Hello Kevin, ( a fine and good sounding name by the way), as a chemist you should be able to run with this stuff a little faster as long as you make the switch from covalent to metallic bonding and just focus on the crystalline, but the carbide compounds should make perfect sense.

But enough of that, the process can be simple, just the deeper knowledge helps sort out the “whys” of it. As to what temperatures you use, well that is touchy because while it should be up to the steel it is often a matter of preference of each individual smith, and giving contrary information can step on toes, so I will tread very lightly here.

Most students are confused about the differences between normalizing and annealing, after all they appear to be the same procedure for many smiths, however the difference rests in what we are trying to accomplish. A thermal cycle or treatment for the specific purpose of softening the metal for subsequent cold working (most often machining) would be annealing. A thermal treatment for the purpose of homogenizing or refining the internal conditions or structures of the steel would be normalizing. One can normalize and end up with the steel a bit harder than before, but this is not the case with annealing.

The normalizing you are using is pretty standard among many bladesmiths, you are heating to a temperature between 1414F and 1500F resulting in recrystallization and some redistribution of carbide. During forging the hammering introduced a lot of chaos in a very uneven fashion, you are now smoothing all that out. If there has been a lot of issues or you have some added alloying an initial higher temperature normalizing may be necessary to break up carbides and even them out.

Always staying below the Currie point may allow you to introduce some new finer grains but it will not do anything for your carbide distribution, so you had better have all your carbides exactly where you want them without normalizing if you wish to take that approach. The vast majority of smiths tend to focus on grain size at the expense of carbides, but it will be the carbides that will determine how sharp your knife will get and how the edge will wear under use, so it is worth it to give them some attention. Many smiths who take this approach will often have accentuated alloy banding appear on their blades since the carbon only tends to group up instead of dispersing as it would with higher temperatures.

For annealing, there are two categories – full annealing (above the Currie point) and “sub-critical” annealing (below the Currie point). Full annealing is the easiest and will work with simple alloys below .85% carbon. Just heat to put everything into solution and then leave it in the vermiculite. This will allow the carbon to slowly come out of solution into alternating sheets of carbide and iron within the steel. This will allow grinding, bending and other simple cold operations but those sheets will not be nice to mills, drills and other fine cutters. Also this method will remake what ever grains you had from normalizing into a new set, and obviously rearrange carbides. This method can be very problematic with steels that have carbon content in excess of .84% or have added alloying.

The sub-critical anneal does not involve recrystallization or heavy movement of carbide grouping but instead will take that pesky carbide and group it up where it will stay out of trouble in fine rounded shapes that are nicer to your cutting tools. But since it will leave things more as is it is best preceded by whatever treatments you want to get grain and carbide just the way you want it. It can be done with any steel but is more important for steels above .84% carbon or have added alloying.

I hope this helps, and does not add to the confusion. Many prefer simple one or two line answers that quickly tell you to do this or do that, you will find I am not wired that way and insist on giving the “why’s” behind it, although I fear it often adds to the confusion.

Thank you for replying Kevin. I was hoping that you would. Great explanation as always.

No confusion here. It makes sense to me. I would rather have more of the "why's" because I like full information.

We will see you at Blade. Thank you again.

Brion

Kevin C - much obliged for how you put this in understandable language. Your note about .84% C sent me off to my "go-to" chart on steel compositions at http://www.zknives.com/knives/steels/steelchart.php - if you have other web references that you like for steel composition or characteristics that you'd like to share - please do!

Hi Kevin,

Thanks for the reply. The carbides do make sense. Part of the problem is the reduction to practice. And learning enough ferric metallurgy to not field like a new graduate student... If you don't mind, I am going to restate things as I understand them and ask a few more questions.

Most students are confused about the differences between normalizing and annealing, after all they appear to be the same procedure for many smiths, however the difference rests in what we are trying to accomplish. A thermal cycle or treatment for the specific purpose of softening the metal for subsequent cold working (most often machining) would be annealing. A thermal treatment for the purpose of homogenizing or refining the internal conditions or structures of the steel would be normalizing. One can normalize and end up with the steel a bit harder than before, but this is not the case with annealing.

Okay I got this - that's why we normalize to reduce the stress we induce and then anneal it so it's soft.

The normalizing you are using is pretty standard among many bladesmiths, you are heating to a temperature between 1414F and 1500F resulting in recrystallization and some redistribution of carbide. During forging the hammering introduced a lot of chaos in a very uneven fashion, you are now smoothing all that out. If there has been a lot of issues or you have some added alloying an initial higher temperature normalizing may be necessary to break up carbides and even them out.

Always staying below the Currie point may allow you to introduce some new finer grains but it will not do anything for your carbide distribution, so you had better have all your carbides exactly where you want them without normalizing if you wish to take that approach. The vast majority of smiths tend to focus on grain size at the expense of carbides, but it will be the carbides that will determine how sharp your knife will get and how the edge will wear under use, so it is worth it to give them some attention. Many smiths who take this approach will often have accentuated alloy banding appear on their blades since the carbon only tends to group up instead of dispersing as it would with higher temperatures.

The above may explain how my full quench knife, which broke, had a large grain structure and yet cut really well on the rope-wood-hair tests. It shattered on bending. So if I when do I mix and mathc these? Say an initial normalization (how many?) to above Tc move carbides around and then 2-3 below to change grain size? I thought that to change grain size you needed to be above the Curie point? That seems to be wrong? I thought the loss of magnetism was due to disorienting the crystals?(Yes, we chemists don't do much metallurgy and no one in the Mat Sci Dept I work at does iron)

Also how tight is that temperature range? Would borrowing a TC be a good idea or are colors okay?

For annealing, there are two categories – full annealing (above the Currie point) and “sub-critical” annealing (below the Currie point). Full annealing is the easiest and will work with simple alloys below .85% carbon. Just heat to put everything into solution and then leave it in the vermiculite. This will allow the carbon to slowly come out of solution into alternating sheets of carbide and iron within the steel. This will allow grinding, bending and other simple cold operations but those sheets will not be nice to mills, drills and other fine cutters. Also this method will remake what ever grains you had from normalizing into a new set, and obviously rearrange carbides. This method can be very problematic with steels that have carbon content in excess of .84% or have added alloying.

That makes a lot of sense and its what we do with mild steel in blacksmithing. What happens at 84 points? I know Greg Nealy mentioned it in class but why do things change there?

The sub-critical anneal does not involve recrystallization or heavy movement of carbide grouping but instead will take that pesky carbide and group it up where it will stay out of trouble in fine rounded shapes that are nicer to your cutting tools. But since it will leave things more as is it is best preceded by whatever treatments you want to get grain and carbide just the way you want it. It can be done with any steel but is more important for steels above .84% carbon or have added alloying.

So this way the changes induced by normalizing are kept... got that.

Thanks so much. This was a lot of what I was looking for: as a working scientist, I hate the answer "it works" when you ask why so your explanation was great. I guess my question is now practically how do I know when to normalize above or below Tc. And should I do both?

And may I ask for some references on the science end? I just ordered the book "Steel Metallurgy for the Non-Metallurgist" - what else would you recommend?

Thanks again,

Kevin

Here's a pretty decent freebie on metallurgy, forge buddy.

http://www.feine-klingen.de/PDFs/verhoeven.pdf

|quoted:

Here's a pretty decent freebie on metallurgy, forge buddy.

Hi Matt,

I ordered his book mentioned above but what is confusing me is the application. Kevin C's example makes sense and helps with book but how do I, as a dappler-smith, know when to do which? From the above, it sounds like I should be normalizing above Tc and then again below to both distribute the carbides and reduce grain size because what I do to steel is probably considered abuse. <img src=' http://www.americanbladesmith.com/ipboard/public/style_emoticons//smile.gi f' class='bbc_emoticon' alt=':)' />

BTW I did buy that treadle hammer. Not as cool as a tire or air hammer but so much better than arm power.

Kevin

|quoted:

Okay I got this - that's why we normalize to reduce the stress we induce and then anneal it so it's soft.

Not "reducing" anything as much as "normalizing" or evening it all out, with steels like L6 or O-1, 5160 or 52100 there may be more stress after a good normalizing but it should be evenly distibuted so as not to cause as many problems, but the real advantage is in the microstructure size and distribution itself, with carbides benefitting the most.

|quoted:

The above may explain how my full quench knife, which broke, had a large grain structure and yet cut really well on the rope-wood-hair tests. It shattered on bending. So if I when do I mix and mathc these? Say an initial normalization (how many?) to above Tc move carbides around and then 2-3 below to change grain size? I thought that to change grain size you needed to be above the Curie point? That seems to be wrong?

Oneo of the reason why large grains can be so insidious is that they can actually increase some desireable properties at the expense of others. Large grained steel ill harden much better than fine grain, this is one big reason why just having great Rockwell numbers simply is not enough. Large grained steel will also cut more agressively under certain circumstances. But large grains destroy inpact toughness and many other very key qualities in a knife. I go with one initial high temp normalizing and then follow it with subsqequently lower ones, how hot the first one is depends on how bad I may think things are in the steel and, of course, what steel it is.

Your point on needing to be well above Ac1 or the Currie point (Ac2) is a good one to address because I have seen some folks somewhat mystified by the effects of low temp cycling on the grains. Recrystallyzation, being diffusive in nature, is not a simple on/off switch but rather a series of events that occur in degrees as temperature increases. The first thing to happen is nucleation of fresh embryonic points where the new austenite grain will grow from. In a sense there is even a potential new grain site even before you start heating since this nucleation will occur in preferred areas of higher energy in the matrix; this is often in the grain boundaries and especially at the corners where three grain boundaries intersect.

After the nucleation of the tiny new grain, the next step is its expansion into a full new austenite grain as it devours the previous phase grain that it was born from. There will be a short period of equilibrium when its grain boundary meets its sibling’s grain boundaries or other materials like carbides etc... This is how vanadium works as an excellent grain refiner; it creates very stubborn carbides in the grain boundaries that encourage the stability of this equilibrium period. Here is also why an even grain size can be even more important than a fine grain size since large grains grow by devouring smaller grains around them due to heavy non-equilibrium in the grain boundaries. If you have a large grain surrounded by many fine grains, on heating you have created a monster and supplied it with all the food it needs to get much worse. But if all the grains had been an equal size when their expanding boundaries encountered each other there would have been a comfortable equilibrium that would have maintained the even and fine grain structure.

Of course the next step is when the heat energy going into the system becomes so great that it overcomes the equilibrium as all the limiting factors are dissolved and there will be a ghastly cannibalistic orgy as grains devour each other and grow at an alarming rate. Each steel, due to its chemistry, has its own comfortable grain size followed by its own grain coarsening temperature. Within the comfortable range time has much less effect than temperature and the grains will resist growing for surprisingly long times, and this is why I am always harping on the myth that has frightened so many smiths that soaking grows grain, if one stays below the grain coarsening temperature they can literally soak for hours at temperature with very little grain growth. If however you exceed the grain coarsening temp you will loose control of it all in the blink of an eye.

Some interesting points about grain coarsening temperatures- modern alloys are what used to be referred to (back when there was an alternative to compare with) as “fine grained steel” for even if you remove the alloying it still benefits from aluminum nitride particles from the deoxidation process at the mill. In the old days steel was deoxidized with silicon and would begin to gradually grow as soon as it was heated to critical but the growth would be quite gradual instead of the steady resistance followed by the sudden snap. And, of course, old steels lacking any alloying would need to be watched much more carefully in regards to grain growth. To the argument that bladesmiths have done fine for 1000 years with their methods I cannot say it enough – alloying changed everything.

In view of this you can see how one doesn’t have to heat all the way to the upper temperatures to affect grain size. Heating just enough to initiate nucleation will seed the matrix with many fine points where new grains will arise.

|quoted:

I thought the loss of magnetism was due to disorienting the crystals?(Yes, we chemists don't do much metallurgy and no one in the Mat Sci Dept I work at does iron)

This one messes with both metallurgists and chemists since it actually falls with the realm of physicists, the loss of ferro-magnetism appears to be the result of changes in the spin of the iron atoms electrons. Fortunately it conveniently corresponds with face centered iron allotropes, whatever occurs with the electrons to permit the shift to the face centered cubic stacking of iron atoms results in the loss of ferro-magnetism. But it is too simplistic to say that merely heating to critical results in the loss of magnetism, as bladesmiths often do, since one needs to remember that there are other phases like delta iron that are even higher in temp but are once again completely magnetic. So the actual cause and effect of the Currie point is worth noting since delta iron is again body centered cubic and not face centered, and room temperature austenitic stainless steels or 400F marquenched blades are not magnetic since they are still fcc. So considering all this, your observation is correct that the Currie point is actually not an indicator of temperature at all but a result of the atomic situation of the iron.

|quoted:

Also how tight is that temperature range? Would borrowing a TC be a good idea or are colors okay?

This, of course, would depend on the steel, but for the most part our preferred simple steels can be done safely by eye and guided with a magnet. We have just enough alloying to make things stable but not enough to over complicate things.

|quoted:

That makes a lot of sense and its what we do with mild steel in blacksmithing. What happens at 84 points? I know Greg Nealy mentioned it in class but why do things change there?

Around .8 % carbon is what is known as the “eutectoid” in steel. It is a place that is similar to the eutectic (the lowest spot where both elements in an alloy are liquid) but deals with the solid state solution of carbon in iron. Just think of it as the most efficient ratio of iron to carbon in order not to have leftovers. Below .8% and there will not be enough carbon to fill the iron without really working and spreading it fairly thin. Above .8% will results in carbon saturated iron with some to spare that you have find a place for or it will get into mischief in the form of brittle and problematic carbides. Slow cool 1050 from critical and you will get pearlite (a nice eutectoid mix of iron and carbon) with leftover patches of carbon depleted iron (ferrite). Slow cool 1080 from above critical and you will get an entirely pearlitic structure. Slow cool 1095 from above critical and you will get pearlite with carbide in excess of what was needed for the pearlite.

Any simple steel will benefit from either heat ranges, depending on what your main focus is in desired effect. It is often more important to use the higher range for hypoeutectoids (below .8%) due to having to spread the sparse butter over the same slice of bread. While with hypereutectoids (above .8%) higher temps are more for seeing to it that the butter is evenly spread without any big globs.

|quoted:

And may I ask for some references on the science end? I just ordered the book "Steel Metallurgy for the Non-Metallurgist" - what else would you recommend?

Thanks again,

Kevin

I would normally recommend “Metallurgy Fundamentals” by Daniel Brandt to help people dive into Verhoeven’s work more easily, I have used it as a text book in teaching many people new to metallurgy, but with your background it may not be as necessary. I cannot recommend the ASM books enough, particularly the “ASM Heat Treater’s Guide” which is invaluable in heat treating almost any steel made. For those with a good grounding in science based disciplines any of the books by George Krauss are excellent resources with “Principles of Heat Treatment of Steel” being my favorite.

Kevin, I want to thank you for the excellent level of questions you ask and how you pose them, it has been some time since I have enjoyed typing this much in disussing these matters that are rather dry to many.