Welcome! Log In Create A New Profile

Advanced

Analyzing current knife steels with JMatPro

Posted by Larrin 
Analyzing current knife steels with JMatPro
October 28, 2015 04:33PM
I had some fun today running some numbers through JMatPro, a thermodynamics simulation program similar to Thermo-Calc. You can learn a little about how these programs work on this site (I am not affiliated): [www.calphad.com]

You can use these programs to calculate the equilibrium phase fraction and wt pct in each phase. There are a variety of limitations to this method; here are some of those limitations:

- It assumes that the steel is at equilibrium (the most stable state) which is often reached very slowly. Much of heat treatment is used to specifically not reach equilibrium (i.e. martensite formation).
- With high alloy steels, there are often large primary carbides left over from casting that are very difficult to get rid of through future processing. These are often not shown in equilibrium calculations because they aren't supposed to be there anymore.
- These simulations come primarily from empirical data which may not be applicable to different combinations of alloys, especially rarely used elements used together and in high amounts.

That being said I have some data for you. I ran these simulations at and around the common austenitizing temperatures used for these grades. I then pulled out carbide, C, Cr, and Mo wt pct at those temperature. That tells you generally what properties you will have after quenching. For example, the amount of carbon in austenite tells you approximately what hardness can be reached and the general proportion of lath to plate martensite. The amount of Cr and Mo in solution tells you the nominal corrosion resistance, more Cr means better corrosion resistance. The amount and type of carbide controls the wear resistance, toughness, and to some extent the size of the carbides.

First let's check some numbers against known carbide volumes: [www.crucible.com]
Carbide Type and Volume
Vanadium Niobium Chromium Total
CPM S35VN 3.0% 0.5% 10.5% 14.0%
CPM S30V 4.0% 10.5% 14.5%
440C 0% 12.0% 12.0%
154 CM 0% 17.5% 17.5%

Here are the numbers I got for S35VN, S30V, 440C, and 154CM:

print screen windows 7

The carbide numbers are actually pretty close. The biggest deviation is 154CM, which is quite a bit higher in Crucible's numbers than the JMatPro calculations. None of these simulations should be taken as gospel. Some are close and some are going to be off. But as an approximation they're probably not too bad. Based on these numbers we would expect these steels to have comparable toughness but the V grades would have superior wear resistance because of harder MC carbides.

Here are some low carbide stainless steels:


Speaking to Sandvik, they said that 14C28N has a little bit higher carbide volume than 13C26, they said that they knew they were at the edge of the carbide volume range for avoiding primary carbides. However, in these calculations, JMatPro predicts lower carbide volume for 14C28N. Again. the numbers aren't perfect. You can see that 19C27 and Niolox have somewhat intermediate carbide volumes, though Niolox carbides may be somewhat smaller because of the niobium addition. I haven't seen any micrographs that confirm that fact. Both of those steels are also on the low end of Cr in solution. JMatPro predicts a little higher C in solution for 19C27, and a little lower for Niolox, which seems to fit with the achievable hardness for those grades.


image uploading site

Elmax looks terrible here, with lots of chromium carbides, very little MC, and low achievable hardness and corrosion resistance. D2 has medium-high carbide volume, but in practice is even higher because of the very persistent primary carbides in D2. Vanadis 4 Extra looks pretty good being a PM grade with a medium amount of carbide and all of it being predicted to be MC carbide, which is why it has such a good combination of toughness and wear resistance. The MC carbides being harder than chromium carbides means much better wear resistance while the difference in toughness is negligible. The PM process allows a little higher carbide volume than conventionally cast while maintaining good toughness and edge stability, but edge stability is still worse than AEB-L/13C26 or 52100 with their 3% carbide. A2 has just a little more carbide volume than 13C26. It is not a bad choice if one wants an air hardening steel but want to avoid stainless for whatever reason. However, the predicted amount of C is too high, meaning the predicted carbide volume is likely too low.


upload image

M4 is very popular in its PM version right now, JMatPro says it has relatively high carbide volume. 52100 is right around 3% which is similar to 13C26, micrographs of the steel confirm that it has very small carbides. However, one myth about 52100 is that it has chromium carbides, it actually is cementite (M3C) which is higher in Cr. The carbides may be slightly harder than a simple carbon steel but not dramatically so. VG-10 has medium-high carbide according to JMatPro. Some have compared VG-10 to 19C27 but they really aren't that similar. The predicted numbers for Vascowear/PD-1 are almost certainly wrong. There is no way it has so much carbon in solution and so little carbide.

I also spent some time creating a hypothetical Nb-modified AEB-L. I will post those results in a day or two.



Edited 1 time(s). Last edit at 10/28/2015 04:35PM by Larrin.
me2
Re: Analyzing current knife steels with JMatPro
October 28, 2015 06:27PM
Really nice stuff there. I did some digging on A2 and apparently wasn't looking in the right place, as I never found what I'd call reliable numbers, but came out of it with about 4%-5% carbide as an educated guess. I have a piece I've been waiting to use until the inspiration hits, which will likely be around 2020.
Re: Analyzing current knife steels with JMatPro
October 29, 2015 09:52AM
Thanks a bunch Larrin, this helps a lot, I've been trying to compile a list like this for a long time, but having such complete results , even if approximate, is a great joy.
There seems to be a small bug with the VG10 values they aren't included.

Also, from what I know for Elmax typical austemp is 1150°C.

Interesting thing with the large difference for 154CM, might have something to do with the large aggregate, as you said. I remember in Crucible's 3V patent data they mentioned that without PM technology, 3V would have quite a different carbide structure, less MC carbide and some K1 carbide instead.
Re: Analyzing current knife steels with JMatPro
October 29, 2015 10:13AM
5% carbide volume for A2 is probably pretty close.

The glitch in VG-10 numbers is that I forgot to add them to the table. I checked the data sheet for Elmax and it has 1050 and 1080 °C, though that doesn't necessarily stop you from going higher. PM technology certainly changes the proportion and type of carbides formed. While PM can cause the formation of non-equilibrium phases or carbides, when comparing the carbide volume numbers, it looks like the PM process is getting us closer to equilibrium carbides rather than farther away. Part of what Crucible did with PM was take advantage of the fact that they formed more MC carbides when using PM rather than chromium carbide. That had a couple advantages: 1) It separated the carbides into two different types, making processing easier and the final carbide size smaller. 2) It increased the average hardness of the carbides, as MC carbide is much harder than high-V chromium carbides.
Re: Analyzing current knife steels with JMatPro
October 29, 2015 10:43AM
Quote
Larrin
when comparing the carbide volume numbers, it looks like the PM process is getting us closer to equilibrium carbides rather than farther away.

That was exactly my impression, too. That's why I think that Crucible numbers for 154CM come from measuring carbide volumes optically, rather than estimations, and I suspect that for CPM154 or RWL34 real numbers would be closer to the JMatPro simulation. Just a guess.

For Elmax, I was just reffering to what I have seen here, see page 7, table 6:
[docs.google.com]
I allway like to cross compare data, when possible. But looking at manufacturer sheets, it does seem like they recommend 1150°C only for M390.

And welcome to the forum, by the way smiling smiley.



Edited 2 time(s). Last edit at 10/29/2015 02:16PM by Bogdan M..
Re: Analyzing current knife steels with JMatPro
October 29, 2015 04:48PM
Crucible's numbers are supposed to be from microscopy. Unfortunately no information is given on the heat treatment or procedure.

Thank you for the welcome.
Re: Analyzing current knife steels with JMatPro
October 29, 2015 09:06PM
Quote
Larrin


Elmax looks terrible here, with lots of chromium carbides, very little MC, and low achievable hardness and corrosion resistance.

Care has to be taken here to realize what these programs calculate, what assumptions they make and the influence it has on the properties calculated. There are a few of them available, they generally run on empirical sets and if you use the various programs you will end up with ranges of properties produced as they all use different sets. The main thing to realize is how they are used and how they are not used. For example Elmax can be hardened to 60+ HRC readily, as can many of the other stainless steels shown in the above to look to have very low C volumes. But again, care needs to be taken as you listed to understand what these programs assume and the meaning it has on the properties produced. In no case would the data predicted be used to over ride actual materials data, in general they are used in the design state, or material selection stage with all properties being verified at some point through actual measurement. The programs are constantly being updated with new empirical data.

The main complain I had, which may not be true any more as it was some time since I asked, they didn't at the time produce ranges for the predicted properties, they simply calculated them. This is a pretty big issue for any empirical result as if you don't know the range of the predicted value you really don't actually know the value in a meaningful sense as you can't tell if two values are actually different from each other or not.
Re: Analyzing current knife steels with JMatPro
October 30, 2015 07:03AM
As I said, Elmax looks terrible in the calculations. In the extreme case, these simulations are for entertainment purposes only, in most cases they give good qualitative comparisons between steels (Elmax has much more carbide than 13C26), and in the best case (perhaps a limited range of compositions) it can make pretty solid predictions. I didn't run any of the JMatPro simulations for final mechanical properties, though some of those are available. It can make estimations of final hardness based on the predicted composition of austenite at a given austenitizing temperature, for example.

When designing alloys, generally a company/engineer will produce a handful of compositions that yield good results in the thermodynamic simulations, and then when heats are produced, use compositions that are somewhat below and somewhat above the predictions for key elements. You can see some of that by searching through patents from Crucible and Bohler-Uddeholm, such as this one: [patft.uspto.gov]

However, some of steel design is also based on published literature that is somewhat separate from thermodynamic predictions. For example, you may know that it is difficult to produce a steel with more than 2% Nb, or that even though thermodynamic simulations predict that VN carbides will form, in practice you will get high-V chromium carbides in the type of steel being produced, etc. Then there are kinetic considerations (not contained in thermodynamics), production considerations, end user considerations, etc.
Re: Analyzing current knife steels with JMatPro
November 01, 2015 03:39PM
Larrin, yes, that is just the critical part, to understand what is being calculated and the limitations of it. Can JMatPro produces ranges on the properties based on the errors/ranges in the empirical data it uses? This, combined with the typical alloy spread in a given alloy is the starting point for interpretation of the results. Run 440C for example with minimal C and maximum Cr and then reverse for a look at the range of possibilities. Now do this with several other simulation programs and now you have sets of ranges. The total set is then a sensible bound for what is known, but that set will be very large.

Again, not trying to make a point such calculations are not useful, just to make it clear that those numbers are not actual exact to the numbers listed, in fact they are likely not certain to even one significant digit when the above is considered.
Re: Analyzing current knife steels with JMatPro
November 02, 2015 06:49AM
Hey Larrin, I don't know if you still have access to the software or time, but if you do , a few 8% tool steels would be interesting to see, too : 3V, Z-Tuff, Sleipner, A8Mod (Uddeholm Viking/Chipper).
Re: Analyzing current knife steels with JMatPro
November 03, 2015 02:45PM
If I have time and I remember I could run them through on Friday.
Re: Analyzing current knife steels with JMatPro
November 04, 2015 07:14AM
Cool, here are the numbers to hopefully save you some time.

		3V	A8	Sl	Z-T
Carbon		0.80	0.50	0.90	0.70
Chromium	7.50	8.00	7.80	7.50
Vanadium	2.75	0.50	0.50	1.00
Molybdenum	1.30	1.50	2.50	2.00
Silicon			1.00	0.90	
Manganese		0.50	0.50	
Nickel 					1.50
Data sheets:
[www.crucible.com]
[www.uddeholm.com]
[www.bucanada.ca]
[www.zapp.com]



Edited 1 time(s). Last edit at 11/04/2015 07:18AM by Bogdan M..
Re: Analyzing current knife steels with JMatPro
November 15, 2015 07:35AM
Re: Analyzing current knife steels with JMatPro
November 15, 2015 01:40PM
Larrin - Thanks so much for this info! For a last year or so, I wanted to get a good view into - Carbon in solution & MC at iso-aust-temperature. This will save me from wasting a lot time conduct blinded experimentation on ht mid-to-hi Cr+Mo steels.

Is there a strain/dislocation curve goes along with cooling curve and phase transformation for any steels? Good to know at least for a couple steels, from there I can wildly extrapolate resulting matrix composition for a targeted steel.

Quote
Larrin
Re: Analyzing current knife steels with JMatPro
November 15, 2015 04:43PM
Larrin - I just looked at OP and noticed your comment on 52100 implies - cementite as predominant and virtually no M7C3 involves Cr.

Below is my latest BSED pic of my 52100 ht experiment. I interpreted those brighter particles as they associated with Cr. Where non-spheroid bright particles (stand alone and interfaces to spheroid particles) as RA. Am I mis-interpreted?

*note: 775C/1425F is not a typo. My Evenheat oven temp is calibrated. So 775C means +- 4C flux from actual. Although temp flux could be higher at different locations depend on proximity to elements.


[i.imgur.com]
not embedded because this pic is large. You might need to download the image to see all the details. OK, this pic is actually reduced by 40% from its original size prior uploaded to imgur.
Re: Analyzing current knife steels with JMatPro
November 16, 2015 06:15AM
bluntcut:

It is hard to interpret a single micrograph without more experience with that particular piece of steel's processing, its etching response, different imaging modes (such as SEI), etc. I couldn't say what those fuzzy areas are; it could even be an etching artifact. Certain carbides being "brighter" wouldn't necessarily mean a different type of carbide. It could be from different crystal orientations. The carbides may be too small to get EDS, but EDS might be able to tell you if some carbides are higher in Cr than others.



Edited 1 time(s). Last edit at 11/16/2015 06:20AM by Larrin.
Re: Analyzing current knife steels with JMatPro
November 16, 2015 06:20AM
bluntcut:

Do you mean stress-strain curves? I think JMatPro can generate them, but most will look the same - high strength and limited ductility. Their usefulness would be pretty low. It can also generate CCT and TTT curves. Their utility might depend on what it is you are trying to determine, such as the required cooling rate to achieve full martensite, or if you are trying to form bainite, or get a pearlitic microstructure, etc.



Edited 1 time(s). Last edit at 11/16/2015 06:21AM by Larrin.
Re: Analyzing current knife steels with JMatPro
November 17, 2015 03:50AM
Cool, thank you very much Larrin. The numbers seem to go very well with what I knew about these steels...
Re: Analyzing current knife steels with JMatPro
November 17, 2015 09:54AM
Larrin,

Your insights are well appreciated!

These "brighter" areas are the same on BSED(back scatter) and SEM. I thought, they could be precipitated type, where Cr might mixed-in. I wish to have direct access to a BSED machine, so using multiple sampling - I can decipher (akin to filtering signal from noise) object & crystal types. Those larger particles are intended Fe3C at aust-temperature.

With new perspective, I will re-examine my sets of BSED&SEM images.

==Luong

Quote
Larrin
bluntcut:

It is hard to interpret a single micrograph without more experience with that particular piece of steel's processing, its etching response, different imaging modes (such as SEI), etc. I couldn't say what those fuzzy areas are; it could even be an etching artifact. Certain carbides being "brighter" wouldn't necessarily mean a different type of carbide. It could be from different crystal orientations. The carbides may be too small to get EDS, but EDS might be able to tell you if some carbides are higher in Cr than others.
Re: Analyzing current knife steels with JMatPro
November 17, 2015 10:16AM
Hi Larrin,

Yes, stress-strain curves. I try to produce high cohesive; low dislocation/strain martensite matrix. Which mean, fine grain + tight grain boundary + low mis-align grain angles + fine carbide. In this quest, I suspect phase-changes - especially diffusion type - stress piling to cause nano cracks somewhere between Ms - Mf. I mostly, tinkering/home-researching with low Cr, C% between .9-1.3 steels. This strategy could be also applicable to mid & hi Cr+Mo steels.

Of course, CCT or TTT curves with matrix of phase vectors would be uber awesome. Each matrix point is phase fraction/vector. Cooling rate for (aust->Mf in seconds: 3, 7, 11, 20, 60); sampling rate at says at every 50C or so.

Grateful for your insights/advises!

==Luong

Quote
Larrin
bluntcut:

Do you mean stress-strain curves? I think JMatPro can generate them, but most will look the same - high strength and limited ductility. Their usefulness would be pretty low. It can also generate CCT and TTT curves. Their utility might depend on what it is you are trying to determine, such as the required cooling rate to achieve full martensite, or if you are trying to form bainite, or get a pearlitic microstructure, etc.
Sorry, only registered users may post in this forum.

Click here to login