before I let any blades out of the shop, I decided to do a little testing. I rough ground 5 blades out of 1/8" thick 1095 from admiral steel to about 1/16" thick at the edge (my usual pre-heat treat thickness), drilled and profiled the tang as I usually would. My heat treat method (now, after testing, quite obviously flawed) was as follows:

-edge heated with a torch to non-magnetic (2/3 of the distance from edge to spine)
-immediately water-quenched in a preheated tank (125 degrees F) to hand warm
-transferred immediately from the quench to a preheated temp. oven (400 deg. F)
-tempered for 3 cycles (1 hour per cycle at 400 F, allowed to air cool to room temp between cycles)

Keep in mind that none of these blades cracked in the quench (as far as I could see or hear). Rockwell hardness, according to the charts, should be around Rc 62.

Each of the blades was clamped in the vise approximately 1/3 of the distance from the tip to plunge and pulled by hand (no snipes or torque wrenches here) in the attempt to do a 90 degree bend. I say attempt because not one of the blades went past 45 degrees before snapping like a twig. Here's what I observed:

-3 of the 5 blades broke in multiple places with little or no permanent deformation
-2 of the blades had a brittle longitudinal fracture along the bounderies of the quench line (possible carbide segregation along this line?)
-as you can observe in the macro pic of the cross-section, there is a course structure near the spine (soft area), a finer structure along the edge, and a very fine structure in the center at the boundary of the heated area (quench line)

I must admit I had some apprehension about posting these results, I wouldn't want anyone to think I'm putting out crap - but, I figure that showing the results of my own testing would help a few of the beginners. I have done bend tests like this with O-1 and have not had any issues. This is why I test everything before I send a blade out - I would rather bust a few blades in the shop than ever send out a flawed product.

Obviously, my ht methods on these blades were severly flawed. If Mr. Cashen or anyone else can add anything to help me refine my methods and put out the best blades I possibly can, I would appreciate it.

Have a good one,
Nathan
Click to expand...

A couple of thoughts for you.

firstly:
WHY do you consider this to be a fatal flaw?
Yes I realize that the blade broke. I am challenging the assumption that the blade should DEFORM prior to BREAKAGE.


Secondly:
WHY did you heat just the edge to what we will assume is a correct hardening temperature?
Instead of hardening the entire blade and then draw back the spine?


This is not about how I or any other bladesmith HT's 1095.
We need to establish what your end goal is before we can determine a process for achieving it.
 
Here's a chart that may lead to some insight, but I'm in agreement with Stephan on this one regarding all of his counter-questions.

web.jpg
 
I'm with Stephen above. What is your goal? If your goal is to use 1095 to create knife that will pass an ABS-style 90 degree bend test, you should bear overall geometry in mind as well. Knives that are generally tested in this manner have blade lengths around 10". You're expecting to get a 90 degree bend in about 3" of blade (once you subtract the bit clamped into the vise). ABS test knives will be trying to make the same bend in about 8" of blade length. The difference here is HUGE. The difference in the radii needed to do a 90-degree bend in these two design is just massive.

Also, 1095 being a hypereuctoid steel may not be the best suited steel for this test. If it were, you'd hear about a lot of folks using 1095 for an ABS test. As it stands, you don't. There's a reason.

Now, assuming that you did get your temps and quench right, I'd argue that 400F is too low a temp for a blade to successfully withstand the bend test you gave these test blades.

To add some more thoughts to this as well...It doesn't appear that you normalized prior to HT. With a hypereuctoid steel, I'd say that this step is critical. You want to ensure that your carbon has been brought evenly into solution throughout the entire piece. this can help to reduce carbide "clumping" that may have occurred at the mill if there was any variance in the anneal (and if Kevin's conversations have taught me ANYTHING it's that I can't trust anything I didn't do myself, and even then it should be suspect...). The course structure toward the spine should only make you question the condition of the steel as purchased IMHO.

My gut is saying that if you didn't crack any of 5 blades in a water quench that something was likely amiss in your HT anyways. I've never met anybody water quenching a hypereuctectoid steel that is THAT lucky...

Just some thoughts...

-d
 
The ht method I used in this instance was based upon a method given in a video by an ABS mastersmith (who I will not name, so please don't ask - this is about me trying to refine my methods, not criticize someone else's). This is why I only heated the edge and not the entire blade. The magnetic state of the steel was checked with a 1" pot magnet as the blade was brought up to temp. (again, according to the prescribed ht method I was using).

To answer your first question, I do not consider a broken blade to be, in and of itself, a fatal flaw. How I heat-treat largely depends on what a blade is meant to do (a 6" skinning knife will be placed under less stress/side-loading than a 15" camper) However, the way that these blades broke indicates an extremely brittle condition and gives evidence of the significant internal stresses that remained in the blade after tempering. The fact that these blades experienced a brittle fracture along multiple planes without any significant deformation indicates to me that these blades may have failed under repeated stress and/or side loading during normal use. A hunting/outdoors blade should, at least in my opinion, be able to withstand the repetitive flexing and side loading which could be exerted on it during "normal" use (cutting through cartilage, hitting bone, clearing shooting lanes, camp chores, etc.).

When I test my blades, I test for side loading/flexing, edge holding, sharpenability, and balance/fit (how does it feel in the hand). If I had finished these blades out, they may have met some of the quality criteria I hold myself to, but failed others - that is unacceptable to me. Therefore, my methods in making these blades were unsuitable for the expectations I place upon blades of this type. What I need to figure out, is how to modify my methods to produce a blade that will meet these expectations. I may give up on using the torch/forge for HT and use the Evenheat. I just hate the idea of foil wrapping an oil/water quenching steel (maybe use turco?)
 
Deker,

Point taken on normalizing. I did not normalize these blades before grinding and heat treating. I will definitely stick the next ones in the kiln and normalize them first.

My objective here was not necessarily to pass a 90 degree bend test. If they had bent back and forth like a wet noodle - I would question my methods just as much as if they broke with little flex. My objective here was to flex them to the point of failure, observe how they failed and determine whether or not the stresses that caused that failure would exceed the stresses that would be placed on the blade during it's intended normal usage.
 
First: 1095 being a hypereutechtoid steel, just bringing it up to the Curie point (nonmagnetic, about 1412 F) then quenching it puts very little of the carbides into proper solution where you can actually use them first of all, you were below critical for 1095, so my guess is you didn't properly austentize it. If you don't hold it at 1500 or slightly more for a proper soak your carbides do not dissolve into the solution, you can have all sorts of interesting issues.
Try reading Kevin's thread on hypereuctechtoid steel (which is what 1095 is)

http://www.bladeforums.com/forums/showthread.php?t=615086

lot's of good info!

-Page
 
Quite a while back I wrote to Kevin Cashen for some help on heat treating 1095 with a basic forge and Parks #50 quench. He was kind enough to to answer my request and the method he describes below it is what I still use if I do my own heat treat on 1095. Kevin posted this publicly so I hope he does not mind me re-posting.

One additional trick I have learned is to use table salt as a heat indicator. It melts at 1475 degrees F. Just sprinkle it on the blade and when it melts wait about 1 additional minute and go to the quench.

Here is Kevin's post:

Now for the 1095. 1095 tends to get a bad wrap for a couple of reasons, the first is that it is not 1084. 1084 has just the right carbon levels to make the most complete resulting phases and it also has the manganese levels to allow it to harden a bit easier. 1095 has lower manganese levels causing it to need a faster quench, but it also has an extra .10% carbon that needs to be dealt with in order to insure it is useful and not detrimental.

The ideal final condition for 1095 is a very fine martensite (the hard stuff) containing around .8% carbon in solution with the extra .15% in very fine carbides evenly dispersed throughout. It is possible to do this with tight heat controls, but in the forge we can come close by using prior heat treatments to set things up. The worst condition for 1095 would be low carbon coarse martensite with all the extra carbon separated out in the grain boundaries; a blade made of this would have lousy edge holding and still be rather brittle.

To get what we want we should start our heat treating in the forging, and watching our heats. Start hot and work cooler as you get closer to finishing. Normalizing will be important. Start hot in normalizing as well, 1700F is not too hot. Allow to air cool as quickly as you can, cooling quickly to Ar1 (where pearlite forms) will be important in not having extra carbon where you don’t want it. Since 1095 is so simple all you have to do is cool below 1000F for good results. Believe it or not I have a large container of all kinds of old used oils that I use in normalizing. If you quickly quench into a sludgy oil just enough to turn the steel black and then reheat, you can refine things quite well without any carbon forming in the grain boundaries and not shock the steel by quenching from such a high temperature. For me the normalizing would consist of several cycles with the heat getting cooler every time. The final heat would be to just nonmagnetic and the quench would be all the way to room temp in preparation for the anneal.

Next take the quenched and hardened blank to a dull red glow but never allow it to go nonmagnetic, do this several times but always staying magnetic, if you go over it kind of takes you back to square one. When done, cool in air. This should take all of the carbon and put it into very fine globs throughout that will allow machining and other operations.

When it is time to harden your 1095 blade, you should carefully heat it to nonmagnetic and hold it there for just a second (this isn’t really a soak but the attempt to hold will allow the temp to actually go slightly above nonmagnetic and achieve a good solution). Quench entirely under the surface of the #50 oil and keep the blade moving tip to tang as quickly as you can, do not just hold the blade still under the oil. If you wish to edge quench do NOT use the #50 as you will only ruin your investment in a good oil, get some cheap ATF (automatic transmission fluid) and ruin that with an edge quench technique. However I would also point out that there really is no strength benefits in an edge quench and the chances of fine pearlite at the edge are significantly increased.

As soon as the blade is down to the temperature of the oil, get it into a tempering oven preheated to at least 350F and heat it for around an hour. The actual tempering temperature will depend on your desired final hardness and how things worked out in the hardening heat. 400F normally makes a very good knife with 1095.

If you wanted to play with interrupted quenching and other techniques you should find plenty of information on that with a search of these forums. But this should give you a blade if fine gained 1095 that will only have pearlite at the very spine and a very nice carbide dispersion with good hardness. If properly done it should be able to slightly out cut a similar blade of 1080.
 
its hard to tell...but by the end picture...looks like the grain was large... did you normalize at all ?

small grain makes a difference. ....!
 
I don't think you need to worry at all about people thinking poorly about your knives from these posts.

It only shows that you are making sure you have a good quality knife.

I have to confess, I was sad to see those great looking blanks in pieces. That looks like a great design.
 
Thanks for the comments guys!! This is exactly what I was looking for - things I hadn't necessarily thought of, like normalizing a piece of supposedly spherodized and annealed barstock. Honestly, I think I will probably give up the guesswork and just work out a method for heat-treating my carbon steel blades in the evenheat kiln (where I do my stainless blades) instead of trying to use simpler methods like a forge. I tried a method I was unfamiliar with on a steel I was unfamiliar with and failure was the end result - I'm just trying to learn what I can from it.

Have a good one,
Nathan
 
I would never recommend hardening the edge only. You would have a martensitic edge and much weaker spheroidized spine. Much better would be heat the entire blade and differentially harden [edge quench] That would give you a martensitic edge and pearlitic spine .
Heating just the edge has the risk of overheating the very edge and especially the tip.That results in a large grain and brittle steel !! If you heat with a torch the heat should be applied only to the spine and let that heat diffuse to the edge.For 1095 remember that annealing risks grain boundary carbides also resulting in brittle steel !
 
Not much to add here, but here are a few thoughts.

Normalize before HT, even if not forged.
Use a fast oil, water just isn't a good quench.
400f temper is OK for a cutter but too low for a bender.
Torch or forge is OK to use as a heat source but takes lots of practice.
 
stephanfowler said:
A couple of thoughts for you.

firstly:
WHY do you consider this to be a fatal flaw?
Yes I realize that the blade broke. I am challenging the assumption that the blade should DEFORM prior to BREAKAGE.


Secondly:
WHY did you heat just the edge to what we will assume is a correct hardening temperature?
Instead of hardening the entire blade and then draw back the spine?


This is not about how I or any other bladesmith HT's 1095.
We need to establish what your end goal is before we can determine a process for achieving it.
Click to expand...



Just to be very clear here, I'm not trying to imply at ALL that what you did is wrong. I do my 1095 entirely differently and I wanted to have a clear understanding of your intent before I started suggesting you do things the way I do.
 
Well heck, if you have an Evenheat by all means use that! :D

What you want to do do avoid scale, etc is to use PBC Anti-scale from Brownell's. Heat the blades to 500-600F, then apply a sprinkling of PBC powder to the whole blade evenly. It will melt and form a barrier to the oxygen. After quenching use hot water and a good abrasive pad to remove the PBC. Voila! I know guys who go to a 600 grit finish before HT now because of PBC.

-d
 
deker said:
Well heck, if you have an Evenheat by all means use that! :D

What you want to do do avoid scale, etc is to use PBC Anti-scale from Brownell's. Heat the blades to 500-600F, then apply a sprinkling of PBC powder to the whole blade evenly. It will melt and form a barrier to the oxygen. After quenching use hot water and a good abrasive pad to remove the PBC. Voila! I know guys who go to a 600 grit finish before HT now because of PBC.

-d
Click to expand...

The above, works well for me also.

I have good luck with PBC over clay. The clay seems to hold together better than without. It holds the edges of the clay to the blade surface.
Fred
 
I think we discribed this in detail in Kevin's "Hypereutectoid " thread. The slow cooling of 1095 and other hypereutectoid steels permits the carbides to diffuse to the grain boundaries causing a brittle condition. With a faster cooling rate such as normalizing or even quenching the carbon has no time to diffuse and the carbides are then distributed evenly.
My photo shows intergranular fracture which you would get when you have grain boundary carbides. Note that the fractures continue even perpendicular to the fracture surface ! Kevin's photo shows a polished and etched sample . http://www.bladeforums.com/forums/showthread.php?t=615086
 
Thank you Mete. I just finished up reading the thread you linked to. I will have to restudy portions again. The concept of the carbon gathering in the boundaries seems simple enough. But all of the concepts involved in heat treating seem overwhelming. Little by little it's sinking in.

Thank you Mete and Kevin
Alden
 
You must log in or register to reply here.
Back
Top