Carbide Games

Kevin R. Cashen

Super Moderator
In the other threads in this series I have covered the different kinds of edges one can have on a knife, all with a valid claim to the title of “sharp” despite, their very different modes of cut. I would be remiss, here in the Heat Treating forum, not to cover one of the most profound effects that heat treatment can bring about in the developments of these edges. In the thread Tempering for different kinds of "sharp" I touched on the effect that carbides can have on the edge, but in this thread I will expand on how we can affect those carbide themselves in heat treatment.

When we exceed levels of around .8% carbon in steel we also exceed the iron's ability to hold extra carbon in solution. This extra carbon will concentrate to form carbide. In simple carbon steel this will be basic iron carbide (Fe3C), but in alloyed steels more complex carbides can occur with elements such as Cr, V, W, and others. But there are some factors that are worth remembering when heat treating any steel.

Not only is there a percentage of carbon that when exceeded will form carbide, but there is also a percentage of carbon in solution that will set a limit on how hard the steel can get. Most of us when hardening steel are quite happy with 65HRC as-quenched, why is this? Why not 69 or 70? Well, for the simple reason that once you reach .6% carbon in solution you start leveling off as you approach the limits of hardness you can obtain from carbon and, by the time you reach that .8% mark, adding more to the solution will actually work against you and begin to lower the hardness. This is why grain growth is actually one of the lesser things to worry about from overheating, as there are a number of bad things that will happen long before the grains begin to enlarge, and the main one is over-saturation of the austenite solution.

But that is another topic, for now we want to just remember the fact that you can control hardness in the heat to harden, almost as much as with tempering, and that there is only so much of the available carbon that you need in order to get that 65HRC you hope for. But, what I would like to talk about is what you can do with any carbon in excess of that needed for hardness. The extra is used in carbides to give you added effects, both beneficial and detrimental. As previously mentioned in other threads, abrasion resistance and hardness are not synonymous, and can be quite independent of each other. Once maximum hardness is obtained, abrasion resistance can be added on top of it with the extra carbide. But how we form and condition that carbide will have profound effects on the kind of edge we can create and how it will perform.

By controlling our hardening heat we can affect how much carbon is put into solution, and some steels will now offer different hardening heats in their spec sheets, one for greater hardness and a lower one for greater toughness. Ideally, we will only put enough carbon into solution to achieve our desired hardness, and leave the rest in well conditioned carbides. But what is “well conditioned?” For stable and consistent edge properties, finer carbides with an even distribution is desirable. This will allow you to get the finest of polished edges that will resist wear very uniformly. But what if we fall short of this condition?

Improper normalizing, poor annealing, and heavy cycling at undesirable temperatures, can result in larger, or segregated, carbide formations. This will have noticeable effects on the cutting edge. Fine polished edges many be difficult, to impossible, and hardness deviations of 20 points or more, microns apart, along the edge will drastically change how that edge breaks down in use.

Alloy banding or, in some steels, dendritic structures will provide areas of carbide concentration that will change edge performance, not necessarily for the better, but decidedly different. In the most extreme examples you can actually have bands of carbide so rich that it has robbed the surrounding material of maximum hardness, which can smear, wear, or even tear out in use. Such a blade will also cut soft fibrous materials quite aggressively, similar to serration, but it would be less than accurate to say it is holding a stable edge, and its cutting abilities would quickly wane on harder materials.

However, in heat treatment it may be a bit too late in the process to deal with carbides during the hardening operation. In its ability to control and condition carbide, normalizing is the greatest tool we have and is key to setting ourselves up for success in all subsequent operations. Normalizing has the ability to dissolve and redistribute carbide and determine how it will influence the final knife edge performance. It also has the ability to undo detrimental carbide conditions left from improper forging or annealing operations.

More than once I have had makers come to me desperate to know why the edge of their 1095 blades will not get sharp or stay sharp. Lab examination revealed the tell-tale microchipping effect. Their grain may even be quite good and the blade not overly hard, and yet the edge is plagued by brittle micro-divots. Metallography of the cross section quickly reveals that at some point, before hardening, all of the steels abundant carbon was put into solution and allowed to cool slow enough to gather it in the grain boundaries. Under the pressure of edge use, each grain, no matter how fine, is waiting to become a void in the edge as the brittle frame holding it in place lets go. This is just one condition that is quickly fixed with proper normalizing.
 
Last edited:
If it weren't for your research and sharing - (you and Ed are the two more prolific Teachers that should get some kind of Life Time Award) - I'd never have a clue why, spend years - literally - and never ever likely figure it out. I give you credit for expanding the depth and broadening of understanding the steel used for Blade Smithing. Thank you sir.
 
Kevin, I normalize by color and still air cooling. I do this mostly because I HT with a torch or in a forge but I have recently changed that by acquiring a HT oven. Now that I have "control" over temperature is there a temperature (or range) that is best for normalizing. Does it change dramatically by steel type?
 
If it weren't for your research and sharing - (you and Ed are the two more prolific Teachers that should get some kind of Life Time Award) - I'd never have a clue why, spend years - literally - and never ever likely figure it out. I give you credit for expanding the depth and broadening of understanding the steel used for Blade Smithing. Thank you sir.

Marc, that is very kind of you. You are most welcome. You put me in good company with Ed, I only wish I could have the energy to keep up with him in helping folks on this forum. I have been doing this long enough that I often run out of fresh things to talk about, so it is great questions from folks like yourself that inspire me to do more. In sophistication of the knowledge of the knife community I find it hard to believe how far we have come in the last ten years alone. With all the information that is available today the possibilities are endless.
 
Kevin, I normalize by color and still air cooling. I do this mostly because I HT with a torch or in a forge but I have recently changed that by acquiring a HT oven. Now that I have "control" over temperature is there a temperature (or range) that is best for normalizing. Does it change dramatically by steel type?

The beauty of normalizing, for as powerful as it can be, is that it is such a simple operation. As mentioned in another thread, steels alloyed enough to require tight controls can be difficult to impossible to normalize in the traditional sense. For all the steels I work with, I still often just do it by eye myself; I have the forge hot and have just finished forging so why not get it done now. I have any number of sophisticated means of controlled heating but if I want to go for more than the forge, I will fire up the kiln to around 100°F higher than the recommended hardening temperature, equalize the blade and then air cool. If working with alloys or a steel that I believe to have segregation problems I may add another 50°F to that. Normalizing is great that way- all you have to due is achieve full solution, nothing fancy, and then air cool. If there are issues from a hotter normalizing you can then fix them with a couple of follow up cycles at a lower heat, after the initial problems are fixed.
 
The beauty of normalizing, for as powerful as it can be, is that it is such a simple operation. As mentioned in another thread, steels alloyed enough to require tight controls can be difficult to impossible to normalize in the traditional sense. For all the steels I work with, I still often just do it by eye myself; I have the forge hot and have just finished forging so why not get it done now. I have any number of sophisticated means of controlled heating but if I want to go for more than the forge, I will fire up the kiln to around 100°F higher than the recommended hardening temperature, equalize the blade and then air cool. If working with alloys or a steel that I believe to have segregation problems I may add another 50°F to that. Normalizing is great that way- all you have to due is achieve full solution, nothing fancy, and then air cool. If there are issues from a hotter normalizing you can then fix them with a couple of follow up cycles at a lower heat, after the initial problems are fixed.
That is what I needed to know, Thanks. Do you have a set number of normalizing cycles you like to do?
 
Oh I'm giddy-I feel like we're going to talk about "thermal cycling" and what it really is and what it really does. :)

If you would like.

“Thermal cycling” is a term that was necessitated by the very large amount of confusion among knifemakers regarding true normalizing. For years knifemakers seemed to believe that the definition of normalizing was any heat treatment done pre-hardening that was not annealing. Indeed, I see a very large number of folks use the term that way today. I often see knifemakers refer to heats to around 1500°F, or below, as “normalizing.” But this is not the case. Normalizing, as defined by industry, is a homogenizing operation involving temperatures high enough both recrystallize, and fully dissolve all the current phases in the steel for complete austenite solution. This makes it a VERY powerful thermal tool, essentially the ultimate reset switch for the steel.

But complete austenite solution will also involve grain enlargement. This, however, is not necessarily a bad thing since the new grains will have a more uniform common size, regardless if what size that is. Grossly uneven grain size is a worse problem than slightly larger grains. Also, knifemakers need to worry less about grain size because it is one of the easiest things to manipulate, or fix, in steel. This is where the other cycles come in.

Manipulating carbide is very hot work, and the purview of normalizing, but adjusting grain size is slightly cooler work that should be done at temps that will allow it without touching the carbide condition we carefully arranged with our normalizing. Reheating at a lower temperature, more along the lines of what we would do in hardening, we can create new grains within the uniform framework created during normalizing. For cycling, it is worth noting that quicker heating, with less hold time, is more effective in refinement. Faster heating increases the rate of subgrain nucleation and thus many more grains in the same given area. For normalizing hold times can be handy because uniform solutions are a matter of equilibrium resulting from time at temperature. But when refining grain size, points of un-equalized higher energy results in more grains and, coupled with undissolved carbide, eliminates subsequent growth of those grains.

So, not only are we seeing a contrast between desired outcome of normalizing and cycling, but we are seeing that they are indeed different in procedure. Cycling also incorporates repeating the action to double up on its effects. There will often be reducing heats on each cycle and this may, or may not, be necessary, depending on the degree of refinement achieved and yet desired. Each grain size also has a grain coarsening temperature- the temperature at which the equalized grains will begin to absorb each other and grow. This temperature will decrease along with grain size, so it may be necessary to lower the temperature as the grains fall in size in order to keep that reduction going.

But be aware that there are always tradeoffs and limits to benefits. Finer grain size will reduce hardenability necessitating faster quenches later on. But also, be aware that while high temperature work will dissolve and scatter carbide, low temperature cycling will bunch it back up, so excessive cycling in the 1300’s F can undo some of the gains made in normalizing. Those who get carried away with this may find a nasty surprise in the form of alloy banding in their finished blade due to this effect, indicating that another normalizing reset may have been in order.
 
Last edited:
I was so right to be giddy, thanks Kevin!

Can generalizations be made, regarding low alloy steels, about appropriate normalizing and grain refinement temps based on carbon content and alloying? Is there a starting point for the knife maker in lieu of a data sheet?
 
I was so right to be giddy, thanks Kevin!

Can generalizations be made, regarding low alloy steels, about appropriate normalizing and grain refinement temps based on carbon content and alloying? Is there a starting point for the knife maker in lieu of a data sheet?
That's a great question. Thanks for asking.
 
I was so right to be giddy, thanks Kevin!

Can generalizations be made, regarding low alloy steels, about appropriate normalizing and grain refinement temps based on carbon content and alloying? Is there a starting point for the knife maker in lieu of a data sheet?

Simple carbon steels are much easier to guess at. I can start by laying down some general guidelines on hardening heats. I see no reason ever to exceed 1500°F when hardening a simple carbon steel, and I would even stay at 1475°F maximum if there is any hold time. Depending on the condition going in you have a range from 1425°F to 1475° to work with for 1070, 1075, 1080, 1084 and 1095. Generally, the lower the carbon content, the higher the temperature you can use.

As mentioned in my post above, normalizing will often be around 100°F higher than hardening to insure total solution. So, something like 1095 could normalize from 1550°F up to 1600°F. 1080 or 1084 would be more around 1600°F and 1070 could go to 1625°F. However, sometimes there may be segregation issues that could require more temperature thrown at it. If you feel you have banding or are getting lower than desire as-quenched hardness, regardless of the quench, you may have to go to 1650°F or better to wipe it out. But, generally, the lower the temperature, the better.

With the addition of any kind of alloying, the numbers can go up surprisingly fast. I was honestly shocked to see how high one could heat 52100 before fully dissolution of prior phases- like forging temperatures hot! And, speaking of forging, the need for serious normalizing can be greatly reduced if you are forging the blades. When forging is approached as just another homogenizing heat treatment, it is even more effective than normalizing in redistributing carbide. Then normalizing’s job is mostly to deal with the uneven strain effects of forging.
 
Simple carbon steels are much easier to guess at. I can start by laying down some general guidelines on hardening heats. I see no reason ever to exceed 1500°F when hardening a simple carbon steel, and I would even stay at 1475°F maximum if there is any hold time. Depending on the condition going in you have a range from 1425°F to 1475° to work with for 1070, 1075, 1080, 1084 and 1095. Generally, the lower the carbon content, the higher the temperature you can use.

As mentioned in my post above, normalizing will often be around 100°F higher than hardening to insure total solution. So, something like 1095 could normalize from 1550°F up to 1600°F. 1080 or 1084 would be more around 1600°F and 1070 could go to 1625°F. However, sometimes there may be segregation issues that could require more temperature thrown at it. If you feel you have banding or are getting lower than desire as-quenched hardness, regardless of the quench, you may have to go to 1650°F or better to wipe it out. But, generally, the lower the temperature, the better.

With the addition of any kind of alloying, the numbers can go up surprisingly fast. I was honestly shocked to see how high one could heat 52100 before fully dissolution of prior phases- like forging temperatures hot! And, speaking of forging, the need for serious normalizing can be greatly reduced if you are forging the blades. When forging is approached as just another homogenizing heat treatment, it is even more effective than normalizing in redistributing carbide. Then normalizing’s job is mostly to deal with the uneven strain effects of forging.

Mr. Kevin, as always your posts are very interesting !! Thank you very much for your time!!!
One question please: which thermal cycle would you recommend for the 5160?
I hope you understand me, I am using a translator, I am writing to you from Argentina.

Thank you very much!!!

Jorge Iruzubieta
 
Back
Top